]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
projects / thirdparty / gcc.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
Add new option -ggnu-pubnames.
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
7a1dd0fa
SA		 12013-10-22 Sterling Augustine <saugustine@google.com>
2
3 * doc/invoke.texi: Document -ggnu-pubnames.
4 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
5 logic.
6 * dwarf2out.c: Include gdb/gdb-index.h.
7 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
8 debug_generate_pub_sections.
9 (is_java, output_pubtables, output_pubname): New functions.
10 (include_pubname_in_output): Handle debug_generate_pub_sections at
11 level 2.
12 (size_of_pubnames): Use new local space_for_flags based on
13 debug_generate_pub_sections.
14 (output_pubnames): Unify pubnames and pubtypes output logic.
15 Genericize comments. Call output_pubname.
16 (dwarf2out_finish): Move logic to output_pubnames and call it.
17
9ae93484
UB		 182013-10-22 Uros Bizjak <ubizjak@gmail.com>
19
20 PR target/58779
21 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
22 Remove CCCmode handling.
23 <case LTU>: Return 'c' suffix for CCCmode.
24 <case GEU>: Return 'nc' suffix for CCCmode.
25 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
26 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
27 (*sub<mode>3_cc_overflow): Ditto.
28 (*subsi3_zext_cc_overflow): Ditto.
29
ade97eb1
SE		 302013-10-22 Steve Ellcey <sellcey@mips.com>
31
32 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
33 (AND (NOT OP1) (NOT OP2)).
34
843d68c6
BS		 352013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
36
37 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
38 meaning of merge-high and merge-low masks for little endian; avoid
39 use of vector-pack masks for little endian for mismatched modes.
40
574733d3
JBG		 412013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
42
43 * config/tilepro/tilepro.c: Include "tree.h".
44
1adb2fb9
AS		 452013-10-22 Andreas Schwab <schwab@suse.de>
46
47 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
48 with PRE_DEC.
49
f2f39141
PC		 502013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
51
52 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
53 [Ed Smith-Rowland]): New entries.
54 ([Stephen M. Webb]): Update.
55
8223bb78
AM		 562013-10-22 Andrew MacLeod <amacleod@redhat.com>
57
58 * tree-ssa-ter.h: Remove duplicate copy of file contents.
59
1e08df0e
MP		 602013-10-21 Marek Polacek <polacek@redhat.com>
61
62 PR middle-end/58809
63 * fold-const.c (fold_range_test): Return 0 if the type is not
64 an integral type.
65
67b5215c
RS		 662013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
67
68 * system.h: Move hwint.h include further down.
69 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
70 gcc_checking_asserts.
71 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
72
0cad6830
BE		 732013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
74
75 Fix volatile issues in optimize_bit_field_compare.
76 * fold-const.c (optimize_bit_field_compare): Bail out if
77 lvolatilep or rvolatilep.
78
5dfd32cd
BE		 792013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
80
81 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
82 and get_inner_reference returning different pmode for non-volatile
83 bit-field members dependent on flag_strict_volatile_bitfields.
84 * stor-layout.c (layout_decl): Remove special handling of
85 flag_strict_volatile_bitfields.
86 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
87 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
88
0bff3610
PM		 892013-10-21 Paulo Matos <pmatos@broadcom.com>
90
91 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
92 calculations.
93
b5c4ff78
JL		 942013-10-21 Jeff Law <law@redhat.com>
95
f2f39141
PC		 96 * tree-ssa-threadedge.c (thread_through_normal_block): New
97 argument VISITED. Remove VISISTED as a local variable. When we
98 have a threadable jump, verify the destination of the jump has not
99 been visised.
100 (thread_across_edge): Allocate VISITED bitmap once at function
101 scope and use it throughout. Make sure to set appropriate bits in
102 VISITED for E (start of jump thread path).
103 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
104 through a joiner if any edge on the path has a recorded jump thread.
b5c4ff78 105
ce0770fc
ILT		 1062013-10-21 Ian Lance Taylor <iant@google.com>
107
108 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
109 don't imply that attributes can solve all problems.
110 (Directory Options): Fix typo.
111
0e7504c8
KT		 1122013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
113
114 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
115 extend and extend_add.
116
2445aeac
RB		 1172013-10-21 Richard Biener <rguenther@suse.de>
118
119 PR tree-optimization/58794
120 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
121 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
122
d480e6a5
RB		 1232013-10-21 Richard Biener <rguenther@suse.de>
124
125 PR middle-end/58742
126 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
127 to (T) X for sign-changing conversions (or no conversion).
128
54f9aba0
UB		 1292013-10-20 Uros Bizjak <ubizjak@gmail.com>
130
131 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
132
d6c6ba3c
JH		 1332013-10-20 Jan Hubicka <jh@suse.cz>
134
135 * config/i386/i386-tune.def: Add comment; organize into categories
136
d606b917
MZ		 1372013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
138
139 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
140 argument. Update function comment.
141 (expand_set_or_movmem_via_rep): New function combining
142 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
143 (expand_movmem_via_rep_mov): Remove.
144 expand_setmem_via_rep_stos): Remove.
145 (expand_movmem_epilogue): Update calls correspondingly.
146 (expand_setmem_epilogue_via_loop): Likewise.
147 (emit_memset): New.
148 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
149 (expand_set_or_movmem_prologue): New function combining
150 expand_movmem_prologue and expand_setmem_prologue.
151 (expand_movmem_prologue): Remove.
152 (expand_setmem_prologue): Remove.
153 (expand_set_or_movmem_constant_prologue): New function combining
154 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
155 (expand_constant_movmem_prologue): Remove.
156 (expand_constant_setmem_prologue): Remove.
157 (promote_duplicated_reg): Allow vector-const0 value.
158 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
159 and ix86_expand_setmem.
160 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
161 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
162
4d648807
DN		 1632013-10-21 Diego Novillo <dnovillo@google.com>
164
165 * asan.c: Include tree.h
166 * bb-reorder.c: Likewise.
167 * cfgcleanup.c: Likewise.
168 * cfgloopmanip.c: Likewise.
169 * data-streamer-in.c: Likewise.
170 * data-streamer-out.c: Likewise.
171 * data-streamer.c: Likewise.
172 * dwarf2cfi.c: Likewise.
173 * graphite-blocking.c: Likewise.
174 * graphite-clast-to-gimple.c: Likewise.
175 * graphite-dependences.c: Likewise.
176 * graphite-interchange.c: Likewise.
177 * graphite-optimize-isl.c: Likewise.
178 * graphite-poly.c: Likewise.
179 * graphite-scop-detection.c: Likewise.
180 * graphite-sese-to-poly.c: Likewise.
181 * graphite.c: Likewise.
182 * ipa-devirt.c: Likewise.
183 * ipa-profile.c: Likewise.
184 * ipa.c: Likewise.
185 * ira.c: Likewise.
186 * loop-init.c: Likewise.
187 * loop-unroll.c: Likewise.
188 * lower-subreg.c: Likewise.
189 * lto/lto-object.c: Likewise.
190 * recog.c: Likewise.
191 * reginfo.c: Likewise.
192 * tree-loop-distribution.c: Likewise.
193 * tree-parloops.c: Likewise.
194 * tree-ssa-strlen.c: Likewise.
195 * tree-streamer.c: Likewise.
196 * value-prof.c: Likewise.
197 * target-globals.c: Likewise.
198 * expr.h: Include tree-core.h instead of tree.h.
199 * gimple.h: Likewise.
200 * ipa-prop.h: Likewise.
201 * ipa-utils.h: Likewise.
202 * lto-streamer.h: Likewise.
203 * streamer-hooks.h: Likewise.
204 * ipa-reference.h: Include cgraph.h instead of tree.h.
205 * cgraph.h: Include basic-block.h instead of tree.h.
206 * tree-streamer.h: Do not include tree.h.
207 * genattrtab.c (write_header): Generate inclusion of
208 tree.h.
209 * genautomata.c (main): Likewise.
210 * genemit.c: Likewise.
211 * genopinit.c: Likewise.
212 * genoutput.c (output_prologue): Likewise.
213 * genpeep.c: Likewise.
214
dd346f1f
BS		 2152013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
216
217 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
218 little endian.
219 (vec_unpacku_hi_v8hi): Likewise.
220 (vec_unpacku_lo_v16qi): Likewise.
221 (vec_unpacku_lo_v8hi): Likewise.
222
915e8e6e
JH		 2232013-10-20 Jan Hubicka <jh@suse.cz>
224
225 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
226 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
227 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
228
0f37323c
RS		 2292013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
230
231 * config/mips/mips.h (ISA_HAS_WSBH): Define.
232 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
233 constants.
234 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
235
84935d25
JDA		 2362013-10-19 John David Anglin <danglin@gcc.gnu.org>
237
238 PR target/58603
239 * system.h: Undef m_slot.
240
d7d1451a
BS		 2412013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
242
243 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
244 all elements for both endian flavors.
245
6eb793dd
UB		 2462013-10-19 Uros Bizjak <ubizjak@gmail.com>
247
248 PR target/58792
249 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
250 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
251 and SI_REG for 64bit SYSV ABI targets.
252
c07757e5
UB		 2532013-10-19 Uros Bizjak <ubizjak@gmail.com>
254
255 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
256 to multi_reg_return. Clarify that we are skipping USEs of multiple
257 return registers. Use bool type where appropriate.
258
41ee845b
JH		 2592013-10-18 Jan Hubicka <jh@suse.cz>
260
261 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
262 for cold functions.
263 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
264 (X86_TUNE_PUSH_MEMORY): Likewise.
265 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
266 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
c07757e5
UB		 267 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
268 New.
41ee845b
JH		 269 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
270 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
271 Remove.
272 (ix86_option_override_internal): Update to use tune features instead
273 of variables.
274
6f978a2a
CH		 2752013-10-18 Cong Hou <congh@google.com>
276
277 PR tree-optimization/58508
278 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
279 statement that contains data refs with zero-step.
280
4484a35a
AM		 2812013-10-18 Andrew MacLeod <amacleod@redhat.com>
282
c07757e5
UB		 283 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
284 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
4484a35a
AM		 285 * gimple-low.c (gimple_check_call_arg,
286 gimple_check_call_matching_types): Move to cgraph.c.
287 * gimple-low.h: Remove prototype.
288 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
289 Relocate from gimple-low.c.
290 * cgraph.h: Add prototype. Don't include basic-block.h.
291 * gimplify.c: Add gimple-low to include list.
292 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
293 * tree-eh.c: Add gimple-low to include list.
294 * tree-nested.c: Likewise.
295 * cfgexpand.c: Add tree-ssa-address.h to include list.
296 * expr.c: Likewise.
297 * gimple-fold.c: Likewise.
298 * gimple-ssa-strength-reduction.c: Likewise.
299 * trans-mem.c: Likewise.
300 * tree-mudflap.c: Likewise.
301 * tree-ssa-loop-ivopts.c: Likewise.
302 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
303 (degenerate_phi_result): Move to tree-phinodes.c.
304 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
305 * tree-ssa-threadedge.c: Likewise.
306 * tree-vrp.c: Likewise.
307 * tree-phinodes.c (degenerate_phi_result): Relocate here.
308 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
309 * tree-phinodes.h (degenerate_phi_result): Add prototype.
310 * tree-ssa-copy.c: Include tree-ssa-dom.h.
311 * tree-ssa-forwprop.c: Likewise.
312 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
313 pass_data_cleanup_cfg_post_optimizing,
314 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
315 * tree-optimize.c: Delete File.
316 * graphite.c: Include tree-cfgcleanup.h.
317 * passes.c: Likewise.
318 * tree-cfg.c: Likewise.
319 * tree-profile.c: Likewise.
320 * tree-ssa-dse.c: Likewise.
321 * tree-ssa-loop-ivcanon.c: Likewise.
322 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
323 * tree-outof-ssa.c: Include sbitmap.h.
324 * tree-ssa-live.c: Likewise.
325 * tree-ssa-propagate.c: Likewise.
326 * tree-ssa-structalias.c: Likewise.
327 * tree-stdarg.c: Likewise.
328 * Makefile.in (OBJS): Delete tree-optimize.o.
329 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
330 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
331 * varasm.c: Include basic-block.h.
332 * cfgloop.h: Include function.h instead of basic-block.h
333 (bb_loop_depth): Move to cfgloop.c.
334 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
335
bfaa17c1
TJ		 3362013-10-18 Teresa Johnson <tejohnson@google.com>
337
338 * predict.c (probably_never_executed): Compare frequency-based
339 count to number of training runs.
340 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
341
f58883bd
KT		 3422013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
343
344 * config/arm/arm.c (cortexa9_extra_costs): New table.
345 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
346
01e127b1
JL		 3472013-10-18 Jeff Law <law@redhat.com>
348
349 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
350
351 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
c07757e5
UB		 352 (register_jump_thread): Add "registered_jump_thread" debug
353 counter support.
01e127b1
JL		 354 * dbgcnt.def (registered_jump_thread): New debug counter.
355
f41b8848
AM		 3562013-10-18 Andrew MacLeod <amacleod@redhat.com>
357
358 * config/rs6000/rs6000.c: Include cgraph.h.
359
adc7a812
TJ		 3602013-10-18 Teresa Johnson <tejohnson@google.com>
361
362 * tree-ssa-tail-merge.c (replace_block_by): Update edge
363 weights during merging.
364
377d9792
AM		 3652013-10-18 Andrew MacLeod <amacleod@redhat.com>
366
367 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
368 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
369 * tree-ssa-operands.h: Remove prototype.
370 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
371 * gimple.c (virtual_operand_p): Relocate from gimple.c.
372 * gimple.h: Add prototype.
373 * gimple-ssa.h: Include tree-ssa-operands.h.
374 * tree-dump.c: Add tree-cfg.h to include list.
375 * tree-ssa-alias.c: Add ipa-reference.h to include list.
376 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
377 * config/i386/i386.c: Don't include tree-flow.h.
378 * config/rs6000/rs6000.c: Likewise.
379
8ad8afaf
JBG		 3802013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
381
382 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
383
c0e081a9
RB		 3842013-10-18 Richard Biener <rguenther@suse.de>
385
386 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
387 or TYPE_UNSIGNED of integral types.
388 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
389 NULL_TREE for zero-precision integral types.
390
054fceb5
JG		 3912013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
392
393 * config/aarch64/arm_neon.h
394 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
395
70e8b89b
ST		 3962013-10-17 Sriraman Tallam <tmsriram@google.com>
397
398 * opth-gen.awk: Define target_flags_explicit.
399
3170766c 4002013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5845f602
MM		 401
402 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
403 fields to the reg_addr array that describes the valid addressing
404 mode for any register, general purpose registers, floating point
405 registers, and Altivec registers.
406 (FIRST_RELOAD_REG_CLASS): Likewise.
407 (LAST_RELOAD_REG_CLASS): Likewise.
408 (struct reload_reg_map_type): Likewise.
409 (reload_reg_map_type): Likewise.
410 (RELOAD_REG_VALID): Likewise.
411 (RELOAD_REG_MULTIPLE): Likewise.
412 (RELOAD_REG_INDEXED): Likewise.
413 (RELOAD_REG_OFFSET): Likewise.
414 (RELOAD_REG_PRE_INCDEC): Likewise.
415 (RELOAD_REG_PRE_MODIFY): Likewise.
416 (reg_addr): Likewise.
417 (mode_supports_pre_incdec_p): New helper functions to say whether
418 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
419 (mode_supports_pre_modify_p): Likewise.
420 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
421 print the valid address mode bits for each mode.
422 (rs6000_debug_print_mode): Likewise.
423 (rs6000_debug_reg_global): Likewise.
424 (rs6000_setup_reg_addr_masks): New function to set up the address
425 mask bits for each type.
426 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
427 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
428 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
429 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
430 PRE_MODIFY are supported.
3170766c
MM		 431 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
432 registers, instead of {src,dest}_av_p.
5845f602
MM		 433 (rs6000_print_options_internal): Tweak the debug output slightly.
434
82e33890
UB		 4352013-10-17 Uros Bizjak <ubizjak@gmail.com>
436
437 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
438 isa attribute.
439
0645c1a2
AM		 4402013-10-17 Andrew MacLeod <amacleod@redhat.com>
441
442 * tree-flow.h (struct omp_region): Move to omp-low.c.
443 Remove omp_ prototypes and variables.
444 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
445 (copy_var_decl): Relocate prototype from tree-flow.h.
446 * gimple.c (copy_var_decl): Relocate from omp-low.c.
447 * tree.h: Move prototype to omp-low.h.
448 * omp-low.h: New File. Relocate prototypes here.
449 * omp-low.c (struct omp_region): Make local here.
450 (root_omp_region): Make static.
451 (copy_var_decl) Move to gimple.c.
452 (new_omp_region): Make static.
453 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
454 * tree-cfg.c: Include omp-low.h.
455 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
456 * gimplify.c: Include omp-low.h.
457 * tree-parloops.c: Likewise.
458
6f1abb55
UB		 4592013-10-17 Uros Bizjak <ubizjak@gmail.com>
460
461 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
462 operands are in memory, prefer to force non-matched operand 1 to
463 the register.
464
1ddbbbc0
MM		 4652013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
466
467 PR target/58673
468 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
469 restrict TImode addresses to single indirect registers if both
470 -mquad-memory and -mvsx-timode are used.
471 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
472 we should emit load/store quad. Remove using %y for quad memory
473 addresses.
474
475 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
476 constraints to allow load/store quad on machines where TImode is
477 not allowed in VSX registers. Use 'n' instead of 'F' constraint
478 for TImode to load integer constants.
479
f541a481
KT		 4802013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
481
482 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
483
51bb310d
MS		 4842013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
485
486 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
487 handling of STACK_REG.
488
f4bed77b
RB		 4892013-10-17 Richard Biener <rguenther@suse.de>
490
491 PR tree-optimization/58143
492 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
493 New function.
494 (rewrite_to_defined_overflow): Likewise.
495 (move_computations_dom_walker::before_dom): Rewrite stmts
496 with undefined signed overflow that are not always executed
497 into unsigned arithmetic.
498
d358fb9d
MM		 4992013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
500
501 PR target/57756
502 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
503 explicit isa flag to be an options variable, instead of using
504 global_options_set. Remove define from rs6000.h.
505 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
506
507 * config/rs6000/rs6000.c (rs6000_option_override_internal):
508 Initialize rs6000_isa_flags_explicit.
509 (rs6000_function_specific_save): Add gcc_options* parameter, so
510 that the powerpc builds after the 2013-10-15 changes.
511 (rs6000_function_specific_restore): Likewise.
512
c71aed47
DD		 5132013-10-16 DJ Delorie <dj@redhat.com>
514
515 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
516 op is a REG before checking REGNO.
517 (rl78_alloc_physical_registers): Verify pattern is a SET before
518 checking SET_SRC.
519
90d3bd51
BS		 5202013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
521
522 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
523 endianness.
524 (vec_unpacks_lo_v4sf): Likewise.
525 (vec_unpacks_float_hi_v4si): Likewise.
526 (vec_unpacks_float_lo_v4si): Likewise.
527 (vec_unpacku_float_hi_v4si): Likewise.
528 (vec_unpacku_float_lo_v4si): Likewise.
529
de75c876
BS		 5302013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
531
532 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
533 (vsx_concat_v2sf): Likewise.
534
21216998
JG		 5352013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
536
537 * config/aarch64/aarch64.md
538 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
539
1ef88893 5402013-10-16 Andrew MacLeod <amacleod@redhat.com>
7ab4168e 541
1ef88893 542 PR tree-optimization/58697
7ab4168e 543 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1ef88893
AM		 544 estimated_loop_iterations_int.
545 (max_stmt_executions_int): Call get_max_loop_iterations_int.
7ab4168e 546 (get_max_loop_iterations_int): New. HWINT version of
1ef88893
AM		 547 get_max_loop_iterations.
548 * cfgloop.h: Add prototypes.
549 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
7ab4168e 550 * loop-unroll.c (decide_peel_once_rolling): Call
1ef88893
AM		 551 get_estimated_loop_iterations_int.
552 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
553 * tree-ssa-loop-niter.h: Tweak prototypes.
554
c22df64f
DM		 5552013-10-16 David Malcolm <dmalcolm@redhat.com>
556
557 * gengtype-parse.c (struct_field_seq): Ignore access-control
558 keywords ("public:" etc).
559
f24bb080
MS		 5602013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
561
562 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
563 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
564
c281bf26
YR		 5652013-10-16 Yvan Roux <yvan.roux@linaro.org>
566
567 * config/arm/arm.opt (mlra): New option.
568 * config/arm/arm.c (arm_lra_p): New function.
569 (TARGET_LRA_P): Define.
570
5806f481
PM		 5712013-10-16 Paulo Matos <pmatos@broadcom.com>
572
573 * tree-core.h (tree_code_name): Remove.
574 * tree.h (get_tree_code_name): New prototype.
575 * tree.c (tree_code_name): Make static.
576 (get_tree_code_name): New function.
577 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
578 tree_class_check_failed, tree_range_check_failed,
579 tree_not_class_check_failed, omp_clause_check_failed,
580 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
581 wrapper get_tree_code_name instead of calling tree_code_name directly.
582 * tree-vrp.c (dump_asserts_for): Likewise.
583 * tree-dump.c (dequeue_and_dump): Likewise.
584 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
585 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
586 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
587 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
588 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
589 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
590 dump_gimple_omp_for): Likewise.
591 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
592 * tree-ssa-pre.c (print_pre_expr): Likewise.
593 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
594 * print-tree.c (print_node_brief, print_node): Likewise.
595 * gimple.c (gimple_check_failed): Likewise.
596 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
597 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
598 * config/mep/mep.c (mep_validate_vliw): Likewise.
599 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
600 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
601
82c9df56
GG		 6022013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
603
604 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
605 for AMD bdver3.
606
b82d0df9
HPN		 6072013-10-16 Hans-Peter Nilsson <hp@axis.com>
608
609 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
610 (MULTILIB_MATCHES): Add multilib for -march=v8.
611
bf7b5747
ST		 6122013-10-15 Sriraman Tallam <tmsriram@google.com>
613
614 PR target/57756
615 * optc-save-gen.awk: Add extra parameter to the save and restore
616 target calls.
617 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
618 * tree.c (build_optimization_node): New parameter. Add extra parameter
619 to call to cl_optimization_save.
620 (build_target_option_node): New parameter. Add extra parameter
621 to call to cl_target_option_save.
622 * tree.h (build_optimization_node): New parameter.
623 (build_target_option_node): New parameter.
624 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
625 build_optimization_node and build_target_option_node.
626 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
627 (handle_pragma_push_options): Ditto.
628 * toplev.c (process_options): Ditto.
629 * opts.c (init_options_struct): Check for opts_set non-null.
630 * target.def (target_option.save): New parameter.
631 (target_option.restore): New parameter.
632 * tm.texi: Generate.
633 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
634 (ix86_pragma_target_parse): Ditto.
635 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
636 parameters.
637 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
638 to build_optimization_node and build_target_option_node.
639 (rs6000_valid_attribute_p): Ditto.
640 (rs6000_pragma_target_parse): Ditto.
641 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
642 data.
643 * config/i386/i386.h:
644 TARGET_64BIT_P: New Macro
645 TARGET_MMX_P: New Macro.
646 TARGET_3DNOW_P: New Macro.
647 TARGET_3DNOW_A_P: New Macro.
648 TARGET_SSE_P: New Macro.
649 TARGET_SSE2_P: New Macro.
650 TARGET_SSE3_P: New Macro.
651 TARGET_SSSE3_P: New Macro.
652 TARGET_SSE4_1_P: New Macro.
653 TARGET_SSE4_2_P: New Macro.
654 TARGET_AVX_P: New Macro.
655 TARGET_AVX2_P: New Macro.
656 TARGET_AVX512F_P: New Macro.
657 TARGET_AVX512PF_P: New Macro.
658 TARGET_AVX512ER_P: New Macro.
659 TARGET_AVX512CD_P: New Macro.
660 TARGET_FMA_P: New Macro.
661 TARGET_SSE4A_P: New Macro.
662 TARGET_FMA4_P: New Macro.
663 TARGET_XOP_P: New Macro.
664 TARGET_LWP_P: New Macro.
665 TARGET_ABM_P: New Macro.
666 TARGET_BMI_P: New Macro.
667 TARGET_BMI2_P: New Macro.
668 TARGET_LZCNT_P: New Macro.
669 TARGET_TBM_P: New Macro.
670 TARGET_POPCNT_P: New Macro.
671 TARGET_SAHF_P: New Macro.
672 TARGET_MOVBE_P: New Macro.
673 TARGET_CRC32_P: New Macro.
674 TARGET_AES_P: New Macro.
675 TARGET_PCLMUL_P: New Macro.
676 TARGET_CMPXCHG16B_P: New Macro.
677 TARGET_FSGSBASE_P: New Macro.
678 TARGET_RDRND_P: New Macro.
679 TARGET_F16C_P: New Macro.
680 TARGET_RTM_P: New Macro.
681 TARGET_HLE_P: New Macro.
682 TARGET_RDSEED_P: New Macro.
683 TARGET_PRFCHW_P: New Macro.
684 TARGET_ADX_P: New Macro.
685 TARGET_FXSR_P: New Macro.
686 TARGET_XSAVE_P: New Macro.
687 TARGET_XSAVEOPT_P: New Macro.
688 TARGET_LP64_P: New Macro.
689 TARGET_X32_P: New Macro.
690 TARGET_FPMATH_DEFAULT_P: New Macro.
691 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
692 * config/i386/i386.c (ix86_option_override_internal): New parameters.
693 opts and opts_set.
694 Change ix86_tune_string to access opts->x_ix86_tune_string.
695 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
696 Change ix86_arch_string to access opts->x_ix86_arch_string.
697 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
698 Change ix86_pmode to access opts->x_ix86_pmode.
699 Change ix86_abi to access opts->x_ix86_abi.
700 Change ix86_cmodel to access opts->x_ix86_cmodel.
701 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
702 Change ix86_isa_flags_explicit to access
703 opts->x_ix86_isa_flags_explicit.
704 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
705 Change ix86_regparm to access opts->x_ix86_regparm.
706 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
707 Change ix86_preferred_stack_boundary_arg to access
708 opts->x_ix86_preferred_stack_boundary_arg.
709 Change ix86_force_align_arg_pointer to access
710 opts->x_ix86_force_align_arg_pointer.
711 Change ix86_incoming_stack_boundar_arg to access
712 opts->x_ix86_incoming_stack_boundar_arg.
713 Change ix86_fpmath to access opts->x_ix86_fpmath.
714 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
715 Change ix86_recip_name to access opts->x_ix86_recip_name.
716 Change ix86_stack_protector_guard to access
717 opts->x_ix86_stack_protector_guard.
718 Change ix86_tune_memcpy_strategy to access
719 opts->x_ix86_tune_memcpy_strategy.
720 Change ix86_tune_memset_strategy to access
721 opts->x_ix86_tune_memset_strategy.
722 Change global_options to access opts.
723 Change global_options_set to access opts_set.
6f1abb55
UB		 724 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
725 Change TARGET_MMX to TARGET_MMX_P (opts->...).
726 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
727 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
728 Change TARGET_SSE to TARGET_SSE_P (opts->...).
729 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
730 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
731 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
732 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
733 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
734 Change TARGET_AVX to TARGET_AVX_P (opts->...).
735 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
736 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
737 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
738 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
739 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
740 Change TARGET_FMA to TARGET_FMA_P (opts->...).
741 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
742 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
743 Change TARGET_XOP to TARGET_XOP_P (opts->...).
744 Change TARGET_LWP to TARGET_LWP_P (opts->...).
745 Change TARGET_ABM to TARGET_ABM_P (opts->...).
746 Change TARGET_BMI to TARGET_BMI_P (opts->...).
747 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
748 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
749 Change TARGET_TBM to TARGET_TBM_P (opts->...).
750 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
751 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
752 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
753 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
754 Change TARGET_AES to TARGET_AES_P (opts->...).
755 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
756 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
757 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
758 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
759 Change TARGET_F16C to TARGET_F16C_P (opts->...).
760 Change TARGET_RTM to TARGET_RTM_P (opts->...).
761 Change TARGET_HLE to TARGET_HLE_P (opts->...).
762 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
763 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
764 Change TARGET_ADX to TARGET_ADX_P (opts->...).
765 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
766 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
767 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
768 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
769 Change TARGET_X32 to TARGET_X32_P (opts->...).
770 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
771 Change TARGET_FLOAT_RETURNS_IN_80387 to
772 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
bf7b5747
ST		 773 (ix86_function_specific_save): New parameter. Use opts-> fields
774 to replace global fields.
775 (ix86_function_specific_restore): Ditto.
776 (ix86_valid_target_attribute_inner_p): New parameters.
777 Fix recursive call.
778 Fix call to ix86_handle_option and set_option.
779 (ix86_valid_target_attribute_tree): New parameters.
780 Change global_options to access opts.
781 Change global_options_set to access opts_set.
782 Fix call to ix86_valid_target_attribute_inner_p.
783 Change ix86_tune_string to access opts->x_ix86_tune_string.
784 Change ix86_arch_string to access opts->x_ix86_arch_string.
785 Change ix86_fpmath to access opts->x_ix86_fpmath
786 Fix call to ix86_option_override_internal.
787 Fix call to ix86_add_new_builtins.
788 Fix calls to build_optimization_node and build_target_option_node.
789 (ix86_valid_target_attribute_p): Remove access to global_options.
790 Use new gcc_options structure func_options.
791 Fix call to ix86_valid_target_attribute_tree.
792 Fix call to build_optimization_node.
793 (get_builtin_code_for_version): Fix call to
794 ix86_valid_target_attribute_tree.
795
459260ec
DM		 7962013-10-15 David Malcolm <dmalcolm@redhat.com>
797
798 * Makefile.in (PICFLAG): New.
799 (enable_host_shared): New.
800 (INTERNAL_CFLAGS): Use PICFLAG.
801 (LIBIBERTY): Use pic build of libiberty.a if configured with
802 --enable-host-shared.
803 * configure.ac: Add --enable-host-shared, setting up new
804 PICFLAG variable.
805 * configure: Regenerate.
6f1abb55 806 * doc/install.texi (--enable-shared): Add note contrasting it with ...
459260ec
DM		 807 (--enable-host-shared): New option.
808
16022041
RS		 8092013-10-15 Richard Biener <rguenther@suse.de>
810
811 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
812 for built-in functions.
813
b114bfb4
ZC		 8142013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
815
816 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
817 (optimize_range_tests_1): New function,
818 extracted from optimize_range_tests.
819 (optimize_range_tests_xor): Similarly.
820 (optimize_range_tests_diff): New function.
821 (optimize_range_tests): Use optimize_range_tests_1.
822
69b8f2f9
CH		 8232013-10-15 Cong Hou <congh@google.com>
824
7ab4168e
MS		 825 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
826 requirement of the reduction pattern so that one operand of the
69b8f2f9
CH		 827 reduction operation can come from outside of the loop.
828
8ff5a4e4
JG		 8292013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
830
831 * config/arm/neon-schedgen.ml: Remove.
832 * config/arm/cortex-a9-neon.md: Remove comment regarding
833 neon-schedgen.ml.
834
52b69ad2
JG		 8352013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
836
837 * config/arm/types: Remove old neon types.
838
e7c6a556
JG		 8392013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
840
841 * config/arm/cortex-a7.md
842 (cortex_a7_neon_type): New.
843 (cortex_a7_neon_mul): Update for new types.
844 (cortex_a7_neon_mla): Likewise.
845 (cortex_a7_neon): Likewise.
846
795dd8d5
JG		 8472013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
848
849 * config/arm/cortex-a15-neon.md
850 (cortex_a15_neon_type): New,
851
852 (cortex_a15_neon_int_1): Remove.
853 (cortex_a15_neon_int_2): Likewise.
854 (cortex_a15_neon_int_3): Likewise.
855 (cortex_a15_neon_int_4): Likewise.
856 (cortex_a15_neon_int_5): Likewise.
857 (cortex_a15_neon_vqneg_vqabs): Likewise.
858 (cortex_a15_neon_vmov): Likewise.
859 (cortex_a15_neon_vaba): Likewise.
860 (cortex_a15_neon_vaba_qqq): Likewise.
861 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
862 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
863 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
864 Likewise.
865 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
866 (cortex_a15_neon_mla_qqq_8_16): Likewise.
867 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
868 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
869 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
870 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
871 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
872 (cortex_a15_neon_shift_1): Likewise.
873 (cortex_a15_neon_shift_2): Likewise.
874 (cortex_a15_neon_shift_3): Likewise.
875 (cortex_a15_neon_vshl_ddd): Likewise.
876 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
877 (cortex_a15_neon_vsra_vrsra): Likewise.
878 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
879 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
880 (cortex_a15_neon_bp_3cycle): Likewise.
881 (cortex_a15_neon_ldm_2): Likewise.
882 (cortex_a15_neon_stm_2): Likewise.
883 (cortex_a15_neon_mcr): Likewise.
884 (cortex_a15_neon_mrc): Likewise.
885 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
886 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
887 (cortex_a15_neon_fp_vmul_ddd): Likewise.
888 (cortex_a15_neon_fp_vmul_qqd): Likewise.
889 (cortex_a15_neon_fp_vmla_ddd): Likewise.
890 (cortex_a15_neon_fp_vmla_qqq): Likewise.
891 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
892 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
893 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
894 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
895 (cortex_a15_neon_bp_simple): Likewise.
896 (cortex_a15_neon_bp_2cycle): Likewise.
897 (cortex_a15_neon_bp_3cycle): Likewise.
898 (cortex_a15_neon_vld1_1_2_regs): Likewise.
899 (cortex_a15_neon_vld1_3_4_regs): Likewise.
900 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
901 (cortex_a15_neon_vld2_4_regs): Likewise.
902 (cortex_a15_neon_vld3_vld4): Likewise.
903 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
904 (cortex_a15_neon_vst1_3_4_regs): Likewise.
905 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
906 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
907 (cortex_a15_neon_vst3_vst4): Rename to...
908 (cortex_a15_neon_vst4): This, update for new attributes.
909 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
910 (cortex_a15_neon_vld3_vld4_lane): Likewise.
911 (cortex_a15_neon_vst1_vst2_lane): Likewise.
912 (cortex_a15_neon_vst3_vst4_lane): Likewise.
913 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
914 (cortex_a15_neon_ldm_2): Likewise.
915 (cortex_a15_neon_stm_2): Likewise.
916 (cortex_a15_neon_mcr): Likewise.
917 (cortex_a15_neon_mcr_2_mcrr): Likewise.
918 (cortex_a15_neon_mrc): Likewise.
919 (cortex_a15_neon_mrrc): Likewise.
920
921 (cortex_a15_neon_abd): New.
922 (cortex_a15_neon_abd_q): Likewise.
923 (cortex_a15_neon_aba): Likewise.
924 (cortex_a15_neon_aba_q): Likewise.
925 (cortex_a15_neon_acc): Likewise.
926 (cortex_a15_neon_acc_q): Likewise.
927 (cortex_a15_neon_arith_basic): Likewise.
928 (cortex_a15_neon_arith_complex): Likewise.
929 (cortex_a15_neon_multiply): Likewise.
930 (cortex_a15_neon_multiply_q): Likewise.
931 (cortex_a15_neon_mla): Likewise.
932 (cortex_a15_neon_mla_q): Likewise.
933 (cortex_a15_neon_sat_mla_long): Likewise.
934 (cortex_a15_neon_shift_acc): Likewise.
935 (cortex_a15_neon_shift_imm_basic): Likewise.
936 (cortex_a15_neon_shift_imm_complex): Likewise.
937 (cortex_a15_neon_shift_reg_basic): Likewise.
938 (cortex_a15_neon_shift_reg_basic_q): Likewise.
939 (cortex_a15_neon_shift_reg_complex): Likewise.
940 (cortex_a15_neon_shift_reg_complex_q): Likewise.
941 (cortex_a15_neon_fp_negabs): Likewise
942 (cortex_a15_neon_fp_arith): Likewise
943 (cortex_a15_neon_fp_arith_q): Likewise
944 (cortex_a15_neon_fp_cvt_int): Likewise
945 (cortex_a15_neon_fp_cvt_int_q): Likewise
946 (cortex_a15_neon_fp_cvt_16): Likewise
947 (cortex_a15_neon_fp_mul): Likewise
948 (cortex_a15_neon_fp_mul_q): Likewise
949 (cortex_a15_neon_fp_mla): Likewise
950 (cortex_a15_neon_fp_mla_q): Likewise
951 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
952 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
953 (cortex_a15_neon_bitops): Likewise.
954 (cortex_a15_neon_bitops_q): Likewise.
955 (cortex_a15_neon_from_gp): Likewise.
956 (cortex_a15_neon_from_gp_q): Likewise.
957 (cortex_a15_neon_tbl3_tbl4): Likewise.
958 (cortex_a15_neon_zip_q): Likewise.
959 (cortex_a15_neon_to_gp): Likewise.
960 (cortex_a15_neon_load_a): Likewise.
961 (cortex_a15_neon_load_b): Likewise.
962 (cortex_a15_neon_load_c): Likewise.
963 (cortex_a15_neon_load_d): Likewise.
964 (cortex_a15_neon_load_e): Likewise.
965 (cortex_a15_neon_load_f): Likewise.
966 (cortex_a15_neon_store_a): Likewise.
967 (cortex_a15_neon_store_b): Likewise.
968 (cortex_a15_neon_store_c): Likewise.
969 (cortex_a15_neon_store_d): Likewise.
970 (cortex_a15_neon_store_e): Likewise.
971 (cortex_a15_neon_store_f): Likewise.
972 (cortex_a15_neon_store_g): Likewise.
973 (cortex_a15_neon_store_h): Likewise.
974 (cortex_a15_vfp_to_from_gp): Likewise.
975
bc5e57e3
JG		 9762013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
977
978 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
979
980 (cortex_a9_neon_vshl_ddd): Remove.
981 (cortex_a9_neon_vst3_vst4): Likewise.
982 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
983
984 (cortex_a9_neon_bit_ops_q): New.
985
986 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
987 (cortex_a9_neon_int_2): Likewise.
988 (cortex_a9_neon_int_3): Likewise.
989 (cortex_a9_neon_int_4): Likewise.
990 (cortex_a9_neon_int_5): Likewise.
991 (cortex_a9_neon_vqneg_vqabs): Likewise.
992 (cortex_a9_neon_vmov): Likewise.
993 (cortex_a9_neon_vaba): Likewise.
994 (cortex_a9_neon_vaba_qqq): Likewise.
995 (cortex_a9_neon_shift_1): Likewise.
996 (cortex_a9_neon_shift_2): Likewise.
997 (cortex_a9_neon_shift_3): Likewise.
998 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
999 (cortex_a9_neon_vsra_vrsra): Likewise.
1000 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1001 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1002 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1003 Likewise.
1004 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1005 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1006 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1007 Likewise.
1008 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1009 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1010 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1011 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1012 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1013 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1014 (cortex_a9_neon_fp_vsum): Likewise.
1015 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1016 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1017 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1018 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1019 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1020 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1021 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1022 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1023 (cortex_a9_neon_bp_simple): Likewise.
1024 (cortex_a9_neon_bp_2cycle): Likewise.
1025 (cortex_a9_neon_bp_3cycle): Likewise.
1026 (cortex_a9_neon_ldr): Likewise.
1027 (cortex_a9_neon_str): Likewise.
1028 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1029 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1030 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1031 (cortex_a9_neon_vld2_4_regs): Likewise.
1032 (cortex_a9_neon_vld3_vld4): Likewise.
1033 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1034 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1035 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1036 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1037 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1038 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1039 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1040 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1041 (cortex_a9_neon_mcr): Likewise.
1042 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1043 (cortex_a9_neon_mrc): Likewise.
1044 (cortex_a9_neon_mrrc): Likewise.
1045
9deda8cb
JG		 10462013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1047
1048 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1049
1050 (cortex_a8_neon_vshl_ddd): Remove.
1051 (cortex_a8_neon_vst3_vst4): Likewise.
1052 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1053
1054 (cortex_a8_neon_bit_ops_q): New.
1055
1056 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1057 (cortex_a8_neon_int_2): Likewise..
1058 (cortex_a8_neon_int_3): Likewise.
1059 (cortex_a8_neon_int_5): Likewise.
1060 (cortex_a8_neon_vqneg_vqabs): Likewise.
1061 (cortex_a8_neon_int_4): Likewise.
1062 (cortex_a8_neon_vaba): Likewise.
1063 (cortex_a8_neon_vaba_qqq): Likewise.
1064 (cortex_a8_neon_shift_1): Likewise.
1065 (cortex_a8_neon_shift_2): Likewise.
1066 (cortex_a8_neon_shift_3): Likewise.
1067 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1068 (cortex_a8_neon_vsra_vrsra): Likewise.
1069 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1070 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1071 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1072 Likewise.
1073 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1074 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1075 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1076 Likewise.
1077 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1078 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1079 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1080 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1081 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1082 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1083 (cortex_a8_neon_fp_vsum): Likewise.
1084 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1085 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1086 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1087 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1088 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1089 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1090 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1091 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1092 (cortex_a8_neon_bp_simple): Likewise.
1093 (cortex_a8_neon_bp_2cycle): Likewise.
1094 (cortex_a8_neon_bp_3cycle): Likewise.
1095 (cortex_a8_neon_ldr): Likewise.
1096 (cortex_a8_neon_str): Likewise.
1097 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1098 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1099 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1100 (cortex_a8_neon_vld2_4_regs): Likewise.
1101 (cortex_a8_neon_vld3_vld4): Likewise.
1102 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1103 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1104 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1105 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1106 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1107 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1108 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1109 (cortex_a8_neon_mcr): Likewise.
1110 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1111 (cortex_a8_neon_mrc): Likewise.
1112 (cortex_a8_neon_mrrc): Likewise.
1113
0f686aa9
JG		 11142013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1115
1116 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1117 (fp): New.
1118 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1119 (aarch64_simd_dup<mode>): Add "type" attribute.
1120 (aarch64_dup_lane<mode>): Likewise.
1121 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1122 (*aarch64_simd_mov<mode>): Likewise.
1123 (aarch64_simd_mov_from_<mode>low): Likewise.
1124 (aarch64_simd_mov_from_<mode>high): Likewise.
1125 (orn<mode>3): Likewise.
1126 (bic<mode>3): Likewise.
1127 (add<mode>3): Likewise.
1128 (sub<mode>3): Likewise.
1129 (mul<mode>3): Likewise.
1130 (*aarch64_mul3_elt<mode>): Likewise.
1131 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1132 (*aarch64_mul3_elt_to_128df): Likewise.
1133 (*aarch64_mul3_elt_to_64v2df): Likewise.
1134 (neg<mode>2): Likewise.
1135 (abs<mode>2): Likewise.
1136 (abd<mode>_3): Likewise.
1137 (aba<mode>_3): Likewise.
1138 (fabd<mode>_3): Likewise.
1139 (*fabd_scalar<mode>3): Likewise.
1140 (and<mode>3): Likewise.
1141 (ior<mode>3): Likewise.
1142 (xor<mode>3): Likewise.
1143 (one_cmpl<mode>2): Likewise.
1144 (aarch64_simd_vec_set<mode>): Likewise.
1145 (aarch64_simd_lshr<mode>): Likewise.
1146 (aarch64_simd_ashr<mode>): Likewise.
1147 (aarch64_simd_imm_shl<mode>): Likewise.
1148 (aarch64_simd_reg_sshl<mode): Likewise.
1149 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1150 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1151 (aarch64_simd_vec_setv2di): Likewise.
1152 (aarch64_simd_vec_set<mode>): Likewise.
1153 (aarch64_mla<mode>): Likewise.
1154 (*aarch64_mla_elt<mode>): Likewise.
1155 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1156 (aarch64_mls<mode>): Likewise.
1157 (*aarch64_mls_elt<mode>): Likewise.
1158 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1159 (<su><maxmin><mode>3): Likewise.
1160 (move_lo_quad_<mode>): Likewise.
1161 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1162 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1163 (vec_pack_trunc_<mode>): Likewise.
1164 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1165 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1166 (*aarch64_<su>mlal_lo<mode>): Likewise.
1167 (*aarch64_<su>mlal_hi<mode>): Likewise.
1168 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1169 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1170 (*aarch64_<su>mlal<mode>): Likewise.
1171 (*aarch64_<su>mlsl<mode>): Likewise.
1172 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1173 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1174 (add<mode>3): Likewise.
1175 (sub<mode>3): Likewise.
1176 (mul<mode>3): Likewise.
1177 (div<mode>3): Likewise.
1178 (neg<mode>2): Likewise.
1179 (abs<mode>2): Likewise.
1180 (fma<mode>4): Likewise.
1181 (*aarch64_fma4_elt<mode>): Likewise.
1182 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1183 (*aarch64_fma4_elt_to_128df): Likewise.
1184 (*aarch64_fma4_elt_to_64v2df): Likewise.
1185 (fnma<mode>4): Likewise.
1186 (*aarch64_fnma4_elt<mode>): Likewise.
1187 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1188 (*aarch64_fnma4_elt_to_128df): Likewise.
1189 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1190 (<frint_pattern><mode>2): Likewise.
1191 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1192 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1193 (vec_unpacks_lo_v4sf): Likewise.
1194 (aarch64_float_extend_lo_v2df): Likewise.
1195 (vec_unpacks_hi_v4sf): Likewise.
1196 (aarch64_float_truncate_lo_v2sf): Likewise.
1197 (aarch64_float_truncate_hi_v4sf): Likewise.
1198 (aarch64_vmls<mode>): Likewise.
1199 (<su><maxmin><mode>3): Likewise.
1200 (<maxmin_uns><mode>3): Likewise.
1201 (reduc_<sur>plus_<mode>): Likewise.
1202 (reduc_<sur>plus_v2di): Likewise.
1203 (reduc_<sur>plus_v2si): Likewise.
1204 (reduc_<sur>plus_<mode>): Likewise.
1205 (aarch64_addpv4sf): Likewise.
1206 (clz<mode>2): Likewise.
1207 (reduc_<maxmin_uns>_<mode>): Likewise.
1208 (reduc_<maxmin_uns>_v2di): Likewise.
1209 (reduc_<maxmin_uns>_v2si): Likewise.
1210 (reduc_<maxmin_uns>_<mode>): Likewise.
1211 (reduc_<maxmin_uns>_v4sf): Likewise.
1212 (aarch64_simd_bsl<mode>_internal): Likewise.
1213 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1214 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1215 (aarch64_get_lane<mode>): Likewise.
1216 (*aarch64_combinez<mode>): Likewise.
1217 (aarch64_combine<mode>): Likewise.
1218 (aarch64_simd_combine<mode>): Likewise.
1219 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1220 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1221 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1222 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1223 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1224 (aarch64_<sur>h<addsub><mode>): Likewise.
1225 (aarch64_<sur><addsub>hn<mode>): Likewise.
1226 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1227 (aarch64_pmul<mode>): Likewise.
1228 (aarch64_<su_optab><optab><mode>): Likewise.
1229 (aarch64_<sur>qadd<mode>): Likewise.
1230 (aarch64_sqmovun<mode>): Likewise.
1231 (aarch64_<sur>qmovn<mode>): Likewise.
1232 (aarch64_s<optab><mode>): Likewise.
1233 (aarch64_sq<r>dmulh<mode>): Likewise.
1234 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1235 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1236 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1237 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1238 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1239 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1240 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1241 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1242 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1243 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1244 (aarch64_sqdmull<mode>): Likewise.
1245 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1246 (aarch64_sqdmull_n<mode>): Likewise.
1247 (aarch64_sqdmull2<mode>_internal): Likewise.
1248 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1249 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1250 (aarch64_<sur>shl<mode>): Likewise.
1251 (aarch64_<sur>q<r>shl<mode>
1252 (aarch64_<sur>shll_n<mode>): Likewise.
1253 (aarch64_<sur>shll2_n<mode>): Likewise.
1254 (aarch64_<sur>shr_n<mode>): Likewise.
1255 (aarch64_<sur>sra_n<mode>): Likewise.
1256 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1257 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1258 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1259 (aarch64_cm<optab><mode>): Likewise.
1260 (aarch64_cm<optab>di): Likewise.
1261 (aarch64_cm<optab><mode>): Likewise.
1262 (aarch64_cm<optab>di): Likewise.
1263 (aarch64_cmtst<mode>): Likewise.
1264 (aarch64_cmtstdi): Likewise.
1265 (aarch64_cm<optab><mode>): Likewise.
1266 (*aarch64_fac<optab><mode>): Likewise.
1267 (aarch64_addp<mode>): Likewise.
1268 (aarch64_addpdi): Likewise.
1269 (sqrt<mode>2): Likewise.
1270 (vec_load_lanesoi<mode>): Likewise.
1271 (vec_store_lanesoi<mode>): Likewise.
1272 (vec_load_lanesci<mode>): Likewise.
1273 (vec_store_lanesci<mode>): Likewise.
1274 (vec_load_lanesxi<mode>): Likewise.
1275 (vec_store_lanesxi<mode>): Likewise.
1276 (*aarch64_mov<mode>): Likewise.
1277 (aarch64_ld2<mode>_dreg): Likewise.
1278 (aarch64_ld2<mode>_dreg): Likewise.
1279 (aarch64_ld3<mode>_dreg): Likewise.
1280 (aarch64_ld3<mode>_dreg): Likewise.
1281 (aarch64_ld4<mode>_dreg): Likewise.
1282 (aarch64_ld4<mode>_dreg): Likewise.
1283 (aarch64_tbl1<mode>): Likewise.
1284 (aarch64_tbl2v16qi): Likewise.
1285 (aarch64_combinev16qi): Likewise.
1286 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1287 (aarch64_st2<mode>_dreg): Likewise.
1288 (aarch64_st2<mode>_dreg): Likewise.
1289 (aarch64_st3<mode>_dreg): Likewise.
1290 (aarch64_st3<mode>_dreg): Likewise.
1291 (aarch64_st4<mode>_dreg): Likewise.
1292 (aarch64_st4<mode>_dreg): Likewise.
1293 (*aarch64_simd_ld1r<mode>): Likewise.
1294 (aarch64_frecpe<mode>): Likewise.
1295 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1296 (aarch64_frecps<mode>): Likewise.
1297
f7379e5e
JG		 12982013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1299
1300 * config/arm/iterators.md (V_elem_ch): New.
1301 (q): Likewise.
1302 (VQH_type): Likewise.
1303 * config/arm/arm.md (is_neon_type): New.
1304 (conds): Use is_neon_type.
1305 (anddi3_insn): Update type attribute.
1306 (xordi3_insn): Likewise.
1307 (one_cmpldi2): Likewise.
1308 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1309 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1310 (*movmisalign<mode>_neon_store): Likewise.
1311 (*movmisalign<mode>_neon_load): Likewise.
1312 (vec_set<mode>_internal): Likewise.
1313 (vec_set<mode>_internal): Likewise.
1314 (vec_setv2di_internal): Likewise.
1315 (vec_extract<mode>): Likewise.
1316 (vec_extract<mode>): Likewise.
1317 (vec_extractv2di): Likewise.
1318 (*add<mode>3_neon): Likewise.
1319 (adddi3_neon): Likewise.
1320 (*sub<mode>3_neon): Likewise.
1321 (subdi3_neon): Likewise.
1322 (fma<VCVTF:mode>4): Likewise.
1323 (fma<VCVTF:mode>4_intrinsic): Likewise.
1324 (*fmsub<VCVTF:mode>4): Likewise.
1325 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1326 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1327 (ior<mode>3): Likewise.
1328 (and<mode>3): Likewise.
1329 (orn<mode>3_neon): Likewise.
1330 (orndi3_neon): Likewise.
1331 (bic<mode>3_neon): Likewise.
1332 (bicdi3_neon): Likewise.
1333 (xor<mode>3): Likewise.
1334 (one_cmpl<mode>2): Likewise.
1335 (abs<mode>2): Likewise.
1336 (neg<mode>2): Likewise.
1337 (negdi2_neon): Likewise.
1338 (*umin<mode>3_neon): Likewise.
1339 (*umax<mode>3_neon): Likewise.
1340 (*smin<mode>3_neon): Likewise.
1341 (*smax<mode>3_neon): Likewise.
1342 (vashl<mode>3): Likewise.
1343 (vashr<mode>3_imm): Likewise.
1344 (vlshr<mode>3_imm): Likewise.
1345 (ashl<mode>3_signed): Likewise.
1346 (ashl<mode>3_unsigned): Likewise.
1347 (neon_load_count): Likewise.
1348 (ashldi3_neon_noclobber): Likewise.
1349 (ashldi3_neon): Likewise.
1350 (signed_shift_di3_neon): Likewise.
1351 (unsigned_shift_di3_neon): Likewise.
1352 (ashrdi3_neon_imm_noclobber): Likewise.
1353 (lshrdi3_neon_imm_noclobber): Likewise.
1354 (<shift>di3_neon): Likewise.
1355 (widen_ssum<mode>3): Likewise.
1356 (widen_usum<mode>3): Likewise.
1357 (quad_halves_<code>v4si): Likewise.
1358 (quad_halves_<code>v4sf): Likewise.
1359 (quad_halves_<code>v8hi): Likewise.
1360 (quad_halves_<code>v16qi): Likewise.
1361 (reduc_splus_v2di): Likewise.
1362 (neon_vpadd_internal<mode>): Likewise.
1363 (neon_vpsmin<mode>): Likewise.
1364 (neon_vpsmax<mode>): Likewise.
1365 (neon_vpumin<mode>): Likewise.
1366 (neon_vpumax<mode>): Likewise.
1367 (*ss_add<mode>_neon): Likewise.
1368 (*us_add<mode>_neon): Likewise.
1369 (*ss_sub<mode>_neon): Likewise.
1370 (*us_sub<mode>_neon): Likewise.
1371 (neon_vadd<mode>_unspec): Likewise.
1372 (neon_vaddl<mode>): Likewise.
1373 (neon_vaddw<mode>): Likewise.
1374 (neon_vhadd<mode>): Likewise.
1375 (neon_vqadd<mode>): Likewise.
1376 (neon_vaddhn<mode>): Likewise.
1377 (neon_vmul<mode>): Likewise.
1378 (neon_vfms<VCVTF:mode>): Likewise.
1379 (neon_vmlal<mode>): Likewise.
1380 (neon_vmls<mode>): Likewise.
1381 (neon_vmlsl<mode>): Likewise.
1382 (neon_vqdmulh<mode>): Likewise.
1383 (neon_vqdmlal<mode>): Likewise.
1384 (neon_vqdmlsl<mode>): Likewise.
1385 (neon_vmull<mode>): Likewise.
1386 (neon_vqdmull<mode>): Likewise.
1387 (neon_vsub<mode>_unspec): Likewise.
1388 (neon_vsubl<mode>): Likewise.
1389 (neon_vsubw<mode>): Likewise.
1390 (neon_vqsub<mode>): Likewise.
1391 (neon_vhsub<mode>): Likewise.
1392 (neon_vsubhn<mode>): Likewise.
1393 (neon_vceq<mode>): Likewise.
1394 (neon_vcge<mode>): Likewise.
1395 (neon_vcgeu<mode>): Likewise.
1396 (neon_vcgt<mode>): Likewise.
1397 (neon_vcgtu<mode>): Likewise.
1398 (neon_vcle<mode>): Likewise.
1399 (neon_vclt<mode>): Likewise.
1400 (neon_vcage<mode>): Likewise.
1401 (neon_vcagt<mode>): Likewise.
1402 (neon_vtst<mode>): Likewise.
1403 (neon_vabd<mode>): Likewise.
1404 (neon_vabdl<mode>): Likewise.
1405 (neon_vaba<mode>): Likewise.
1406 (neon_vabal<mode>): Likewise.
1407 (neon_vmax<mode>): Likewise.
1408 (neon_vmin<mode>): Likewise.
1409 (neon_vpaddl<mode>): Likewise.
1410 (neon_vpadal<mode>): Likewise.
1411 (neon_vpmax<mode>): Likewise.
1412 (neon_vpmin<mode>): Likewise.
1413 (neon_vrecps<mode>): Likewise.
1414 (neon_vrsqrts<mode>): Likewise.
1415 (neon_vqabs<mode>): Likewise.
1416 (neon_vqneg<mode>): Likewise.
1417 (neon_vcls<mode>): Likewise.
1418 (clz<mode>2): Likewise.
1419 (popcount<mode>2): Likewise.
1420 (neon_vrecpe<mode>): Likewise.
1421 (neon_vrsqrte<mode>): Likewise.
1422 (neon_vget_lane<mode>_sext_internal): Likewise.
1423 (neon_vget_lane<mode>_zext_internal): Likewise.
1424 (neon_vdup_n<mode>): Likewise.
1425 (neon_vdup_n<mode>): Likewise.
1426 (neon_vdup_nv2di): Likewise.
1427 (neon_vdup_lane<mode>_interal): Likewise.
1428 (*neon_vswp<mode>): Likewise.
1429 (neon_vcombine<mode>): Likewise.
1430 (float<mode><V_cvtto>2): Likewise.
1431 (floatuns<mode><V_cvtto>2): Likewise.
1432 (fix_trunc<mode><V_cvtto>2): Likewise.
1433 (fixuns_trunc<mode><V_cvtto>2
1434 (neon_vcvt<mode>): Likewise.
1435 (neon_vcvt<mode>): Likewise.
1436 (neon_vcvtv4sfv4hf): Likewise.
1437 (neon_vcvtv4hfv4sf): Likewise.
1438 (neon_vcvt_n<mode>): Likewise.
1439 (neon_vcvt_n<mode>): Likewise.
1440 (neon_vmovn<mode>): Likewise.
1441 (neon_vqmovn<mode>): Likewise.
1442 (neon_vqmovun<mode>): Likewise.
1443 (neon_vmovl<mode>): Likewise.
1444 (neon_vmul_lane<mode>): Likewise.
1445 (neon_vmul_lane<mode>): Likewise.
1446 (neon_vmull_lane<mode>): Likewise.
1447 (neon_vqdmull_lane<mode>): Likewise.
1448 (neon_vqdmulh_lane<mode>): Likewise.
1449 (neon_vqdmulh_lane<mode>): Likewise.
1450 (neon_vmla_lane<mode>): Likewise.
1451 (neon_vmla_lane<mode>): Likewise.
1452 (neon_vmlal_lane<mode>): Likewise.
1453 (neon_vqdmlal_lane<mode>): Likewise.
1454 (neon_vmls_lane<mode>): Likewise.
1455 (neon_vmls_lane<mode>): Likewise.
1456 (neon_vmlsl_lane<mode>): Likewise.
1457 (neon_vqdmlsl_lane<mode>): Likewise.
1458 (neon_vext<mode>): Likewise.
1459 (neon_vrev64<mode>): Likewise.
1460 (neon_vrev32<mode>): Likewise.
1461 (neon_vrev16<mode>): Likewise.
1462 (neon_vbsl<mode>_internal): Likewise.
1463 (neon_vshl<mode>): Likewise.
1464 (neon_vqshl<mode>): Likewise.
1465 (neon_vshr_n<mode>): Likewise.
1466 (neon_vshrn_n<mode>): Likewise.
1467 (neon_vqshrn_n<mode>): Likewise.
1468 (neon_vqshrun_n<mode>): Likewise.
1469 (neon_vshl_n<mode>): Likewise.
1470 (neon_vqshl_n<mode>): Likewise.
1471 (neon_vqshlu_n<mode>): Likewise.
1472 (neon_vshll_n<mode>): Likewise.
1473 (neon_vsra_n<mode>): Likewise.
1474 (neon_vsri_n<mode>): Likewise.
1475 (neon_vsli_n<mode>): Likewise.
1476 (neon_vtbl1v8qi): Likewise.
1477 (neon_vtbl2v8qi): Likewise.
1478 (neon_vtbl3v8qi): Likewise.
1479 (neon_vtbl4v8qi): Likewise.
1480 (neon_vtbl1v16qi): Likewise.
1481 (neon_vtbl2v16qi): Likewise.
1482 (neon_vcombinev16qi): Likewise.
1483 (neon_vtbx1v8qi): Likewise.
1484 (neon_vtbx2v8qi): Likewise.
1485 (neon_vtbx3v8qi): Likewise.
1486 (neon_vtbx4v8qi): Likewise.
1487 (*neon_vtrn<mode>_insn): Likewise.
1488 (*neon_vzip<mode>_insn): Likewise.
1489 (*neon_vuzp<mode>_insn): Likewise.
1490 (neon_vld1<mode>): Likewise.
1491 (neon_vld1_lane<mode>): Likewise.
1492 (neon_vld1_lane<mode>): Likewise.
1493 (neon_vld1_dup<mode>): Likewise.
1494 (neon_vld1_dup<mode>): Likewise.
1495 (neon_vld1_dupv2di): Likewise.
1496 (neon_vst1<mode>): Likewise.
1497 (neon_vst1_lane<mode>): Likewise.
1498 (neon_vst1_lane<mode>): Likewise.
1499 (neon_vld2<mode>): Likewise.
1500 (neon_vld2<mode>): Likewise.
1501 (neon_vld2_lane<mode>): Likewise.
1502 (neon_vld2_lane<mode>): Likewise.
1503 (neon_vld2_dup<mode>): Likewise.
1504 (neon_vst2<mode>): Likewise.
1505 (neon_vst2<mode>): Likewise.
1506 (neon_vst2_lane<mode>): Likewise.
1507 (neon_vst2_lane<mode>): Likewise.
1508 (neon_vld3<mode>): Likewise.
1509 (neon_vld3qa<mode>): Likewise.
1510 (neon_vld3qb<mode>): Likewise.
1511 (neon_vld3_lane<mode>): Likewise.
1512 (neon_vld3_lane<mode>): Likewise.
1513 (neon_vld3_dup<mode>): Likewise.
1514 (neon_vst3<mode>): Likewise.
1515 (neon_vst3qa<mode>): Likewise.
1516 (neon_vst3qb<mode>): Likewise.
1517 (neon_vst3_lane<mode>): Likewise.
1518 (neon_vst3_lane<mode>): Likewise.
1519 (neon_vld4<mode>): Likewise.
1520 (neon_vld4qa<mode>): Likewise.
1521 (neon_vld4qb<mode>): Likewise.
1522 (neon_vld4_lane<mode>): Likewise.
1523 (neon_vld4_lane<mode>): Likewise.
1524 (neon_vld4_dup<mode>): Likewise.
1525 (neon_vst4<mode>): Likewise.
1526 (neon_vst4qa<mode>): Likewise.
1527 (neon_vst4qb<mode>): Likewise.
1528 (neon_vst4_lane<mode>): Likewise.
1529 (neon_vst4_lane<mode>): Likewise.
1530 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1531 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1532 (neon_vec_<US>mult_lo_<mode>): Likewise.
1533 (neon_vec_<US>mult_hi_<mode>): Likewise.
1534 (neon_vec_<US>shiftl_<mode>): Likewise.
1535 (neon_unpack<US>_<mode>): Likewise.
1536 (neon_vec_<US>mult_<mode>): Likewise.
1537 (vec_pack_trunc_<mode>): Likewise.
1538 (neon_vec_pack_trunc_<mode>): Likewise.
1539 (neon_vabd<mode>_2): Likewise.
1540 (neon_vabd<mode>_3): Likewise.
1541
a9e66678
JG		 15422013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1543
1544 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1545 (load_pair): Update type attribute.
1546 (store_pair): Update type attribute.
1547 * config/aarch64/iterators.md (q): New.
1548
20445ed6
JG		 15492013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1550
1551 * config/arm/types.md: Add new types for Neon insns.
1552
afb4ac68
AI		 15532013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1554 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1555 Sergey Lega <sergey.s.lega@intel.com>
1556 Anna Tikhonova <anna.tikhonova@intel.com>
1557 Ilya Tocar <ilya.tocar@intel.com>
1558 Andrey Turetskiy <andrey.turetskiy@intel.com>
1559 Ilya Verbin <ilya.verbin@intel.com>
1560 Kirill Yukhin <kirill.yukhin@intel.com>
1561 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1562
1563 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1564 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1565 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1566 (rcp14<mode>): New.
1567 (srcp14<mode>): Ditto.
1568 (rsqrt14<mode>): Ditto.
1569 (rsqrt14<mode>): Ditto.
1570 (avx512f_vmscalef<mode>): Ditto.
1571 (avx512f_scalef<mode>): Ditto.
1572 (avx512f_getexp<mode>): Ditto.
1573 (avx512f_sgetexp<mode>): Ditto.
1574 (avx512f_fixupimm<mode>): Ditto.
1575 (avx512f_sfixupimm<mode>): Ditto.
1576 (avx512f_rndscale<mode>): Ditto.
1577 (*avx512er_exp2<mode>): Ditto.
1578 (*avx512er_rcp28<mode>): Ditto.
1579 (avx512er_rsqrt28<mode>): Ditto.
1580 (avx512f_getmant<mode>): Ditto.
1581 (avx512f_getmant<mode>): Ditto.
1582 (avx512f_rndscale<mode>): Fix formatting.
1583
a1b499a3
JL		 15842013-10-15 Martin Jambor <mjambor@suse.cz>
1585
1586 * ipa-utils.h (ipa_edge_within_scc): Declare.
1587 * ipa-cp.c (edge_within_scc): Moved...
1588 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
afb4ac68 1589
2e2206fa
AI		 15902013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1591 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1592 Sergey Lega <sergey.s.lega@intel.com>
1593 Anna Tikhonova <anna.tikhonova@intel.com>
1594 Ilya Tocar <ilya.tocar@intel.com>
1595 Andrey Turetskiy <andrey.turetskiy@intel.com>
1596 Ilya Verbin <ilya.verbin@intel.com>
1597 Kirill Yukhin <kirill.yukhin@intel.com>
1598 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1599
1600 * config/i386/predicates.md (const_8_to_15_operand): New.
1601 (const_16_to_31_operand): Ditto.
1602 * config/i386/sse.md (V8FI): New.
1603 (V16FI): Ditto.
1604 (reduc_splus_v8df): Ditto.
1605 (reduc_splus_v16sf): Ditto.
1606 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1607 (vec_extract_hi_<mode>): Ditto.
1608 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1609 (vec_set_lo_<mode>): Ditto.
1610 (vec_set_hi_<mode>): Ditto.
1611 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1612 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1613 (avx512f_pshufd_1): Ditto.
1614 (avx512f_broadcast<mode>): Ditto.
1615 (avx512f_broadcast<mode>): Ditto.
1616 (define_split): Split vec_extract_lo into move.
1617 (ssequartermode): Ditto.
1618 (ssedoublemode): Extened with wider modes.
1619 (vec_extract_lo_<mode>): Ditto.
1620
0fe65b75
AI		 16212013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1622 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1623 Sergey Lega <sergey.s.lega@intel.com>
1624 Anna Tikhonova <anna.tikhonova@intel.com>
1625 Ilya Tocar <ilya.tocar@intel.com>
1626 Andrey Turetskiy <andrey.turetskiy@intel.com>
1627 Ilya Verbin <ilya.verbin@intel.com>
1628 Kirill Yukhin <kirill.yukhin@intel.com>
1629 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1630
1631 * config/i386/predicates.md (register_or_constm1_operand): New.
1632 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1633 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1634 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1635 UNSPEC_SCATTER_PREFETCH
1636 (VI48_512): New.
1637 (avx512f_ucmp<mode>3): Ditto.
1638 (avx512f_vternlog<mode>): Ditto.
1639 (avx512f_align<mode>): Ditto.
1640 (<shift_insn><mode>3): Ditto.
1641 (avx512f_<rotate>v<mode>): Ditto.
1642 (avx512f_<rotate><mode>): Ditto.
1643 (avx512f_eq<mode>3): Ditto.
1644 (avx512f_eq<mode>3_1): Ditto.
1645 (avx512f_gt<mode>3): Ditto.
1646 (avx512f_testm<mode>3): Ditto.
1647 (avx512f_testnm<mode>3): Ditto.
1648 (avx512pf_gatherpf<mode>): Ditto.
1649 (*avx512pf_gatherpf<mode>_mask): Ditto.
1650 (*avx512pf_gatherpf<mode>): Ditto.
1651 (avx512pf_scatterpf<mode>): Ditto.
1652 (*avx512pf_scatterpf<mode>_mask): Ditto.
1653 (*avx512pf_scatterpf<mode>): Ditto.
1654 (avx512f_vec_dup_gpr<mode>): Ditto.
1655 (clz<mode>2): Ditto.
1656 (conflict<mode>): Ditto.
1657 (REDUC_SMINMAX_MODE): Extened with wider modes.
1658 (reduc_<code>_<mode>): Ditto.
1659 (vlshr<mode>3): Ditto.
1660 (vashl<mode>3): Ditto.
1661
ab931c71
AI		 16622013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1663 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1664 Sergey Lega <sergey.s.lega@intel.com>
1665 Anna Tikhonova <anna.tikhonova@intel.com>
1666 Ilya Tocar <ilya.tocar@intel.com>
1667 Andrey Turetskiy <andrey.turetskiy@intel.com>
1668 Ilya Verbin <ilya.verbin@intel.com>
1669 Kirill Yukhin <kirill.yukhin@intel.com>
1670 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1671
1672 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1673 UNSPEC_SCATTER.
1674 (VI48F_512): New.
1675 (avx512fmaskmode): Ditto.
1676 (bcstscalarsuff): Ditto.
1677 (avx512f_blendm<mode>): Ditto.
1678 (cmp_imm_predicate): Ditto.
1679 (avx512f_cmp<mode>3): Ditto.
1680 (avx512f_vec_dup<mode>): Ditto.
1681 (avx512f_vec_dup_mem<mode>): Ditto.
1682 (avx512f_vpermi2var<mode>3): Ditto.
1683 (avx512f_vpermt2var<mode>3): Ditto.
1684 (vec_init<mode>): Ditto.
1685 (avx512f_gathersi<mode>): Ditto.
1686 (*avx512f_gathersi<mode>): Ditto.
1687 (*avx512f_gathersi<mode>_2): Ditto.
1688 (avx512f_gatherdi<mode>): Ditto.
1689 (*avx512f_gatherdi<mode>): Ditto.
1690 (*avx512f_gatherdi<mode>_2): Ditto.
1691 (avx512f_scattersi<mode>): Ditto.
1692 (*avx512f_scattersi<mode>): Ditto.
1693 (avx512f_scatterdi<mode>): Ditto.
1694 (*avx512f_scatterdi<mode>): Ditto.
1695 (sseintprefix): Extened with wider modes.
1696 (VEC_GATHER_IDXSI): Ditto.
1697 (VEC_GATHER_IDXDI): Ditto.
1698 (VEC_GATHER_SRCDI): Ditto.
1699
c9acb877
MGD		 17002013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1701 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1702
1703 * config/arm/t-aprofile: New file.
1704 * config.gcc: Handle --with-multilib-list option.
1705
ecbda484
BS		 17062013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1707
6f1abb55
UB		 1708 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1709 the input for r2.
ecbda484 1710
24f161fd
RB		 17112013-10-15 Richard Biener <rguenther@suse.de>
1712
1713 * tree-loop-distribution.c (build_empty_rdg): Inline into
1714 single user.
1715 (rdg_flag_vertex): Inline into single user.
1716 (rdg_flag_vertex_and_dependent): Likewise.
1717 (build_rdg_partition_for_vertex): Remove processed bitmap.
1718 (rdg_build_partitions): Simplify.
1719
5de989ed
RB		 17202013-10-15 Richard Biener <rguenther@suse.de>
1721
1722 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1723 Restructure forwarding through conversions and copies to
1724 avoid performing copy-propagation the wrong way. Adjust
1725 recursion invocations.
1726 (forward_propagate_addr_expr): Add argument stating if we
1727 are recursing from a single-use.
1728 (ssa_forward_propagate_and_combine): Adjust.
1729
47e0da37
DM		 17302013-10-14 David Malcolm <dmalcolm@redhat.com>
1731
1732 * dumpfile.h (gcc::dump_manager): New class, to hold state
1733 relating to dumpfile management.
1734 (get_dump_file_name): Remove in favor of method of dump_manager.
1735 (dump_initialized_p): Likewise.
1736 (dump_start): Likewise.
1737 (dump_finish): Likewise.
1738 (dump_switch_p): Likewise.
1739 (dump_register): Likewise.
1740 (get_dump_file_info): Likewise.
1741 * context.c (gcc::context::context): Construct the dump_manager
1742 instance.
1743 * context.h (gcc::context::get_dumps): New.
1744 (gcc::context::m_dumps): New.
1745 * coverage.c (coverage_init): Port to dump_manager API.
6f1abb55 1746 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
47e0da37
DM		 1747 (extra_dump_files_in_use): Likewise.
1748 (extra_dump_files_alloced): Likewise.
1749 (gcc::dump_manager::dump_manager): New.
1750 (dump_register): Convert to...
1751 (gcc::dump_manager::dump_register): ...method, replacing
1752 function-static next_dump with m_next_dump field.
1753 (get_dump_file_info): Convert to...
1754 (gcc::dump_manager::get_dump_file_info): ...method.
1755 (get_dump_file_name): Convert to...
1756 (gcc::dump_manager::get_dump_file_name): ...method.
1757 (dump_start): Convert to...
1758 (gcc::dump_manager::dump_start): ...method.
1759 (dump_finish): Convert to...
1760 (gcc::dump_manager::dump_finish): ...method.
1761 (dump_begin): Replace body with...
1762 (gcc::dump_manager::dump_begin): ...new method.
1763 (dump_phase_enabled_p): Convert to...
1764 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1765 (dump_phase_enabled_p): Convert to...
1766 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1767 (dump_initialized_p): Convert to...
1768 (gcc::dump_manager::dump_initialized_p): ...method.
1769 (dump_flag_name): Replace body with...
1770 (gcc::dump_manager::dump_flag_name): ...new method.
1771 (dump_enable_all): Convert to...
1772 (gcc::dump_manager::dump_enable_all): ...new method.
1773 (opt_info_enable_passes): Convert to...
1774 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1775 (dump_switch_p_1): Convert to...
1776 (gcc::dump_manager::dump_switch_p_1): ...new method.
1777 (dump_switch_p): Convert to...
1778 (gcc::dump_manager::dump_switch_p): ...new method.
1779 (opt_info_switch_p): Port to dump_manager API.
1780 (enable_rtl_dump_file): Likewise.
1781 * opts-global.c (handle_common_deferred_options): Port to new
1782 dump_manager API.
1783 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1784 (pass_manager::register_one_dump_file): Likewise.
1785 (pass_manager::register_pass): Likewise.
1786 (pass_init_dump_file): Likewise.
1787 (pass_fini_dump_file): Likewise.
1788 * statistics.c (statistics_early_init): Likewise.
1789
83fd5d11
RB		 17902013-10-14 Richard Biener <rguenther@suse.de>
1791
1792 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1793 iterative_hash_canonical_type, gimple_canonical_type_hash,
1794 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1795 gimple_register_canonical_type, print_gimple_types_stats,
1796 free_gimple_type_tables): Move to lto/lto.c
1797 (gt-gimple.h): Do not include.
1798 * gimple.h (gimple_register_canonical_type,
1799 print_gimple_types_stats, free_gimple_type_tables): Remove.
1800 * Makefile.in (GTFILES): Remove gimple.c.
1801
a6dbd0f7
TS		 18022013-10-14 Travis Snoozy <quandary@remstate.com>
1803
1804 PR target/58716
1805 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1806 scanning for msp430x targets.
1807
b5ccb9ed
EB		 18082013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1809
1810 PR bootstrap/58509
1811 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1812 (registers_ok_for_ldd_peep): Move around.
1813 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1814 * config/sparc/sparc.md (widening peepholes): Use it.
1815
fd7e5084
RB		 18162013-10-14 Richard Biener <rguenther@suse.de>
1817
1818 PR middle-end/58712
1819 PR middle-end/55358
1820 * gimple.c (iterative_hash_canonical_type): Make sure to
1821 record the hash into the correct hashtable slot.
1822
3246afaa
EB		 18232013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1824
1825 PR rtl-optimization/58662
1826 * combine.c (try_combine): Take into account death nodes on I2 when
1827 splitting a PARALLEL of two independent SETs. Fix dump message.
1828
5d30dc5b
OE		 18292013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1830
1831 PR target/51244
1832 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1833 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1834 * config/sh/t-sh (sh_treg_combine.o): New entry.
1835 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1836 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1837 (register_sh_passes): New function. Register sh_treg_combine pass.
1838 (sh_option_override): Invoke it.
1839 (sh_canonicalize_comparison): Handle op0_preserve_value.
1840 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1841 opportunities. Canonicalize branch condition.
1842 (nott): Allow only if pseudos can be created for non-SH2A.
1843
37d1c5db
L		 18442013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1845
1846 PR target/58690
1847 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1848 (ix86_expand_movmem): Replace copy_addr_to_reg with
1849 ix86_copy_addr_to_reg.
1850 (ix86_expand_setmem): Likewise.
1851
b86a70ff
AM		 18522013-10-12 Alexander Monakov <amonakov@ispras.ru>
1853
1854 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1855 provided by ix86_fp_compare_mode instead of CCFPUmode.
1856
6e1d5961
JG		 18572013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1858
1859 * config/aarch64/arm_neon.h
1860 (vtbx<1,3>_<psu>8): Fix register constriants.
1861
ef3cfba2
JL		 18622013-10-11 Jeff Law <law@redhat.com>
1863
1864 PR tree-optimization/58640
adcfd489
UB		 1865 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1866 threading paths that cross over two loop entry points.
ef3cfba2 1867
6e8b7d9c
BS		 18682013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1869
1870 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1871 handle vector float as well.
1872 (*vsx_le_perm_load_v4si): Likewise.
1873 (*vsx_le_perm_store_v2di): Likewise.
1874 (*vsx_le_perm_store_v4si): Likewise.
1875
f200869a
BS		 18762013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1877
1878 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1879 directly to circumvent subtract from splat{31} workaround.
1880 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1881 prototype.
1882 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1883 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1884 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1885 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1886 separate big and little endian logic.
1887 (*altivec_vperm_<mode>_internal): New define_insn.
1888 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1889 separate big and little endian logic.
1890 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1891 (vec_permv16qi): Add little endian logic.
1892
8d28e3fc
MG		 18932013-10-11 Marc Glisse <marc.glisse@inria.fr>
1894
1895 * doc/extend.texi (returns_nonnull): Remove arguments.
1896
3c87b77b
AI		 18972013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1898 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1899 Sergey Lega <sergey.s.lega@intel.com>
1900 Anna Tikhonova <anna.tikhonova@intel.com>
1901 Ilya Tocar <ilya.tocar@intel.com>
1902 Andrey Turetskiy <andrey.turetskiy@intel.com>
1903 Ilya Verbin <ilya.verbin@intel.com>
1904 Kirill Yukhin <kirill.yukhin@intel.com>
1905 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1906
1907 * config/i386/sse.md (VI48F_256_512): New.
1908 (avx2_permvar<mode>): Change to ...
1909 (<avx2_avx512f>_permvar<mode>): This.
1910
50e60d7d
AI		 19112013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1912 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1913 Sergey Lega <sergey.s.lega@intel.com>
1914 Anna Tikhonova <anna.tikhonova@intel.com>
1915 Ilya Tocar <ilya.tocar@intel.com>
1916 Andrey Turetskiy <andrey.turetskiy@intel.com>
1917 Ilya Verbin <ilya.verbin@intel.com>
1918 Kirill Yukhin <kirill.yukhin@intel.com>
1919 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1920
6f1abb55 1921 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
50e60d7d
AI		 1922 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1923 * config/i386/sse.md (VI4_AVX): New.
1924 (sf2simodelower): Ditto.
1925 (sse2_cvtps2dq): Change to ...
1926 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1927
f62ce24f
AI		 19282013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1929 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1930 Sergey Lega <sergey.s.lega@intel.com>
1931 Anna Tikhonova <anna.tikhonova@intel.com>
1932 Ilya Tocar <ilya.tocar@intel.com>
1933 Andrey Turetskiy <andrey.turetskiy@intel.com>
1934 Ilya Verbin <ilya.verbin@intel.com>
1935 Kirill Yukhin <kirill.yukhin@intel.com>
1936 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1937
1938 * config/i386/sse.md (V_512): New.
1939 (VI_512): Ditto.
1940 (vcond<V_512:mode><VF_512:mode>): Ditto.
1941 (vcond<V_512:mode><VI_512:mode>): Ditto.
1942 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1943
558d9f79
AI		 19442013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1945 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1946 Sergey Lega <sergey.s.lega@intel.com>
1947 Anna Tikhonova <anna.tikhonova@intel.com>
1948 Ilya Tocar <ilya.tocar@intel.com>
1949 Andrey Turetskiy <andrey.turetskiy@intel.com>
1950 Ilya Verbin <ilya.verbin@intel.com>
1951 Kirill Yukhin <kirill.yukhin@intel.com>
1952 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1953
1954 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1955 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1956 (FMAMODE): Ditto.
1957 (fma<mode>4): Removed condition.
1958 (fms<mode>4): Ditto.
1959 (fnma<mode>4): Ditto.
1960 (fnms<mode>4): Ditto.
1961 (fma4i_fmadd_<mode>): Ditto.
1962 (*fma_fmadd_<mode>): Ditto.
1963 (*fma_fmsub_<mode>): Ditto.
1964 (*fma_fnmadd_<mode>): Ditto.
1965 (*fma_fnmsub_<mode>): Ditto.
1966 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1967 (*fma_fmaddsub_<mode>): Ditto.
1968 (*fma_fmsubadd_<mode>): Ditto.
1969 (*fmai_fmadd_<mode>): Ditto.
1970 (*fmai_fmsub_<mode>): Ditto.
1971 (*fmai_fnmadd_<mode>): Ditto.
1972 (*fmai_fnmsub_<mode>): Ditto.
1973
e8d08206
AI		 19742013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1975 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1976 Sergey Lega <sergey.s.lega@intel.com>
1977 Anna Tikhonova <anna.tikhonova@intel.com>
1978 Ilya Tocar <ilya.tocar@intel.com>
1979 Andrey Turetskiy <andrey.turetskiy@intel.com>
1980 Ilya Verbin <ilya.verbin@intel.com>
1981 Kirill Yukhin <kirill.yukhin@intel.com>
1982 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1983
1984 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1985 (VI124_256): Changed to ...
1986 (VI124_256_48_512): This.
1987 (ssepackmode): Extended with wider modes.
1988 (<code><mode>3): Changed iterator.
1989 (*avx2_<code><mode>3): Ditto.
1990 (vec_pack_trunc_<mode>): Ditto.
1991
16821545
AI		 19922013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1993 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1994 Sergey Lega <sergey.s.lega@intel.com>
1995 Anna Tikhonova <anna.tikhonova@intel.com>
1996 Ilya Tocar <ilya.tocar@intel.com>
1997 Andrey Turetskiy <andrey.turetskiy@intel.com>
1998 Ilya Verbin <ilya.verbin@intel.com>
1999 Kirill Yukhin <kirill.yukhin@intel.com>
2000 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2001
2002 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2003 (VI8F_256_512): Ditto.
2004 (abs<mode>2): Changed iterator.
2005 (avx2_perm<mode>): Changed to ...
2006 (<avx2_avx512f>_perm<mode>): This.
2007 (avx2_perm<mode>_1): Changed to ...
2008 (<avx2_avx512f>_perm<mode>_1): This.
2009
5348cff8
AI		 20102013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2011 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2012 Sergey Lega <sergey.s.lega@intel.com>
2013 Anna Tikhonova <anna.tikhonova@intel.com>
2014 Ilya Tocar <ilya.tocar@intel.com>
2015 Andrey Turetskiy <andrey.turetskiy@intel.com>
2016 Ilya Verbin <ilya.verbin@intel.com>
2017 Kirill Yukhin <kirill.yukhin@intel.com>
2018 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2019
2020 * config/i386/sse.md (VI48_AVX512F): New.
2021 (VI48_AVX2): Changed to ...
2022 (VI48_AVX2_48_AVX512F): This.
2023 (avx2_ashrv<mode>): Changed to ...
2024 (<avx2_avx512f>_ashrv<mode>): This.
2025 (avx2_<shift_insn>v<mode>): Changed to ...
2026 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2027
f5f41d88
AI		 20282013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2029 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2030 Sergey Lega <sergey.s.lega@intel.com>
2031 Anna Tikhonova <anna.tikhonova@intel.com>
2032 Ilya Tocar <ilya.tocar@intel.com>
2033 Andrey Turetskiy <andrey.turetskiy@intel.com>
2034 Ilya Verbin <ilya.verbin@intel.com>
2035 Kirill Yukhin <kirill.yukhin@intel.com>
2036 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2037
2038 * config/i386/sse.md (VI4_AVX512F): New.
2039 (VI8_AVX2_AVX512F): Ditto.
2040 (mul<mode>3): Extended with wider modes.
2041 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2042 (mul<mode>3): Ditto.
2043 (vec_widen_<s>mult_odd_<mode>): Ditto.
2044
3bdf6340
AI		 20452013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2046 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2047 Sergey Lega <sergey.s.lega@intel.com>
2048 Anna Tikhonova <anna.tikhonova@intel.com>
2049 Ilya Tocar <ilya.tocar@intel.com>
2050 Andrey Turetskiy <andrey.turetskiy@intel.com>
2051 Ilya Verbin <ilya.verbin@intel.com>
2052 Kirill Yukhin <kirill.yukhin@intel.com>
2053 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2054
2055 * config/i386/sse.md (VI2_AVX512F): New.
2056 (VI124_AVX512F): Ditto.
2057 (sseunpackmode): Extended with wider modes.
2058 (sseunpackfltmode): Ditto.
2059 (vec_unpacks_float_hi_<mode>): Ditto.
2060 (vec_unpacks_float_lo_<mode>): Ditto.
2061 (vec_unpacku_float_hi_<mode>): Ditto.
2062 (vec_unpacku_float_lo_<mode>): Ditto.
2063 (vec_unpacks_lo_<mode>): Ditto.
2064 (vec_unpacks_hi_<mode>): Ditto.
2065 (vec_unpacku_lo_<mode>): Ditto.
2066 (vec_unpacku_hi_<mode>): Ditto.
2067
67f783cb
AI		 20682013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2069 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2070 Sergey Lega <sergey.s.lega@intel.com>
2071 Anna Tikhonova <anna.tikhonova@intel.com>
2072 Ilya Tocar <ilya.tocar@intel.com>
2073 Andrey Turetskiy <andrey.turetskiy@intel.com>
2074 Ilya Verbin <ilya.verbin@intel.com>
2075 Kirill Yukhin <kirill.yukhin@intel.com>
2076 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2077
2078 * config/i386/i386.md (multdiv): New.
2079 (multdiv_mnemonic): Ditto.
2080 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2081 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2082 (<sse>_vmdiv<mode>3): Removed.
2083
ec5e777c
AI		 20842013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2085 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2086 Sergey Lega <sergey.s.lega@intel.com>
2087 Anna Tikhonova <anna.tikhonova@intel.com>
2088 Ilya Tocar <ilya.tocar@intel.com>
2089 Andrey Turetskiy <andrey.turetskiy@intel.com>
2090 Ilya Verbin <ilya.verbin@intel.com>
2091 Kirill Yukhin <kirill.yukhin@intel.com>
2092 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2093
2094 * config/i386/sse.md (V): Extended with wider modes.
2095 (VF2): Ditto.
2096 (ssehalfvecmode): Ditto.
2097 (i128): Ditto.
2098 (ssepackfltmode): Ditto.
2099 (avx_vec_concat<mode>): Ditto.
2100 (V_256_512): New iterator.
2101 (VF2_512_256): Ditto.
2102 (si2dfmode): New attribute.
2103 (si2dfmodelower): Ditto.
2104 (sf2dfmode): Ditto.
2105 (concat_tg_mode): Ditto.
2106 (floatv4siv4df2): Changed to ...
2107 (float<si2dfmodelower><mode>2): This.
2108 (avx_cvtps2pd256): Changed to ...
2109 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2110 (vec_pack_trunc_v4df): Changed to ...
2111 (vec_pack_trunc_<mode>): This.
2112 (avx_vpermil<mode>): Changed to ...
2113 (<sse2_avx_avx512f>_vpermil<mode>): This.
2114 (<fixsuffix>fix_truncv8dfv8si2): New.
2115 (vec_pack_sfix_trunc_v8df): Ditto.
2116 (avx512f_rndscale<mode>): Ditto.
2117 (avx512f_roundpd512): Ditto.
2118 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2119
a9ccbba2
AI		 21202013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2121 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2122 Sergey Lega <sergey.s.lega@intel.com>
2123 Anna Tikhonova <anna.tikhonova@intel.com>
2124 Ilya Tocar <ilya.tocar@intel.com>
2125 Andrey Turetskiy <andrey.turetskiy@intel.com>
2126 Ilya Verbin <ilya.verbin@intel.com>
2127 Kirill Yukhin <kirill.yukhin@intel.com>
2128 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2129
2130 * config/i386/i386.md (any_fix): New iterator.
2131 (fixsuffix): New attribute.
2132 * config/i386/sse.md (VF1): Extened with wider modes.
2133 (VI): Ditto.
2134 (VI_AVX2): Ditto.
2135 (VI8): Ditto.
2136 (sseintvecmodelower): Ditto.
2137 (ssescalarmode): Ditto.
2138 (ssescalarnum): Ditto.
2139 (VF1_128_256): New.
2140 (ssexmmmode): Ditto.
2141 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2142 (<sse>_rcp<mode>2): Change iterator.
2143 (rsqrt<mode>2): Ditto.
2144 (<sse>_rsqrt<mode>2): Ditto.
2145 (avx2_vec_dup<mode>): Ditto.
2146 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2147 (round<mode>2_sfix): Ditto.
2148 (avx2_pbroadcast<mode>): Ditto.
2149 (*andnot<mode>3): Handle XI mode.
2150 (*<code><mode>3): Ditto.
2151 (AVXTOSSEMODE): Removed.
2152 (avx_vpermil<mode>): Changed to ...
2153 (<sse2_avx_avx512f>_vpermil<mode>): This.
2154
2b1ebb0c
AI		 21552013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2156 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2157 Sergey Lega <sergey.s.lega@intel.com>
2158 Anna Tikhonova <anna.tikhonova@intel.com>
2159 Ilya Tocar <ilya.tocar@intel.com>
2160 Andrey Turetskiy <andrey.turetskiy@intel.com>
2161 Ilya Verbin <ilya.verbin@intel.com>
2162 Kirill Yukhin <kirill.yukhin@intel.com>
2163 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2164
2165 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2166 (<sse>_comi): Ditto.
2167 (<sse>_ucomi): Ditto.
2168 (sse_cvtss2siq_2): Ditto.
2169 (sse2_cvtsd2si): Ditto.
2170 (sse2_cvtsd2siq): Ditto.
2171 (sse2_cvttsd2si): Ditto.
2172 (sse2_cvttsd2siq): Ditto.
2173 (<shift_insn><mode>3): Ditto.
2174 (sse2_cvtsi2sdq): Update constraint and prefix.
2175 (sse_cvtsi2ss): Update prefix.
2176 (sse_cvtsi2ssq): Ditto.
2177
d89124ed
JJ		 21782013-10-11 Jakub Jelinek <jakub@redhat.com>
2179
2180 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2181 ignore internal calls.
2182
e6ad28c3
RB		 21832013-10-11 Richard Biener <rguenther@suse.de>
2184
6f1abb55
UB		 2185 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2186 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2187 TREE_OVERFLOW is set.
e6ad28c3 2188
d1227447
TS		 21892013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2190
1ac4b8e4
TS		 2191 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2192
7d05cebb
TS		 2193 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2194 * gimple.h: Likewise.
2195
adcfd489 2196 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6545f0b3 2197
1a79fb8a
TS		 2198 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2199 macros.
2200
3afd2873
TS		 2201 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2202
adcfd489 2203 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
d1227447
TS		 2204 * configure: Regenerate.
2205
acf0174b
JJ		 22062013-10-11 Jakub Jelinek <jakub@redhat.com>
2207
2208 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2209 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2210 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2211 similarly pp_right_paren.
2212 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2213 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2214 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2215 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2216 clauses.
2217 (omp_declare_simd_clauses_equal,
2218 omp_remove_redundant_declare_simd_attrs): New functions.
2219 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2220 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2221 * tree.h (OMP_LOOP_CHECK): Define.
2222 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2223 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2224 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2225 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2226 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2227 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2228 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2229 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2230 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2231 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2232 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2233 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2234 (omp_remove_redundant_declare_simd_attrs): New prototype.
2235 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2236 GIMPLE_OMP_TEAMS): New codes.
2237 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2238 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2239 fields.
2240 (target_nesting_level): New variable.
2241 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2242 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2243 for collapse > 1 static schedule unless ordered.
2244 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2245 (determine_parallel_type): Adjust get_ws_args_for caller.
2246 (install_var_field): Handle mask & 4 for double indirection.
2247 (scan_sharing_clauses): Ignore shared clause on teams construct.
2248 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2249 (create_omp_child_function): If inside target or declare target
adcfd489 2250 constructs, set "omp declare target" attribute on the child function.
acf0174b
JJ		 2251 (find_combined_for): New function.
2252 (scan_omp_parallel): Handle combined loops.
2253 (scan_omp_target, scan_omp_teams): New functions.
2254 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2255 restrictions and set ctx->cancellable for cancellable constructs.
2256 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2257 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2258 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
adcfd489 2259 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
acf0174b 2260 (omp_clause_aligned_alignment): New function.
6f1abb55 2261 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
acf0174b
JJ		 2262 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2263 on teams constructs. Handle user defined reductions and new
2264 OpenMP 4.0 clauses.
2265 (lower_reduction_clauses): Don't set placeholder to address of ref
2266 if it has already the right type.
2267 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2268 (expand_parallel_call): Use the new non-_start suffixed builtins,
2269 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2270 and GOMP_parallel_end after the call.
2271 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2272 (expand_omp_for_init_counts): Handle combined loops.
2273 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2274 loops.
2275 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2276 end of cancellable loops.
2277 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2278 Likewise. Handle collapse > 1 loops.
2279 (expand_omp_simd): Handle combined loops.
2280 (expand_omp_for): Add inner_stmt argument, adjust callers of
2281 expand_omp_for* functions, use expand_omp_for_static*chunk even
2282 for collapse > 1 unless ordered.
2283 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2284 of cancellable sections.
2285 (expand_omp_single): Remove need_barrier variable, just rely on
2286 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2287 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2288 (expand_omp_atomic_load, expand_omp_atomic_store,
2289 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2290 (expand_omp_target): New function.
2291 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2292 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2293 (build_omp_regions_1): Immediately close region for
2294 GF_OMP_TARGET_KIND_UPDATE.
2295 (maybe_add_implicit_barrier_cancel): New function.
2296 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2297 cancellation.
2298 (lower_omp_single): Likewise. Add clobber after the barrier.
2299 (lower_omp_taskgroup): New function.
2300 (lower_omp_for): Handle combined loops. Adjust
2301 lower_rec_input_clauses caller. Handle cancellation.
2302 (lower_depend_clauses): New function.
2303 (lower_omp_taskreg): Lower depend clauses. Adjust
2304 lower_rec_input_clauses caller. Add clobber after the call. Handle
2305 cancellation.
2306 (lower_omp_target, lower_omp_teams): New functions.
2307 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2308 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2309 and GOMP_cancellation_point calls.
2310 (lower_omp): Fold stmts inside of target region.
2311 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2312 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2313 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2314 (BT_FN_VOID_OMPFN_PTR_UINT,
2315 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2316 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2317 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2318 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2319 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2320 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2321 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2322 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2323 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2324 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2325 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2326 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2327 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2328 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2329 BUILT_IN_GOMP_PARALLEL_END.
2330 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2331 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2332 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2333 GF_OMP_FOR_KIND_DISTRIBUTE.
2334 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2335 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2336 (dump_gimple_omp_return): Print lhs if it has any.
2337 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2338 gimple_omp_atomic_seq_cst_p.
2339 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2340 and GIMPLE_OMP_TEAMS.
2341 * langhooks.c (lhd_omp_mappable_type): New function.
2342 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2343 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2344 hook.
2345 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2346 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6f1abb55 2347 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
acf0174b
JJ		 2348 (struct gimplify_omp_ctx): Add combined_loop field.
2349 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2350 on stmts inside of target region.
2351 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2352 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2353 ORT_TARGET and ORT_TARGET_DATA.
2354 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2355 Handle GOVD_MAP.
2356 (omp_notice_threadprivate_variable): Complain about threadprivate
2357 variables in target region.
2358 (omp_notice_variable): Complain about vars with non-mappable type
2359 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2360 (omp_check_private): Ignore ORT_TARGET* regions.
2361 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2362 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2363 (find_combined_omp_for): New function.
2364 (gimplify_omp_for): Handle gimplification of combined loops.
2365 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2366 OMP_TEAMS.
2367 (gimplify_omp_target_update): New function.
2368 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2369 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2370 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2371 (gimplify_body): If fndecl has "omp declare target" attribute, add
2372 implicit ORT_TARGET context around it.
2373 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2374 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2375 * tree-nested.c (convert_nonlocal_reference_stmt,
2376 convert_local_reference_stmt, convert_gimple_call): Handle
2377 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2378 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2379 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2380 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2381 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2382 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2383 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2384 BUILT_IN_GOMP_LOOP_END_CANCEL,
2385 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2386 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2387 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2388 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2389 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2390 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2391 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2392 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2393 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2394 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2395 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2396 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2397 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2398 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2399 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2400 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2401 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2402 gimple_build_omp_teams): New functions.
2403 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2404 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2405 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2406 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2407 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2408 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2409 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2410 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2411 GF_OMP_ATOMIC_SEQ_CST): New.
2412 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2413 gimple_build_omp_teams): New prototypes.
2414 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2415 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2416 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2417 GIMPLE_OMP_SINGLE as end of range.
2418 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2419 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2420 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2421 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2422 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2423 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2424 gimple_omp_target_kind, gimple_omp_target_set_kind,
2425 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2426 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2427 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2428 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2429 gimple_omp_teams_set_clauses): New inlines.
2430 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2431 and GIMPLE_OMP_TASKGROUP.
adcfd489 2432 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
acf0174b
JJ		 2433 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2434 enum omp_clause_proc_bind_kind): New.
2435 (union omp_clause_subcode): Add depend_kind, map_kind and
2436 proc_bind_kind fields.
2437 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2438 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2439 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2440 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2441 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2442
67fa7880
TJ		 24432013-10-10 Teresa Johnson <tejohnson@google.com>
2444
2445 * predict.c (tree_estimate_probability): Add new parameter
2446 for estimate_bb_frequencies.
2447 (estimate_bb_frequencies): Add new parameter to force estimation.
2448 (rebuild_frequencies): When max frequency in function is small,
2449 recompute counts from frequencies.
2450 * predict.h (estimate_bb_frequencies): New parameter.
2451
b591a8b7
DM		 24522013-10-10 David Malcolm <dmalcolm@redhat.com>
2453
2454 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2455 optimizations are disabled.
2456
782f0db2
DM		 24572013-10-10 David Malcolm <dmalcolm@redhat.com>
2458
2459 * coverage.c (coverage_finish): Fix leak of da_file_name.
2460
13ef00fa
JH		 24612013-10-10 Jan Hubicka <jh@suse.cz>
2462
2463 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2464 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2465 Bobcat and generic.
2466
a1d50386
JJ		 24672013-10-10 Jakub Jelinek <jakub@redhat.com>
2468
2469 PR middle-end/58670
2470 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2471 if any labels are in FALLTHRU_BB, use a special label emitted
2472 immediately after the asm goto insn rather than label_rtx
2473 of the LABEL_DECL.
2474 (expand_asm_stmt): Adjust caller.
2475 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2476 edge if the last insn in predecessor is a jump with single successor,
2477 but it isn't simplejump_p.
2478
e2c2fde2
RB		 24792013-10-10 Richard Biener <rguenther@suse.de>
2480
2481 PR tree-optimization/58656
6f1abb55 2482 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
e2c2fde2 2483
de6a5608
AM		 24842013-10-10 Andrew MacLeod <amacleod@redhat.com>
2485
adcfd489
UB		 2486 * gimplify.c: Include expr.h and tm_p.h for targets with special
2487 va-arg padding requirements.
de6a5608 2488
b184c8f1
AM		 24892013-10-10 Andrew MacLeod <amacleod@redhat.com>
2490
2491 * tree-flow.h: Move some prototypes to gimple.h.
2492 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2493 * gimple.h: Relocate some prototypes from tree-flow.h
2494 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2495 Move to gimplify.c.
2496 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2497 (build_va_arg_indirect_ref): Relocate and make static.
2498 (std_gimplify_va_arg_expr): Relocate here.
2499 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2500 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2501
13b72c22
AK		 25022013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2503
2504 * doc/md.texi: Document the mnemonic attribute.
2505
5db40447
AK		 25062013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2507
2508 PR target/57377
2509 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2510 (set_attr_alternative x ...) when searching for user defined
2511 mnemonic attribute.
2512
ed018d4b
AM		 25132013-10-10 Andrew MacLeod <amacleod@redhat.com>
2514
2515 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2516
4d5b5e9f
ER		 25172013-10-09 Easwaran Raman <eraman@google.com>
2518
2519 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2520 * cfgexpand.c (defer_stack_allocation): ...use here
2521 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2522
c9ef86a1
ZC		 25232013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2524
2525 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2526 (operand_equal_for_value_replacement): New function, extracted from
2527 value_replacement and enhanced to catch more cases.
2528 (value_replacement): Use operand_equal_for_value_replacement.
2529
2431114f
AM		 25302013-10-09 Andrew MacLeod <amacleod@redhat.com>
2531
adcfd489 2532 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2431114f
AM		 2533 get_max_loop_iterations.
2534
1dd2a9a0
KT		 25352013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2536
2537 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2538 Place comment above function.
2539
c1bf2a39
AM		 25402013-10-09 Andrew MacLeod <amacleod@redhat.com>
2541
2542 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2543 are not related to tree-cfg.c.
2544 * tree-ssa-address.h: New file. Relocate prototypes.
2545 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2546 (addr_for_mem_ref): New. Combine call to get_address_description and
2547 return addr_for_mem_ref.
2548 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2549 * tree-ssa-live.h: Adjust prototypes.
2550 * passes.c: Include tree-ssa-live.h.
2551 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2552 * graphite.c (graphite_transform_loops): Make static.
2553 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
adcfd489 2554 make_pass_graphite, pass_data_graphite_transforms,
c1bf2a39
AM		 2555 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2556 * ipa-pure-const.c (warn_function_noreturn): Make static.
2557 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2558 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2559 Relocate from tree-cfg.c
2560 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2561 static.
2562 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2563 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2564 Move to ipa-pure-const.c.
adcfd489
UB		 2565 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2566 Relocate from tree-optimize.c.
c1bf2a39
AM		 2567 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2568 make_pass_fixup_cfg): Move to tree-cfg.c.
2569 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2570 Relocate some prototypes.
2571 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2572 * tree-dump.c (dump_function_to_file): Remove prototype.
2573 Add tree-flow.h to the include file.
2574 * tree-dump.h: Remove prototype.
2575 * tree-parloops.h: New File. Add prototypes.
2576 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2577 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2578 from tree-ssa-loop.c.
2579 * tree-predcom.c (run_tree_predictive_commoning,
2580 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2581 Relocate here from tree-ssa-loop.c.
adcfd489 2582 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
c1bf2a39
AM		 2583 ssa_name_values.release ().
2584 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2585 (ssa_name_values): Relocate from tree-flow.h.
2586 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2587 * tree-ssa-loop.c (run_tree_predictive_commoning,
2588 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2589 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2590 make_pass_graphite, pass_data_graphite_transforms,
2591 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2592 tree_parallelize_loops, pass_data_parallelize_loops,
2593 make_pass_parallelize_loops): Move to other files.
2594 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2595 moved here.
2596 * tree.h: Remove prototypes from tree-address.c.
2597
cc524fc7
AM		 25982013-10-09 Andrew MacLeod <amacleod@redhat.com>
2599
2600 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2601 (struct int_tree_map): Move to tree-hasher.h
2602 (SCALE, LABEL, PERCENT): Move to gimple.h
2603 * tree-flow-inline.h: Delete. Move functions to other files.
2604 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2605 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2606 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2607 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2608 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2609 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2610 * tree-hasher.h: Don't include tree-flow.h.
2611 (struct int_tree_map): Relocate from tree-flow.h.
2612 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2613 tree-flow-inline.h and make static.
adcfd489
UB		 2614 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2615 tree-flow-inline.h.
cc524fc7
AM		 2616 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2617 tree-flow-inline.h and make static.
2618 * tree.h (is_global_var, may_be_aliased): Relocate from
2619 tree-flow-inline.h.
2620 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2621 * value-prof.c: No longer include tree-flow-inline.h.
2622 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2623
71343877
AM		 26242013-10-09 Andrew MacLeod <amacleod@redhat.com>
2625
2626 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2627 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2628 (enum move_pos): Move to tree-ssa-loop-im.h
2629 * cfgloop.h: Move some prototypes.
2630 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2631 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2632 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2633 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2634 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2635 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2636 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2637 (enum move_pos): Relocate here.
2638 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2639 tree-ssa-loop.c.
2640 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
adcfd489
UB		 2641 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2642 make_pass_lim): Relocate here from tree-ssa-loop.c.
2643 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
71343877
AM		 2644 tree-ssa-loop.c.
2645 (loop_edge_to_cancel, unloop_loops): Make static.
2646 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2647 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2648 (tree_complete_unroll, gate_tree_complete_unroll,
2649 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2650 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2651 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2652 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2653 (stmt_invariant_in_loop_p): Remove unused function.
2654 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2655 * tree-ssa-loop-manip.h: New file. Add prototypes.
2656 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2657 (gcov_type_to_double_int): Move to cfgloop.h.
2658 (double_int_cmp, bound_index,
2659 estimate_numbers_of_iterations_loop): Make static.
2660 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2661 (max_loop_iterations): Factor out get_max_loop_iterations.
2662 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2663 cfgloop.c.
2664 * tree-ssa-loop-niter.h: New file. Add prototypes.
2665 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2666 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2667 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2668 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2669 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2670 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2671 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2672 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2673 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2674 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2675 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2676 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2677 pass_data_complete_unroll, make_pass_complete_unroll,
2678 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2679 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2680 tree-ssa-loop-ivcanon.c.
2681 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2682 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2683 tree-ssa-loop-prefetch.c.
2684 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2685 tree-ssa-loop-im.c.
2686 (get_lsm_tmp_name): Relocate and add suffix parameter.
2687 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2688 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2689 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2690 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
adcfd489 2691 (get_estimated_loop_iterations): Factor out accessor from
71343877 2692 estimated_loop_iterations in tree-ssa-loop-niter.c.
adcfd489
UB		 2693 (get_max_loop_iterations): Factor out accessor from
2694 _max_loop_iterations in tree-ssa-niter.c.
71343877
AM		 2695 * loop-unroll.c (decide_unroll_constant_iterations,
2696 decide_unroll_runtime_iterations, decide_peel_simple,
2697 decide_unroll_stupid): Use new get_* accessors.
2698
826cacfe
MG		 26992013-10-09 Marc Glisse <marc.glisse@inria.fr>
2700
2701 PR tree-optimization/20318
2702 * doc/extend.texi (returns_nonnull): New function attribute.
2703 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2704 attribute.
2705 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2706 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2707
378f8976
EB		 27082013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2709
2710 PR middle-end/58570
2711 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2712 false if both components are bitfields.
2713
0fe04f5c
AV		 27142013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2715
2716 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2717 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2718 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2719 * config/aarch64/aarch64.h
2720 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2721 * config/aarch64/aarch64-simd-builtins.def
2722 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2723
bed9bae4
AV		 27242013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2725
2726 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2727
7f3d8b19
AV		 27282013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2729
2730 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2731
7df625a6
AV		 27322013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2733
2734 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2735 (vneg_f64): New intrinsic.
2736 (vneg_s8): Asm replaced with C.
2737 (vneg_s16): Likewise.
2738 (vneg_s32): Likewise.
2739 (vneg_s64): New intrinsic.
2740 (vnegq_f32): Asm replaced with C.
2741 (vnegq_f64): Likewise.
2742 (vnegq_s8): Likewise.
2743 (vnegq_s16): Likewise.
2744 (vnegq_s32): Likewise.
2745 (vnegq_s64): Likewise.
2746
2b86fca7
RL		 27472013-10-09 Renlin Li <Renlin.Li@arm.com>
2748
2749 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2750
e179df83
AK		 27512013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2752
2753 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2754 packed stack special handling.
2755 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2756 back to fixed stack slots for FPRs saved due to stdarg.
2757
74129172
AK		 27582013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2759
2760 * config/s390/s390.c (s390_frame_info): Restructure function.
2761
6455a49e
AK		 27622013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2763
2764 * config/s390/s390.c (struct s390_frame_layout): New field
2765 gpr_save_slots.
2766 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2767 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2768 regs_ever_clobbered to char*.
adcfd489
UB		 2769 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2770 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6455a49e
AK		 2771 (s390_register_info): Enable FPR save slots. Move/Copy some
2772 functionality into ...
2773 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2774 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2775 function.
2776 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2777 not imply a stack frame.
2778 (s390_init_frame_layout): Remove variable clobbered_regs.
2779 (s390_update_register_info): Remove function.
2780 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2781 cannot be used for register renaming.
2782 (s390_hard_regno_scratch_ok): New function.
2783 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2784 (s390_initial_elimination_offset): Change offset calculation of
2785 the return address pointer.
adcfd489 2786 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6455a49e 2787 (restore_gprs): Set frame related flag.
adcfd489 2788 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6455a49e
AK		 2789 (s390_emit_prologue): Call s390_register_info instead of
2790 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2791 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2792 (s390_optimize_prologue): Call s390_optimize_register_info.
2793 Try to remove also FPR slot save/restore INSNs. Remove frame
2794 related flags from restore INSNs.
2795
59d96342
DD		 27962013-10-08 DJ Delorie <dj@redhat.com>
2797
0dc7adad
DD		 2798 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2799 (movhi): Likewise.
2800
59d96342
DD		 2801 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2802 avoid conflict with the MI use of %c.
2803 * config/rl78/rl78-real.md: change %c to %C throughout.
2804 * config/rl78/rl78-virt.md: Likewise.
2805
a81169d8
JH		 28062013-10-08 Jan Hubicka <jh@suse.cz>
2807
2808 * config/i386/i386.c (ix86_option_override_internal): Switch
2809 to SSE math for -ffast-math when target ISA supports SSE2.
2810
481d1b81
AM		 28112013-10-08 Andrew MacLeod <amacleod@redhat.com>
2812
2813 * tree-flow.h: Remove some prototypes.
2814 * tree.h: Remove some protypes, add a couple.
2815 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2816 using_eh_for_cleanups_p): Add interface routines for front ends.
2817 * tree-eh.h: New file. Add protoptyes.
2818 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2819 (add_stmt_to_eh_lp_fn): Make static.
2820 (lower_try_finally): Use new using_eh_for_cleanups_p.
2821 * emit-rtl.c: Include tree-eh.h.
2822 * gimple.h: Include tree-eh.h.
2823
0e6a0e48
MG		 28242013-10-08 Marc Glisse <marc.glisse@inria.fr>
2825
2826 PR tree-optimization/58480
2827 * tree-vrp.c (infer_nonnull_range): New function.
2828 (infer_value_range): Call infer_nonnull_range.
2829
ef6179d1
DC		 28302013-10-08 Dehao Chen <dehao@google.com>
2831
2832 PR tree-optimization/58619
2833 * tree-inline.c (copy_phis_for_bb): Combine location data
2834 only if non-null.
2835
07d964d5
ZC		 28362013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2837
2838 PR target/58423
2839 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2840 RTX_FRAME_RELATED_P on INSN.
2841
19e9b2a3
BS		 28422013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2843
2844 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2845 (altivec_expand_vec_perm_const): Call it.
2846
0cf68694
BS		 28472013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2848
2849 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2850 sequences for LE VSX loads and stores at expand time.
2851 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2852 prototype.
2853 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2854 (rs6000_gen_le_vsx_permute): New.
2855 (rs6000_gen_le_vsx_load): New.
2856 (rs6000_gen_le_vsx_store): New.
2857 (rs6000_gen_le_vsx_move): New.
2858 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2859 (*vsx_le_perm_load_v4si): New.
2860 (*vsx_le_perm_load_v8hi): New.
2861 (*vsx_le_perm_load_v16qi): New.
2862 (*vsx_le_perm_store_v2di): New.
2863 (*vsx_le_perm_store_v4si): New.
2864 (*vsx_le_perm_store_v8hi): New.
2865 (*vsx_le_perm_store_v16qi): New.
2866 (*vsx_xxpermdi2_le_<mode>): New.
2867 (*vsx_xxpermdi4_le_<mode>): New.
2868 (*vsx_xxpermdi8_le_V8HI): New.
2869 (*vsx_xxpermdi16_le_V16QI): New.
2870 (*vsx_lxvd2x2_le_<mode>): New.
2871 (*vsx_lxvd2x4_le_<mode>): New.
2872 (*vsx_lxvd2x8_le_V8HI): New.
2873 (*vsx_lxvd2x16_le_V16QI): New.
2874 (*vsx_stxvd2x2_le_<mode>): New.
2875 (*vsx_stxvd2x4_le_<mode>): New.
2876 (*vsx_stxvd2x8_le_V8HI): New.
2877 (*vsx_stxvd2x16_le_V16QI): New.
2878
9520e1eb
RL		 28792013-10-07 Renlin Li <Renlin.Li@arm.com>
2880
2881 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2882
36c0bd4f
AK		 28832013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2884
2885 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2886 loop to work also for 31bit ABI.
2887 Save the stack pointer for frame_size > 0.
2888
ee163e72
AK		 28892013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2890
2891 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2892 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2893 constraint letters from expanders.
2894 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2895 retry count to general_operand.
2896 ("tabort"): Give operand 0 a mode.
2897 ("tabort_1"): Add mode and constraint letter for operand 0.
2898 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2899
45d99234
JL		 29002013-10-04 Jeff Law <law@redhat.com>
2901
2902 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2903
adcfd489
UB		 2904 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2905 out of ...
45d99234
JL		 2906 (thread_across_edge): Here. Call it.
2907
7c327f7b
CC		 29082013-10-04 Cary Coutant <ccoutant@google.com>
2909
2910 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
adcfd489 2911 discarding a location list expression (or a piece of one).
7c327f7b 2912
6867e128
JH		 29132013-10-03 Jan Hubicka <jh@suse.cz>
2914
65efa7e7
UB		 2915 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2916 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3700bce3 2917 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6867e128 2918
c7f36d55
JH		 29192013-10-03 Jan Hubicka <jh@suse.cz>
2920
65efa7e7 2921 * config/i386/i386.c (ix86_option_override_internal): Do not enable
c7f36d55
JH		 2922 accumulate-outgoing-args when producing unwind info.
2923
4f0bee4c
WM		 29242013-10-03 Wei Mi <wmi@google.com>
2925
2926 * lra-constraints.c (insert_move_for_subreg): New function
2927 extracted from simplify_operand_subreg.
2928 (simplify_operand_subreg): Add reload for paradoxical subreg.
2929
111c3f39
RX		 29302013-10-03 Rong Xu <xur@google.com>
2931
65efa7e7
UB		 2932 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2933 the candidate of builtin_expect such that we should fix the
2934 size/time estimation.
2935 (estimate_function_body_sizes): Do the acutally size/time fix-up
2936 for builtin_expect.
111c3f39 2937
942df739
RX		 29382013-10-03 Rong Xu <xur@google.com>
2939
65efa7e7
UB		 2940 * predict.c (tree_predict_by_opcode): Get the probability
2941 for builtin_expect from param builtin_expect_probability.
2942 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2943 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2944 * doc/invoke.texi: Add documentation for builtin-expect-probability.
942df739 2945
2284b034
MG		 29462013-10-03 Marc Glisse <marc.glisse@inria.fr>
2947
2948 PR c++/19476
2949 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2950 * calls.c (alloca_call_p): Use get_callee_fndecl.
2951 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2952 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2953 Likewise.
2954 (vrp_visit_stmt): Remove duplicated code.
2955
0609bdf2
MM		 29562013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2957
2958 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2959 ceildf2, btruncdf2, instead of vsx_* name.
2960
2961 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2962 iterators to only do V2DF and V4SF here. Move the DF code to
2963 rs6000.md where it is combined with SF mode. Replace <VSv> with
2964 just 'v' since only vector operations are handled with these insns
2965 after moving the DF support to rs6000.md.
2966 (vsx_sub<mode>3): Likewise.
2967 (vsx_mul<mode>3): Likewise.
2968 (vsx_div<mode>3): Likewise.
2969 (vsx_fre<mode>2): Likewise.
2970 (vsx_neg<mode>2): Likewise.
2971 (vsx_abs<mode>2): Likewise.
2972 (vsx_nabs<mode>2): Likewise.
2973 (vsx_smax<mode>3): Likewise.
2974 (vsx_smin<mode>3): Likewise.
2975 (vsx_sqrt<mode>2): Likewise.
2976 (vsx_rsqrte<mode>2): Likewise.
2977 (vsx_fms<mode>4): Likewise.
2978 (vsx_nfma<mode>4): Likewise.
2979 (vsx_copysign<mode>3): Likewise.
2980 (vsx_btrunc<mode>2): Likewise.
2981 (vsx_floor<mode>2): Likewise.
2982 (vsx_ceil<mode>2): Likewise.
2983 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2984 (vsx_sminsf3): Likewise.
2985 (vsx_fmadf4): Likewise.
2986 (vsx_fmsdf4): Likewise.
2987 (vsx_nfmadf4): Likewise.
2988 (vsx_nfmsdf4): Likewise.
2989 (vsx_cmpdf_internal1): Likewise.
2990
2991 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2992 simpler to select whether a target has SPE or traditional floating
2993 point support in iterators.
2994 (TARGET_DF_SPE): Likewise.
2995 (TARGET_SF_FPR): Likewise.
2996 (TARGET_DF_FPR): Likewise.
2997 (TARGET_SF_INSN): Macros to say whether floating point support
2998 exists for a given operation for expanders.
2999 (TARGET_DF_INSN): Likewise.
3000
3001 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3002 combining of SF/DF mode operations, using both traditional and VSX
3003 registers.
3004 (Fvsx): Likewise.
3005 (Ff): Likewise.
3006 (Fv): Likewise.
3007 (Fs): Likewise.
3008 (Ffre): Likewise.
3009 (FFRE): Likewise.
3010 (abs<mode>2): Combine SF/DF modes using traditional floating point
3011 instructions. Add support for using the upper DF registers with
3012 VSX support, and SF registers with power8-vector support. Update
3013 expanders for operations supported by both the SPE and traditional
3014 floating point units.
3015 (abs<mode>2_fpr): Likewise.
3016 (nabs<mode>2): Likewise.
3017 (nabs<mode>2_fpr): Likewise.
3018 (neg<mode>2): Likewise.
3019 (neg<mode>2_fpr): Likewise.
3020 (add<mode>3): Likewise.
3021 (add<mode>3_fpr): Likewise.
3022 (sub<mode>3): Likewise.
3023 (sub<mode>3_fpr): Likewise.
3024 (mul<mode>3): Likewise.
3025 (mul<mode>3_fpr): Likewise.
3026 (div<mode>3): Likewise.
3027 (div<mode>3_fpr): Likewise.
3028 (sqrt<mode>3): Likewise.
3029 (sqrt<mode>3_fpr): Likewise.
3030 (fre<Fs>): Likewise.
3031 (rsqrt<mode>2): Likewise.
3032 (cmp<mode>_fpr): Likewise.
3033 (smax<mode>3): Likewise.
3034 (smin<mode>3): Likewise.
3035 (smax<mode>3_vsx): Likewise.
3036 (smin<mode>3_vsx): Likewise.
3037 (negsf2): Delete SF operations that are merged with DF.
3038 (abssf2): Likewise.
3039 (addsf3): Likewise.
3040 (subsf3): Likewise.
3041 (mulsf3): Likewise.
3042 (divsf3): Likewise.
3043 (fres): Likewise.
3044 (fmasf4_fpr): Likewise.
3045 (fmssf4_fpr): Likewise.
3046 (nfmasf4_fpr): Likewise.
3047 (nfmssf4_fpr): Likewise.
3048 (sqrtsf2): Likewise.
3049 (rsqrtsf_internal1): Likewise.
3050 (smaxsf3): Likewise.
3051 (sminsf3): Likewise.
3052 (cmpsf_internal1): Likewise.
3053 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3054 (negdf2): Delete DF operations that are merged with SF.
3055 (absdf2): Likewise.
3056 (nabsdf2): Likewise.
3057 (adddf3): Likewise.
3058 (subdf3): Likewise.
3059 (muldf3): Likewise.
3060 (divdf3): Likewise.
3061 (fred): Likewise.
3062 (rsqrtdf_internal1): Likewise.
3063 (fmadf4_fpr): Likewise.
3064 (fmsdf4_fpr): Likewise.
3065 (nfmadf4_fpr): Likewise.
3066 (nfmsdf4_fpr): Likewise.
3067 (sqrtdf2): Likewise.
3068 (smaxdf3): Likewise.
3069 (smindf3): Likewise.
3070 (cmpdf_internal1): Likewise.
3071 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3072 (btrunc<mode>2): Delete separate expander, and combine with the
3073 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3074 (btrunc<mode>2_fpr): Likewise.
3075 (ceil<mode>2): Likewise.
3076 (ceil<mode>2_fpr): Likewise.
3077 (floor<mode>2): Likewise.
3078 (floor<mode>2_fpr): Likewise.
3079 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3080 Add support for using the upper registers with VSX and
3081 power8-vector. Move insns to be closer to the define_expands. On
3082 VSX systems, prefer the traditional form of FMA over the VSX
3083 version, since the traditional form allows the target not to
3084 overlap with the inputs.
3085 (fms<mode>4_fpr): Likewise.
3086 (nfma<mode>4_fpr): Likewise.
3087 (nfms<mode>4_fpr): Likewise.
3088
5bea0c6c 30892013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
65efa7e7 3090 Richard Earnshaw <richard.earnshaw@arm.com>
5bea0c6c
KT		 3091
3092 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3093 (struct mult_cost_table): Likewise.
3094 (struct mem_cost_table): Likewise.
3095 (struct fp_cost_table): Likewise.
3096 (struct vector_cost_table): Likewise.
3097 (cpu_cost_table): Likewise.
3098 * config/arm/arm.opt (mold-rts-costs): New option.
3099 (mnew-generic-costs): Likewise.
3100 * config/arm/arm.c (generic_extra_costs): New table.
3101 (cortexa15_extra_costs): Likewise.
3102 (arm_slowmul_tune): Use NULL as new costs.
3103 (arm_fastmul_tune): Likewise.
3104 (arm_strongarm_tune): Likewise.
3105 (arm_xscale_tune): Likewise.
3106 (arm_9e_tune): Likewise.
3107 (arm_v6t2_tune): Likewise.
3108 (arm_cortex_a5_tune): Likewise.
3109 (arm_cortex_a9_tune): Likewise.
3110 (arm_v6m_tune): Likewise.
3111 (arm_fa726te_tune): Likewise.
3112 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3113 (arm_cortex_tune): Use generict_extra_costs.
3114 (shifter_op_p): New function.
3115 (arm_unspec_cost): Likewise.
3116 (LIBCALL_COST): Define.
3117 (arm_new_rtx_costs): New function.
3118 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3119 table is available. Use old costs otherwise unless mnew-generic-costs
3120 is specified.
3121 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3122 (cpu_cost_table): Declare.
3123
24c56925
MS		 31242013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3125
3126 PR target/58460
3127 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3128 (*subs_mul_imm_<mode>)
3129 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3130 (*sub_<shift>_<mode>)
3131 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3132 Remove k constraint.
3133
2ab8f063
IB		 31342013-10-03 Ian Bolton <ian.bolton@arm.com>
3135
3136 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3137 code.
3138 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3139
79221839
TJ		 31402013-10-02 Teresa Johnson <tejohnson@google.com>
3141
3142 * predict.c (probably_never_executed): New function.
3143 (probably_never_executed_bb_p): Invoke probably_never_executed.
3144 (probably_never_executed_edge_p): Ditto.
3145 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3146 Treat profile insanities conservatively.
3147
d441afe0
JDA		 31482013-10-02 John David Anglin <danglin@gcc.gnu.org>
3149
3150 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3151
80f466c4
VM		 31522013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3153
3154 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3155 use it. Use smaller increase for scratch. Don't increase reject
3156 for early clobber scratch.
3157 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3158 setting eliminated regs except setting fp from hfp.
3159 (lra_eliminate): Check lra_insn_recog_data on NULL.
3160
6e228b4b
MM		 31612013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3162
3163 PR target/58587
3164 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
65efa7e7 3165 setting -mvsx-timode by default until the underlying problem is fixed.
6e228b4b
MM		 3166 (RS6000_CPU, power7 defaults): Likewise.
3167
11b54a5a
UB		 31682013-10-02 Uros Bizjak <ubizjak@gmail.com>
3169
3170 * config/x-linux (host-linux.o): Remove header dependencies.
3171 Use $(COMPILE) and $(POSTCOMPILE).
3172 * config/t-linux-android (linux-android.o): Ditto.
3173
bbc02b69
UB		 31742013-10-02 Uros Bizjak <ubizjak@gmail.com>
3175
3176 * Makefile.in (expmed.o-warn): Remove.
3177
440917de
AM		 31782013-10-02 Andrew MacLeod <amacleod@redhat.com>
3179
3180 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3181 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3182
99206ca9
TJ		 31832013-10-02 Teresa Johnson <tejohnson@google.com>
3184
3185 * dojump.c (do_jump_1): Divide probability between
3186 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3187
56e82b14
TT		 31882013-10-02 Tom Tromey <tromey@redhat.com>
3189
3190 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3191
3d9c733e
AM		 31922013-10-02 Andrew MacLeod <amacleod@redhat.com>
3193
3194 * tree-flow.h: Remove some prototypes.
3195 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3196 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3197 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3198 mark_virtual_phi_result_for_renaming): Relocate here.
3199 * tree-into-ssa.h: Add prototypes.
bbc02b69 3200 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3d9c733e
AM		 3201 single_pred_before_succ_order.
3202 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3203 (nonfreeing_call_p) Move to gimple.c.
3204 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3205 tree-ssa-phiopt.c.
3206 * basic-block.h (single_pred_before_succ_order): Add prototype.
3207 * gimple.c (nonfreeing_call_p): Relocate here.
3208 * gimple.h: Add prototype.
3209 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3210 * tree-ssa-dom.h: New file. Relocate prototypes here.
3211 * tree-ssa.h: Include tree-ssa-dom.h.
3212
78cedfb1
UB		 32132013-10-02 Uros Bizjak <ubizjak@gmail.com>
3214
3215 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3216 Use $(COMPILE) and $(POSTCOMPILE).
3217
3218 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3219
744730a4
AM		 32202013-10-02 Andrew MacLeod <amacleod@redhat.com>
3221
3222 * tree-flow.h: Remove some prototypes.
3223 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3224 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3225 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3226 propagate_tree_value*): Move from here to...
3227 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3228 propagate_tree_value*): Relocate here.
3229 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3230 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3231 * gimple-fold.c: Remove gimple-fold.h from include list.
3232 * tree-vrp.c: Remove gimple-fold.h from include list.
3233 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3234 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3235 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3236 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3237 * sese.c: Add tree-ssa-propagate.h to include list.
3238
826a536d
RB		 32392013-10-02 Richard Biener <rguenther@suse.de>
3240
3241 * tree-loop-distribution.c: Include tree-vectorizer.h for
3242 find_loop_location.
3243 (enum partition_kind): Remove PKIND_REDUCTION.
78cedfb1 3244 (struct partition_s): Remove has_writes member, add reduction_p member.
826a536d
RB		 3245 (partition_alloc): Adjust.
3246 (partition_builtin_p): Likewise.
3247 (partition_has_writes): Remove.
3248 (partition_reduction_p): New function.
3249 (partition_merge_into): Likewise.
3250 (generate_code_for_partition): Commonize builtin partition
3251 handling tail.
3252 (rdg_cannot_recompute_vertex_p): Remove.
3253 (already_processed_vertex_p): Likewise.
3254 (rdg_flag_vertex): Do not set has_writes.
3255 (classify_partition): Adjust.
3256 (rdg_build_partitions): Do not set has_writes, treat all
3257 partitions as useful.
78cedfb1 3258 (distribute_loop): Record number of library calls generated. Adjust.
826a536d
RB		 3259 (tree_loop_distribution): Report number of loops and library
3260 calls generated as opt-info.
3261
4b403ece
AM		 32622013-10-02 Andrew MacLeod <amacleod@redhat.com>
3263
3264 * tree-flow.h: Include new .h files. Move prototypes.
3265 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3266 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3267 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3268 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3269 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3270 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3271 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3272 * tree.h (get_ref_base_and_extent): Move prototype out.
78cedfb1 3273 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4b403ece
AM		 3274 tree-dfa.h.
3275 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3276 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3277 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3278 * tree-scalar-evolution.c: Include tree.h.
3279 * sese.c: Include tree.h.
3280 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3281 * dwarf2out.c: Include tree-dfa.h.
3282 * tree-chrec.c: Include tree.h.
3283 * tree-data-ref.c: Include tree.h.
3284
1d2151c6
YZ		 32852013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3286
3287 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3288 Fix whitespace.
3289
441ad147
RO		 32902013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3291
3292 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3293 Use $(COMPILE) and $(POSTCOMPILE).
3294 (sol2-cxx.o): Likewise.
3295 (sol2-stubs.o): Likewise.
3296 (sol2.o): Likewise.
3297 * config/x-solaris (host-solaris.o): Likewise.
3298
3299 * config/sparc/t-sparc (sparc.o): Remove.
3300 (sparc-c.o): Remove header dependencies.
3301 Use $(COMPILE) and $(POSTCOMPILE).
3302 * config/sparc/x-sparc: Likewise.
3303
2deaf8b0
JR		 33042013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3305
3306 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3307 * config/arc/arc700.md: Likewise.
3308 * config/arc/arc-modes.def: Likewise.
3309 * config/arc/arc-simd.h: Likewise.
3310 * config/arc/t-arc-uClibc: Likewise.
3311 * config/arc/t-arc-newlib: Likewise.
3312
f6fe771a
RL		 33132013-10-02 Renlin Li <renlin.li@arm.com>
3314
78cedfb1
UB		 3315 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3316 plus_constant.
f6fe771a
RL		 3317 (aarch64_expand_epilogue): Likewise.
3318
0916f876
YZ		 33192013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3320 Yufeng Zhang <yufeng.zhang@arm.com>
3321
3322 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3323 declaration.
3324 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3325 'base_in' represent a conversion and legal_cast_p_1 holds; set
3326 'base_in' with the returned value from get_unwidened.
3327
ebfcd719
KT		 33282013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3329
3330 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3331 plus_constant is not used.
3332
157ca3e9
WM		 33332013-10-01 Wei Mi <wmi@google.com>
3334
78cedfb1 3335 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
157ca3e9
WM		 3336 * config/i386/i386.md: Add define_peephole2 to
3337 break partial reg stall for cvtss2sd/cvtsd2ss.
3338
cd4dd8f0
JR		 33392013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3340
3341 * config/arc/arc.c (pass_arc_ifcvt::clone):
78cedfb1 3342 Update for ctxt_ -> m_ctxt change.
cd4dd8f0 3343
1465e5cf
JL		 33442013-10-01 Jeff Law <law@redhat.com>
3345
3346 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3347 outgoing_edge and intermediate_edge fields. Instead store the path.
3348 (redirection_data::hash): Hash on the last edge's destination index.
3349 (redirection_data::equal): Check the entire thread path.
3350 (lookup_redirectio_data): Corresponding changes.
3351 (create_edge_and_update_destination_phis): Likewise.
3352 (thread_single_edge): Likewise.
3353
25c606cb 33542013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
0ccbc132 3355 Diego Novillo <dnovillo@google.com>
25c606cb
JR		 3356
3357 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3358 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3359 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3360 versions of these patterns.
3361
0ccbc132
JR		 3362 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3363 (__builtin_arc_aligned): Likewise.
3364
f55d4a20
JR		 3365 * config/arc/arc.md: Expand adc_0 comment stating the intended
3366 purpose and why it isn't ready.
3367 Replace commented out call_value_via_label_mixed with a
3368 plain comment about bl_s.
3369
5719867d 3370 * config/arc/arc.c (stdio.h): Don't include directly.
5719867d
JR		 3371 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3372 Remove if (1) condition.
3373 (arc_encode_section_info): Fix comment.
3374
6462fab0
JR		 33752013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3376
3377 * config/arc/arc.c (arc_conditional_register_usage):
3378 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3379 Also set reg_alloc_order for DMA config regs.
3380
5d5f6720
JR		 33812013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3382 Jeremy Bennett <jeremy.bennett@embecosm.com>
3383
3384 * doc/install.texi (--with-cpu): Mention ARC.
3385 (arc-*-elf32): New paragraph.
3386 (arc-linux-uclibc): Likewise.
3387 * doc/md.texi (Machine Constraints): Add ARC part.
3388 * doc/invoke.texi: (menu): Add ARC Options.
3389 (Machine Dependent Options) <ARC Options>: Add synopsis.
3390 (node ARC Options): Add.
3391 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3392 (ARC Built-in Functions): New section defining
3393 generic ARC built-in functions.
3394 (ARC SIMD Built-in Functions): New section defining SIMD specific
3395 built-in functions.
3396 (Declaring Attributes of Functions): Extended
3397 description of short_call and long_call attributes for ARC and
3398 added index entries.
3399
526b7aee
SV		 34002013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3401 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3402 Joern Rennecke <joern.rennecke@embecosm.com>
3403 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3404 Brendan Kehoe <brendan@zen.org>
3405 Michael Eager <eager@eagercon.com>
3406 Simon Cook <simon.cook@embecosm.com>
3407 Jeremy Bennett <jeremy.bennett@embecosm.com>
3408
3409 * config/arc, common/config/arc: New directories.
3410
53426f6c
JR		 34112013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3412 Brendan Kehoe <brendan@zen.org>
3413 Simon Cook <simon.cook@embecosm.com>
3414
3415 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3416
1df9f5a9
AM		 34172013-10-01 Andrew MacLeod <amacleod@redhat.com>
3418
3419 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3420 * tree-ssa-coalesce.h: New. Move prototype to here.
3421 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3422 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3423 (gimple_can_coalesce_p): Move to...
3424 * gimple.c (gimple_can_coalesce_p): Here.
3425
80560f95
AM		 34262013-10-01 Andrew MacLeod <amacleod@redhat.com>
3427
3428 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3429 (dump_decl_set): Move to gimple.c.
3430 * gimple.h: Don't include tree-ssa-operands.h.
3431 (dump_decl_set): Add prototype.
3432 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3433 Move to gimple-ssa.h.
3434 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3435 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3436 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3437 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3438 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3439 than PHI_ARG_DEF.
3440 (dump_decl_set): Relocate here.
3441 * gimple-ssa.h: New file.
3442 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3443 Relocate from gimple.h.
3444 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3445 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3446 swap_tree_operands and remove non-ssa path.
3447 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3448 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3449 swap_ssa_operands.
3450 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3451 vect_is_simple_reduction_1): Use swap_ssa_operands.
3452 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3453 (enum need_phi_state): Move to tree-into-ssa.c.
3454 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3455 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3456 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3457 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3458 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3459 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3460 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3461 num_imm_uses): Move to ssa-iterators.h.
3462 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
7ab4168e 3463 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
80560f95
AM		 3464 tree-phinodes.h.
3465 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3466 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3467 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3468 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3469 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3470 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3471 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3472 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3473 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3474 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3475 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3476 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3477 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3478 (set_phi_nodes): Move to tree-phinodes.h.
3479 * tree-ssa-operands.h (enum ssa_op_iter_type,
3480 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3481 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3482 (dump_decl_set): Remove prototype.
3483 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3484 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3485 (set_phi_nodes): Relocate from tree-flow-inline.h.
3486 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3487 tree-flow-inline.h
3488 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3489 include list. Temporarily add gimple.h to include list.
3490 * ssa-iterators.h: New file.
3491 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3492 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3493 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3494 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3495 Relocate from tree-ssa-operands.h.
3496 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3497 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3498 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3499 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3500 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3501 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3502 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3503 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3504 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3505 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3506 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3507 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3508 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3509 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3510 Relocate from tree-flow-inline.h.
3511 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3512
8da00d65
VP		 35132013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3514
3515 * aarch64-simd.md
3516 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3517 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3518 Insert '\t' to output template.
3519 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3520 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3521 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3522 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3523
e2ebe1c2
UB		 35242013-10-01 Uros Bizjak <ubizjak@gmail.com>
3525
3526 * doc/install.texi (Host/target specific installation notes for GCC):
3527 Put @anchor before @heading.
3528 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3529 Use @email for email addresses.
3530
aee2d611
JL		 35312013-10-01 Jeff Law <law@redhat.com>
3532
3533 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3534 a vec. Only delete the path if we create one without successfully
3535 registering a jump thread.
3536 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3537 as a pointer.
3538 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3539 (paths): New vector of jump threading paths.
3540 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3541 (THREAD_PATH): New accessor macro for the entire thread path.
3542 (lookup_redirection_data): Get intermediate and final outgoing edge
3543 from the thread path.
3544 (create_edge_and_update_destination_phis): Copy the threading path.
3545 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3546 jump threading path.
3547 (ssa_redirect_edges): Get edges and block types from the jump threading
3548 path. Free the path vector.
3549 (thread_block): Get edges from the jump threading path. Look at the
3550 entire path to see if we thread to a loop exit. If we cancel a jump
3551 thread request, then free the path vector.
3552 (thread_single_edge): Get edges and block types from the jump threading
3553 path. Free the path vector.
3554 (thread_through_loop_header): Get edges and block types from the jump
3555 threading path. Free the path vector.
3556 (mark_threaded_blocks): Iterate over the vector of paths and store
3557 the path on the appropriate edge. Get edges and block types from the
3558 jump threading path.
3559 (mark_threaded_blocks): Get edges and block types from the jump
3560 threading path. Free the path vector.
3561 (thread_through_all_blocks): Use the vector of paths rather than
3562 a vector of 3-edge sets.
3563 (register_jump_thread): Accept pointer to a path vector rather
3564 than the path vector itself. Store the path vector for later use.
3565 Simplify.
3566
966f97ac 35672013-10-01 Jakub Jelinek <jakub@redhat.com>
8da00d65 3568 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
966f97ac
JJ		 3569
3570 PR target/58574
3571 * config/s390/s390.c (s390_split_branches): Modify check for table
3572 jump insns.
3573 (s390_chunkify_start): Rearrange table jump insn check in order to
3574 deal with compare and branch insns correctly.
3575
3a323a38
KV		 35762013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3577
3578 PR target/58578
3579 Revert
3580 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3581 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3582 define_insn_and_split.
3583 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3584 (shiftsi3_compare): New pattern.
3585 (rrx): New pattern.
3586 * config/arm/unspecs.md (UNSPEC_RRX): New.
3587
90444831
AM		 35882013-10-01 Alan Modra <amodra@gmail.com>
3589
3590 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3591 casing inout operands.
3592
818625cf
RB		 35932013-10-01 Richard Biener <rguenther@suse.de>
3594
3595 PR tree-optimization/58553
3596 * tree-loop-distribution.c (struct partition_s): Add niter member.
3597 (classify_partition): Populate niter member for the partition
3598 and properly identify whether the relevant store happens before
3599 or after the loop exit.
3600 (generate_memset_builtin): Use niter member from the partition.
3601 (generate_memcpy_builtin): Likewise.
3602
30f641cd
RS		 36032013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3604
3605 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3606 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3607
65d3284b
RS		 36082013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3609
3610 * basic-block.h (edge_list): Prefix member names with "m_".
3611 * context.h (context): Likewise.
3612 * domwalk.h (dom_walker): Likewise.
3613 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3614 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3615 * hash-table.h (hash_table): Likewise.
3616 * machmode.h (bit_field_mode_iterator): Likewise.
3617 * pass_manager.h (pass_list): Likewise.
3618 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3619 * tree-pass.h (pass_data): Likewise.
3620 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3621 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3622 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3623 * asan.c (pass_data_asan): Update accordingly.
3624 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3625 (control_dependences::control_dependences): Likewise.
3626 (control_dependences::~control_dependences): Likewise.
3627 (control_dependences::~control_dependences): Likewise.
3628 (control_dependences::get_edges_dependent_on): Likewise.
3629 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3630 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3631 * context.c (gcc::context::context): Likewise.
3632 * cprop.c (pass_rtl_cprop::clone): Likewise.
3633 * domwalk.c (dom_walker::walk): Likewise.
3634 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3635 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3636 * mode-switching.c (pass_mode_switching::clone): Likewise.
3637 * passes.c (opt_pass::opt_pass): Likewise.
3638 (pass_manager::pass_manager): Likewise.
3639 * predict.c (pass_strip_predict_hints::clone): Likewise.
3640 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3641 (pass_split_all_insns::clone): Likewise.
3642 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3643 Likewise.
3644 (bit_field_mode_iterator::next_mode): Likewise.
3645 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3646 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3647 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3648 * tree-complex.c (pass_lower_complex::clone): Likewise.
3649 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3650 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3651 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3652 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3653 (pass_fold_builtins::clone): Likewise.
3654 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3655 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3656 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3657 (pass_cd_dce::clone): Likewise.
3658 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3659 (pass_phi_only_cprop::clone): Likewise.
3660 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3661 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3662 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3663 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3664 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3665 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3666 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3667 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3668 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3669 * tree-vrp.c (pass_vrp::clone): Likewise.
3670 * tsan.c (pass_tsan::clone): Likewise.
3671
f66d0891
JJ		 36722013-09-30 Jakub Jelinek <jakub@redhat.com>
3673
d2a365a8
JJ		 3674 PR middle-end/58564
3675 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3676 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3677
f66d0891
JJ		 3678 PR middle-end/58564
3679 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3680 optimization, punt if sign_bit_p looked through any zero extension.
3681
05357ac3
TJ		 36822013-09-30 Teresa Johnson <tejohnson@google.com>
3683
3684 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3685 Update redirected out edge count in joiner case.
3686 (ssa_redirect_edges): Common the joiner and non-joiner cases
3687 so that joiner case gets profile updates.
3688
ca406576
RB		 36892013-09-30 Richard Biener <rguenther@suse.de>
3690
3691 PR tree-optimization/58554
e2ebe1c2
UB		 3692 * tree-loop-distribution.c (classify_partition): Require
3693 unconditionally executed stores for memcpy and memset recognition.
ca406576
RB		 3694 (tree_loop_distribution): Calculate dominance info.
3695
92d649c4
VK		 36962013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3697
3698 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3699 (NO_PROFILE_COUNTERS): Likewise.
3700 (PROFILE_HOOK): Likewise.
3701 (FUNCTION_PROFILER): Likewise.
3702 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3703
bd9534e2
IS		 37042013-09-30 Iain Sandoe <iain@codesourcery.com>
3705
3706 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3707 calls and defines in TARGET_MACHO conditional.
3708 (load_macho_picbase_di): Likewise.
3709 (reload_macho_picbase): Likewise.
3710 (reload_macho_picbase_si): Likewise.
3711 (reload_macho_picbase_di): Likewise.
3712 (nonlocal_goto_receiver): Likewise.
3713
4494fbc9
NC		 37142013-09-30 Nick Clifton <nickc@redhat.com>
3715
3716 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3717 that use the MSP430X ISA.
3718 (msp430_option_override): Scan -mmcu command line option for any
3719 MCU name that supports the MSP430X ISA.
3720 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3721 -mmcu options which enable the MSP430X ISA.
3722
791d9044
RB		 37232013-09-30 Richard Biener <rguenther@suse.de>
3724
3725 PR middle-end/58532
3726 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3727 before looking for setjmp-like calls.
3728
e1c5c877
IS		 37292013-09-29 Iain Sandoe <iain@codesourcery.com>
3730
3731 PR target/10901
3732 * config/darwin-protos.h (machopic_get_function_picbase): New.
3733 * config/darwin.c (machopic_get_function_picbase): New.
3734 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3735 label for a new func. (load_macho_picbase_di): Likewise.
3736 (reload_macho_picbase): New expand.
3737 (reload_macho_picbase_si): New insn.
3738 (reload_macho_picbase_di): New insn.
3739 (nonlocal_goto_receiver): New define and split.
3740 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3741 (unspecv enum): Add UNSPECV_NLGR.
3742
749278c5
IS		 37432013-09-29 Iain Sandoe <iain@codesourcery.com>
3744
3745 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3746 that altivec registers are correctly sized on Darwin.
3747
2c43a51c
IS		 37482013-09-29 Iain Sandoe <iain@codesourcery.com>
3749
3750 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3751 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3752 * config/x-darwin (host-darwin.o): Likewise.
3753 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3754 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3755 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3756
7a5911d3
UB		 37572013-09-29 Uros Bizjak <ubizjak@gmail.com>
3758
3759 * doc/invoke.texi: Fix usage of @tie{} command.
3760
3f67a73d
EB		 37612013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3762
3763 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3764
9ac2f538
JH		 37652013-09-28 Jan Hubicka <jh@suse.cz>
3766
3767 * config/i386/x86-tune.def: Add documentation for each of the options;
3768 add whitespace.
3769
b0ff06da
JH		 37702013-09-28 Jan Hubicka <jh@suse.cz>
3771
7a5911d3
UB		 3772 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3773 generic.
b0ff06da
JH		 3774 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3775 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3776 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3777 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3778 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3779
c3284718
RS		 37802013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3781
3782 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3783 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3784 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3785 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3786 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3787 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3788 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3789 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3790 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3791 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3792 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3793 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3794 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3795 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3796 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3797 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3798 gimple.h, godump.c, graphite-clast-to-gimple.c,
3799 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3800 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3801 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3802 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3803 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3804 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3805 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3806 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3807 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3808 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3809 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3810 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3811 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3812 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3813 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3814 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3815 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3816 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3817 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3818 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3819 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3820 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3821 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3822 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3823 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3824 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3825 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3826 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3827 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3828 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3829 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3830 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3831 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3832 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3833 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3834 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3835 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3836 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3837 whitespace before "(".
3838
c6285bd7
SL		 38392013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3840
3841 * expr.h (extract_bit_field): Remove packedp parameter.
3842 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3843 from forward declaration.
3844 (store_split_bit_field): Remove packedp arg from calls to
3845 extract_fixed_bit_field.
3846 (extract_bit_field_1): Remove packedp parameter and packedp
3847 argument from recursive calls and calls to extract_fixed_bit_field.
3848 (extract_bit_field): Remove packedp parameter and corresponding
3849 arg to extract_bit_field_1.
3850 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3851 to issue warnings.
3852 (extract_split_bit_field): Remove packedp arg from call to
3853 extract_fixed_bit_field.
3854 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3855 (copy_blkmode_from_reg): Likewise.
3856 (copy_blkmode_to_reg): Likewise.
3857 (read_complex_part): Likewise.
3858 (store_field): Likewise.
3859 (expand_expr_real_1): Likewise.
3860 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3861 to extract_bit_field.
3862 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3863 call to extract_bit_field.
3864 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3865 call to extract_bit_field.
3866 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3867 and special packedp behavior from -fstrict-volatile-bitfields
3868 documentation.
3869
522d4efc
JBG		 38702013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3871
3872 * lra-eliminations.c (init_elim_table): Guard value_p.
3873
19cdb489
MM		 38742013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3875
3876 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3877 DFmode, DImode, and SFmode in the upper VSX registers based on the
3878 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3879 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3880 of the VSX register handling.
3881
3882 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3883 (f32_sv): Likewise.
3884 (zero_extendsidi2_lfiwzx): Add support for loading into the
3885 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3886 only do VSX memory options on Altivec registers.
3887 (extendsidi2_lfiwax): Likewise.
3888 (extendsfdf2_fpr): Likewise.
3889 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3890 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3891 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3892 (movdi_internal64): Likewise.
3893
d6d11272
XDL		 38942013-09-27 Xinliang David Li <davidxl@google.com>
3895
3896 * opts.c (finish_options): Adjust parameters
3897 according to vect cost model.
3898 (common_handle_option): Set dynamic vect cost
3899 model for FDO.
3900 targhooks.c (default_add_stmt_cost): Compute stmt cost
3901 unconditionally.
3902 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3903 Use helper function.
3904 * tree-vectorizer.h (unlimited_cost_model): New function.
3905 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3906 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3907 function.
3908 (vect_enhance_data_refs_alignment): Ditto.
3909 * flag-types.h: New enum.
3910 * common/config/i386/i386-common.c (ix86_option_init_struct):
3911 No need to initialize vect_cost_model flag.
3912 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3913 unconditionally.
3914
ac1857a3
DN		 39152013-09-27 Diego Novillo <dnovillo@google.com>
3916
3917 * gimple.h (enum ssa_mode): Remove.
3918
0d0e2af6
PM		 39192013-09-27 Paulo Matos <pmatos@broadcom.com>
3920
3921 * cfgloop.h (number_of_loops): Fix typo in check for null.
3922
09dc585e
JJ		 39232013-09-27 Jakub Jelinek <jakub@redhat.com>
3924
3925 PR middle-end/58551
3926 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3927 are children of outermost saved_cfun's loop, and set it up to
3928 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3929 if loop != loop0 and SESE region contains bbs that belong to loop0.
3930
ec5a3504
RS		 39312013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3932
3933 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3934 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3935 (set_address_segment, set_address_base, set_address_index)
3936 (set_address_disp): Accept the argument unconditionally.
3937 (baseness): Remove must_be_base_p and must_be_index_p checks.
3938 (decompose_normal_address): Classify as much as possible in the
3939 main loop.
3940
f91aec98
RS		 39412013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3942
3943 * cse.c (count_reg_usage): Handle INT_LIST.
3944 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3945 * reginfo.c (reg_scan_mark_refs): Likewise.
3946 * reload1.c (eliminate_regs_1): Likewise.
3947
a9195970
IS		 39482013-09-27 Iain Sandoe <iain@codesourcery.com>
3949
3950 PR middle-end/58547
3951 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3952 signed.
3953
e6f1c509
RB		 39542013-09-27 Richard Biener <rguenther@suse.de>
3955
3956 PR tree-optimization/58459
3957 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3958 restriction not propagating into loops.
3959
84f48495
FW		 39602013-09-26 Florian Weimer <fw@deneb.enyo.de>
3961
3962 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3963 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3964 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3965
74fc8b8a
RB		 39662013-09-26 Richard Biener <rguenther@suse.de>
3967
7a5911d3 3968 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
74fc8b8a 3969
b4ada065
RB		 39702013-09-26 Richard Biener <rguenther@suse.de>
3971
3972 * alias.h (component_uses_parent_alias_set): Rename to ...
3973 (component_uses_parent_alias_set_from): ... this.
3974 * alias.c (component_uses_parent_alias_set): Rename to ...
3975 (component_uses_parent_alias_set_from): ... this and return
3976 the desired parent.
3977 (reference_alias_ptr_type_1): Use the result from
3978 component_uses_parent_alias_set_from instead of stripping
3979 components one at a time.
3980 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3981
78bca40d
AM		 39822013-09-26 Andrew MacLeod <amacleod@redhat.com>
3983
7a5911d3
UB		 3984 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3985 Move prototypes to...
78bca40d
AM		 3986 * tree-ssa-ter.h: New File. Move prototypes here.
3987 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3988 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3989 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7a5911d3 3990 from is_replaceable_p.
78bca40d
AM		 3991 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3992 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3993 (process_replaceable): Use ter_is_replaceable_p.
3994 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3995 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3996 newly refactored ssa_is_replaceable_p.
3997 * cfgexpand.c: Include tree-outof-ssa.h.
3998 * ssaexpand.h: Delete.
3999
ff2a63a7
AM		 40002013-09-26 Andrew MacLeod <amacleod@redhat.com>
4001
4002 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4003 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4004 tree-ssa.c
4005 (create_gimple_tmp): Delete.
4006 (get_expr_type, build_assign, build_type_cast): Move to...
4007 * gimple-builder.c: New File.
4008 (get_expr_type): Relocate from gimple.c.
4009 (build_assign, build_type_cast): Change to only create ssanames.
4010 * gimple.h: Move prototypes to...
4011 * gimple-builder.h: New File. Here.
4012 * tree-ssa.h: And here.
4013 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4014 count_uses_and_derefs): Relocate from gimple.c.
4015 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4016 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4017 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4018 gimple_replace_ssa_lhs.
4019 * asan.c: Include gimple-builder.h.
4020 * Makefile.in: Add gimple-builder.o.
4021
a2544177
RB		 40222013-09-26 Richard Biener <rguenther@suse.de>
4023
4024 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4025 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4026 (loe_visit_block): Use gcc_checking_assert.
4027 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4028 gimple_assign_ssa_name_copy_p.
4029 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4030 change.
4031
6d1ca9a0
DE		 40322013-09-26 David Edelsohn <dje.gcc@gmail.com>
4033
4034 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4035 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4036
46e950db
RB		 40372013-09-26 Richard Biener <rguenther@suse.de>
4038
4039 PR tree-optimization/58539
4040 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4041 the fact that debug statements are not taking part in loop-closed
4042 SSA construction.
4043
40ada30a
NC		 40442013-09-26 Nick Clifton <nickc@redhat.com>
4045
4046 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4047 time warning message.
4048 (msp430_print_operand_raw): Delete unused letter parameter.
4049 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4050 (msp430_print_operand_address): New function.
4051 (msp430_print_operand): Move address printing code from here to
4052 new function.
4053 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4054 assembler.
4055 (zero_extendpsisi2): Likewise.
4056 (extendpsisi2): New pattern.
4057 (andneghi3): New pattern.
4058
38e8f663
YR		 40592013-09-26 Yvan Roux <yvan.roux@linaro.org>
4060
4061 * config/aarch64/aarch64.opt (mlra): New option.
4062 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4063 (TARGET_LRA_P): Define.
4064
4167e8d4
EB		 40652013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4066
4067 * expr.c (expand_assignment): Remove obsolete comment.
4068
5254eac4
JL		 40692013-09-25 Jeff Law <law@redhat.com>
4070
4071 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4072 tree-ssa-threadupdate.h.
4073 (register_jump_thread): Similarly.
4074 * tree-ssa-threadupdate.h: New header file.
4075 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4076 * tree-vrp.c: Likewise.
4077 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4078 (thread_around_empty_blocks): Change type of path vector argument to
4079 an edge,type pair from just an edge. Initialize both elements when
4080 appending to a jump threading path. Tweak references to elements
4081 appropriately.
4082 (thread_across_edge): Similarly. Release memory for the elements
4083 as needed.
4084 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4085 (dump_jump_thread_path): New function broken out from
4086 register_jump_thread.
4087 (register_jump_thread): Use dump_jump_thread_path. Change type of
4088 path vector entries. Search the path for NULL edges and dump
4089 the path if one is found. Tweak the conversion of path to 3-edge
4090 form to use the block copy type information embedded in the path.
4091
20f114a3
YR		 40922013-09-25 Yvan Roux <yvan.roux@linaro.org>
4093
4094 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4095
3936bafc
YR		 40962013-09-25 Yvan Roux <yvan.roux@linaro.org>
4097 Vladimir Makarov <vmakarov@redhat.com>
4098
4099 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4100 from the least significant bit.
4101 (strip_address_mutations): Add bitfield operations handling.
4102 (must_be_index_p): Add shifting and rotate operations handling.
4103 (set_address_base): Use must_be_base_p predicate.
4167e8d4 4104 (set_address_index): Use must_be_index_p predicate.
3936bafc 4105
b86f6e9e
AI		 41062013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4107 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4108 Sergey Lega <sergey.s.lega@intel.com>
4109 Anna Tikhonova <anna.tikhonova@intel.com>
4110 Ilya Tocar <ilya.tocar@intel.com>
4111 Andrey Turetskiy <andrey.turetskiy@intel.com>
4112 Ilya Verbin <ilya.verbin@intel.com>
4113 Kirill Yukhin <kirill.yukhin@intel.com>
4114 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4115
4116 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4117 Use new names.
4118 (ix86_expand_vector_move_misalign): Support new unaligned load and
4119 stores and use new names.
4120 (CODE_FOR_sse2_storedqu): Rename to ...
4121 (CODE_FOR_sse2_storedquv16qi): ... this.
4122 (CODE_FOR_sse2_loaddqu): Rename to ...
4123 (CODE_FOR_sse2_loaddquv16qi): ... this.
4124 (CODE_FOR_avx_loaddqu256): Rename to ...
4125 (CODE_FOR_avx_loaddquv32qi): ... this.
4126 (CODE_FOR_avx_storedqu256): Rename to ...
4127 (CODE_FOR_avx_storedquv32qi): ... this.
4128 * config/i386/i386.md (fpint_logic): New.
4129 * config/i386/sse.md (VMOVE): Extend for AVX512.
4130 (VF): Ditto.
4131 (VF_128_256): New.
4132 (VF_512): Ditto.
4133 (VI_UNALIGNED_LOADSTORE): Ditto.
4134 (sse2_avx_avx512f): Ditto.
4135 (sse2_avx2): Extend for AVX512.
4136 (sse4_1_avx2): Ditto.
4137 (avx2_avx512f): New.
4138 (sse): Extend for AVX512.
4139 (sse2): Ditto.
4140 (sse4_1): Ditto.
4141 (avxsizesuffix): Ditto.
4142 (sseintvecmode): Ditto.
4143 (ssePSmode): Ditto.
4144 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4145 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4146 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4147 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4148 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4149 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4150 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4151 (STORENT_MODE): Extend for AVX512.
4152 (*absneg<mode>2): Replace constraint "x" with "v".
4153 (*mul<mode>3): Ditto.
4154 (*ieee_smin<mode>3): Ditto.
4155 (*ieee_smax<mode>3): Ditto.
4156 (avx_cmp<mode>3): Replace VF with VF_128_256.
4157 (*<sse>_maskcmp<mode>3_comm): Ditto.
4158 (<sse>_maskcmp<mode>3): Ditto.
4159 (<sse>_andnot<mode>3): Extend for AVX512.
4160 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4161 (<code><mode>3, fpint_logic): New.
4162 (*<code><mode>3): Extend for AVX512.
4163 (avx512flogicsuff): New.
4164 (avx512f_<logic><mode>): Ditto.
4165 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4166 VF_128_256.
4167 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4168 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4169 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4170 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4171 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4172 (xop_vpermil2<mode>3): Ditto.
4173 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4174 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4175 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4176 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4177
4d44d03c
TT		 41782013-09-25 Tom Tromey <tromey@redhat.com>
4179
4180 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4181 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4182 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4183 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4184 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4185 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4186 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4187 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4188 (GRAPHITE_HTAB_H): Remove.
4189
86aaf68d
TT		 41902013-09-25 Tom Tromey <tromey@redhat.com>
4191
4192 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4193
c987deb8
TT		 41942013-09-25 Tom Tromey <tromey@redhat.com>
4195
4196 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4197
9439747e
TT		 41982013-09-25 Tom Tromey <tromey@redhat.com>
4199
4200 * config/i386/t-i386 (i386.o): Remove.
4201 (i386-c.o): Use COMPILE and POSTCOMPILE.
4202
aefc31a1
TT		 42032013-09-25 Tom Tromey <tromey@redhat.com>
4204
4205 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4206
b6541edc
TT		 42072013-09-25 Tom Tromey <tromey@redhat.com>
4208
4209 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4210 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4211 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4212 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4213 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4214 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4215 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4216 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4217 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4218 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4219 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4220 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4221 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4222 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4223 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4224 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4225 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4226 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4227 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4228 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4229 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4230 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4231 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4232 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4233 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4234 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4235 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4236 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4237 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4238 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4239 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4240 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4241 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4242 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4243 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4244 (graphite-clast-to-gimple.o, graphite-dependences.o)
4245 (graphite-interchange.o, graphite-poly.o)
4246 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4247 (graphite-optimize-isl.o, tree-vect-loop.o)
4248 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4249 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4250 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4251 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4252 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4253 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4254 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4255 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4256 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4257 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4258 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4259 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4260 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4261 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4262 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4263 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4264 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4265 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4266 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4267 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4268 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4269 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4270 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4271 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4272 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4273 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4274 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4275 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4276 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4277 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4278 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4279 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4280 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4281 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4282 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4283 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4284 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4285 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4286 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4287 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4288 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4289 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4290 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4291 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4292 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4293 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4294 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4295 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4296 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4297 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4298 (gcov-dump.o): Remove.
4299 (default-c.o): Use COMPILE and POSTCOMPILE.
4300 (CFLAGS-gcc.o): New variable.
4301 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4302
8c796d19
TT		 43032013-09-25 Tom Tromey <tromey@redhat.com>
4304
4305 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4306 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4307 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4308 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4309 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4310 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4311 (c-family/array-notation-common.o, c-family/stub-objc.o)
4312 (c-family/c-ubsan.o): Remove.
4313
0a6c2227
TT		 43142013-09-25 Tom Tromey <tromey@redhat.com>
4315
4316 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4317
f3bc55f0
TT		 43182013-09-25 Tom Tromey <tromey@redhat.com>
4319
4320 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4321 to add -DENABLE_SHARED_LIBGCC.
4322 (gcc.o): Don't use subshell.
4323
c11c795e
TT		 43242013-09-25 Tom Tromey <tromey@redhat.com>
4325
4326 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4327 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4328 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4329 * configure, config.in: Rebuild.
4330
043378c3
TT		 43312013-09-25 Tom Tromey <tromey@redhat.com>
4332
4333 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4334 (COMPILE, POSTCOMPILE): New variables.
4335 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4336 (DEPFILES): New variable.
4337 Include ".Po" files.
4338 * configure.ac: Add checks for dependency checking.
4339 * configure, aclocal.m4: Regenerate.
4340
eab34643
TT		 43412013-09-25 Tom Tromey <tromey@redhat.com>
4342
4343 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7a5911d3 4344 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
eab34643 4345
d2db5e91
TT		 43462013-09-25 Tom Tromey <tromey@redhat.com>
4347
4348 * Makefile.in (generated_files): Add options.h,
4349 target-hooks-def.h, insn-opinit.h,
4350 common/common-target-hooks-def.h, pass-instances.def,
4351 c-family/c-target-hooks-def.h.
4352
af33629e
JL		 43532013-09-25 Jeff Law <law@redhat.com>
4354
4355 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4356 than foo[foo.length () - 1] to access last member in a vec.
4357 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4358
ed0d3051
RB		 43592013-09-25 Richard Biener <rguenther@suse.de>
4360
4361 PR middle-end/58521
4362 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4363
123485ca
JH		 43642013-09-25 Jan Hubicka <jh@suse.cz>
4365
4366 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4367 test.
4368
d70b0f1f
MP		 43692013-09-25 Marek Polacek <polacek@redhat.com>
4370
4371 PR sanitizer/58420
4372 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4373 when determining the type name.
4374
e3f02798
OE		 43752013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4376
4377 * config/sh/sh.md: Fix formatting.
4378
7a5911d3 43792013-09-24 Xinliang David Li <davidxl@google.com>
4f17aa0b 4380
7a5911d3
UB		 4381 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4382 max peel iterations parameter.
4f17aa0b
XDL		 4383 * param.def: New parameter.
4384 * doc/invoke.texi: Document New parameter.
4385
0498471b
CL		 43862013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4387
4388 * gimple-pretty-print.c: Various whitespace tweaks.
4389 * tree-core.h: Likewise.
4390 * tree-pretty-print.c: Likewise.
4391 * tree-ssa-alias.c: Likewise.
4392 * tree-ssa-copy.c: Likewise.
4393 * tree-ssanames.c: Likewise.
4394 * tree-ssanames.h: Likewise.
4395 * tree-vrp.c: Likewise.
4396
5a8c2b57
AM		 43972013-09-24 Alan Modra <amodra@gmail.com>
4398
4399 PR middle-end/57134
4400 PR middle-end/57586
7a5911d3
UB		 4401 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4402 for output operands that disallow regs. Don't use EXPAND_WRITE on
4403 inout operands.
5a8c2b57 4404
f40333af
RB		 44052013-09-24 Richard Biener <rguenther@suse.de>
4406
4407 PR middle-end/58513
4408 * tree.c (reference_alias_ptr_type): Move ...
4409 * alias.c (reference_alias_ptr_type): ... here and implement
4410 in terms of the new reference_alias_ptr_type_1.
4411 (ref_all_alias_ptr_type_p): New helper.
4412 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4413 use ref_all_alias_ptr_type_p.
4414 (get_deref_alias_set): Add flag_strict_aliasing check here.
4415 (reference_alias_ptr_type_1): New function, split out from ...
4416 (get_alias_set): ... here.
4417 (alias_ptr_types_compatible_p): New function.
4418 * alias.h (reference_alias_ptr_type): Declare.
4419 (alias_ptr_types_compatible_p): Likewise.
4420 * tree.h (reference_alias_ptr_type): Remove.
4421 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4422 to compare MEM_REF alias types.
4423
583e8bf5
RB		 44242013-09-24 Richard Biener <rguenther@suse.de>
4425
4426 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4427
5e8586d7
MM		 44282013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4429
19c3e797
MM		 4430 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4431 reload helper function arrays into a single array reg_addr.
4432 (reload_fpr_gpr): Likewise.
4433 (reload_gpr_vsx): Likewise.
4434 (reload_vsx_gpr): Likewise.
4435 (struct rs6000_reg_addr): Likewise.
4436 (reg_addr): Likewise.
4437 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4438 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4439 (rs6000_init_hard_regno_mode_ok): Likewise.
4440 (rs6000_secondary_reload_direct_move): Likewise.
4441 (rs6000_secondary_reload): Likewise.
4442
5e8586d7
MM		 4443 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4444 constraints: wu, ww, and wy. Repurpose wv constraint added during
4445 power8 changes. Put wg constraint in alphabetical order.
4446
4447 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4448 for future work to add ISA 2.07 VSX single precision support.
4449 (-mvsx-scalar-double): Change default from -1 to 1, update
4450 documentation comment.
4451 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4452 (-mupper-regs-df): New debug switch to control whether DF values
4453 can go in the traditional Altivec registers.
4454 (-mupper-regs-sf): New debug switch to control whether SF values
4455 can go in the traditional Altivec registers.
4456
4457 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4458 and wy constraints.
4459 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4460 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4461 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4462 (rs6000_debug_legitimate_address_p): Print if we are running
4463 before, during, or after reload.
4464 (rs6000_secondary_reload): Add a comment.
4465 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4466
4467 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4468 constraints. Update documentation string.
4469 (wd constraint): Likewise.
4470 (wf constraint): Likewise.
4471 (wg constraint): Likewise.
4472 (wn constraint): Likewise.
4473 (ws constraint): Likewise.
4474 (wt constraint): Likewise.
4475 (wx constraint): Likewise.
4476 (wz constraint): Likewise.
4477 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4478 instructions.
4479 (ww constraint): Likewise.
4480 (wy constraint): Likewise.
358e1993
MM		 4481 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4482 in the previous submissions.
5e8586d7
MM		 4483 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4484
e5af9ddd
RS		 44852013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4486
4487 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4488 (REG_BR_PROB): Say that the probability is stored in an int_list.
4489 * reg-notes.def: Update commentary to mention INT_LIST.
4490 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4491 (INT_LIST): New rtx.
4492 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4493 * rtlanal.c (int_reg_note_p): New function.
4494 (alloc_reg_note): Assert that the note does not have an int argument.
4495 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4496 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4497 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4498 rather than an INSN_LIST. Handle INT_LIST.
4499 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4500 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4501 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4502 Manipulate them as ints rather than rtxes.
4503 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4504 * regmove.c (copy_src_to_dest): Likewise.
4505 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4506
4507 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4508 into the cases that need it.
4509 * config/arm/arm.c (arm_unwind_emit): Likewise.
4510
4511 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4512 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4513 * loop-doloop.c (add_test, doloop_modify): Likewise.
4514 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4515 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4516 * predict.c (combine_predictions_for_insn): Likewise.
4517 * print-rtl.c (print_rtx): Handle INT_LIST.
4518 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4519 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4520 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4521 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4522 (ix86_split_fp_branch, predict_jump): Likewise.
4523 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4524 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4525 * config/spu/spu.c (ea_load_store_inline): Likewise.
4526
4527 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4528 value of a REG_BR_PROB note.
4529 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4530 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4531 * emit-rtl.c (try_split): Likewise.
4532 * predict.c (br_prob_note_reliable_p): Likewise.
4533 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4534 * reorg.c (mostly_true_jump): Likewise.
4535 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4536 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4537 * config/i386/i386.c (ix86_print_operand): Likewise.
4538 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4539 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4540 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4541 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4542 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4543 * config/sparc/sparc.c (output_cbranch): Likewise.
4544 * config/spu/spu.c (get_branch_target): Likewise.
4545 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4546 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4547
450ad0cd
JH		 45482013-09-23 Jan Hubicka <jh@suse.cz>
4549
4550 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4551 for ipa-devirt.
4552 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4553 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4554 of external calls
4555 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4556 (gimple_fold_call): Dump inconsistent devirtualizations; add
4557 sanity check for type based devirtualizations.
4558 * ipa-prop.c: Include ipa-utils.h
4559 (ipa_intraprocedural_devirtualization): Add sanity check.
4560 (try_make_edge_direct_virtual_call): Likewise.
4561
cefce769
EB		 45622013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4563
4564 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4565 assignment statements.
4566
a895a2b8
KV		 45672013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4568
4569 * gimple-pretty-print.c (dump_ssaname_info): New function.
4570 (dump_gimple_phi): Call it.
4571 (pp_gimple_stmt_1): Likewise.
4572 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4573 (range_info_def): Declare.
4574 * tree-pretty-print.c (pp_double_int): New function.
4575 (dump_generic_node): Call it.
4576 * tree-pretty-print.h (pp_double_int): Declare.
4577 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4578 * tree-ssanames.h (range_info_def): New structure.
4579 (value_range_type): Move definition here.
4580 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4581 Declare.
4582 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4583 initialization.
4584 (set_range_info): New function.
4585 (get_range_info): Likewise.
4586 (duplicate_ssa_name_range_info): Likewise.
4587 (duplicate_ssa_name_fn): Check pointer type and call
4588 duplicate_ssa_name_range_info.
4589 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4590 * tree-vrp.c (value_range_type): Remove definition, now in
4591 tree-ssanames.h.
7a5911d3 4592 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
a895a2b8
KV		 4593 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4594 (SSA_NAME_RANGE_INFO): New macro.
4595
984af6ac
RB		 45962013-09-23 Richard Biener <rguenther@suse.de>
4597
4598 PR tree-optimization/58464
4599 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7a5911d3 4600 (phi_trans_add): Change to add conditionally on being not present.
984af6ac
RB		 4601 (phi_translate_1): Remove recursion detection here.
4602 (phi_translate): Pre-seed the cache with NULL to catch
4603 recursion here in a more generic way.
4604 (bitmap_find_leader): Adjust comment.
4605 (get_representative_for): Dump value-numbers.
4606 (create_expression_by_pieces): Likewise.
4607 (insert_into_preds_of_block): Likewise.
4608
0e26cf79
CB		 46092013-09-23 Christian Bruel <christian.bruel@st.com>
4610
4611 PR target/58475
4612 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4613 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4614
4d0e5b54
JG		 46152013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4616
4617 Revert r202780:
4618 2013-09-20 Renlin Li <renlin.li@arm.com>
4619
7a5911d3
UB		 4620 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4621 plus_constant.
4d0e5b54
JG		 4622 (aarch64_expand_epilogue): Likewise.
4623 (aarch64_legitimize_reload_address): Likewise.
4624
ca081cc8
EB		 46252013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4626
4627 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4628 NULL_TREE before pushing them onto the vector. Likewise for labels.
4629
8fc53a5f
EB		 46302013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4631
4632 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4633 comparison operators when -fno-trapping-math is in effect.
4634 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4635 comparison operators in TFmode and assert that unsupported operators
4636 cannot reach here.
4637 (ia64_print_operand): Likewise.
4638
46392013-09-21 Jan Hubicka <jh@suse.cz>
0ca6c49f
JH		 4640
4641 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4642 (sse_typeless_stores): Enable for core
4643 (sse_load0_by_pxor): Likewise.
4644 (four_jump_limit): Disable for core.
4645 (pad_returns): Likewise.
4646 (avoid_vector_decode): Likewise.
4647 (fuse_cmp_and_branch): Enable for cores.
4648 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4649
578e51e6
JDA		 46502013-09-20 John David Anglin <danglin@gcc.gnu.org>
4651
96ce28a3
JDA		 4652 PR middle-end/56791
4653 * config/pa/pa.c (pa_option_override): Disable auto increment and
4654 decrement instructions until reload is completed.
4655
578e51e6
JDA		 4656 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4657 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4658 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4659
6fcd3a13 46602013-09-20 DJ Delorie <dj@redhat.com>
7a5911d3 4661 Nick Clifton <nickc@redhat.com>
6fcd3a13
DD		 4662
4663 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4664 (need_to_save): Save bank 0 on interrupts.
4665 (characterize_address): Strip far address wrappers.
4666 (rl78_as_legitimate_address): Likewise.
4667 (transcode_memory_rtx): Likewise.
4668 (rl78_peep_movhi_p): Disable this peephole after devirt.
4669 (rl78_propogate_register_origins): Forget all origins when a
4670 CLOBBER is seen.
4671 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4672 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4673 peephole2's.
4674 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4675 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4676 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4677 relocating.
4678 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4679
4a08db33
JDA		 46802013-09-20 John David Anglin <danglin@gcc.gnu.org>
4681
4682 * config/pa/pa.md: In "scc" insn patterns, change output template to
4683 handle const0_rtx in reg_or_0_operand operands.
4684
7691132c
JBG		 46852013-09-20 Martin Husemann <martin@NetBSD.org>
4686
4687 PR target/56875
4688 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4689 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4690
a3cc13cc
RB		 46912013-09-20 Richard Biener <rguenther@suse.de>
4692
4693 PR middle-end/58484
4694 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4695 remembering SSA name version and block index.
4696 (new_scev_info_str): Adjust.
4697 (hash_scev_info): Likewise. Also hash the block index.
4698 (eq_scev_info): Adjust.
4699 (find_var_scev_info): Likewise.
4700 (struct instantiate_cache_entry): Remove.
4701 (struct instantiate_cache_type): Use a htab to map name, block
4702 to chrec.
4703 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4704 (get_instantiated_value_entry): Likewise.
4705 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4706 (instantiate_scev_name): Adjust.
4707
925f3871
JL		 47082013-09-20 Jeff Law <law@redhat.com>
4709
4710 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4711
2888c331
YZ		 47122013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4713
4714 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4715 Call aarch64_simd_expand_args to update op[argc].
4716
0a811e96
BS		 47172013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4718
7a5911d3
UB		 4719 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4720 plugin argument.
0a811e96 4721
0078f462
BS		 47222013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4723
7a5911d3
UB		 4724 * gengtype.c (file_rules): Added rule for *.cc files.
4725 (get_output_file_with_visibility): Give fatal message when no
4726 rules found.
0078f462 4727
d20e48b0
RL		 47282013-09-20 Renlin Li <renlin.li@arm.com>
4729
4730 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4731 (aarch64_expand_epilogue): Likewise.
4732 (aarch64_legitimize_reload_address): Likewise.
4733
59bb1544
BE		 47342013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4735
4736 PR middle-end/57748
4737 * expr.c (expand_assignment): Remove misalignp code path.
4738
a5430019
MP		 47392013-09-20 Marek Polacek <polacek@redhat.com>
4740
4741 PR sanitizer/58413
4742 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4743 TYPE_PRECISION. Add asserts.
4744
9fed7f3a
RB		 47452013-09-20 Richard Biener <rguenther@suse.de>
4746
4747 PR tree-optimization/58453
4748 * tree-loop-distribution.c (distribute_loop): Apply the cost
4749 model for -ftree-loop-distribute-patterns, too.
4750
a820c834
RB		 47512013-09-20 Richard Biener <rguenther@suse.de>
4752
4753 PR middle-end/58473
4754 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4755 make type comparison less strict.
4756
3cbe17f7
AM		 47572013-09-20 Alan Modra <amodra@gmail.com>
4758
4759 * configure: Regenerate.
4760 * aclocal.m4: Regenerate.
4761
87fccdbb
MP		 47622013-09-20 Marek Polacek <polacek@redhat.com>
4763
4764 PR other/58467
4765 * doc/extend.texi: Document that attribute used is meant to be used
4766 on variables with static storage duration.
4767
8928eff3
JJ		 47682013-09-19 Jakub Jelinek <jakub@redhat.com>
4769
4770 PR tree-optimization/58472
4771 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4772 simd_lane_access set inv_p = false.
4773 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4774 the simduid magic VAR_DECL.
4775
b7b28c53
JH		 47762013-09-19 Jan Hubicka <jh@suse.cz>
4777
4778 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4779
83ae86f5
JL		 47802013-09-17 Jeff Law <law@redhat.com>
4781
4782 * tree-ssa-dom.c (record_temporary_equivalences): New function
4783 split out of dom_opt_dom_walker::after_dom_children.
4784 (dom_opt_dom_walker::thread_across_edge): Move common code
4785 in here from dom_opt_dom_walker::after_dom_children.
4786 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4787
9d532162
JH		 47882013-09-19 Jan Hubicka <jh@suse.cz>
4789
4790 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4791 (TARGET_GENERIC): Use PROCESOR_GENERIC
4792 (enum processor_type): Unify generic32 and 64.
4793 * i386.md (cpu): Likewise.
4794 * x86-tune.def (use_leave): Enable for generic32.
4795 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4796 * athlon.md: Change generic64 to generic in all occurences.
4797 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4798 (ix86_target_macros_internal): Likewise.
4799 * driver-i386.c (host_detect_local_cpu): Likewise.
7a5911d3
UB		 4800 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4801 to ..
9d532162
JH		 4802 (generic_memcpy, generic_memset, generic_cost): This one.
4803 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4804 (m_GENERIC32, m_GENERIC64): Remove.
4805 (m_GENERIC): Turn into one flag.
4806 (processor_target): Unify generic tunnings.
4807 (ix86_option_override_internal): Replace generic32/64 by generic.
4808 (ix86_issue_rate): Likewise.
4809 (ix86_adjust_cost): Likewise.
4810
188c7d00
JH		 48112013-09-19 Jan Hubicka <jh@suse.cz>
4812
4813 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4814 of speculative flag.
4815
4befd127
JJ		 48162013-09-19 Jakub Jelinek <jakub@redhat.com>
4817
4818 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4819 GOMP_sections_start, even if !exit_reachable.
4820
8a26ad39
VM		 48212013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4822
4823 * lra-constraints.c (need_for_all_save_p): Use macro
4824 HARD_REGNO_CALL_PART_CLOBBERED.
4825 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4826 set up pseudo conflict hard regs.
4827
09487185
MM		 48282013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4829
4830 PR target/58452
4831 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4832 operands.
4833
080cbf9e
VM		 48342013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4835
4836 PR rtl-optimization/58438
4837 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4838 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4839 from previous subpasses.
4840
a0f70fcb
RE		 48412013-09-18 Richard Earnshaw <rearnsha@arm.com>
4842
4843 * arm.c (arm_get_frame_offsets): Validate architecture supports
4844 LDRD/STRD before accepting the tuning preference.
4845 (arm_expand_prologue): Likewise.
4846 (arm_expand_epilogue): Likewise.
4847
0547c9b6
RB		 48482013-09-18 Richard Biener <rguenther@suse.de>
4849
4850 PR tree-optimization/58417
4851 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4852 have chrecs with symbols defined in the loop as operands.
4853 (chrec_fold_multiply): Likewise.
4854 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4855 parameters before folding binary operations.
4856 (struct instantiate_cache_entry_hasher): Remove.
4857 (struct instantiate_cache_type): Use a pointer-map.
4858 (instantiate_cache_type::instantiate_cache_type): New function.
4859 (instantiate_cache_type::get): Likewise.
4860 (instantiate_cache_type::set): Likewise.
4861 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4862 (get_instantiated_value_entry): Likewise.
4863 (global_cache): New global.
4864 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4865 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4866 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4867 (instantiate_scev_name): Adjust.
4868 (instantiate_scev): Construct global instead of local cache.
4869 (resolve_mixers): Likewise.
4870
309dc1aa
DM		 48712013-09-18 Daniel Morris <danielm@ecoscentric.com>
4872 Paolo Carlini <paolo.carlini@oracle.com>
4873
4874 PR c++/58458
4875 * doc/implement-cxx.texi: Fix references to the C++ standards.
4876
995a1b4a
JJ		 48772013-09-18 Jakub Jelinek <jakub@redhat.com>
4878
4879 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4880 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4881 simd_lane_access drs, update also DR_ALIGNED_TO.
4882
ce6923c5
MP		 48832013-09-18 Marek Polacek <polacek@redhat.com>
4884
4885 PR sanitizer/58411
4886 * doc/extend.texi: Document no_sanitize_undefined attribute.
4887 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4888 no_sanitize_undefined attribute.
4889
d30d00a2
NC		 48902013-09-18 Nick Clifton <nickc@redhat.com>
4891
4892 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4893 (ASM_DECLARE_FUNCTION_NAME): Define.
4894
4d9192b5
TS		 48952013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4896
4897 * compare-elim.c (find_comparison_dom_walker): New class
4898 (find_comparisons_in_bb): Rename to
4899 find_comparison_dom_walker::before_dom_children
4900 (find_comparisons): Adjust
4901 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4902 adjust.
4903 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4904 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4905 (init_walk_dominator_tree): Remove declaration.
4906 (fini_walk_dominator_tree): Remove declaration.
4907 * fwprop.c (single_def_use_dom_walker): New class
4908 (single_def_use_enter_block): Convert to
4909 single_def_use_dom_walker::before_dom_children.
4910 (single_def_use_leave_block): Convert to
4911 single_def_use_dom_walker::after_dom_children.
4912 (build_single_def_use_links): Adjust.
4913 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4914 class.
4915 (find_candidates_in_block): Convert to
4916 find_candidates_dom_walker::before_dom_children.
4917 (execute_strength_reduction): Adjust.
4918 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4919 (sese_dom_walker): New class.
4920 (sese_dom_walker::sese_dom_walker): New constructor.
4921 (sese_dom_walker::~sese_dom_walker): New destructor.
4922 (build_sese_conditions_before): Convert to
4923 sese_dom_walker::before_dom_children.
4924 (build_sese_conditions_after): Convert to
4925 sese_dom_walker::after_dom_children.
4926 (build_poly_scop): Adjust
4927 * tree-into-ssa.c (rewrite_dom_walker): New class
4928 (rewrite_enter_block): Convert to
4929 rewrite_dom_walker::before_dom_children.
4930 (rewrite_leave_block): Convert to
4931 rewrite_dom_walker::after_dom_children.
4932 (rewrite_update_dom_walker): New class.
4933 (rewrite_update_enter_block): Convert to
4934 rewrite_update_dom_walker::before_dom_children.
4935 (rewrite_update_leave_block): Convert to
4936 rewrite_update_dom_walker::after_dom_children.
4937 (rewrite_blocks, rewrite_into_ssa): Adjust.
4938 (mark_def_dom_walker): New class.
4939 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4940 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4941 (mark_def_sites_blocks): Convert to
4942 mark_def_dom_walker::before_dom_children.
4943 (mark_def_site_blocks): Remove.
4944 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4945 (tree_ssa_dominator_optimize): Adjust.
4946 (dom_thread_across_edge): Convert to method
4947 dom_opt_dom_walker::thread_across_edge.
4948 (dom_opt_enter_block): Convert to member function
4949 dom_opt_dom_walker::before_dom_children.
4950 (dom_opt_leave_block): Convert to member function
4951 dom_opt_dom_walker::after_dom_children.
4952 * tree-ssa-dse.c (dse_dom_walker): New class.
4953 (dse_enter_block): Convert to member function
4954 dse_dom_walker::before_dom_children.
4955 (tree_ssa_dse): Adjust.
4956 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4957 (determine_invariantness_stmt): Convert to method
4958 invariantness_dom_walker::before_dom_children.
4959 (determine_invariantness): Remove
4960 (move_computations_dom_walker): New class.
4961 (move_computations_stmt): Convert to method
4962 move_computations_dom_walker::before_dom_children.
4963 (move_computations, tree_ssa_lim): Adjust.
4964 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4965 (nt_init_block): Make method
4966 notrappping_dom_walker::before_dom_children.
4967 (nt_fini_block): Make
4968 method nontrapping_dom_walker::after_dom_children.
4969 (get_non_trapping): Adjust.
4970 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4971 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4972 (eliminate_leave_block): Make method.
4973 eliminate_dom_walker::after_dom_children.
4974 (eliminate): Adjust
4975 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4976 (strlen_enter_block): Make method
4977 strlen_dom_walker::before_dom_children.
4978 (strlen_leave_block): Make
4979 method strlen_dom_walker::after_dom_children.
4980 (tree_ssa_strlen): Adjust.
4981 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4982 (tree_ssa_uncprop): Adjust.
4983 (uncprop_leave_block): Make method
4984 uncprop_dom_walker::after_dom_children.
4985 (uncprop_leave_block): Make method
4986 uncprop_dom_walker::before_dom_children.
4987
3b3cc26b
BC		 49882013-09-18 Bin Cheng <bin.cheng@arm.com>
4989
7a5911d3
UB		 4990 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4991 branch in same basic block. Check both src and dest of the move insn.
3b3cc26b 4992
b5ba3a9b
NC		 49932013-09-17 Nick Clifton <nickc@redhat.com>
4994
4995 * config/rl78/rl78-real.md (bf): New pattern.
4996 (bt): New pattern.
4997 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4998 (rl78_print_operand): Do not put a # before a %B.
4999 * config/rl78/rl78.opt: Tweak doc strings.
5000
09d5094b
DD		 50012013-09-17 DJ Delorie <dj@redhat.com>
5002
5003 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5004 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5005 (rl78_option_override): Likewise, if -mallregs.
5006 (is_virtual_register): Likewise.
5007 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5008 (REGNO_OK_FOR_BASE_P): Likewise.
5009 * config/rl78/rl78.opt (-mallregs): New.
5010
2ba1dca1
NC		 50112013-09-17 Nick Clifton <nickc@redhat.com>
5012
5013 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5014 For interrupt functions: save all call clobbered registers if the
5015 interrupt handler is not a leaf function.
5016 (rl78_expand_prologue): Always recompute the frame information.
5017 For interrupt functions: only select bank 0 if one of the bank 0
5018 registers is going to be psuhed.
5019
2e7c3f21
DD		 50202013-09-17 DJ Delorie <dj@redhat.com>
5021
5022 * config/rl78/constraints.md: For each W* constraint, rename to C*
5023 and create a W* constraint that checks for an optional ES: prefix
5024 pattern also.
5025 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5026 (es_addr): New. Used to wrap far addresses.
5027 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5028 (rl78_es_base): New.
5029 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5030 wrapped far addresses.
5031 (rl78_print_operand_1): Unwrap far addresses before processing.
5032 (rl78_lo16): Wrap far addresses in unspecs.
5033 (rl78_es_addr): New.
5034 (rl78_es_base): New.
5035 (insn_ok_now): Check for not-yet-wrapped far addresses.
5036 (transcode_memory_rtx): Properly re-wrap far addresses.
5037
5c26a69a
SH		 50382013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5039
5040 * config/sparc/t-rtems: Add leon3 multilibs.
5041
181f5f3e
CH		 50422013-09-17 Cong Hou <congh@google.com>
5043
5044 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5045 when checking the dot production pattern. The type of rhs operand
5046 of multiply is now checked correctly.
5047
8d34e421
JL		 50482013-09-17 Jeff Law <law@redhat.com>
5049
5050 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5051 edge implied equivalences into successor phis.
5052 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5053 here from tree-ssa-threadedge.c.
5054 (mark_threaded_blocks): When threading through a joiner, if both
5055 successors of the joiner's clone reach the same block, verify the
5056 PHI arguments are equal. If not, cancel the jump threading request.
5057 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5058 tree-ssa-threadupdate.c
5059 (thread_across_edge): Don't check PHI argument equality when
5060 threading through joiner block here.
7a5911d3 5061
c152901f
AM		 50622013-09-17 Andrew MacLeod <amacleod@redhat.com>
5063
5064 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5065 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7a5911d3
UB		 5066 (warn_uninit, warn_uninitialized_vars,
5067 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5068 Move to tree-ssa-uninit.c.
78cedfb1 5069 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7a5911d3 5070 (has_undefined_value_p): New. Pass dependant parts of
c152901f
AM		 5071 ssa_undefined_value_p.
5072 (uninit_undefined_value_p): Use has_undefined_value_p.
7a5911d3
UB		 5073 (warn_uninit, warn_uninitialized_vars,
5074 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5075 Move from tree-ssa.c.
5076 * tree-ssa.h: Adjust prototypes.
c152901f 5077
b631d45a
JH		 50782013-09-17 Jan Hubicka <jh@suse.cz>
5079
5080 PR middle-end/58332
5081 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5082 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5083 FUNCTION_NOT_OPTIMIZED.
5084 * ipa-inline-analysis.c (compute_inline_parameters): Function
5085 not optimized is not inlinable unless it is alwaysinline.
5086 (inline_analyze_function): Force calls in not optimized
5087 function not inlinable.
5088
5b79657a
JH		 50892013-09-17 Jan Hubicka <jh@suse.cz>
5090
5091 PR middle-end/58329
5092 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5093 to return NULL.
5094 * ipa.c (function_and_variable_visibility): Likewise.
5095 * ipa-profile.c (ipa_profile): Likewise.
5096
27442c24
BE		 50972013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5098
5099 PR ipa/58398
5100 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5101 attribute, and don't inline the resolver in this case.
5102
753b9bc3
TJ		 51032013-09-17 Teresa Johnson <tejohnson@google.com>
5104
5105 * coverage.c (get_coverage_counts): Add missing newline.
5106
e3d3cfb4
KT		 51072013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5108
5109 PR tree-optimization/58088
5110 * fold-const.c (mask_with_trailing_zeros): New function.
5111 (fold_binary_loc): Make sure we don't recurse infinitely
5112 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5113 Use mask_with_trailing_zeros where appropriate.
5114
9a0ac98f
YR		 51152013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5116
5117 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5118 of 'prev' var to get better distance estimation.
5119
78bbd765
EB		 51202013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5121
5122 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5123 * tree-inline.c (is_parameter_of): New predicate.
5124 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5125 a parameter has been remapped.
5126 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5127 (optimize_inline_calls): Initialize transform_parameter.
5128 (copy_gimple_seq_and_replace_locals): Likewise.
5129 (tree_function_versioning): Likewise.
5130 (maybe_inline_call_in_expr): Likewise.
5131
cad055a4
NC		 51322013-09-17 Nick Clifton <nickc@redhat.com>
5133
5134 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5135 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5136 interrupt handlers.
5137 (is_attr_func): New function.
5138 (msp430_is_interrupt_func): New function.
5139 (is_naked_func): New function.
5140 (is_reentrant_func): New function.
5141 (is_critical_func): New function.
5142 (msp430_start_function): Add annotations for function attributes.
5143 (msp430_attr): New function.
5144 (msp430_attribute_table): New.
5145 (msp430_function_section): New function.
5146 (TARGET_ASM_FUNCTION_SECTION): Define.
5147 (msp430_builtin): New enum.
5148 (msp430_init_builtins): New function.
5149 (msp430_builtin_devl): New function.
5150 (msp430_expand_builtin): New function.
5151 (TARGET_INIT_BUILTINS): Define.
5152 (TARGET_EXPAND_BUILTINS): Define.
5153 (TARGET_BUILTIN_DECL): Define.
5154 (msp430_expand_prologue): Add support for naked, interrupt,
5155 critical and reentrant functions.
5156 (msp430_expand_epilogue): Likewise.
5157 (msp430_print_operand): Handle 'O' character.
5158 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5159 NO_TRAMPOLINES.
5160 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5161 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5162 (pushm): Use a 'n' rather than an 'i' constraint.
5163 (msp_return): Add generation of the interrupt return instruction.
5164 (disable_interrupts): New pattern.
5165 (enable_interrupts): New pattern.
5166 (push_intr_state): New pattern.
5167 (pop_intr_state): New pattern.
5168 (bic_SR): New pattern.
5169 (bis_SR): New pattern.
5170 * doc/extend.texi: Document MSP430 function attributes and builtin
5171 functions.
5172
deb6c11a
RB		 51732013-09-17 Richard Biener <rguenther@suse.de>
5174
5175 PR tree-optimization/58432
5176 * tree-loop-distribution.c (tree_loop_distribution): Also
5177 scan PHIs for outside loop uses and seed a partition from them.
5178
c068654b
BC		 51792013-09-17 Bin Cheng <bin.cheng@arm.com>
5180
5181 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5182 (restructure_reference): Call backtrace_base_for_ref.
5183
567969e4
AM		 51842013-09-17 Alan Modra <amodra@gmail.com>
5185
5186 PR target/57589
7a5911d3
UB		 5187 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5188 patch.
567969e4 5189
5c0029de
DD		 51902013-09-16 DJ Delorie <dj@redhat.com>
5191
5192 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5193 vregs location for RL78/G10.
5194 (rl78_expand_prologue): Avoid SEL on G10.
5195 (rl78_expand_epilogue): Likewise.
5196 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5197 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5198 __RL78_G10__ when appropriate.
5199 (ASM_SPEC): Pass -mg10 along to the assembler.
5200 * config/rl78/rl78.md (sel_rb): Disable for G10.
5201 * config/rl78/rl78.opt: Add -mg10 option.
5202 * config/rl78/t-rl78: Add -mg10 multilib.
5203
ea0f3e87
XDL		 52042013-09-16 Xinliang David Li <davidxl@google.com>
5205
5206 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5207 * omp-low.c (omp_max_vf): Ditto.
5208 (expand_omp_simd): Ditto.
5209 * tree-vectorizer.c (vectorize_loops): Ditto.
5210 (gate_vect_slp): Ditto.
5211 (gate_increase_alignment): Ditto.
5212 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5213 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5214 (gate_tree_vectorize): Name change.
5215 (tree_vectorize): Ditto.
5216 (pass_vectorize::gate): Call new function.
5217 (pass_vectorize::execute): Ditto.
7a5911d3 5218 * opts.c: O3 default setting change.
ea0f3e87
XDL		 5219 (finish_options): Check new flag.
5220 * doc/invoke.texi: Document new flags.
5221 * common.opt: New flags.
5222
4f3da2e5
AS		 52232013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5224
5225 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5226 (Condition Code): Don't reference it.
5227 * doc/tm.texi: Regenerate.
5228
aa940f7c
VM		 52292013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5230
5231 PR middle-end/58418
5232 * lra-constraints.c (undo_optional_reloads): Consider all optional
5233 reload even if it did not get a hard reg.
5234
e645e942
TJ		 52352013-09-16 Teresa Johnson <tejohnson@google.com>
5236
5237 * dumpfile.c (dump_loc): Remove newline emission.
5238 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5239 emission to dump_printf_loc calls where missing.
5240 (vect_mark_for_runtime_alias_test): Ditto.
5241 (vect_analyze_data_ref_dependence): Ditto.
5242 (vect_analyze_data_ref_dependences): Ditto.
5243 (vect_slp_analyze_data_ref_dependence): Ditto.
5244 (vect_slp_analyze_data_ref_dependences): Ditto.
5245 (vect_compute_data_ref_alignment): Ditto.
5246 (vect_update_misalignment_for_peel): Ditto.
5247 (vect_verify_datarefs_alignment): Ditto.
5248 (vector_alignment_reachable_p): Ditto.
5249 (vect_get_data_access_cost): Ditto.
5250 (vect_enhance_data_refs_alignment): Ditto.
5251 (vect_find_same_alignment_drs): Ditto.
5252 (vect_analyze_data_refs_alignment): Ditto.
5253 (vect_analyze_group_access): Ditto.
5254 (vect_analyze_data_ref_access): Ditto.
5255 (vect_analyze_data_ref_accesses): Ditto.
5256 (vect_prune_runtime_alias_test_list): Ditto.
5257 (vect_analyze_data_refs): Ditto.
5258 (vect_create_addr_base_for_vector_ref): Ditto.
5259 (vect_create_data_ref_ptr): Ditto.
5260 (vect_grouped_store_supported): Ditto.
5261 (vect_grouped_load_supported): Ditto.
5262 * value-prof.c (check_counter): Ditto.
5263 (check_ic_target): Ditto.
5264 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5265 (vect_recog_widen_mult_pattern): Ditto.
5266 (vect_recog_widen_sum_pattern): Ditto.
5267 (vect_recog_over_widening_pattern): Ditto.
5268 (vect_recog_widen_shift_pattern): Ditto.
5269 (vect_recog_rotate_pattern): Ditto.
5270 (vect_recog_vector_vector_shift_pattern): Ditto.
5271 (vect_recog_divmod_pattern): Ditto.
5272 (vect_recog_mixed_size_cond_pattern): Ditto.
5273 (vect_recog_bool_pattern): Ditto.
5274 (vect_pattern_recog_1): Ditto.
5275 (vect_pattern_recog): Ditto.
5276 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5277 (vect_is_simple_iv_evolution): Ditto.
5278 (vect_analyze_scalar_cycles_1): Ditto.
5279 (vect_get_loop_niters): Ditto.
5280 (vect_analyze_loop_1): Ditto.
5281 (vect_analyze_loop_form): Ditto.
5282 (vect_analyze_loop_operations): Ditto.
5283 (vect_analyze_loop_2): Ditto.
5284 (vect_analyze_loop): Ditto.
5285 (report_vect_op): Ditto.
5286 (vect_is_slp_reduction): Ditto.
5287 (vect_is_simple_reduction_1): Ditto.
5288 (vect_get_known_peeling_cost): Ditto.
5289 (vect_estimate_min_profitable_iters): Ditto.
5290 (vect_model_reduction_cost): Ditto.
5291 (vect_model_induction_cost): Ditto.
5292 (get_initial_def_for_induction): Ditto.
5293 (vect_create_epilog_for_reduction): Ditto.
5294 (vectorizable_reduction): Ditto.
5295 (vectorizable_induction): Ditto.
5296 (vectorizable_live_operation): Ditto.
5297 (vect_loop_kill_debug_uses): Ditto.
5298 (vect_transform_loop): Ditto.
5299 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5300 (vect_stmt_relevant_p): Ditto.
5301 (process_use): Ditto.
5302 (vect_mark_stmts_to_be_vectorized): Ditto.
5303 (vect_model_simple_cost): Ditto.
5304 (vect_model_promotion_demotion_cost): Ditto.
5305 (vect_model_store_cost): Ditto.
5306 (vect_get_store_cost): Ditto.
5307 (vect_model_load_cost): Ditto.
5308 (vect_get_load_cost): Ditto.
5309 (vect_init_vector_1): Ditto.
5310 (vect_get_vec_def_for_operand): Ditto.
5311 (vect_finish_stmt_generation): Ditto.
5312 (vectorizable_call): Ditto.
5313 (vectorizable_conversion): Ditto.
5314 (vectorizable_assignment): Ditto.
5315 (vectorizable_shift): Ditto.
5316 (vectorizable_operation): Ditto.
5317 (vectorizable_store): Ditto.
5318 (vectorizable_load): Ditto.
5319 (vectorizable_condition): Ditto.
5320 (vect_analyze_stmt): Ditto.
5321 (vect_transform_stmt): Ditto.
5322 (vect_is_simple_use): Ditto.
5323 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5324 (vect_can_advance_ivs_p): Ditto.
5325 (vect_update_ivs_after_vectorizer): Ditto.
5326 (vect_do_peeling_for_loop_bound): Ditto.
5327 (vect_gen_niters_for_prolog_loop): Ditto.
5328 (vect_update_inits_of_drs): Ditto.
5329 (vect_create_cond_for_alias_checks): Ditto.
5330 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5331 (vect_build_slp_tree_1): Ditto.
5332 (vect_supported_load_permutation_p): Ditto.
5333 (vect_analyze_slp_instance): Ditto.
5334 (vect_analyze_slp): Ditto.
5335 (vect_make_slp_decision): Ditto.
5336 (vect_detect_hybrid_slp): Ditto.
5337 (vect_bb_vectorization_profitable_p): Ditto.
5338 (vect_slp_analyze_bb_1): Ditto.
5339 (vect_update_slp_costs_according_to_vf): Ditto.
5340 (vect_get_mask_element): Ditto.
5341 (vect_transform_slp_perm_load): Ditto.
5342 (vect_schedule_slp_instance): Ditto.
5343 (vect_schedule_slp): Ditto.
5344 (vect_slp_transform_bb): Ditto.
5345 * profile.c (read_profile_edge_counts): Ditto.
5346 (compute_branch_probabilities): Ditto.
5347 * coverage.c (get_coverage_counts): Ditto.
5348
46f851f3
DN		 53492013-09-16 Diego Novillo <dnovillo@google.com>
5350
7a5911d3 5351 * tree-core.h: Add missing comment lines from refactoring of tree.h.
46f851f3 5352
1632a686
JH		 53532013-09-16 Jan Hubicka <jh@suse.cz>
5354
5355 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7a5911d3 5356 abstract functions; for static functions check the presence of body.
1632a686 5357
828e70c1
JG		 53582013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5359
5360 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5361 * config/aarch64/aarch64-simd.md
5362 (aarch64_mla_elt<mode>): New.
5363 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5364 (aarch64_mls_elt<mode>): Likewise.
5365 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5366 (aarch64_fma4_elt<mode>): Likewise.
5367 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5368 (aarch64_fma4_elt_to_128v2df): Likewise.
5369 (aarch64_fma4_elt_to_64df): Likewise.
5370 (fnma<mode>4): Likewise.
5371 (aarch64_fnma4_elt<mode>): Likewise.
5372 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5373 (aarch64_fnma4_elt_to_128v2df): Likewise.
5374 (aarch64_fnma4_elt_to_64df): Likewise.
5375 * config/aarch64/iterators.md (VDQSF): New.
5376 * config/aarch64/arm_neon.h
5377 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5378 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5379
779aea46
JG		 53802013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5381
5382 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5383 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5384 (aarch64_mul3_elt_to_128df): Likewise.
5385 (aarch64_mul3_elt_to_64v2df): Likewise.
5386 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5387 (VMUL): New.
5388 (VMUL_CHANGE_NLANES) Likewise.
5389 (h_con): Likewise.
5390 (f): Likewise.
5391 * config/aarch64/arm_neon.h
5392 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5393
a407a750
JG		 53942013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5395
5396 * config/aarch64/arm_neon.h
5397 (vcvtx_high_f32_f64): Fix parameters.
5398
f77e2d2d 53992013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
a5bd71fa 5400 Uros Bizjak <ubizjak@gmail.com>
f77e2d2d 5401
a5bd71fa 5402 * config/alpha.c: Include tree-ssanames.h.
f77e2d2d 5403
36875e8f
RB		 54042013-09-16 Richard Biener <rguenther@suse.de>
5405
5406 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5407 (dot_rdg_1): Handle control_dd.
5408 (create_edge_for_control_dependence): New function.
5409 (create_rdg_edges): Add control dependences if asked for.
5410 (build_rdg): Likewise.
5411 (generate_loops_for_partition): If there are not necessary
5412 control stmts remove all their dependencies.
5413 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5414 (distribute_loop): Pass on control dependences.
5415 (tree_loop_distribution): Compute control dependences and remove
5416 restriction on number of loop nodes.
5417
96e24d49
JJ		 54182013-09-16 Jakub Jelinek <jakub@redhat.com>
5419
5420 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5421 for internal calls.
5422
40dbb05c
RS		 54232013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5424
5425 * cse.c (try_const_anchors): Punt on CC modes.
5426
97191ad0
JBG		 54272013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5428
5429 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5430
403a3fb7
JDA		 54312013-09-14 John David Anglin <danglin@gcc.gnu.org>
5432
5433 PR target/58382
5434 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5435 calls to word_mode.
5436
fb5610fb
IS		 54372013-09-14 Iain Sandoe <iain@codesourcery.com>
5438
5439 PR target/48094
5440 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
f77e2d2d
JBG		 5441 seen.
5442 (darwin_objc1_section): Likewise.
fb5610fb
IS		 5443 (darwin_file_end): Emit Image Info section when required.
5444
afdec9bd
JH		 54452013-09-14 Jan Hubicka <jh@suse.cz>
5446
5447 * tree-into-ssa.c (gate_into_ssa): New.
5448 (pass_data_build_ssa): Use it.
5449 * cgraph.h (expand_thunk): Update prototype.
5450 * cgraphunit.c (analyze_function): Expand thunks early.
5451 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5452 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5453 set lowered flag; do not add new function.
5454 (assemble_thunks_and_aliases): Update.
5455 * tree-ssa.c (gate_init_datastructures): New gate.
5456 (pass_data_init_datastructures): Use it.
5457
492d1e0a
IS		 54582013-09-14 Iain Sandoe <iain@codesourcery.com>
5459
5460 PR target/58269
f77e2d2d 5461 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
492d1e0a
IS		 5462 xmm register set described in the psABI.
5463
d13c6b70
EG		 54642013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5465
f77e2d2d
JBG		 5466 * dwarf2out.c (should_emit_struct_debug): Add check
5467 for type_decl variable is not NULL.
d13c6b70 5468
88b0e79e
JC		 54692013-09-13 Jacek Caban <jacek@codeweavers.com>
5470
5471 * config.gcc: Use new winnt-c.c target hooks
5472 * config/t-winnt: New file
5473 * config/winnt-c.c: New file
5474 * doc/tm.texi.in: Document new hook
5475 * doc/tm.texi: Regenerated
5476
19ba6aab
JH		 54772013-09-13 Jan Hubicka <jh@suse.cz>
5478
5479 PR middle-end/58094
5480 * ipa-inline.c (check_callers): New function.
5481 (check_caller_edge): Remove.
5482 (want_inline_function_to_all_callers_p): Also permit alises that are
5483 called dirrectly.
5484 (inline_to_all_callers): Terminate the walk when devirtualization
5485 introduce new calls.
5486
a93c18c8
JH		 54872013-09-13 Jan Hubicka <jh@suse.cz>
5488
5489 * ipa-inline-analysis.c (struct growth_data): Add node.
5490 (do_estimate_growth_1): Fix detection of recursion.
5491
f3b331d1
JJ		 54922013-09-13 Jakub Jelinek <jakub@redhat.com>
5493
5494 PR tree-optimization/58392
5495 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5496 to avoid shadowing of outer loop variable. If
5497 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5498 replace_by_duplicate_decl simduid of loops that have it set and
5499 set dest_cfun->has_simduid_loops and/or
5500 dest_cfun->has_force_vect_loops.
5501 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5502 instead of maybe_lookup_decl.
5503 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5504 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5505 call. Copy over force_vect and copy and remap simduid. Set
5506 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5507 (copy_cfg_body): Remove blocks_to_copy argument. Use
5508 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5509 caller. Don't set cfun->has_simduid_loops and/or
5510 cfun->has_force_vect_loops here.
5511 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5512 caller.
5513 (expand_call_inline, tree_function_versioning): Adjust copy_body
5514 callers.
5515
ae6d0907
MJ		 55162013-09-13 Martin Jambor <mjambor@suse.cz>
5517
5518 PR bootstrap/58388
5519 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5520 the assert if the edge was a speculative one.
5521
2fd5894f
RB		 55222013-09-13 Richard Biener <rguenther@suse.de>
5523
5524 * tree-data-ref.h (known_dependences_p): Move here ...
5525 * tree-loop-distribution.c (known_dependences_p): ... from here.
5526 (dump_rdg_component, debug_rdg_component): Remove.
5527 (dump_rdg): Adjust.
5528 (generate_loops_for_partition): Use gimple_uid instead of
5529 relying on matching stmt visit order.
5530 (rdg_build_partitions): Take starting stmt vector.
5531 (ldist_gen): Merge into ...
5532 (distribute_loop): ... this function. Do not compute starting
5533 vertices vector.
5534 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5535
86eb4bd7
KT		 55362013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5537
5538 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5539 Set type attribute correctly. Set predicable_short_it attribute.
5540 (cmpsi_shiftsi): Remove %? from output template.
5541
83a95546
RB		 55422013-09-13 Richard Biener <rguenther@suse.de>
5543
5544 * tree-loop-distribution.c (struct rdg_component,
5545 rdg_defs_used_in_other_loops_p, free_rdg_components,
5546 rdg_build_components): Remove.
5547 (stmts_from_loop): Do not record virtual PHIs.
5548 (generate_loops_for_partition): Skip virtual PHIs.
5549 (build_rdg_partition_for_component): Rename to ...
5550 (build_rdg_partition_for_vertex): ... this and adjust.
5551 (rdg_build_partitions): Take a vector of starting vertices
5552 instead of components. Remove unnecessary leftover handling.
5553 (ldist_gen): Do not build components or record other stores.
5554 (distribute_loop): Do not distribute loops containing stmts
5555 with side-effects.
5556
c0ecf7f1
CB		 55572013-09-13 Christian Bruel <christian.bruel@st.com>
5558
5559 PR target/58314
5560 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5561
34f56efc
KT		 55622013-09-13 Kai Tietz <ktietz@redhat.com>
5563
740f2bf3
KT		 5564 * config.gcc: Separate cases for mingw and cygwin targets,
5565 and add 64-bit cygwin target case.
5566
34f56efc
KT		 5567 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5568 dll-export inline-functions.
5569 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5570
5571
55722013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL		 5573
5574 PR middle-end/58387
5575 Revert:
5576 2013-09-06 Jeff Law <law@redhat.com>
5577
5578 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5579 edge implied equivalences into successor phis.
5580
aadb52ad
DD		 55812013-09-12 DJ Delorie <dj@redhat.com>
5582
5583 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5584 separators.
5585
90ae7019
BM		 55862013-09-12 Brooks Moses <bmoses@google.com>
5587
5588 PR driver/42955
5589 * Makefile.in: Do not install driver binaries in $(target)/bin.
5590
17b2557c
DD		 55912013-09-12 DJ Delorie <dj@redhat.com>
5592
c5215a9b
DD		 5593 * config/rl78/rl78.opt (mrelax): New.
5594 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5595 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5596
17b2557c
DD		 5597 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5598 between SP and FP.
5599 (rl78_expand_epilogue): Likewise.
5600
ecafcf05
VM		 56012013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5602
5603 PR middle-end/58335
84f6d9fa 5604 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM		 5605 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5606 pointer elimination with using remove_reg_equal_offset_note.
5607
8369f38a
DD		 56082013-09-12 DJ Delorie <dj@redhat.com>
5609
f6a83b4a
DD		 5610 * config/msp430/: New port.
5611 * config.gcc (msp430): Added.
5612 * doc/invoke.texi: Document MSP430 options.
5613 * doc/install.texi: Document msp430-elf
5614 * doc/md.texi: Document msp430-elf
5615 * doc/contrib.texi: Document msp430-elf
5616
8369f38a
DD		 5617 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5618 PSImode.
5619
a854f856
MJ		 56202013-09-12 Martin Jambor <mjambor@suse.cz>
5621
5622 PR ipa/58389
5623 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5624 reference descriptor is NULL.
5625 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5626 edge to NULL.
5627
b12ebd96
AM		 56282013-09-12 Andrew MacLeod <amacleod@redhat.com>
5629
5630 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5631 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5632 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5633 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5634 make_temp_ssa_name): move to tree-ssanames.h
5635 * tree-ssa-alias.h: Move prototype.
5636 * tree-ssa.h: Include tree-ssanames.h.
5637 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5638 * tree-ssanames.h: New. Move items from tree-flow*.h
5639 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5640
e1fb4ad3
RB		 56412013-09-12 Richard Biener <rguenther@suse.de>
5642
5643 PR tree-optimization/58404
5644 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5645 propagate non-invariant addresses into dereferences wrapped
5646 in component references.
5647
154c9188
RB		 56482013-09-12 Richard Biener <rguenther@suse.de>
5649
5650 PR tree-optimization/58402
5651 * passes.def: Move pass_late_warn_uninitialized later.
5652
7a300452
AM		 56532013-09-12 Andrew MacLeod <amacleod@redhat.com>
5654
5655 * tree-ssa.h: New. Move content from tree-flow.h and
5656 tree-flow-inline.h.
5657 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5658 Move prototypes belonging to tree-ssa.c.
f77e2d2d 5659 * tree-flow-inline.h (redirect_edge_var_map_def,
7a300452
AM		 5660 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5661 tree-ssa.h.
5662 * gimple.h: Adjust prototypes.
f77e2d2d
JBG		 5663 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5664 to...
5665 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7a300452
AM		 5666 * tree.h: Move prototype to tree-ssa.h.
5667 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5668 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5669 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5670 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5671 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5672 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5673 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5674 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
f77e2d2d
JBG		 5675 graphite-blocking.c, graphite-clast-to-gimple.c,
5676 graphite-dependences.c, graphite-interchange.c,
5677 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5678 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5679 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5680 ipa-reference.c, ipa-split.c, ipa-utils.c,
7a300452
AM		 5681 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5682 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5683 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5684 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5685 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5686 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5687 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5688 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5689 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5690 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5691 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5692 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5693 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5694 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5695 value-prof.c, var-tracking.c,
5696 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5697
97463b2b
RB		 56982013-09-12 Richard Biener <rguenther@suse.de>
5699
5700 PR tree-optimization/58396
5701 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5702 (build_rdg): Take a loop-nest parameter, fix memleaks.
5703 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5704
80ecd89d
YR		 57052013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5706
5707 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5708 for SLM.
5709
893e85fa
CM		 57102013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5711
5712 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5713 The FMA instruction names should have a 'v' prefix.
5714
174ec470
RB		 57152013-09-12 Richard Biener <rguenther@suse.de>
5716
5717 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5718 (dot_rdg): Use popen instead of system in optional code.
5719 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5720 (already_processed_vertex_p): Adjust.
5721 (has_anti_or_output_dependence, predecessor_has_mem_write,
5722 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5723 rdg_flag_uses): Remove.
5724 (rdg_flag_vertex): Simplify.
5725 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5726 remove recursion.
5727 (build_rdg_partition_for_component): Process the first vertex
5728 of a component only.
5729 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5730
38c89763
AM		 57312013-09-12 Alan Modra <amodra@gmail.com>
5732
5733 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5734
3e3d9d17
DD		 57352013-09-11 DJ Delorie <dj@redhat.com>
5736 Nick Clifton <nickc@redhat.com>
5737
5738 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5739 (rl78_stack_based_mem): New.
5740 * config/rl78/constraints.md (Iv08): New.
5741 (Iv16): New.
5742 (Iv24): New.
5743 (Is09): New.
5744 (Is17): New.
5745 (Is25): New.
5746 (ISsi): New.
5747 (IShi): New.
5748 (ISqi): New.
5749 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5750 (movhi): Likewise.
5751 (movsi): Change from expand to insn-and-split.
5752 (ashrsi3): Clobber AX.
5753 (lshrsi3): New.
5754 (ashlsi3): New.
5755 (cbranchsi4): New.
5756 * config/rl78/rl78.md (CC_REG): Fix.
5757 (addsi3): Allow memory and immediate operands.
5758 (addsi3_internal): Split into...
5759 (addsi3_internal_virt): ...new, and ...
5760 (addsi3_internal_real): ...new.
5761 (subsi): New.
5762 (subsi3_internal_virt): New.
5763 (subsi3_internal_real): New.
5764 (mulsi3): Add memory operand.
5765 (mulsi3_rl78): Likewise.
5766 (mulsi3_g13): Likewise.
5767 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5768 (cbranchqi4_real): Add more constraint options.
5769 (cbranchhi4_real): Expand pattern.
5770 (cbranchhi4_real_signed): New.
5771 (cbranchhi4_real_inverted): New.
5772 (cbranchsi4_real_lt): New.
5773 (cbranchsi4_real_ge): New.
5774 (cbranchsi4_real_signed): New.
5775 (cbranchsi4_real): New.
5776 (peephole2): New.
893e85fa
CM		 5777 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5778 constant shifts.
3e3d9d17
DD		 5779 (lshrsi3_virt): Likewise.
5780 (ashlsi3_virt): Likewise.
5781 (cbranchqi4_virt_signed): New.
5782 (cbranchhi4_virt_signed): New.
5783 (cbranchsi4_virt): New.
5784 * config/rl78/rl78.c: Whitespace fixes throughout.
5785 (move_elim_pass): New.
5786 (pass_data_rl78_move_elim): New.
5787 (pass_rl78_move_elim): New.
5788 (make_pass_rl78_move_elim): New.
5789 (rl78_devirt_info): Run devirt earlier.
5790 (rl78_move_elim_info): New.
5791 (rl78_asm_file_start): Register it.
5792 (rl78_split_movsi): New.
5793 (rl78_as_legitimate_address): Allow virtual base registers when
5794 appropriate.
5795 (rl78_addr_space_convert): Remove spurious debug stuff.
5796 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5797 (rl78_print_operand): More cases for not printing '#'.
5798 (rl78_expand_compare): Remove most of the logic.
5799 (content_memory): New.
5800 (clear_content_memory): New.
5801 (get_content_index): New.
5802 (get_content_name): New.
5803 (display_content_memory): New.
5804 (update_content): New.
5805 (record_content): New.
5806 (already_contains): New.
5807 (insn_ok_now): Re-recog insns with virtual registers.
5808 (add_postponed_content_update): New.
5809 (process_postponed_content_update): New.
5810 (gen_and_emit_move): New.
893e85fa
CM		 5811 (transcode_memory_rtx): Record new location content.
5812 Use gen_and_emit_move.
3e3d9d17
DD		 5813 (force_into_acc): New.
5814 (move_to_acc): Use gen_and_emit_move.
5815 (move_from_acc): Likewise.
5816 (move_acc_to_reg): Likewise.
5817 (move_to_x): Likewise.
5818 (move_to_hl): Likewise.
5819 (move_to_de): Likewise.
5820 (rl78_alloc_physical_registers_op1): Record location content.
5821 (has_constraint): New.
5822 (rl78_alloc_physical_registers_op2): Record location content.
5823 Optimize use of HL.
5824 (rl78_alloc_physical_registers_ro1): Likewise.
5825 (rl78_alloc_physical_registers_cmp): Likewise.
5826 (rl78_alloc_physical_registers_umul): Likewise.
5827 (rl78_alloc_address_registers_macax): New.
5828 (rl78_alloc_physical_registers): Initialize and set location
5829 content memory as needed.
5830 (rl78_reorg): Make sure split2 is called.
5831 (rl78_rtx_costs): New.
5832
088845a5
RS		 58332013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5834
5835 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5836 for (not (neg ...)) and (neg (not ...)) cases.
5837
5c2961cf
RB		 58382013-09-11 Richard Biener <rguenther@suse.de>
5839
5840 PR middle-end/58377
5841 * passes.def: Split critical edges before late uninit warning passes.
5842 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5843
427e6a14
JJ		 58442013-09-11 Jakub Jelinek <jakub@redhat.com>
5845
5846 PR tree-optimization/58385
5847 * fold-const.c (build_range_check): If both low and high are NULL,
5848 use omit_one_operand_loc to preserve exp side-effects.
5849
df45c0cc
KT		 58502013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5851
5852 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5853
80ab0b19
RB		 58542013-09-11 Richard Biener <rguenther@suse.de>
5855
5856 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5857 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5858 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5859 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5860 stmts_from_loop, known_dependences_p, build_empty_rdg,
5861 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5862 * tree-loop-distribution.c: ... here.
5863 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5864 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5865 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5866 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5867 * tree-loop-distribution.c: ... here.
5868 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5869 (struct partition_s): Add loops member.
5870 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5871 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5872 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5873
85a77221
AI		 58742013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5875 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5876 Sergey Lega <sergey.s.lega@intel.com>
5877 Anna Tikhonova <anna.tikhonova@intel.com>
5878 Ilya Tocar <ilya.tocar@intel.com>
5879 Andrey Turetskiy <andrey.turetskiy@intel.com>
5880 Ilya Verbin <ilya.verbin@intel.com>
5881 Kirill Yukhin <kirill.yukhin@intel.com>
5882 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5883
5884 * config/i386/constraints.md (k): New.
5885 (Yk): Ditto.
5886 * config/i386/i386.c (const regclass_map): Add new mask registers.
5887 (dbx_register_map): Ditto.
5888 (dbx64_register_map): Ditto.
5889 (svr4_dbx_register_map): Ditto.
5890 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5891 disabled.
5892 (ix86_preferred_reload_class): Disable constants for mask registers.
5893 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5894 (ix86_hard_regno_mode_ok): Support new mask registers.
5895 (x86_order_regs_for_local_alloc): Ditto.
5896 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5897 (FIXED_REGISTERS): Add new mask registers.
5898 (CALL_USED_REGISTERS): Ditto.
5899 (REG_ALLOC_ORDER): Ditto.
5900 (VALID_MASK_REG_MODE): New.
5901 (FIRST_MASK_REG): Ditto.
5902 (LAST_MASK_REG): Ditto.
5903 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5904 (MAYBE_MASK_CLASS_P): New.
5905 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5906 (REG_CLASS_CONTENTS): Ditto.
5907 (MASK_REGNO_P): New.
5908 (ANY_MASK_REG_P): Ditto.
5909 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM		 5910 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5911 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5912 mask registers.
85a77221
AI		 5913 (attribute "type"): Add mskmov, msklog.
5914 (attribute "length_immediate"): Support them.
5915 (attribute "memory"): Ditto.
5916 (attribute "prefix_0f"): Ditto.
5917 (*movhi_internal): Support new mask registers.
5918 (*movqi_internal): Ditto.
5919 (define_split): Split out clobber pattern is a logic
5920 insn on mask registers.
5921 (*k<logic><mode>): New.
5922 (*andhi_1): Extend to support mask regs.
5923 (*andqi_1): Extend to support mask regs.
5924 (kandn<mode>): New.
5925 (define_split): Split and-not to and and not if operands
5926 are not mask regs.
5927 (*<code><mode>_1): Separate HI mode to new pattern...
5928 (*<code>hi_1): This.
5929 (*<code>qi_1): Extend to support mask regs.
5930 (kxnor<mode>): New.
5931 (kortestzhi): Ditto.
5932 (kortestchi): Ditto.
5933 (kunpckhi): Ditto.
5934 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5935 (*one_cmplhi2_1): ...Here, now with mask registers support.
5936 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM		 5937 (HI/QImode arithmetics splitter): Don't split if mask registers
5938 are used.
85a77221
AI		 5939 (HI/QImode not splitter): Ditto.
5940 * config/i386/predicated.md (mask_reg_operand): New.
5941 (general_reg_operand): Ditto.
5942
7e570821
AI		 59432013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5944
5945 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5946 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5947
9e1376e9
JL		 59482013-09-10 Jeff Law <law@redhat.com>
5949
5950 PR tree-optimization/58380
5951 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5952 of threading through a buried loop header.
5953
5954 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5955 return value for single successor case.
5956
5bbcb888
JH		 59572013-09-10 Jan Hubicka <jh@suse.cz>
5958
5959 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5960
d742ff4b
RE		 59612013-09-10 Richard Earnshaw <rearnsha@arm.com>
5962
5963 PR target/58361
5964 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5965 support conditional execution.
5966 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5967
b0681c9e
VM		 59682013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5969
5970 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5971 constraint pass.
f77e2d2d
JBG		 5972 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5973 Check destination too to check move insn.
b0681c9e 5974 (undo_optional_reloads): Add check that the original peudo did not
f77e2d2d
JBG		 5975 changed its allocation and the optional reload was inherited on last
5976 inheritance pass. Break loop after deciding to keep optional reload.
b0681c9e
VM		 5977 (lra_undo_inheritance): Add check that inherited pseudo still in
5978 memory.
5979
c9b1c957
JG		 59802013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5981
5982 * config/aarch64/aarch64.md (generic_sched): New.
5983 * config/aarch64/aarch64-generic.md (load): Make conditional
5984 on generic_sched attribute.
5985 (nonload): Likewise.
5986
82d618d3
JH		 59872013-09-10 Jan Hubicka <jh@suse.cz>
5988
5989 * lto-cgraph.c: Include ipa-utils.h.
5990 (compute_ltrans_boundary): Also add possible targets into the boundary.
5991
668b6894
JH		 59922013-09-10 Jan Hubicka <jh@suse.cz>
5993
5994 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5995 VAR_DECL of vtable rather than full expression.
5996
4f90d3e0
JH		 59972013-09-10 Jan Hubicka <jh@suse.cz>
5998 Paolo Carlini <paolo.carlini@oracle.com>
5999
6000 * cgraphunit.c (analyze_functions): Save input_location, set it
6001 to UNKNOWN_LOCATION and restore it at the end.
6002
39e87baf
MJ		 60032013-09-10 Martin Jambor <mjambor@suse.cz>
6004
6005 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6006 represented by a thunk.
6007
581aedec
JL		 60082013-09-10 Jeff Law <law@redhat.com>
6009
6010 PR tree-optimization/58343
6011 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6012 jump threading requests through loop headers buried in the middle
6013 of a jump threading path.
6014
6015 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6016 in return value/type.
6017
84cf4ab6
JJ		 60182013-09-10 Jakub Jelinek <jakub@redhat.com>
6019
6020 PR rtl-optimization/58365
6021 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6022 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6023 it differs.
6024
447dd906
RB		 60252013-09-10 Richard Biener <rguenther@suse.de>
6026
6027 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6028 * tree-data-ref.c (create_rdg_vertices): Collect all data
6029 references, signal failure to the caller, use data-ref API.
893e85fa
CM		 6030 (build_rdg): Compute data references only once. Maintain lifetime
6031 of data references and data dependences from within RDG.
447dd906
RB		 6032 (free_rdg): Free dependence relations.
6033 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6034 inventing extra dependences.
6035 (distribute_loop): Update for RDG API changes.
6036
f14a7324
KT		 60372013-09-10 Kai Tietz <ktietz@redhat.com>
6038
6039 * doc/invoke.texi (fms-extensions): Document changed
6040 behavior for ms-abi targets.
6041 * config/i386/i386.c (ix86_option_override_internal):
6042 Set default value of option -fms-extension for ms-abi targets.
6043
ff131f26
MZ		 60442013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6045
6046 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6047
33115a2c
AM		 60482013-09-10 Alan Modra <amodra@gmail.com>
6049
6050 PR target/58330
6051 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6052
b6872a13
AM		 60532013-09-10 Alan Modra <amodra@gmail.com>
6054
6055 * config/rs6000/predicates.md (add_cint_operand): New.
6056 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6057 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6058 using add_cint_operand.
6059 (largetoc_high_plus_aix): Likewise.
6060
28fd0ba2
JJ		 60612013-09-09 Jakub Jelinek <jakub@redhat.com>
6062
6063 PR tree-optimization/58364
6064 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6065 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6066 the current range can't be an unconditional true or false.
6067
552d99fe
JG		 60682013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6069
6070 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6071
f43245d1
UB		 60722013-09-09 Uros Bizjak <ubizjak@gmail.com>
6073
6074 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6075
c9b0866a
PC		 60762013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6077
6078 PR c++/43452
6079 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6080
78d8b9f0
IB		 60812013-09-09 Ian Bolton <ian.bolton@arm.com>
6082
6083 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6084 NO_REGS for immediate that can't be moved directly into FP_REGS.
6085
274b2532
KT		 60862013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6087
6088 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6089 comparison with negated operand.
c9b0866a
PC		 6090 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6091 RTL form.
274b2532 6092
467a3558
RB		 60932013-09-09 Richard Biener <rguenther@suse.de>
6094
6095 PR middle-end/58326
6096 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6097 of a subloop record all its block as affecting loop-closed SSA form.
6098
088c5368
RS		 60992013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6100
6101 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6102 of an rtx/bitpos pair.
6103 (store_fixed_bit_field): Update accordingly.
6104
69db2d57
RS		 61052013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6106
6107 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6108 GEN_INT.
6109 * builtins.c (expand_errno_check): Likewise.
6110 * dwarf2cfi.c (init_return_column_size): Likewise.
6111 * except.c (sjlj_mark_call_sites): Likewise.
6112 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6113 * lra-constraints.c (emit_inc): Likewise.
6114 * ree.c (combine_set_extension): Likewise.
6115 * regmove.c (fixup_match_2): Likewise.
6116 * reload1.c (inc_for_reload): Likewise.
6117
69a59f0f
RS		 61182013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6119
6120 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6121 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6122 Use gen_int_mode with the mode of the associated simplify_* call.
6123 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6124 * expmed.c (expand_shift_1): Likewise.
6125 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6126 * loop-iv.c (iv_number_of_iterations): Likewise.
6127 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6128 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6129
2f1cd2eb
RS		 61302013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6131
6132 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6133 of the associated expand_* call.
6134 (asan_emit_stack_protection): Likewise.
6135 * builtins.c (round_trampoline_addr): Likewise.
6136 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6137 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6138 (emit_store_flag): Likewise.
6139 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6140 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6141 Likewise.
6142 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6143 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6144 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6145 * modulo-sched.c (generate_prolog_epilog): Likewise.
6146 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6147 (expand_ctz, expand_ffs, expand_unop): Likewise.
6148
4789c0ce
RS		 61492013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6150
6151 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6152 of the associated gen_rtx_* call.
6153 * caller-save.c (init_caller_save): Likewise.
6154 * combine.c (find_split_point, make_extraction): Likewise.
6155 (make_compound_operation): Likewise.
6156 * dwarf2out.c (mem_loc_descriptor): Likewise.
6157 * explow.c (plus_constant, probe_stack_range): Likewise.
6158 * expmed.c (expand_mult_const): Likewise.
6159 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6160 * reload1.c (init_reload): Likewise.
6161 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6162 * var-tracking.c (adjust_mems): Likewise.
6163 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6164 rather than gen_rtx_fmt_ee.
6165
a811f7cb
JH		 61662013-09-09 Jan Hubicka <jh@suse.cz>
6167
6168 PR middle-end/58294
6169 * value-prof.c (gimple_ic): Copy also abnormal edges.
6170
dcad1dd3
RS		 61712013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6172
6173 * asan.c (asan_shadow_cst): Use gen_int_mode.
6174
daf5c770
JH		 61752013-09-08 Jan Hubicka <jh@suse.cz>
6176
6177 * ipa-profile.c: Add toplevel comment.
6178 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6179 (contains_hot_call_p): New function.
6180 (ipa_propagate_frequency): Set frequencies based on counts when
6181 profile is read.
6182 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6183 profile; do not tamper with profile after inlining if it is read.
6184
a12cd2db
JH		 61852013-09-08 Jan Hubicka <jh@suse.cz>
6186
6187 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6188 speculative edges.
6189
856a1588
JH		 61902013-09-08 Jan Hubicka <jh@suse.cz>
6191
6192 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6193 summary generation.
6194
4f4b0b73
JL		 61952013-09-08 Jeff Law <law@redhat.com>
6196
6197 PR bootstrap/58340
6198 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6199 of 'found'.
6200
cc4029ee
AK		 62012013-09-08 Andi Kleen <ak@linux.intel.com>
6202
6203 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6204
7b395ddd
JH		 62052013-09-08 Jan Hubicka <jh@suse.cz>
6206
6207 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6208
c34c46dd
RS		 62092013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6210
6211 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6212 for non-debug insns.
6213 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6214 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6215 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6216 accordingly.
6217
3462aa02
JH		 62182013-09-08 Jan Hubicka <jh@suse.cz>
6219
6220 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6221 targets and devirtualize to BUILT_IN_UNREACHABLE.
6222 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6223 * ipa.c (walk_polymorphic_call_targets): New function.
6224 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6225 functions; use the new timevar.
6226 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6227 was removed from the program.
6228 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6229 not consider it in the walk when its vtable is dead.
6230 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6231 record_binfo.
6232 (devirt_variable_node_removal_hook): New function.
6233 (possible_polymorphic_call_targets): Also register
6234 devirt_variable_node_removal_hook.
6235 (ipa_devirt): Do not do non-speculative devirtualization.
6236 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6237
26e5b0fd
JH		 62382013-09-08 Jan Hubicka <jh@suse.cz>
6239
6240 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6241 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6242 varpool_remove_variable_insertion_hook): Declare.
6243 * varpool.c (varpool_node_hook_list): New structure.
6244 (first_varpool_node_removal_hook,
6245 first_varpool_variable_insertion_hook): New variables.
6246 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6247 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6248 varpool_remove_variable_insertion_hook,
6249 varpool_call_variable_insertion_hooks): New functions.
6250 (varpool_remove_node): Use it.
6251
97f1f314 62522013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 6253
97f1f314
PC		 6254 PR c++/54941
6255 * diagnostic.c (diagnostic_build_prefix): When s.file is
6256 "<built-in>" don't output line and column numbers.
6df2ca09 6257
bcb650cb
JH		 62582013-09-06 Jan Hubicka <jh@suse.cz>
6259
6260 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6261 * lto-streamer-out.c: Stream thunks, too.
6262 * lto-streamer-in.c (input_function): Pop cfun here
6263 (lto_read_body): Instead of here.
6264
87f3fea8
CT		 62652013-09-06 Caroline Tice <cmtice@google.com>
6266
6267 * doc/install.texi: Add documentation for the --enable-vtable-verify
6268 and the --disable-libvtv configure options.
6269
5e8d9b45
JL		 62702013-09-06 Jeff Law <law@redhat.com>
6271
6272 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6273 edge implied equivalences into successor phis.
6274
3a001aff
JR		 62752013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6276
6277 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6278
62792013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ		 6280
6281 * resource.c (mark_target_live_regs): Compute resources taking
6282 into account if a call is predicated or not.
6283
2c641f8b
EB		 62842013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6285
6286 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6287 the compiler in the function names.
6288
a81b0a3d
JH		 62892013-09-06 Jan Hubicka <jh@suse.cz>
6290
6291 PR middle-end/58094
6292 * ipa-inline.c (has_caller_p): New function.
6293 (want_inline_function_to_all_callers_p): Use it.
6294 (sum_callers, inline_to_all_callers): Break out from ...
6295 (ipa_inline): ... here.
6296
68e4ca7e
JH		 62972013-09-06 Jan Hubicka <jh@suse.cz>
6298
2c641f8b
EB		 6299 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6300 only when AVX is enabled.
68e4ca7e 6301
3c73e2fe
JG		 63022013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6303
6304 * config/aarch64/aarch64.md
6305 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6306 is fpsimd_<load/store>2.
6307 (load_pair<mode>): Likewise.
6308 (store_pair<mode>): Likewise.
6309
8c48eecd
JG		 63102013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6311
6312 * config/arm/types.md (type): Add "mrs" type.
6313 * config/aarch64/aarch64.md
6314 (aarch64_load_tp_hard): Make type "mrs".
6315 * config/arm/arm.md
6316 (load_tp_hard): Make type "mrs".
6317 * config/arm/cortex-a15.md: Update with new attributes.
6318 * config/arm/cortex-a5.md: Update with new attributes.
6319 * config/arm/cortex-a53.md: Update with new attributes.
6320 * config/arm/cortex-a7.md: Update with new attributes.
6321 * config/arm/cortex-a8.md: Update with new attributes.
6322 * config/arm/cortex-a9.md: Update with new attributes.
6323 * config/arm/cortex-m4.md: Update with new attributes.
6324 * config/arm/cortex-r4.md: Update with new attributes.
6325 * config/arm/fa526.md: Update with new attributes.
6326 * config/arm/fa606te.md: Update with new attributes.
6327 * config/arm/fa626te.md: Update with new attributes.
6328 * config/arm/fa726te.md: Update with new attributes.
6329
9ea187a5
JG		 63302013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6331
6332 * config/aarch64/aarch64.md
6333 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6334 (*movtf_aarch64): Likewise.
6335 * config/arm/arm.md
6336 (thumb1_movdi_insn): Use "multiple" for type where more than one
6337 instruction is used for a move.
6338 (*arm32_movhf): Likewise.
6339 (*thumb_movdf_insn): Likewise.
6340
292b89b3
JG		 63412013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6342
6343 * config/arm/types.md (type): Rename fcpys to fmov.
6344 * config/arm/vfp.md
6345 (*arm_movsi_vfp): Rename type fcpys as fmov.
6346 (*thumb2_movsi_vfp): Likewise
6347 (*movhf_vfp_neon): Likewise
6348 (*movhf_vfp): Likewise
6349 (*movsf_vfp): Likewise
6350 (*thumb2_movsf_vfp): Likewise
6351 (*movsfcc_vfp): Likewise
6352 (*thumb2_movsfcc_vfp): Likewise
6353 * config/aarch64/aarch64-simd.md
6354 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6355 * config/aarch64/aarch64.md
6356 (*movsi_aarch64): Replace type mov_reg with fmovs.
6357 (*movdi_aarch64): Likewise
6358 (*movsf_aarch64): Likewise
6359 (*movdf_aarch64): Likewise
6360 * config/arm/arm.c
6361 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6362 * config/arm/iwmmxt.md
6363 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6364 * config/arm/arm1020e.md: Update with new attributes.
6365 * config/arm/cortex-a15-neon.md: Update with new attributes.
6366 * config/arm/cortex-a5.md: Update with new attributes.
6367 * config/arm/cortex-a53.md: Update with new attributes.
6368 * config/arm/cortex-a7.md: Update with new attributes.
6369 * config/arm/cortex-a8-neon.md: Update with new attributes.
6370 * config/arm/cortex-a9.md: Update with new attributes.
6371 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6372 * config/arm/cortex-r4f.md: Update with new attributes.
6373 * config/arm/marvell-pj4.md: Update with new attributes.
6374 * config/arm/vfp11.md: Update with new attributes.
6375
6a24a83e
JG		 63762013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6377
6378 * config/aarch64/aarch64.md
6379 (*madd<mode>): Fix type attribute.
6380 (*maddsi_uxtw): Likewise.
6381 (*msub<mode>): Likewise.
6382 (*msubsi_uxtw): Likewise.
6383 (<su_optab>maddsidi4): Likewise.
6384 (<su_optab>msubsidi4): Likewise.
6385
b86923f0
JG		 63862013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6387
6388 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6389 * config/arm/arm.md (core_cycles): Remove fdiv.
6390 * config/arm/vfp.md:
6391 (*sqrtsf2_vfp): Update for attribute changes.
6392 (*sqrtdf2_vfp): Likewise.
6393 * config/aarch64/aarch64.md:
6394 (sqrt<mode>2): Update for attribute changes.
6395 * config/arm/arm1020e.md: Update with new attributes.
6396 * config/arm/cortex-a15-neon.md: Update with new attributes.
6397 * config/arm/cortex-a5.md: Update with new attributes.
6398 * config/arm/cortex-a53.md: Update with new attributes.
6399 * config/arm/cortex-a7.md: Update with new attributes.
6400 * config/arm/cortex-a8-neon.md: Update with new attributes.
6401 * config/arm/cortex-a9.md: Update with new attributes.
6402 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6403 * config/arm/cortex-r4f.md: Update with new attributes.
6404 * config/arm/marvell-pj4.md: Update with new attributes.
6405 * config/arm/vfp11.md: Update with new attributes.
6406
7b49c9e1
JG		 64072013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6408
6409 * config/arm/types.md
6410 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6411 * config/aarch64/aarch64.md
6412 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6413 new attributes.
6414 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6415 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6416 (float<GPI:mode><GPF:mode>2): Likewise.
6417 * config/arm/vfp.md
6418 (*truncsisf2_vfp): Update with new attributes.
6419 (*truncsidf2_vfp): Likewise.
6420 (fixuns_truncsfsi2): Likewise.
6421 (fixuns_truncdfsi2): Likewise.
6422 (*floatsisf2_vfp): Likewise.
6423 (*floatsidf2_vfp): Likewise.
6424 (floatunssisf2): Likewise.
6425 (floatunssidf2): Likewise.
6426 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6427 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6428 * config/arm/arm1020e.md: Update with new attributes.
6429 * config/arm/cortex-a15-neon.md: Update with new attributes.
6430 * config/arm/cortex-a5.md: Update with new attributes.
6431 * config/arm/cortex-a53.md: Update with new attributes.
6432 * config/arm/cortex-a7.md: Update with new attributes.
6433 * config/arm/cortex-a8-neon.md: Update with new attributes.
6434 * config/arm/cortex-a9.md: Update with new attributes.
6435 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6436 * config/arm/cortex-r4f.md: Update with new attributes.
6437 * config/arm/marvell-pj4.md: Update with new attributes.
6438 * config/arm/vfp11.md: Update with new attributes.
6439
e7df8af8
JG		 64402013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6441
6442 * config/aarch64/arm_neon.h
6443 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6444 (vqtbx<1,2,3,4><q>_s8): Likewise.
6445
594726e4
JG		 64462013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6447
f43245d1
UB		 6448 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6449 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG		 6450 (add<mode>3): Add type attribute.
6451 (add<mode>3): Likewise.
6452 (usadd<mode>3): Likewise.
6453 (ssadd<mode>3): Likewise.
6454 (sub<mode>3): Likewise.
6455 (sub<mode>3): Likewise.
6456 (ussub<mode>3): Likewise.
6457 (sssub<mode>3): Likewise.
6458 (ssmulsa3): Likewise.
6459 (usmulusa3): Likewise.
6460 (arm_usatsihi): Likewise.
6461 * config/arm/vfp.md
6462 (*movdi_vfp): Add types for all instructions.
6463 (*movdi_vfp_cortexa8): Likewise.
6464 (*movhf_vfp_neon): Likewise.
6465 (*movhf_vfp): Likewise.
6466 (*movdf_vfp): Likewise.
6467 (*thumb2_movdf_vfp): Likewise.
6468 (*thumb2_movdfcc_vfp): Likewise.
6469 * config/arm/arm.md: Add type attribute to all insn patterns.
6470 (*thumb1_adddi3): Add type attribute.
6471 (*arm_adddi3): Likewise.
6472 (*adddi_sesidi_di): Likewise.
6473 (*adddi_zesidi_di): Likewise.
6474 (*thumb1_addsi3): Likewise.
6475 (addsi3_compare0): Likewise.
6476 (*addsi3_compare0_scratch): Likewise.
6477 (*compare_negsi_si): Likewise.
6478 (cmpsi2_addneg): Likewise.
6479 (*addsi3_carryin_<optab>): Likewise.
6480 (*addsi3_carryin_alt2_<optab>): Likewise.
6481 (*addsi3_carryin_clobercc_<optab>): Likewise.
6482 (*subsi3_carryin): Likewise.
6483 (*subsi3_carryin_const): Likewise.
6484 (*subsi3_carryin_compare): Likewise.
6485 (*subsi3_carryin_compare_const): Likewise.
6486 (*arm_subdi3): Likewise.
6487 (*thumb_subdi3): Likewise.
6488 (*subdi_di_zesidi): Likewise.
6489 (*subdi_di_sesidi): Likewise.
6490 (*subdi_zesidi_di): Likewise.
6491 (*subdi_sesidi_di): Likewise.
6492 (*subdi_zesidi_ze): Likewise.
6493 (thumb1_subsi3_insn): Likewise.
6494 (*arm_subsi3_insn): Likewise.
6495 (*anddi3_insn): Likewise.
6496 (*anddi_zesidi_di): Likewise.
6497 (*anddi_sesdi_di): Likewise.
6498 (*ne_zeroextracts): Likewise.
6499 (*ne_zeroextracts): Likewise.
6500 (*ite_ne_zeroextr): Likewise.
6501 (*ite_ne_zeroextr): Likewise.
6502 (*anddi_notdi_di): Likewise.
6503 (*anddi_notzesidi): Likewise.
6504 (*anddi_notsesidi): Likewise.
6505 (andsi_notsi_si): Likewise.
6506 (thumb1_bicsi3): Likewise.
6507 (*iordi3_insn): Likewise.
6508 (*iordi_zesidi_di): Likewise.
6509 (*iordi_sesidi_di): Likewise.
6510 (*thumb1_iorsi3_insn): Likewise.
6511 (*xordi3_insn): Likewise.
6512 (*xordi_zesidi_di): Likewise.
6513 (*xordi_sesidi_di): Likewise.
6514 (*arm_xorsi3): Likewise.
6515 (*andsi_iorsi3_no): Likewise.
6516 (*smax_0): Likewise.
6517 (*smax_m1): Likewise.
6518 (*arm_smax_insn): Likewise.
6519 (*smin_0): Likewise.
6520 (*arm_smin_insn): Likewise.
6521 (*arm_umaxsi3): Likewise.
6522 (*arm_uminsi3): Likewise.
6523 (*minmax_arithsi): Likewise.
6524 (*minmax_arithsi_): Likewise.
6525 (*satsi_<SAT:code>): Likewise.
6526 (arm_ashldi3_1bit): Likewise.
6527 (arm_ashrdi3_1bit): Likewise.
6528 (arm_lshrdi3_1bit): Likewise.
6529 (*arm_negdi2): Likewise.
6530 (*thumb1_negdi2): Likewise.
6531 (*arm_negsi2): Likewise.
6532 (*thumb1_negsi2): Likewise.
6533 (*negdi_extendsid): Likewise.
6534 (*negdi_zero_extend): Likewise.
6535 (*arm_abssi2): Likewise.
6536 (*thumb1_abssi2): Likewise.
6537 (*arm_neg_abssi2): Likewise.
6538 (*thumb1_neg_abss): Likewise.
6539 (one_cmpldi2): Likewise.
6540 (extend<mode>di2): Likewise.
6541 (*compareqi_eq0): Likewise.
6542 (*arm_extendhisi2addsi): Likewise.
6543 (*arm_movdi): Likewise.
6544 (*thumb1_movdi_insn): Likewise.
6545 (*arm_movt): Likewise.
6546 (*thumb1_movsi_insn): Likewise.
6547 (pic_add_dot_plus_four): Likewise.
6548 (pic_add_dot_plus_eight): Likewise.
6549 (tls_load_dot_plus_eight): Likewise.
6550 (*thumb1_movhi_insn): Likewise.
6551 (*thumb1_movsf_insn): Likewise.
6552 (*movdf_soft_insn): Likewise.
6553 (*thumb_movdf_insn): Likewise.
6554 (cbranchsi4_insn): Likewise.
6555 (cbranchsi4_scratch): Likewise.
6556 (*negated_cbranchsi4): Likewise.
6557 (*tbit_cbranch): Likewise.
6558 (*tlobits_cbranch): Likewise.
6559 (*tstsi3_cbranch): Likewise.
6560 (*cbranchne_decr1): Likewise.
6561 (*addsi3_cbranch): Likewise.
6562 (*addsi3_cbranch_scratch): Likewise.
6563 (*arm_cmpdi_insn): Likewise.
6564 (*arm_cmpdi_unsig): Likewise.
6565 (*arm_cmpdi_zero): Likewise.
6566 (*thumb_cmpdi_zero): Likewise.
6567 (*deleted_compare): Likewise.
6568 (*mov_scc): Likewise.
6569 (*mov_negscc): Likewise.
6570 (*mov_notscc): Likewise.
6571 (*cstoresi_eq0_thumb1_insn): Likewise.
6572 (cstoresi_nltu_thumb1): Likewise.
6573 (cstoresi_ltu_thu): Likewise.
6574 (thumb1_addsi3_addgeu): Likewise.
6575 (*arm_jump): Likewise.
6576 (*thumb_jump): Likewise.
6577 (*check_arch2): Likewise.
6578 (arm_casesi_internal): Likewise.
6579 (thumb1_casesi_dispatch): Likewise.
6580 (*arm_indirect_jump): Likewise.
6581 (*thumb1_indirect_jump): Likewise.
6582 (nop): Likewise.
6583 (*and_scc): Likewise.
6584 (*ior_scc): Likewise.
6585 (*compare_scc): Likewise.
6586 (*cond_move): Likewise.
6587 (*cond_arith): Likewise.
6588 (*cond_sub): Likewise.
6589 (*cmp_ite0): Likewise.
6590 (*cmp_ite1): Likewise.
6591 (*cmp_and): Likewise.
6592 (*cmp_ior): Likewise.
6593 (*ior_scc_scc): Likewise.
6594 (*ior_scc_scc_cmp): Likewise.
6595 (*and_scc_scc): Likewise.
6596 (*and_scc_scc_cmp): Likewise.
6597 (*and_scc_scc_nod): Likewise.
6598 (*negscc): Likewise.
6599 (movcond_addsi): Likewise.
6600 (movcond): Likewise.
6601 (*ifcompare_plus_move): Likewise.
6602 (*if_plus_move): Likewise.
6603 (*ifcompare_move_plus): Likewise.
6604 (*if_move_plus): Likewise.
6605 (*ifcompare_arith_arith): Likewise.
6606 (*if_arith_arith): Likewise.
6607 (*ifcompare_arith_move): Likewise.
6608 (*if_arith_move): Likewise.
6609 (*ifcompare_move_arith): Likewise.
6610 (*if_move_arith): Likewise.
6611 (*ifcompare_move_not): Likewise.
6612 (*if_move_not): Likewise.
6613 (*ifcompare_not_move): Likewise.
6614 (*if_not_move): Likewise.
6615 (*ifcompare_shift_move): Likewise.
6616 (*if_shift_move): Likewise.
6617 (*ifcompare_move_shift): Likewise.
6618 (*if_move_shift): Likewise.
6619 (*ifcompare_shift_shift): Likewise.
6620 (*ifcompare_not_arith): Likewise.
6621 (*ifcompare_arith_not): Likewise.
6622 (*if_arith_not): Likewise.
6623 (*ifcompare_neg_move): Likewise.
6624 (*if_neg_move): Likewise.
6625 (*ifcompare_move_neg): Likewise.
6626 (*if_move_neg): Likewise.
6627 (prologue_thumb1_interwork): Likewise.
6628 (*cond_move_not): Likewise.
6629 (*sign_extract_onebit): Likewise.
6630 (*not_signextract_onebit): Likewise.
6631 (stack_tie): Likewise.
6632 (align_4): Likewise.
6633 (align_8): Likewise.
6634 (consttable_end): Likewise.
6635 (consttable_1): Likewise.
6636 (consttable_2): Likewise.
6637 (consttable_4): Likewise.
6638 (consttable_8): Likewise.
6639 (consttable_16): Likewise.
6640 (*thumb1_tablejump): Likewise.
6641 (prefetch): Likewise.
6642 (force_register_use): Likewise.
6643 (thumb_eh_return): Likewise.
6644 (load_tp_hard): Likewise.
6645 (load_tp_soft): Likewise.
6646 (tlscall): Likewise.
6647 (*arm_movtas_ze): Likewise.
6648 (*arm_rev): Likewise.
6649 (*arm_revsh): Likewise.
6650 (*arm_rev16): Likewise.
6651 * config/arm/thumb2.md
6652 (*thumb2_smaxsi3): Likewise.
6653 (*thumb2_sminsi3): Likewise.
6654 (*thumb32_umaxsi3): Likewise.
6655 (*thumb2_uminsi3): Likewise.
6656 (*thumb2_negdi2): Likewise.
6657 (*thumb2_abssi2): Likewise.
6658 (*thumb2_neg_abss): Likewise.
6659 (*thumb2_movsi_insn): Likewise.
6660 (tls_load_dot_plus_four): Likewise.
6661 (*thumb2_movhi_insn): Likewise.
6662 (*thumb2_mov_scc): Likewise.
6663 (*thumb2_mov_negs): Likewise.
6664 (*thumb2_mov_negs): Likewise.
6665 (*thumb2_mov_nots): Likewise.
6666 (*thumb2_mov_nots): Likewise.
6667 (*thumb2_movsicc_): Likewise.
6668 (*thumb2_movsfcc_soft_insn): Likewise.
6669 (*thumb2_indirect_jump): Likewise.
6670 (*thumb2_and_scc): Likewise.
6671 (*thumb2_ior_scc): Likewise.
6672 (*thumb2_ior_scc_strict_it): Likewise.
6673 (*thumb2_cond_move): Likewise.
6674 (*thumb2_cond_arith): Likewise.
6675 (*thumb2_cond_ari): Likewise.
6676 (*thumb2_cond_sub): Likewise.
6677 (*thumb2_negscc): Likewise.
6678 (*thumb2_movcond): Likewise.
6679 (thumb2_casesi_internal): Likewise.
6680 (thumb2_casesi_internal_pic): Likewise.
6681 (*thumb2_alusi3_short): Likewise.
6682 (*thumb2_mov<mode>_shortim): Likewise.
6683 (*thumb2_addsi_short): Likewise.
6684 (*thumb2_subsi_short): Likewise.
6685 (thumb2_addsi3_compare0): Likewise.
6686 (*thumb2_cbz): Likewise.
6687 (*thumb2_cbnz): Likewise.
6688 (*thumb2_one_cmplsi2_short): Likewise.
6689 (*thumb2_negsi2_short): Likewise.
6690 (*orsi_notsi_si): Likewise.
6691 * config/arm/arm1020e.md: Update with new attributes.
6692 * config/arm/arm1026ejs.md: Update with new attributes.
6693 * config/arm/arm1136jfs.md: Update with new attributes.
6694 * config/arm/arm926ejs.md: Update with new attributes.
6695 * config/arm/cortex-a15.md: Update with new attributes.
6696 * config/arm/cortex-a5.md: Update with new attributes.
6697 * config/arm/cortex-a53.md: Update with new attributes.
6698 * config/arm/cortex-a7.md: Update with new attributes.
6699 * config/arm/cortex-a8.md: Update with new attributes.
6700 * config/arm/cortex-a9.md: Update with new attributes.
6701 * config/arm/cortex-m4.md: Update with new attributes.
6702 * config/arm/cortex-r4.md: Update with new attributes.
6703 * config/arm/fa526.md: Update with new attributes.
6704 * config/arm/fa606te.md: Update with new attributes.
6705 * config/arm/fa626te.md: Update with new attributes.
6706 * config/arm/fa726te.md: Update with new attributes.
6707
1c83b673
JG		 67082013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6709
6710 * config/aarch64/aarch64-simd.md
6711 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6712 <vwx> iterator to ensure correct register choice.
6713 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6714 (aarch64_sqdmull_n<mode>): Likewise.
6715 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6716 * config/aarch64/arm_neon.h
6717 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6718 (vml<as><q>_n_<su>16): Likewise.
6719 (vml<as>l_high_lane<q>_<su>16): Likewise.
6720 (vml<as>l_high_n_<su>16): Likewise.
6721 (vml<as>l_lane<q>_<su>16): Likewise.
6722 (vml<as>l_n_<su>16): Likewise.
6723 (vmul<q>_lane<q>_<su>16): Likewise.
6724 (vmul<q>_n_<su>16): Likewise.
6725 (vmull_lane<q>_<su>16): Likewise.
6726 (vmull_n_<su>16): Likewise.
6727 (vmull_high_lane<q>_<su>16): Likewise.
6728 (vmull_high_n_<su>16): Likewise.
6729 (vqrdmulh<q>_n_s16): Likewise.
6730
bb1ae543
TB		 67312013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6732
6733 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6734 have the correct lane parameter.
6735
f23c0742
RB		 67362013-09-06 Richard Biener <rguenther@suse.de>
6737
6738 * cfganal.c (control_dependences::~control_dependences):
6739 Properly free all of the vector.
6740
c1d49770
KY		 67412013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6742
6743 PR target/58269
6744 * config/i386/i386.c (ix86_conditional_register_usage):
6745 Proper initialize extended SSE registers.
6746
6da47f52
JH		 67472013-09-06 Jan Hubicka <jh@suse.cz>
6748
6749 PR tree-optimization/58311
6750 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6751
67522013-09-06 Jan Hubicka <jh@suse.cz>
6753
6754 * Makefile.in (tree-sra.o): Update dependencies.
6755 * tree-sra.c: Include ipa-utils.h
6756 (scan_function): Use recursive_call_p.
6757 (has_caller_p): New function.
6758 (cgraph_for_node_and_aliases): Count also callers of aliases.
6759
fc11f321
JH		 67602013-09-06 Jan Hubicka <jh@suse.cz>
6761
6762 PR middle-end/58094
6763 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6764 * tree-tailcall.c: Include ipa-utils.h.
6765 (find_tail_calls): Use it.
6766 * ipa-pure-const.c (check_call): Likewise.
6767 * ipa-utils.c (recursive_call_p): New function.
6768 * ipa-utils.h (recursive_call_p): Dclare.
6769 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6770 (symtab_semantically_equivalent_p): New function.
6771 * Makefile.in (tree-tailcall.o): Update dependencies.
6772
9a6af450
EB		 67732013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6774
6775 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6776 non-inlinable part.
6777
35f32ad4
RB		 67782013-09-06 Richard Biener <rguenther@suse.de>
6779
6780 * lto-streamer.h (lto_global_var_decls): Remove.
6781 * Makefile.in (OBJS): Remove lto-symtab.o.
6782 (lto-symtab.o): Remove.
6783 (GTFILES): Remove lto-symtab.c
6784 * lto-symtab.c: Move to lto/
6785
d12a76f3
AK		 67862013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6787
6788 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6789 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6790 (UNSPEC_FPINT_RINT): New constant definitions.
6791 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6792 definition with 2 attributes.
6793 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6794 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6795 definitions.
6796
67972013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6798
f43245d1 6799 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK		 6800 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6801 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6802 attribute to "z196_alone".
6803 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6804 "zEC12_simple".
6805
c8e9d8c3
RB		 68062013-09-06 Richard Biener <rguenther@suse.de>
6807
6808 * basic-block.h (class control_dependences): New.
6809 * tree-ssa-dce.c (control_dependence_map): Remove.
6810 (cd): New global.
6811 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6812 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6813 find_pdom, find_control_dependence, find_all_control_dependences):
6814 Move to cfganal.c.
f43245d1
UB		 6815 (mark_control_dependent_edges_necessary,
6816 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6817 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB		 6818 * cfganal.c (set_control_dependence_map_bit,
6819 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6820 find_all_control_dependences): Move from tree-ssa-dce.c and
6821 implement as methods of control_dependences class.
6822 (control_dependences::control_dependences): New.
6823 (control_dependences::~control_dependences): Likewise.
6824 (control_dependences::get_edges_dependent_on): Likewise.
6825 (control_dependences::get_edge): Likewise.
6826
7d2ba471
JH		 68272013-09-04 Jan Hubicka <jh@suse.cz>
6828
6829 * tree.c (types_same_for_odr): Drop overactive check.
6830 * ipa-devirt.c (hash_type_name): Likewise.
6831
e18412fc
JH		 68322013-09-04 Jan Hubicka <jh@suse.cz>
6833
6834 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6835 (analyze_functions): ... here.
6836
d352b245
JH		 68372013-09-04 Jan Hubicka <jh@suse.cz>
6838
6839 PR middle-end/58201
6840 * cgraphunit.c (analyze_functions): Clear AUX fields
6841 after processing; initialize assembler name has.
6842
770da076
JL		 68432013-09-05 Jeff Law <law@redhat.com>
6844
6845 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6846 from thread_around_empty_block. Record threading path into PATH.
6847 Recurse if threading through the initial block is successful.
6848 (thread_across_edge): Corresponding changes to slightly simplify.
6849
fe6f68e2
JG		 68502013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6851
6852 * config/aarch64/aarch64.md
6853 (type): Remove frecpe, frecps, frecpx.
6854 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6855 fix to be a TARGET_SIMD instruction.
6856 (aarch64_frecps): Remove.
6857 * config/aarch64/aarch64-simd.md
6858 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6859 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6860
6e4150e1
JG		 68612013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6862 Sofiane Naci <sofiane.naci@arm.com>
6863
f43245d1 6864 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 6865 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB		 6866 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6867 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6868 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6869 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6870 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6871 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6872 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 6873 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 6874 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG		 6875 shift* types.
6876 * config/arm/arm-fixed.md: Update for attribute changes
6877 all occurrences of arlo_* types.
6878 * config/arm/thumb2.md: Update for attribute changes all occurrences
6879 of arlo_* types.
6880 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6881 (cortexa7_older_only): Likewise.
6882 (cortexa7_younger): Likewise.
6883 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6884 (1020alu_shift_op): Likewise.
6885 (1020alu_shift_reg_op): Likewise.
6886 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6887 (alu_shift_op): Likewise.
6888 (alu_shift_reg_op): Likewise.
f43245d1 6889 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG		 6890 (11_alu_shift_op): Likewise.
6891 (11_alu_shift_reg_op): Likewise.
6892 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6893 (9_alu_shift_reg_op): Likewise.
6894 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6895 attribute changes.
6896 (cortex_a15_alu_shift): Likewise.
6897 (cortex_a15_alu_shift_reg): Likewise.
6898 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6899 attribute changes.
6900 (cortex_a5_alu_shift): Likewise.
f43245d1
UB		 6901 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6902 attribute changes.
6e4150e1 6903 (cortex_a53_alu_shift): Likewise.
f43245d1
UB		 6904 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6905 attribute changes.
6e4150e1
JG		 6906 (cortex_a7_alu_reg): Likewise.
6907 (cortex_a7_alu_shift): Likewise.
f43245d1
UB		 6908 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6909 attribute changes.
6e4150e1
JG		 6910 (cortex_a8_alu_shift): Likewise.
6911 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 6912 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 6913 (cortex_a9_dp_shift): Likewise.
f43245d1
UB		 6914 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6915 attribute changes.
6e4150e1
JG		 6916 * config/arm/cortex-r4.md
6917 (cortex_r4_alu): Update for attribute changes.
6918 (cortex_r4_mov): Likewise.
6919 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 6920 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 6921 (526_alu_shift_op): Likewise.
f43245d1
UB		 6922 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6923 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 6924 (626te_alu_shift_op): Likewise.
f43245d1 6925 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG		 6926 (726te_alu_shift_op): Likewise.
6927 (726te_alu_shift_reg_op): Likewise.
6928 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6929 (mp626_alu_shift_op): Likewise.
6930 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6931 (pj4_alu_conds): Likewise.
6932 (pj4_shift): Likewise.
6933 (pj4_shift_conds): Likewise.
6934 (pj4_alu_shift): Likewise.
6935 (pj4_alu_shift_conds): Likewise.
6936 * config/aarch64/aarch64.md: Update for attribute change
6937 all occurrences of arlo_* and shift* types.
6938
e6523306
MS		 69392013-09-05 Mike Stump <mikestump@comcast.net>
6940
6941 * tree.h: Move documentation for tree_function_decl to tree-core.h
6942 with the declaration.
6943
80a18161
PB		 69442013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6945
6946 PR target/58139
6947 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6948 looking for widest mode.
6949
6994430a
EB		 69502013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6951
6952 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6953
e67f39f7
RB		 69542013-09-05 Richard Biener <rguenther@suse.de>
6955
6956 PR tree-optimization/58137
6957 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6958 Do not create vectors of pointers.
6959 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6960 types for the components of the vector initializer.
6961 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6962 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6963
568cda29
MJ		 69642013-09-05 Martin Jambor <mjambor@suse.cz>
6965
6966 * ipa-prop.c (remove_described_reference): Accept missing references,
6967 return false if that hppens, otherwise return true.
6968 (cgraph_node_for_jfunc): New function.
6969 (try_decrement_rdesc_refcount): Likewise.
6970 (try_make_edge_direct_simple_call): Use them.
6971 (ipa_edge_removal_hook): Remove references from rdescs.
6972 (ipa_edge_duplication_hook): Clone rdescs and their references
6973 when the new edge has the same caller as the old one.
6974 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6975 reference before removing any edges.
6976
5a200acb
RE		 69772013-09-05 Richard Earnshaw <rearnsha@arm.com>
6978
6979 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6980 initial store.
6981 * thumb2.md (thumb2_storewb_parisi): New pattern.
6982
5922847b
YZ		 69832013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6984
6985 * config/aarch64/aarch64-option-extensions.def: Add
6986 AARCH64_OPT_EXTENSION of 'crc'.
6987 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6988 (AARCH64_ISA_CRC): Ditto.
6989 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6990 description of the CRC extension.
6991
996746aa
AI		 69922013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6993
6994 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6995 * config/rs6000/linux.h: Ditto.
6996 * alpha/linux.h: Ditto.
6997 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6998 no_c99_libc_has_function.
6999 * config/c6x/uclinux-elf.h: Ditto.
7000 * config/lm32/uclinux-elf.h: Ditto.
7001 * config/m68k/uclinux.h: Ditto.
7002 * config/moxie/uclinux.h: Ditto.
7003 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7004 (crisv32-*-linux*, cris-*-linux*): Ditto.
7005 * config/bfin/bfin.c: Include "tm_p.h".
7006
c370d99f
RB		 70072013-09-05 Richard Biener <rguenther@suse.de>
7008
7009 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7010 check for a definition without a basic-block.
7011
003bb7f3
JG		 70122013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7013 Sofiane Naci <sofiane.naci@arm.com>
7014
7015 * config/aarch64/aarch64.md
7016 (*movti_aarch64): Rename r_2_f and f_2_r.
7017 (*movsf_aarch64): Likewise.
7018 (*movdf_aarch64): Likewise.
7019 (*movtf_aarch64): Likewise.
7020 (aarch64_movdi_<mode>low): Likewise.
7021 (aarch64_movdi_<mode>high): Likewise.
7022 (aarch64_mov<mode>high_di): Likewise.
7023 (aarch64_mov<mode>low_di): Likewise.
7024 (aarch64_movtilow_tilow): Likewise.
7025 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7026 values to config/arm/types.md
7027 (attribute "conds"): Update for attribute change.
7028 (anddi3_insn): Likewise.
7029 (iordi3_insn): Likewise.
7030 (xordi3_insn): Likewise.
7031 (one_cmpldi2): Likewise.
7032 * config/arm/types.md (type): Add Neon types.
7033 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7034 use "type" attribute.
7035 (movmisalign<mode>_neon_store): Likewise.
7036 (movmisalign<mode>_neon_load): Likewise.
7037 (vec_set<mode>_internal): Likewise.
7038 (vec_setv2di_internal): Likewise.
7039 (vec_extract<mode>): Likewise.
7040 (vec_extractv2di): Likewise.
7041 (add<mode>3_neon): Likewise.
7042 (adddi3_neon): Likewise.
7043 (sub<mode>3_neon): Likewise.
7044 (subdi3_neon): Likewise.
7045 (mul<mode>3_neon): Likewise.
7046 (mul<mode>3add<mode>_neon): Likewise.
7047 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7048 (fma<VCVTF:mode>4)): Likewise.
7049 (fma<VCVTF:mode>4_intrinsic): Likewise.
7050 (fmsub<VCVTF:mode>4)): Likewise.
7051 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7052 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7053 (ior<mode>3): Likewise.
7054 (and<mode>3): Likewise.
7055 (anddi3_neon): Likewise.
7056 (orn<mode>3_neon): Likewise.
7057 (orndi3_neon): Likewise.
7058 (bic<mode>3_neon): Likewise.
7059 (bicdi3_neon): Likewise.
7060 (xor<mode>3): Likewise.
7061 (one_cmpl<mode>2): Likewise.
7062 (abs<mode>2): Likewise.
7063 (neg<mode>2): Likewise.
7064 (umin<mode>3_neon): Likewise.
7065 (umax<mode>3_neon): Likewise.
7066 (smin<mode>3_neon): Likewise.
7067 (smax<mode>3_neon): Likewise.
7068 (vashl<mode>3): Likewise.
7069 (vashr<mode>3_imm): Likewise.
7070 (vlshr<mode>3_imm): Likewise.
7071 (ashl<mode>3_signed): Likewise.
7072 (ashl<mode>3_unsigned): Likewise.
7073 (neon_load_count): Likewise.
7074 (ashldi3_neon_noclobber): Likewise.
7075 (signed_shift_di3_neon): Likewise.
7076 (unsigned_shift_di3_neon): Likewise.
7077 (ashrdi3_neon_imm_noclobber): Likewise.
7078 (lshrdi3_neon_imm_noclobber): Likewise.
7079 (widen_ssum<mode>3): Likewise.
7080 (widen_usum<mode>3): Likewise.
7081 (quad_halves_<code>v4si): Likewise.
7082 (quad_halves_<code>v4sf): Likewise.
7083 (quad_halves_<code>v8hi): Likewise.
7084 (quad_halves_<code>v16qi): Likewise.
7085 (reduc_splus_v2di): Likewise.
7086 (neon_vpadd_internal<mode>): Likewise.
7087 (neon_vpsmin<mode>): Likewise.
7088 (neon_vpsmax<mode>): Likewise.
7089 (neon_vpumin<mode>): Likewise.
7090 (neon_vpumax<mode>): Likewise.
7091 (ss_add<mode>_neon): Likewise.
7092 (us_add<mode>_neon): Likewise.
7093 (ss_sub<mode>_neon): Likewise.
7094 (us_sub<mode>_neon): Likewise.
7095 (neon_vadd<mode>_unspec): Likewise.
7096 (neon_vaddl<mode>): Likewise.
7097 (neon_vaddw<mode>): Likewise.
7098 (neon_vhadd<mode>): Likewise.
7099 (neon_vqadd<mode>): Likewise.
7100 (neon_vaddhn<mode>): Likewise.
7101 (neon_vmul<mode>): Likewise.
7102 (neon_vmla<mode>): Likewise.
7103 (neon_vmlal<mode>): Likewise.
7104 (neon_vmls<mode>): Likewise.
7105 (neon_vmlsl<mode>): Likewise.
7106 (neon_vqdmulh<mode>): Likewise.
7107 (neon_vqdmlal<mode>): Likewise.
7108 (neon_vqdmlsl<mode>): Likewise.
7109 (neon_vmull<mode>): Likewise.
7110 (neon_vqdmull<mode>): Likewise.
7111 (neon_vsub<mode>_unspec): Likewise.
7112 (neon_vsubl<mode>): Likewise.
7113 (neon_vsubw<mode>): Likewise.
7114 (neon_vqsub<mode>): Likewise.
7115 (neon_vhsub<mode>): Likewise.
7116 (neon_vsubhn<mode>): Likewise.
7117 (neon_vceq<mode>): Likewise.
7118 (neon_vcge<mode>): Likewise.
7119 (neon_vcgeu<mode>): Likewise.
7120 (neon_vcgt<mode>): Likewise.
7121 (neon_vcgtu<mode>): Likewise.
7122 (neon_vcle<mode>): Likewise.
7123 (neon_vclt<mode>): Likewise.
7124 (neon_vcage<mode>): Likewise.
7125 (neon_vcagt<mode>): Likewise.
7126 (neon_vtst<mode>): Likewise.
7127 (neon_vabd<mode>): Likewise.
7128 (neon_vabdl<mode>): Likewise.
7129 (neon_vaba<mode>): Likewise.
7130 (neon_vabal<mode>): Likewise.
7131 (neon_vmax<mode>): Likewise.
7132 (neon_vmin<mode>): Likewise.
7133 (neon_vpaddl<mode>): Likewise.
7134 (neon_vpadal<mode>): Likewise.
7135 (neon_vpmax<mode>): Likewise.
7136 (neon_vpmin<mode>): Likewise.
7137 (neon_vrecps<mode>): Likewise.
7138 (neon_vrsqrts<mode>): Likewise.
7139 (neon_vqabs<mode>): Likewise.
7140 (neon_vqneg<mode>): Likewise.
7141 (neon_vcls<mode>): Likewise.
7142 (clz<mode>2): Likewise.
7143 (popcount<mode>2): Likewise.
7144 (neon_vrecpe): Likewise.
7145 (neon_vrsqrte): Likewise.
7146 (neon_vget_lane<mode>_sext_internal): Likewise.
7147 (neon_vget_lane<mode>_zext_internal): Likewise.
7148 (neon_vdup_n<mode>): Likewise.
7149 (neon_vdup_nv2di): Likewise.
7150 (neon_vdpu_lane<mode>_internal): Likewise.
7151 (neon_vswp<mode>): Likewise.
7152 (float<mode><V_cvtto>2): Likewise.
7153 (floatuns<mode><V_cvtto>2): Likewise.
7154 (fix_trunc<mode><V_cvtto>)2): Likewise
7155 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7156 (neon_vcvt<mode>): Likewise.
7157 (neon_vcvtv4sfv4hf): Likewise.
7158 (neon_vcvtv4hfv4sf): Likewise.
7159 (neon_vcvt_n<mode>): Likewise.
7160 (neon_vmovn<mode>): Likewise.
7161 (neon_vqmovn<mode>): Likewise.
7162 (neon_vqmovun<mode>): Likewise.
7163 (neon_vmovl<mode>): Likewise.
7164 (neon_vmul_lane<mode>): Likewise.
7165 (neon_vmull_lane<mode>): Likewise.
7166 (neon_vqdmull_lane<mode>): Likewise.
7167 (neon_vqdmulh_lane<mode>): Likewise.
7168 (neon_vmla_lane<mode>): Likewise.
7169 (neon_vmlal_lane<mode>): Likewise.
7170 (neon_vqdmlal_lane<mode>): Likewise.
7171 (neon_vmls_lane<mode>): Likewise.
7172 (neon_vmlsl_lane<mode>): Likewise.
7173 (neon_vqdmlsl_lane<mode>): Likewise.
7174 (neon_vext<mode>): Likewise.
7175 (neon_vrev64<mode>): Likewise.
7176 (neon_vrev32<mode>): Likewise.
7177 (neon_vrev16<mode>): Likewise.
7178 (neon_vbsl<mode>_internal): Likewise.
7179 (neon_vshl<mode>): Likewise.
7180 (neon_vqshl<mode>): Likewise.
7181 (neon_vshr_n<mode>): Likewise.
7182 (neon_vshrn_n<mode>): Likewise.
7183 (neon_vqshrn_n<mode>): Likewise.
7184 (neon_vqshrun_n<mode>): Likewise.
7185 (neon_vshl_n<mode>): Likewise.
7186 (neon_vqshl_n<mode>): Likewise.
7187 (neon_vqshlu_n<mode>): Likewise.
7188 (neon_vshll_n<mode>): Likewise.
7189 (neon_vsra_n<mode>): Likewise.
7190 (neon_vsri_n<mode>): Likewise.
7191 (neon_vsli_n<mode>): Likewise.
7192 (neon_vtbl1v8qi): Likewise.
7193 (neon_vtbl2v8qi): Likewise.
7194 (neon_vtbl3v8qi): Likewise.
7195 (neon_vtbl4v8qi): Likewise.
7196 (neon_vtbx1v8qi): Likewise.
7197 (neon_vtbx2v8qi): Likewise.
7198 (neon_vtbx3v8qi): Likewise.
7199 (neon_vtbx4v8qi): Likewise.
7200 (neon_vtrn<mode>_internal): Likewise.
7201 (neon_vzip<mode>_internal): Likewise.
7202 (neon_vuzp<mode>_internal): Likewise.
7203 (neon_vld1<mode>): Likewise.
7204 (neon_vld1_lane<mode>): Likewise.
7205 (neon_vld1_dup<mode>): Likewise.
7206 (neon_vld1_dupv2di): Likewise.
7207 (neon_vst1<mode>): Likewise.
7208 (neon_vst1_lane<mode>): Likewise.
7209 (neon_vld2<mode>): Likewise.
7210 (neon_vld2_lane<mode>): Likewise.
7211 (neon_vld2_dup<mode>): Likewise.
7212 (neon_vst2<mode>): Likewise.
7213 (neon_vst2_lane<mode>): Likewise.
7214 (neon_vld3<mode>): Likewise.
7215 (neon_vld3qa<mode>): Likewise.
7216 (neon_vld3qb<mode>): Likewise.
7217 (neon_vld3_lane<mode>): Likewise.
7218 (neon_vld3_dup<mode>): Likewise.
7219 (neon_vst3<mode>): Likewise.
7220 (neon_vst3qa<mode>): Likewise.
7221 (neon_vst3qb<mode>): Likewise.
7222 (neon_vst3_lane<mode>): Likewise.
7223 (neon_vld4<mode>): Likewise.
7224 (neon_vld4qa<mode>): Likewise.
7225 (neon_vld4qb<mode>): Likewise.
7226 (neon_vld4_lane<mode>): Likewise.
7227 (neon_vld4_dup<mode>): Likewise.
7228 (neon_vst4<mode>): Likewise.
7229 (neon_vst4qa<mode>): Likewise.
7230 (neon_vst4qb<mode>): Likewise.
7231 (neon_vst4_lane<mode>): Likewise.
7232 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7233 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7234 (neon_vec_<US>mult_lo_<mode>): Likewise.
7235 (neon_vec_<US>mult_hi_<mode>): Likewise.
7236 (neon_vec_<US>shiftl_<mode>): Likewise.
7237 (neon_unpack<US>_<mode>): Likewise.
7238 (neon_vec_<US>mult_<mode>): Likewise.
7239 (vec_pack_trunc_<mode>): Likewise.
7240 (neon_vec_pack_trunk_<mode>): Likewise.
7241 (neon_vabd<mode>_2): Likewise.
7242 (neon_vabd<mode>_3): Likewise.
7243 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7244 (thumb2_movsi_vfp): Likewise.
7245 (movdi_vfp): Likewise.
7246 (movdi_vfp_cortexa8): Likewise.
7247 (movhf_vfp_neon): Likewise.
7248 (movhf_vfp): Likewiwse.
7249 (movsf_vfp): Likewiwse.
7250 (thumb2_movsf_vfp): Likewiwse.
7251 (movdf_vfp): Likewise.
7252 (thumb2_movdf_vfp): Likewise.
7253 (movsfcc_vfp): Likewise.
7254 (thumb2_movsfcc_vfp): Likewise.
7255 (movdfcc_vfp): Likewise.
7256 (thumb2_movdfcc_vfp): Likewise.
7257 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7258 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7259 (v10_v2c): Likewise.
7260 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7261 attribute change.
7262 (cortex_a15_neon_int_2): Likewise.
7263 (cortex_a15_neon_int_3): Likewise.
7264 (cortex_a15_neon_int_4): Likewise.
7265 (cortex_a15_neon_int_5): Likewise.
7266 (cortex_a15_neon_vqneg_vqabs): Likewise.
7267 (cortex_a15_neon_vmov): Likewise.
7268 (cortex_a15_neon_vaba): Likewise.
7269 (cortex_a15_neon_vaba_qqq): Likewise.
7270 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7271 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7272 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7273 scalar_64_32_long_scalar): Likewise.
7274 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7275 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7276 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7277 lotype_qdd_64_32_long): Likewise.
7278 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7279 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7280 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7281 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7282 (cortex_a15_neon_shift_1): Likewise.
7283 (cortex_a15_neon_shift_2): Likewise.
7284 (cortex_a15_neon_shift_3): Likewise.
7285 (cortex_a15_neon_vshl_ddd): Likewise.
7286 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7287 (cortex_a15_neon_vsra_vrsra): Likewise.
7288 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7289 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7290 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7291 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7292 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7293 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7294 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7295 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7296 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7297 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7298 (cortex_a15_neon_bp_simple): Likewise.
7299 (cortex_a15_neon_bp_2cycle): Likewise.
7300 (cortex_a15_neon_bp_3cycle): Likewise.
7301 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7302 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7303 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7304 (cortex_a15_neon_vld2_4_regs): Likewise.
7305 (cortex_a15_neon_vld3_vld4): Likewise.
7306 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7307 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7308 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7309 (cortex_a15_neon_vst3_vst4): Likewise.
7310 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7311 (cortex_a15_neon_vld3_vld4_lane" 10
7312 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7313 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7314 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7315 (cortex_a15_neon_ldm_2): Likewise.0
7316 (cortex_a15_neon_stm_2): Likewise.
7317 (cortex_a15_neon_mcr): Likewise.
7318 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7319 (cortex_a15_neon_mrc): Likewise.
7320 (cortex_a15_neon_mrrc): Likewise.
7321 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7322 change.
7323 (cortex_a15_alu_shift): Likewise.
7324 (cortex_a15_alu_shift_reg): Likewise.
7325 (cortex_a15_mult32): Likewise.
7326 (cortex_a15_mult64): Likewise.
7327 (cortex_a15_block): Likewise.
7328 (cortex_a15_branch): Likewise.
7329 (cortex_a15_load1): Likewise.
7330 (cortex_a15_load3): Likewise.
7331 (cortex_a15_store1): Likewise.
7332 (cortex_a15_store3): Likewise.
7333 (cortex_a15_call): Likewise.
f43245d1 7334 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG		 7335 (cortex_a5_f2r): Likewise.
7336 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7337 change.
7338 (cortex_a53_f2r): Likewise.
7339 * config/arm/cortex-a7.md
7340 (cortex_a7_branch): Update for attribute change.
7341 (cortex_a7_call): Likewise.
7342 (cortex_a7_alu_imm): Likewise.
7343 (cortex_a7_alu_reg): Likewise.
7344 (cortex_a7_alu_shift): Likewise.
7345 (cortex_a7_mul): Likewise.
7346 (cortex_a7_load1): Likewise.
7347 (cortex_a7_store1): Likewise.
7348 (cortex_a7_load2): Likewise.
7349 (cortex_a7_store2): Likewise.
7350 (cortex_a7_load3): Likewise.
7351 (cortex_a7_store3): Likewise.
7352 (cortex_a7_load4): Likewise.
7353 (cortex_a7_store4): Likewise.
7354 (cortex_a7_fpalu): Likewise.
7355 (cortex_a7_fconst): Likewise.
7356 (cortex_a7_fpmuls): Likewise.
7357 (cortex_a7_neon_mul): Likewise.
7358 (cortex_a7_fpmacs): Likewise.
7359 (cortex_a7_neon_mla: Likewise.
7360 (cortex_a7_fpmuld: Likewise.
7361 (cortex_a7_fpmacd: Likewise.
7362 (cortex_a7_fpfmad: Likewise.
7363 (cortex_a7_fdivs: Likewise.
7364 (cortex_a7_fdivd: Likewise.
7365 (cortex_a7_r2f: Likewise.
7366 (cortex_a7_f2r: Likewise.
7367 (cortex_a7_f_flags: Likewise.
7368 (cortex_a7_f_loads: Likewise.
7369 (cortex_a7_f_loadd: Likewise.
7370 (cortex_a7_f_stores: Likewise.
7371 (cortex_a7_f_stored: Likewise.
7372 (cortex_a7_neon): Likewise.
7373 * config/arm/cortex-a8-neon.md
7374 (cortex_a8_neon_mrc): Update for attribute change.
7375 (cortex_a8_neon_mrrc): Likewise.
7376 (cortex_a8_neon_int_1): Likewise.
7377 (cortex_a8_neon_int_2): Likewise.
7378 (cortex_a8_neon_int_3): Likewise.
7379 (cortex_a8_neon_int_4): Likewise.
7380 (cortex_a8_neon_int_5): Likewise.
7381 (cortex_a8_neon_vqneg_vqabs): Likewise.
7382 (cortex_a8_neon_vmov): Likewise.
7383 (cortex_a8_neon_vaba): Likewise.
7384 (cortex_a8_neon_vaba_qqq): Likewise.
7385 (cortex_a8_neon_vsma): Likewise.
7386 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7387 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 7388 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7389 long_scalar): Likewise.
003bb7f3
JG		 7390 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7391 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7392 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7393 long_scalar_qdd_64_32_long): Likewise.
7394 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7395 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7396 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7397 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7398 (cortex_a8_neon_shift_1): Likewise.
7399 (cortex_a8_neon_shift_2): Likewise.
7400 (cortex_a8_neon_shift_3): Likewise.
7401 (cortex_a8_neon_vshl_ddd): Likewise.
7402 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7403 (cortex_a8_neon_vsra_vrsra): Likewise.
7404 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7405 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7406 (cortex_a8_neon_fp_vsum): Likewise.
7407 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7408 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7409 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7410 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7411 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7412 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7413 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7414 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7415 (cortex_a8_neon_bp_simple): Likewise.
7416 (cortex_a8_neon_bp_2cycle): Likewise.
7417 (cortex_a8_neon_bp_3cycle): Likewise.
7418 (cortex_a8_neon_ldr): Likewise.
7419 (cortex_a8_neon_str): Likewise.
7420 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7421 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7422 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7423 (cortex_a8_neon_vld2_4_regs): Likewise.
7424 (cortex_a8_neon_vld3_vld4): Likewise.
7425 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7426 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7427 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7428 (cortex_a8_neon_vst3_vst4): Likewise.
7429 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7430 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7431 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7432 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7433 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7434 (cortex_a8_neon_mcr): Likewise.
7435 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 7436 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG		 7437 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7438 change.
7439 (ca9_neon_mrrc): Likewise.
7440 (cortex_a9_neon_int_1): Likewise.
7441 (cortex_a9_neon_int_2): Likewise.
7442 (cortex_a9_neon_int_3): Likewise.
7443 (cortex_a9_neon_int_4): Likewise.
7444 (cortex_a9_neon_int_5): Likewise.
7445 (cortex_a9_neon_vqneg_vqabs): Likewise.
7446 (cortex_a9_neon_vmov): Likewise.
7447 (cortex_a9_neon_vaba): Likewise.
7448 (cortex_a9_neon_vaba_qqq): Likewise.
7449 (cortex_a9_neon_vsma): Likewise.
7450 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7451 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 7452 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7453 long_scalar): Likewise.
003bb7f3
JG		 7454 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7455 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7456 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7457 long_scalar_qdd_64_32_long): Likewise.
7458 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7459 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7460 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7461 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7462 (cortex_a9_neon_shift_1): Likewise.
7463 (cortex_a9_neon_shift_2): Likewise.
7464 (cortex_a9_neon_shift_3): Likewise.
7465 (cortex_a9_neon_vshl_ddd): Likewise.
7466 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7467 (cortex_a9_neon_vsra_vrsra): Likewise.
7468 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7469 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7470 (cortex_a9_neon_fp_vsum): Likewise.
7471 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7472 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7473 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7474 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7475 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7476 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7477 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7478 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7479 (cortex_a9_neon_bp_simple): Likewise.
7480 (cortex_a9_neon_bp_2cycle): Likewise.
7481 (cortex_a9_neon_bp_3cycle): Likewise.
7482 (cortex_a9_neon_ldr): Likewise.
7483 (cortex_a9_neon_str): Likewise.
7484 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7485 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7486 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7487 (cortex_a9_neon_vld2_4_regs): Likewise.
7488 (cortex_a9_neon_vld3_vld4): Likewise.
7489 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7490 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7491 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7492 (cortex_a9_neon_vst3_vst4): Likewise.
7493 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7494 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7495 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7496 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7497 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7498 (cortex_a9_neon_mcr): Likewise.
7499 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7500 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7501 (cortex_a9_fps): Likewise.
7502 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7503 change.
7504 (cortex_m4_fmuls): Likewise.
7505 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7506 change.
7507 (cortex_r4_mrc): Likewise.
7508 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 7509 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 7510 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB		 7511 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7512 attribute change.
003bb7f3
JG		 7513 (pj4_core_to_vfp): Likewise.
7514 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7515 attribute change.
7516 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7517 (vfp_fstore): Likewise.
7518 * doc/md.texi: Change references to neon_type to refer to type.
7519
fbe0dc78
DS		 75202013-09-04 Dodji Seketeli <dodji@redhat.com>
7521
7522 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7523
c167bc5b
DE		 75242013-09-04 David Edelsohn <dje.gcc@gmail.com>
7525
7526 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7527 lglobl if not weak.
7528
035cb59f
ER		 75292013-09-04 Easwaran Raman <eraman@google.com>
7530
7531 PR middle-end/57370
7532 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7533 (build_and_add_sum): Use it.
7534 (appears_later_in_bb): Simplify code.
7535
71191083
TJ		 75362013-09-04 Teresa Johnson <tejohnson@google.com>
7537
7538 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7539
0c488213
JG		 75402013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7541
7542 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7543
61a74079
JH		 75442013-09-04 Jan Hubicka <jh@suse.cz>
7545
7546 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7547 * ipa-devirt.c: Include diganostic.h
7548 (odr_type_d): Add types and types_set.
7549 (hash_type_name): Work for types with vtables during LTO.
7550 (odr_hasher::remove): Fix comment; destroy types_set.
7551 (add_type_duplicate): New function,
7552 (get_odr_type): Use it.
7553 (dump_type_inheritance_graph): Dump type duplicates.
7554 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7555 graph.
7556 * tree.c (types_same_for_odr): Give exact answers on types with
7557 virtual tables.
7558
24ba8195
DS		 75592013-09-04 Dodji Seketeli <dodji@redhat.com>
7560
7561 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7a5911d3 7562 explaining their differences.
24ba8195 7563
0a04869e
SKS		 75642013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7565
7566 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7567
3b18bc42
JL		 75682013-09-03 Jeff Law <law@redhat.com>
7569
7570 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7571 when not threading through a joiner block. Pass joiner/no joiner
7572 state to register_jump_thread.
7573 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7574 state from argument rather than implying on path length.
7575 Dump the entire jump thread path into debugging dump.
7576 * tree-flow.h (register_jump_thread): Update prototype.
7577
ae84e151
XDL		 75782013-08-29 Xinliang David Li <davidxl@google.com>
7579
7580 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7581 Remove a trivial gcc_assert.
7582
c716e67f
XDL		 75832013-08-29 Xinliang David Li <davidxl@google.com>
7584
7585 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7586 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7587 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7588 Delay base decl alignment adjustment.
7589 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7590 * tree-vectorizer.h: New data structure.
7591 (set_dr_misalignment): New function.
7592 (dr_misalignment): Ditto.
7593 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7594 (vectorizable_load): Ditto.
7595 (ensure_base_align): New function.
7596 (vectorize_loops): Add dbg_cnt support.
7597 (execute_vect_slp): Ditto.
7598 * dbgcnt.def: New debug counter.
7599 * Makefile: New dependency.
7600
db8800bc
MI		 76012013-09-03 Meador Inge <meadori@codesourcery.com>
7602
7603 Revert:
7604
7605 2013-08-30 Meador Inge <meadori@codesourcery.com>
7606
7607 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7608
0d9b4f5c
DE		 76092013-09-03 David Edelsohn <dje.gcc@gmail.com>
7610
7611 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7612 function descriptor.
7613
29e810b9
RB		 76142013-09-03 Richard Biener <rguenther@suse.de>
7615
7616 * tree-affine.c (add_elt_to_tree): Fix association issue,
7617 avoid useless converts and make sure to always return a
7618 properly typed result.
7619
dbfc2894
RB		 76202013-09-03 Richard Biener <rguenther@suse.de>
7621
7622 PR middle-end/57656
7623 * fold-const.c (negate_expr_p): Fix division case.
7624 (negate_expr): Likewise.
7625
fa3bf4ce
RB		 76262013-09-03 Richard Biener <rguenther@suse.de>
7627
7628 PR lto/58285
7629 * tree-streamer-out.c: Include tm.h.
7630 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7631
a64bbb3f
JH		 76322013-09-03 Jan Hubicka <jh@suse.cz>
7633
7634 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7635
93ed07e2
AM		 76362013-09-03 Alan Modra <amodra@gmail.com>
7637
7638 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7639 multilibs to big-endian target and vice versa.
7640 * config/rs6000/t-linux64: Use := assignment on all vars.
7641 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7642 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7643 * config/rs6000/t-linux64le: New file.
7644 * config/rs6000/t-linux64bele: New file.
7645 * config/rs6000/t-linux64lebe: New file.
7646
07d6cd64
JH		 76472013-09-02 Jan Hubicka <jh@suse.cz>
7648
7649 * ipa-inline-transform.c (inline_transform): Do not
7650 optimize_inline_calls when not optimizing.
7651
8e4c9a10
JH		 76522013-09-02 Jan Hubicka <jh@suse.cz>
7653
7654 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7655 duplicated nodes for assembler names.
7656 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7657 hard registers from assembler name hash.
7658
c91061e6
JH		 76592013-09-02 Jan Hubicka <jh@suse.cz>
7660
7661 * ipa-split.c (execute_split_functions): Split externally visible
7662 functions called once.
7663
2fd0985c
MJ		 76642013-09-02 Martin Jambor <mjambor@suse.cz>
7665
7666 PR ipa/58106
7667 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7668 linked list. When finding the correct duplicate, also consider also
7669 the caller in additon to its inlined_to node.
7670
91bd4114
JG		 76712013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7672
7673 * config/aarch64/aarch64-simd-builtins.def
7674 (dup_lane_scalar): Remove.
7675 * config/aarch64/aarch64-simd.md
7676 (aarch64_simd_dup): Add 'w->w' alternative.
7677 (aarch64_dup_lane<mode>): Allow for VALL.
7678 (aarch64_dup_lane_scalar<mode>): Remove.
7679 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7680 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7681 (aarch64_get_lane_unsigned<mode>): Likewise.
7682 (aarch64_get_lane<mode>): Likewise.
7683 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7684 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7685 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7686 (VCON): Change container of V2SF.
7687 (vswap_width_name): Likewise.
7688 * config/aarch64/arm_neon.h
7689 (__aarch64_vdup_lane_any): New.
7690 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7691 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7692 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7693
d617d2d8
EB		 76942013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7695
7696 PR middle-end/56382
7697 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7698 the source or the destination is a single hard register.
7699
63fdb7be
RB		 77002013-09-02 Richard Biener <rguenther@suse.de>
7701
7702 PR middle-end/57511
7703 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7704 non-linear SCEVs.
7705
78de2333
RB		 77062013-09-02 Richard Biener <rguenther@suse.de>
7707
7708 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7709 arithmetic to sizetype.
7710
85ff4ec6
BC		 77112013-09-02 Bin Cheng <bin.cheng@arm.com>
7712
7713 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7714 Find auto-increment use both before and after candidate.
7715
576016fe
MP		 77162013-09-02 Marek Polacek <polacek@redhat.com>
7717
7718 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7719
08f835dc
JH		 77202013-09-01 Jan Hubicka <jh@suse.cz>
7721
7722 * Makefile.in: Add ipa-profile.o
7723 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7724 * cgraph.c (struct cgraph_propagate_frequency_data,
7725 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7726 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7727 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB		 7728 * ipa-inline-analysis.c: Include ipa-utils.h;
7729 drop duplicated cfgloop.h.
08f835dc
JH		 7730 (inline_update_callee_summaries): Update.
7731 * ipa-profile.c: New file.
7732 * ipa-utils.h (ipa_propagate_frequency): Declare.
7733 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 7734 data-streamer.h, value-prof.h.
08f835dc
JH		 7735 (symtab_remove_unreachable_nodes): Update profile.
7736 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7737 account_time_size, cmp_counts, dump_histogram,
7738 ipa_profile_generate_summary, ipa_profile_write_summary,
7739 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7740 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 7741 Move to ipa-profile.c.
08f835dc 7742
cec8583c
JDA		 77432013-09-01 John David Anglin <danglin@gcc.gnu.org>
7744
7745 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7746
bbc9396b
JH		 77472013-09-01 Jan Hubicka <jh@suse.cz>
7748
7749 * common.opt (fdevirtualize-speculatively): New function.
7750 * invoke.texi (fdevirtualize-speculatively): Document.
7751 * ipa-devirt.c: Include ipa-inline.h
7752 (likely_target_p): New function.
7753 (ipa_devirt): New function.
7754 (gate_ipa_devirt): New function.
7755 (pass_data_ipa_devirt): New static var.
7756 (pass_ipa_devirt): Likewise.
7757 (make_pass_ipa_devirt): New function.
7758 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7759 (common_handle_option): Disable devirtualization when
7760 value range profiling is available.
7761 * passes.def (pass_ipa_devirt): Add.
7762 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 7763 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 7764
0cea211e
IS		 77652013-09-01 Iain Sandoe <iain@codesourcery.com>
7766
7767 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7768 include sanitize(undefined).
7769
5e351e96
DN		 77702013-08-31 Diego Novillo <dnovillo@google.com>
7771
7772 * Makefile.in (TREE_CORE_H): Define.
7773 (TREE_H): Use.
7774 (GTFILES): Add tree-core.h.
7775 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7776 size the array.
7777 * tree-core.h: New file.
7778 Move all data structures, enum, typedefs, global
7779 declarations and constants from ...
7780 * tree.h: ... here.
7781
63bf9a90
JH		 77822013-08-31 Jan Hubicka <jh@suse.cz>
7783
7784 * bulitins.c (expand_builtin): Do not early exit for gcov
7785 instrumented functions.
7786
5c0abd6a
MP		 77872013-08-31 Marek Polacek <polacek@redhat.com>
7788
7789 * ubsan.c: Include tm_p.h.
7790
4843f032
JH		 77912013-08-31 Jan Hubicka <jh@suse.cz>
7792
8356c89c
JH		 7793 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7794 warning.
7795
4843f032
JH		 7796 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7797 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 7798 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH		 7799 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7800 (ipa_merge_profiles): New function.
7801 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7802 (lto_input_function_body): Likewise.
5e581212
JH		 7803 * ipa-utils.h (ipa_merge_profiles): Declare.
7804 * lto-streamer.h (lto_input_function_body): Update prototype.
7805 (emit_label_in_global_context_p): Remove.
7806 * lto-symtab.c: Include ipa-utils.h
7807 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 7808
57292ce9
JH		 78092013-08-31 Jan Hubicka <jh@suse.cz>
7810
7811 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7812
b9cbfeeb
JH		 78132013-08-31 Jan Hubicka <jh@suse.cz>
7814
7815 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7816
41dedebd
UB		 78172013-08-31 Uros Bizjak <ubizjak@gmail.com>
7818
7819 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7820 "cmp" RTX before signed_comparison_operator check to account
7821 for "code" changes.
7822
06d65050
JH		 78232013-08-30 Jan Hubicka <jh@suse.cz>
7824
41dedebd 7825 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH		 7826 (detect_type_change_1): Rename to ...
7827 (detect_type_change): ... this one; early return on non-polymorphic
7828 types.
41dedebd 7829 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH		 7830 use of detect_type_change.
7831 (compute_complex_assign_jump_func): Add param_type parameter;
7832 update use of detect_type_change_ssa.
7833 (compute_complex_ancestor_jump_func): Likewise.
7834 (ipa_get_callee_param_type): New function.
7835 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7836 update calls to the jump function computation functions.
7837
600b5b1d
TJ		 78382013-08-30 Teresa Johnson <tejohnson@google.com>
7839 Steven Bosscher <steven@gcc.gnu.org>
7840
7841 * cfgrtl.c (fixup_new_cold_bb): New routine.
7842 (commit_edge_insertions): Invoke fixup_partitions.
7843 (find_partition_fixes): New routine.
7844 (fixup_partitions): Ditto.
7845 (verify_hot_cold_block_grouping): Update comments.
7846 (rtl_verify_edges): Invoke find_partition_fixes.
7847 (rtl_verify_bb_pointers): Update comments.
7848 (rtl_verify_bb_layout): Ditto.
7849 * basic-block.h (probably_never_executed_edge_p): Declare.
7850 (fixup_partitions): Ditto.
7851 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7852 * bb-reorder.c (sanitize_hot_paths): New function.
7853 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7854 sanitize_hot_paths.
7855 * predict.c (probably_never_executed_edge_p): New routine.
7856 * cfg.c (check_bb_profile): Add partition insanity warnings.
7857
0a6ea5c9
MI		 78582013-08-30 Meador Inge <meadori@codesourcery.com>
7859
7860 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7861
de5a5fa1
MP		 78622013-08-30 Marek Polacek <polacek@redhat.com>
7863
7864 * Makefile.in (ubsan.o): Add.
7865 (c-family/c-ubsan.o): Add.
7866 (builtins.o): Add ubsan.h dependency.
7867 * ubsan.h: New file.
7868 * ubsan.c: New file.
7869 * common.opt: Add -fsanitize=undefined option.
7870 (flag_sanitize): Add variable.
7871 (fsanitize=): Add option. Add Driver.
7872 (fsanitize=thread): Remove option.
7873 (fsanitize=address): Likewise.
7874 (static-libubsan): New option.
7875 * doc/invoke.texi: Document the new flag and -static-libubsan.
7876 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7877 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7878 * builtin-attrs.def (ATTR_COLD): Define.
7879 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7880 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7881 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7882 * flag-types.h (sanitize_code): New enum.
7883 * opts.c (common_handle_option): Parse command line arguments
7884 of -fsanitize=. Add -fsanitize=unreachable option.
7885 * varasm.c (get_variable_section): Adjust.
7886 (assemble_noswitch_variable): Likewise.
7887 (assemble_variable): Likewise.
7888 (output_constant_def_contents): Likewise.
7889 (categorize_decl_for_section): Likewise.
7890 (place_block_symbol): Likewise.
7891 (output_object_block): Likewise.
7892 * builtins.def: Likewise.
7893 * toplev.c (compile_file): Likewise.
7894 (process_options): Likewise.
7895 * cppbuiltin.c: Likewise.
7896 * tsan.c (tsan_pass): Likewise.
7897 (tsan_gate): Likewise.
7898 (tsan_gate_O0): Likewise.
7899 * cfgexpand.c (partition_stack_vars): Likewise.
7900 (expand_stack_vars): Likewise.
7901 (defer_stack_allocation): Likewise.
7902 (expand_used_vars): Likewise.
7903 * cfgcleanup.c (old_insns_match_p): Likewise.
7904 * asan.c (asan_finish_file): Likewise.
7905 (asan_instrument): Likewise.
7906 (gate_asan): Likewise.
7907 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7908 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7909 (asan_global_struct): Use pointer_sized_int_node instead
7910 calling build_nonstandard_integer_type.
7911 (initialize_sanitizer_builtins): Likewise.
7912 (asan_finish_file): Likewise.
7913 * gcc.c: Document %{%:function(args):X}.
7914 (static_spec_functions): Add sanitize.
7915 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7916 store funcval != NULL there.
7917 (do_spec_1): Adjust handle_spec_function caller.
7918 (handle_braces): Allow %:function(args) as condition.
7919 (sanitize_spec_function): New function.
7920 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7921 (LIBUBSAN_SPEC): Likewise.
7922 (LIBUBSAN_EARLY_SPEC): Likewise.
7923 (SANITIZER_SPEC): Handle libubsan.
7924 (SANITIZER_EARLY_SPEC): Likewise.
7925 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7926 instead of fsanitize=address.
7927 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7928 instead of fsanitize=address*.
7929 * builtins.c: Include ubsan.h.
7930 (fold_builtin_0): Instrument __builtin_unreachable.
7931 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7932 instead of flag_asan.
7933 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7934 (pointer_sized_int_node): Define.
41dedebd 7935 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 7936
f07f30cf
MS		 79372013-08-30 Mike Stump <mikestump@comcast.net>
7938
7939 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7940 with RE patterns.
7941
8a41354f
JH		 79422013-08-29 Jan Hubicka <jh@suse.cz>
7943
7944 * cgraph.c (cgraph_function_body_availability): Handle weakref
7945 correctly.
7946 * passes.def: Remove pass_fixup_cfg.
7947 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7948 track when we need to remove functions.
7949 (gate_ipa_inline): Execute inlining always; add comment why.
7950 (pass_data_ipa_inline): Remove TODO_remove_functions.
7951 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7952 do not produce summaries.
7953 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7954 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7955 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7956 constant pool and vtable.
7957
dc8d7a0f
TB		 79582013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7959
41dedebd
UB		 7960 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7961 New arm_neon.h's internal macros to specify 64-bit constants.
7962 Avoid using stdint.h's macros.
dc8d7a0f 7963
ac7eacd2
JR		 79642013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7965
7966 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7967 before applying REG_P.
7968
8b29fd4e
JJ		 79692013-08-30 Jakub Jelinek <jakub@redhat.com>
7970
7971 PR tree-optimization/58277
7972 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7973 after seeing too many stmts with vdef in between dombb and current
7974 bb, invalidate everything.
7975
26d75703
RB		 79762013-08-30 Richard Biener <rguenther@suse.de>
7977
7978 * fold-const.c (fold_single_bit_test): Fix overflow test.
7979
f7e088e7
EB		 79802013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7981
7982 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7983 and which can live in a register, always retrieve the value on entry.
7984 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7985 passed by invisible reference specially.
7986 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7987 (vt_add_function_parameter): Correctly deal with a parameter passed by
7988 invisible reference.
7989
79902013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH		 7991
7992 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7993
6e8dad05
RB		 79942013-08-30 Richard Biener <rguenther@suse.de>
7995
7996 PR tree-optimization/58228
7997 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7998 allow invariant loads in nested loop vectorization.
7999
062ef2c8
RB		 80002013-08-30 Richard Biener <rguenther@suse.de>
8001
8002 PR tree-optimization/58223
8003 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8004 (has_anti_or_output_dependence): ... this and adjust to also
8005 look for output dependences.
8006 (mark_nodes_having_upstream_mem_writes): Adjust.
8007 (rdg_flag_uses): Likewise.
8008
7a764c60
RB		 80092013-08-30 Richard Biener <rguenther@suse.de>
8010
8011 PR tree-optimization/58010
8012 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8013 assert that we have a loop-closed PHI.
8014
82e9d642
JH		 80152013-08-29 Jan Hubicka <jh@suse.cz>
8016
8017 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8018 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8019 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8020 (lto_free_function_in_decl_state_for_node): New function.
8021
9cc1fb4b
XDL		 80222013-08-29 Xinliang David Li <davidxl@google.com>
8023
41dedebd 8024 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL		 8025 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8026 Emit alignment peeling message with default -fopt-info.
8027 (vect_loop_versioning): Emit loop version info message.
41dedebd 8028 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL		 8029 (execute_vect_slp): Ditto.
8030
440a5082
EB		 80312013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8032
8033 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8034 of the clone from the DECL_NAME of the original function.
8035
3fa3690d
OE		 80362013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8037
8038 * passes.c (register_pass): Add overload.
8039 * tree-pass.h (register_pass): Forward declare it. Add comment.
8040
0170f33c
JH		 80412013-08-29 Jan Hubicka <jh@suse.cz>
8042
41dedebd
UB		 8043 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8044 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8045 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8046 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 8047 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8048 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 8049 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 8050 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8051
103ff0d6
TJ		 80522013-08-29 Teresa Johnson <tejohnson@google.com>
8053
8054 * dumpfile.c (dump_loc): Output column number.
8055 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8056 * doc/invoke.texi: Document optall -fopt-info flag.
8057 * profile.c (read_profile_edge_counts): Use new dump framework.
8058 (compute_branch_probabilities): Ditto.
8059 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8060 when pass not in any opt group.
8061 * pass_manager.h (pass_manager::get_pass_profile): New method.
8062 * value-prof.c (check_counter): Use new dump framework.
8063 (check_ic_target): Ditto.
8064 * coverage.c (get_coverage_counts): Ditto.
8065 (coverage_init): Setup new dump framework.
8066
301bbc16
RB		 80672013-08-29 Richard Biener <rguenther@suse.de>
8068
8069 PR tree-optimization/58246
8070 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8071 handle the dominance check inside a basic-block.
8072
1b275000
RB		 80732013-08-29 Richard Biener <rguenther@suse.de>
8074
8075 PR middle-end/57287
8076 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8077 of default defs that appear in abnormal PHI nodes.
8078
6b1184ba
RB		 80792013-08-29 Richard Biener <rguenther@suse.de>
8080
8081 PR tree-optimization/57685
8082 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8083 single-use operands to avoid exponential complexity.
8084
1b08b734
DC		 80852013-08-28 Dehao Chen <dehao@google.com>
8086
8087 * ipa-inline.c (edge_badness): Fix integer underflow.
8088
48a3fa69
UB		 80892013-08-28 Uros Bizjak <ubizjak@gmail.com>
8090
8091 * gtm-builtins.def (_ITM_free): Declare leaf.
8092
215f73e6
JJ		 80932013-08-28 Jakub Jelinek <jakub@redhat.com>
8094
8095 PR target/58067
8096 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8097 (*tls_local_dynamic_base_64_largepic): Likewise.
8098 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8099 Remove predicate from call operand.
8100 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8101 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8102
34554d1a
JL		 81032013-08-28 Jeff Law <law@redhat.com>
8104
8105 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8106 checks for the number of predecessors and successors allowed.
8107 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8108 which require copying a joiner block if there is a request which
8109 is a subpath that requires no joiner block copying.
8110
c01c111b
JH		 81112013-08-28 Jan Hubicka <jh@suse.cz>
8112
8113 * lto-streamer-out.c (DFS_write_tree_body): Drop
8114 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8115 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8116 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8117 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8118 Do not read DECL_ERROR_ISSUED.
8119 (unpack_ts_decl_with_vis_value_fields): Do not read
8120 DECL_DEFER_OUTPUT.
48a3fa69
UB		 8121 (lto_input_ts_binfo_tree_pointers): Do not read
8122 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH		 8123 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8124 write DECL_ERROR_ISSUED..
8125 (pack_ts_decl_with_vis_value_fields): Do not write
8126 DECL_DEFER_OUTPUT.
8127 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 8128 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH		 8129 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8130 * tree.h (tree_decl_common): Update comment.
8131 (DECL_ERROR_ISSUED): Remove.
8132
4b128ece
JJ		 81332013-08-28 Jakub Jelinek <jakub@redhat.com>
8134
8135 PR middle-end/58257
8136 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8137
12211b99 81382013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH		 8139
8140 * builtins.def (free): Declare leaf.
8141
8dce4dbc
DM		 81422013-08-27 David Malcolm <dmalcolm@redhat.com>
8143
8144 * gdbhooks.py: New.
8145 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8146 * configure: Regenerate.
8147
b8f6e610
MJ		 81482013-08-27 Martin Jambor <mjambor@suse.cz>
8149
8150 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8151 (ipa_ancestor_jf_data): Likewise.
8152 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8153 (ipa_get_jf_pass_through_type_preserved): New function.
8154 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8155 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 8156 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ		 8157 (ipa_get_jf_ancestor_result): Likewise.
8158 (propagate_vals_accross_pass_through): Use
8159 ipa_get_jf_pass_through_result to do all the value mappings.
8160 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8161 type_preserved flag.
8162 (ipa_set_jf_cst_copy): New function.
8163 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8164 (ipa_set_jf_arith_pass_through): Likewise.
8165 (ipa_set_ancestor_jf): Likewise.
8166 (compute_complex_assign_jump_func): Set type_preserved instead of
8167 punting.
8168 (ipa_compute_jump_functions_for_edge): Likewise.
8169 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8170 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 8171 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ		 8172 (ipa_write_jump_function): Stream the type_preserved flags.
8173 (ipa_read_jump_function): Likewise.
8174
74bf76ed
JJ		 81752013-08-27 Jakub Jelinek <jakub@redhat.com>
8176 Aldy Hernandez <aldyh@redhat.com>
8177
8178 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8179 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8180 * function.h (struct function): Add has_force_vect_loops and
8181 has_simduid_loops.
12211b99 8182 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ		 8183 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8184 handle it.
8185 * gimple.def: Update CLAUSES comments.
8186 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8187 (gimple_build_omp_for): Add argument to prototype.
8188 (gimple_omp_for_kind): New.
8189 (gimple_omp_for_set_kind): New.
8190 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8191 GOVD_DATA_SHARE_CLASS.
8192 (enum omp_region_type): Add ORT_SIMD.
8193 (gimple_add_tmp_var): Handle ORT_SIMD.
8194 (gimplify_var_or_parm_decl): Same.
8195 (is_gimple_stmt): Same.
8196 (omp_firstprivatize_variable): Same.
8197 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8198 (omp_notice_variable): Handle ORT_SIMD.
12211b99 8199 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ		 8200 (omp_check_private): Handle ORT_SIMD.
8201 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8202 OMP_CLAUSE_SAFELEN.
8203 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8204 Handle OMP_CLAUSE_LASTPRIVATE.
8205 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8206 OMP_CLAUSE_SAFELEN.
8207 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8208 (gimplify_expr): Handle OMP_SIMD.
8209 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8210 (expand_GOMP_SIMD_VF): New.
8211 (expand_GOMP_SIMD_LAST_LANE): New.
8212 * internal-fn.def (GOMP_SIMD_LANE): New.
8213 (GOMP_SIMD_VF): New.
8214 (GOMP_SIMD_LAST_LANE): New.
8215 * omp-low.c: Include target.h.
8216 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8217 OMP_CLAUSE_SAFELEN.
8218 (check_omp_nesting_restrictions): Same.
8219 (omp_max_vf): New.
8220 (lower_rec_simd_input_clauses): New.
8221 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8222 OMP_CLAUSE_LINEAR.
8223 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8224 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8225 (expand_omp_build_assign): New.
8226 (expand_omp_for_init_counts): New.
8227 (expand_omp_for_init_vars): New.
8228 (extract_omp_for_update_vars): New.
8229 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8230 and rewrite accordingly.
8231 (expand_omp_simd): New.
8232 (expand_omp_for): Use expand_omp_simd.
8233 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8234 (lower_omp_for): Do not lower the body.
8235 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8236 in their own loops.
8237 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB		 8238 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8239 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ		 8240 (gate_tree_if_conversion): Similarly.
8241 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8242 gimple_build_omp_for.
8243 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8244 * tree-parloops (create_parallel_loop): Pass kind argument to
8245 gimple_build_omp_for.
8246 * tree-pretty-print.c (dump_omp_clause): Add cases for
8247 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8248 OMP_CLAUSE__SIMDUID_.
8249 (dump_generic_node): Handle OMP_SIMD.
8250 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8251 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8252 unroll OMP_SIMD loops here.
12211b99 8253 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 8254 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 8255 loop->safelen.
74bf76ed
JJ		 8256 (vect_analyze_data_refs): Handle simd loops.
8257 * tree-vect-loop.c (vectorizable_live_operation): Handle
8258 IFN_GOMP_SIMD*.
12211b99 8259 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ		 8260 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8261 (vectorizable_load): Same.
12211b99 8262 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ		 8263 (struct simduid_to_vf): New.
8264 (simduid_to_vf::hash): New.
8265 (simduid_to-vf::equal): New.
8266 (struct simd_array_to_simduid): New.
8267 (simd_array_to_simduid::hash): New.
8268 (simd_array_to_simduid::equal): New.
8269 (adjust_simduid_builtins): New.
8270 (struct note_simd_array_uses_struct): New.
8271 (note_simd_array_uses_cb): New.
8272 (note_simd_array_uses): New.
8273 (vectorize_loops): Handle simd hints and adjust simd builtins
8274 accordingly.
8275 * tree-vectorizer.h (struct _stmt_vec_info): Add
8276 simd_lane_access_p field.
8277 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8278 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8279 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8280 (omp_clause_code_name): Same.
8281 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8282 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8283 * tree.def (OMP_SIMD): New entry.
12211b99
UB		 8284 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8285 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ		 8286 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8287 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8288 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8289 (OMP_CLAUSE_LINEAR_STEP): New.
8290 (OMP_CLAUSE_SAFELEN_EXPR): New.
8291 (OMP_CLAUSE__SIMDUID__DECL): New.
8292 (find_omp_clause): New prototype.
8293
d0cf4e84
L		 82942013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8295
8296 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8297 Haswell processor detection.
8298
992592ec
CW		 82992013-08-27 Christian Widmer <shadow@umbrox.de>
8300
8301 PR target/57927
8302 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8303 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8304 AVX2 capable processors.
8305
9116eb22
TB		 83062013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8307
8308 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8309 of vget_low_* with implementations in terms of other intrinsics.
8310
1c5abb53
MG		 83112013-08-27 Marc Glisse <marc.glisse@inria.fr>
8312
8313 PR middle-end/57219
8314 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8315 values to -1, 0 and 1.
8316
2e100703
VP		 83172013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8318
8319 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8320 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8321 (<optab><mode>3_insn): Remove.
8322 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8323 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8324 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8325 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8326 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8327 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8328 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8329 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8330 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8331 (ror<mode>3_insn): Likewise.
8332 * config/aarch64/predicates.md (aarch64_simd_register): New.
8333
4ded8276
RB		 83342013-08-27 Richard Biener <rguenther@suse.de>
8335
8336 PR tree-optimization/57521
8337 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8338 one edge is non-critical.
8339 (find_phi_replacement_condition): Make sure to use a non-critical
8340 edge. Cleanup and remove old bug workarounds.
8341 (bb_postdominates_preds): Remove.
8342 (if_convertible_loop_p_1): Do not compute post-dominators.
8343 (combine_blocks): Do not free post-dominators.
8344 (main_tree_if_conversion): Likewise.
8345 (pass_data_if_conversion): Add TODO_verify_ssa.
8346
5aa11061
DD		 83472013-08-27 DJ Delorie <dj@redhat.com>
8348
8349 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8350
fd91cfe3
YZ		 83512013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8352
8353 * function.c (assign_parm_find_data_types): Set passed_mode and
8354 nominal_mode to the TYPE_MODE of nominal_type for the built
8355 pointer type in case of the struct-pass-by-reference.
8356
907555ce
JR		 83572013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8358
8359 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8360 (UINT16_TYPE): Change default to "unsigned int".
8361
67518c93
JR		 8362 * config/avr/avr.opt (mfract-convert-truncate): New option.
8363 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8364 is set, round negative fractional integers according to n1169
8365 when converting to integer types.
8366
d6d989fc
JH		 83672013-08-26 Jan Hubicka <jh@suse.cz>
8368
8369 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8370 methods can not be called indirectly when their address is not taken.
8371
2aa3da06
JH		 83722013-08-26 Jan Hubicka <jh@suse.cz>
8373
12211b99
UB		 8374 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8375 ctor_for_folding.
2aa3da06 8376
0987ffe7
JH		 83772013-08-26 Jan Hubicka <jh@suse.cz>
8378
8379 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8380 can be unshared.
8381
befe8647
JR		 83822013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8383
8384 * reload.c (find_valid_class): Allow classes that do not include
8385 FIRST_PSEUDO_REGISTER - 1.
8386
a21e735e
JH		 83872013-08-26 Jan Hubicka <jh@suse.cz>
8388
8389 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8390 fix edge count/frequency when speculation failed; fix type check
8391 for the direct call.
8392
e067bd43
JH		 83932013-08-26 Jan Hubicka <jh@suse.cz>
8394
8395 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8396
e8aec975
JH		 83972013-08-26 Jan Hubicka <jh@suse.cz>
8398
8399 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8400 to be changed by edge redirection.
8401
d0b66480
JH		 84022013-08-26 Jan Hubicka <jh@suse.cz>
8403
12211b99
UB		 8404 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8405 formating; add sanity check.
d0b66480
JH		 8406 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8407 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8408 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8409 (resolve_noninline_speculation): Update callee keys, too.
8410
0f9aaac7
JH		 84112013-08-26 Jan Hubicka <jh@suse.cz>
8412
8413 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8414 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8415
4b37444e
JR		 84162013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8417
8418 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8419 into proper place.
8420
5339fc0c
UB		 84212013-08-26 Uros Bizjak <ubizjak@gmail.com>
8422
8423 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8424 (x86_64_elf_select_section): Ditto.
8425 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8426 arguments.
8427 (ix86_pass_by_reference): Ditto.
5339fc0c
UB		 8428 (output_set_got): Ditto.
8429 (ix86_unary_operator_ok): Ditto.
8430 (ix86_expand_builtin): Ditto.
8431
d5c3d3ef
JH		 84322013-08-23 Jan Hubicka <jh@suse.cz>
8433
5339fc0c 8434 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 8435
5b1e7435
JH		 84362013-08-23 Jan Hubicka <jh@suse.cz>
8437
8438 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8439 (tree_decl_with_vis): Add FINAL field.
8440
ead69dac
JL		 84412013-08-23 Jeff Law <law@redhat.com>
8442
8443 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8444 the debugging dump when the expression is fully redundant.
8445
025311c4
GDR		 84462013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8447
8448 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8449 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8450 * pretty-print.c (pp_formatted_text_data): Likewise.
8451 (pp_write_text_to_stream): Likewise.
8452 (pp_write_text_as_dot_label_to_stream): Likewise.
8453 (pp_append_r): Likewise.
8454 (pp_format): Likewise.
8455 (pp_flush): Likewise.
8456 (pp_clear_output_area): Likewise.
8457 (pp_append_text): Likewise.
8458 (pp_formatted_text): Likewise.
8459 (pp_remaining_character_count_for_line): Likewise.
8460 (pp_newline): Likewise.
8461 (pp_character): Likewise.
8462 (output_buffer::~output_buffer): Define.
8463 (pretty_printer::~pretty_printer): Destruct output buffer.
8464 * pretty-print.h (output_buffer::~output_buffer): Declare.
8465 (pretty_printer::~pretty_printer): Declare virtual.
8466
0cadbfaa
MG		 84672013-08-24 Marc Glisse <marc.glisse@inria.fr>
8468
8469 PR other/57324
8470 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8471 HOST_WIDE_INT_M1U): New macros.
8472 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8473 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8474 unsigned -1 for lshift.
8475 * cse.c (cse_insn): Likewise.
8476 * double-int.c (rshift_double, lshift_double): Likewise.
8477 * builtins.c (fold_builtin_bitop): Likewise.
8478 * combine.c (force_to_mode): Likewise.
8479 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8480 * simplify-rtx.c (simplify_const_unary_operation,
8481 simplify_const_binary_operation): Likewise.
8482 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8483 check_va_list_escapes): Likewise.
8484 * rtlanal.c (nonzero_bits1): Likewise.
8485 * expmed.c (expand_smod_pow2): Likewise.
8486 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8487
009e5353
JH		 84882013-08-23 Jan Hubicka <jh@suse.cz>
8489
8490 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8491 as having address taken.
8492
c4be6568
JH		 84932013-08-23 Jan Hubicka <jh@suse.cz>
8494
64cbf23d
JH		 8495 * ipa-utils.h (method_class_type): Declare.
8496 * ipa-devirt.c (method_class_type): Export.
8497
c4be6568
JH		 8498 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8499 do not walk base classes of anonymous types.
8500
97aba8e9
KK		 85012013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8502
8503 PR rtl-optimization/58220
8504 PR regression/58221
8505 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8506 handle SEQUENCE insns properly.
8507
0fc80001
GDR		 85082013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8509
8510 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8511 definition.
8512 (pp_newline_and_indent): Likewise.
8513 (pp_separate_with): Likewise.
8514 * pretty-print.c (pp_newline_and_flush): Define.
8515 (pp_newline_and_indent): Likewise.
8516 (pp_separate_with): Likewise.
8517
520a5868
JJ		 85182013-08-23 Jakub Jelinek <jakub@redhat.com>
8519
8520 PR target/58218
8521 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8522 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8523
e500c62a
KY		 85242013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8525
f43245d1
UB		 8526 * config/i386/predicates.md (ext_sse_reg_operand): New.
8527 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY		 8528 predicate to determine if EVEX is needed.
8529 (*movsi_internal): Ditto.
8530 (*movdf_internal): Ditto.
8531 (*movsf_internal): Ditto.
f43245d1 8532 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 8533
6b00d7dd
JJ		 85342013-08-23 Jakub Jelinek <jakub@redhat.com>
8535
8536 PR tree-optimization/58209
8537 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8538 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8539 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8540 emit POINTER_PLUS_EXPR.
8541 (create_tailcall_accumulator): For pointer result type accumulate in
8542 sizetype type.
8543
4f2a9f90
PC		 85442013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8545
8546 * configure.ac: Add backslashes missing from the last change.
8547 * configure: Regenerate.
8548
7bf4274e
JH		 85492013-08-22 Jan Hubicka <jh@susue.cz>
8550
4f2a9f90
PC		 8551 * ipa.c (function_and_variable_visibility): First remember function
8552 was global and then make it local.
7bf4274e 8553
1bea243d
JB		 85542013-08-22 Julian Brown <julian@codesourcery.com>
8555
8556 * configure.ac: Add aarch64 to list of arches which use "nop" in
8557 debug_line test.
8558 * configure: Regenerate.
8559
bcba7ecc
AK		 85602013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8561
8562 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8563 gnu_libc_has_function.
8564 * config/s390/tpf.h: Likewise.
8565
4c495b0d
JH		 85662013-08-22 Jan Hubicka <jh@susue.cz>
8567
8568 * timevar.c (validate_phases): Add cast.
8569
5e302bca
JH		 85702013-08-22 Jan Hubicka <jh@susue.cz>
8571
8572 * timevar.c (validate_phases): Use size_t for memory.
8573 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8574
da6ca2b5
GDR		 85752013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8576
8577 * pretty-print.h (output_buffer::output_buffer): Declare.
8578 (pretty_printer::pretty_printer): Likewise.
8579 (pp_construct): Remove.
8580 * pretty-print.c (output_buffer::output_buffer): Define.
8581 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8582 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8583 (print_gimple_expr): Likewise.
8584 (print_gimple_seq): Likewise.
8585 (gimple_dump_bb): Likewise.
8586 * sched-vis.c (dump_value_slim): Likewise.
8587 (dump_insn_slim): Likewise.
8588 (dump_rtl_slim): Likewise.
8589 (str_pattern_slim): Likewise.
8590 * tree-mudflap.c (mf_varname_tree): Likewise.
8591 * graph.c (print_graph_cfg): Likewise.
8592 (start_graph_dump): Likewise.
8593 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8594 placement-new.
8595 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8596 pretty printer initialization.
8597 * coretypes.h (diagnostic_context): Remove superflous type alias
8598 declaration.
8599 (pretty_printer): Likewise. Declare directly as a class.
8600 (pretty_print_info): Remove declaration as class.
8601 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8602 and pp_clear_output_area.
8603 (asan_add_global): Likewise.
8604
0e1474e5
JH		 86052013-08-22 Jan Hubicka <jh@suse.cz>
8606
8607 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8608 * ipa-utils.h (update_type_inheritance_graph): Declare.
8609 (possible_polymorphic_call_target_p): Declare.
8610 (possible_polymorphic_call_target_p): New.
8611 * ipa-devirt.c: Update toplevel comments.
8612 (cached_polymorphic_call_targets): Move up.
8613 (odr_type_d): Move ID down.
8614 (polymorphic_type_binfo_p): Update comment.
8615 (odr_hasher::remove): Likewise;
8616 (get_odr_type): Set anonymous_namespace.
8617 (dump_odr_type): Dump it.
8618 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8619 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8620 (record_binfo): Add comment.
5339fc0c
UB		 8621 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8622 built.
0e1474e5
JH		 8623 (devirt_node_removal_hook): Do not iCE when cache is freed.
8624 (possible_polymorphic_call_target_p): New predicate.
8625 (update_type_inheritance_graph): New function.
8626
3f97cb0b
AI		 86272013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8628 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8629 Sergey Lega <sergey.s.lega@intel.com>
8630 Anna Tikhonova <anna.tikhonova@intel.com>
8631 Ilya Tocar <ilya.tocar@intel.com>
8632 Andrey Turetskiy <andrey.turetskiy@intel.com>
8633 Ilya Verbin <ilya.verbin@intel.com>
8634 Kirill Yukhin <kirill.yukhin@intel.com>
8635 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8636
8637 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8638 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8639 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8640 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8641 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8642 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8643 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8644 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8645 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8646 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8647 OPT_mavx512pf, OPT_mavx512er cases.
8648 * config/i386/constraints.md (v): New constraint.
8649 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8650 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8651 (bit_AVX512CD): New.
8652 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8653 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8654 * config/i386/i386-c.c (ix86_target_macros_internal):
8655 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8656 __AVX512PF__.
8657 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8658 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8659 * config/i386/i386.c (regclass_map, dbx_register_map)
8660 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8661 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8662 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8663 -mavx512pf options.
8664 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8665 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8666 -mavx512pf options. Fix formatting.
8667 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8668 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8669 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8670 -mavx512cd, -mavx512pf options.
8671 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8672 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8673 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8674 ALL_SSE_REGS.
8675 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8676 (ix86_set_reg_reg_cost): Ditto.
8677 (x86_order_regs_for_local_alloc): Ditto.
8678 (MAX_VECT_LEN): Extend to 64-byte.
8679 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8680 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8681 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8682 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8683 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8684 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8685 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8686 (SSE_REG_MODE_P): Support new modes.
8687 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8688 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8689 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8690 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8691 (REG_CLASS_CONTENTS): Add new registers.
8692 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8693 (EXT_REX_SSE_REGNO_P): New.
8694 (HI_REGISTER_NAMES): Add new registers.
8695 * config/i386/i386.md: Define constants for new registers.
8696 (mode): Add new 512-bit modes.
8697 (prefix): Support evex prefix.
8698 (isa): Support avx512f, noavx512f, fma_avx512f.
8699 (ssemodesuffix): Add new 512-bit modes.
8700 (movxi): New.
8701 (*movxi_internal_avx512f): Ditto.
8702 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8703 Support MODE_XI.
8704 (*movsi_internal): Likewise.
8705 (*movdf_internal): Likewise.
8706 (*movsf_internal): Likewise.
8707 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8708 (<code><mode>3): Likewise.
5339fc0c
UB		 8709 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8710 New.
3f97cb0b
AI		 8711 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8712 with the new constraint "v".
8713 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8714 modes.
8715 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8716 with the new constraint "v".
8717 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8718 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8719 (*<plusminus_insn><mode>3): Likewise.
8720 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8721 (*mul<mode>3): Likewise.
8722 (<sse>_vmmul<mode>3): Likewise.
8723 (<sse>_div<mode>3): Likewise.
8724 (<sse>_vmdiv<mode>3): Likewise.
8725 (<sse>_sqrt<mode>2): Likewise.
8726 (<sse>_vmsqrt<mode>2): Likewise.
8727 (*<code><mode>3_finite): Likewise.
8728 (*<code><mode>3) <smaxmin>: Likewise.
8729 (<sse>_vm<code><mode>3): Likewise.
8730 (*<code><mode>3) <any_logic>: Likewise.
8731 (*fma_fmadd_<mode>): Likewise.
8732 (*fma_fmsub_<mode>): Likewise.
8733 (*fma_fnmadd_<mode>): Likewise.
8734 (*fma_fnmsub_<mode>): Likewise.
8735 (*fma_fmaddsub_<mode>): Likewise.
8736 (*fma_fmsubadd_<mode>): Likewise.
8737 (*fmai_fmadd_<mode>): Likewise.
8738 (*fmai_fmsub_<mode>): Likewise.
8739 (*fmai_fnmadd_<mode>): Likewise.
8740 (*fmai_fnmsub_<mode>): Likewise.
8741 (sse_cvtsi2ss): Likewise.
8742 (sse_cvtsi2ssq): Likewise.
8743 (sse_cvtss2si): Likewise.
8744 (sse_cvtss2si_2): Likewise.
8745 (sse_cvtss2siq): Likewise.
8746 (sse_cvtss2siq_2): Likewise.
8747 (sse_cvttss2si): Likewise.
8748 (sse_cvtss2siq_2): Likewise.
8749 (float<sseintvecmodelower><mode>2): Likewise.
8750 (sse2_cvtsd2si_2): Likewise.
8751 (sse2_cvtsd2siq_2): Likewise.
8752 (*<plusminus_insn><mode>3): Likewise.
8753 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8754 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8755 (ashr<mode>3): Likewise.
8756 (<shift_insn><mode>3): Likewise.
8757 (avx2_<code><mode>3): Likewise.
8758 (*avx2_<code><mode>3): Likewise.
8759 (*andnot<mode>3): Likewise.
8760 (*<code><mode>3) <any_logic>: Likewise.
8761 (abs<mode>2): Likewise.
8762 (avx2_permvar<mode>): Likewise.
8763 (avx2_perm<mode>_1): Likewise.
8764 (*avx_vpermilp<mode>): Likewise.
8765 (avx_vpermilvar<mode>3): Likewise.
8766 (avx2_ashrv<mode>): Likewise.
8767 (avx2_<shift_insn>v<mode>): Likewise.
8768 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8769 -mavx512cd.
8770 * doc/rtl.texi: Document XImode.
8771
dd5e8423
JL		 87722013-08-21 Jeff Law <law@redhat.com>
8773
b9ebee5d
JL		 8774 * tree-flow.h (register_jump_thread): Pass vector of edges
8775 instead of each important edge.
8776 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8777 thread path into a vector and pass that to register_jump_thread.
8778 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8779 passed in edge vector to the current 3-edge form.
8780
dd5e8423
JL		 8781 Revert:
8782 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8783
9bb6628e 8784 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL		 8785 delete_unmarked_insns removed anything.
8786
445dc8df
JR		 87872013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8788
8789 * reload.h (struct reg_equivs): Rename to ..
8790 (struct reg_equivs_s): .. this.
8791
e0df53dd
ML		 87922013-08-20 Martin Liska <marxin.liska@gmail.com>
8793
8794 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8795
e7d1d3eb
RO		 87962013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8797
5339fc0c 8798 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 8799
5562e26e
JL		 88002013-08-21 Jeff Law <law@redhat.com>
8801
8802 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8803 simplify assignments too. If the RHS collapses to a singleton
8804 range, then return the value for the range.
8805
c7ecdec6
KY		 88062013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8807
7a5911d3
UB		 8808 * config/i386/sse.md (V16): Rename to...
8809 (VMOVE): this.
8810 (mov<mode>): Update iterator name.
8811 (*mov<mode>_internal): Ditto.
8812 (push<mode>1): Ditto.
8813 (movmisalign<mode>): Ditto.
c7ecdec6 8814
bfa3b50a
JH		 88152013-08-20 Jan Hubicka <jh@suse.cz>
8816
8817 PR bootstrap/58186
8818 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8819 entry for direct edges.
8820 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8821
6cd4d135
DM		 88222013-08-20 David Malcolm <dmalcolm@redhat.com>
8823
8824 Revert my last two changes, r201865 and r201864:
8825
8826 Revert r201865:
8827 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8828
8829 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8830 instances can own GC refs.
8831
8832 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8833 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8834 (gcc::context::gt_pch_nx): Likewise.
8835 (gcc::context::gt_pch_nx): Likewise.
8836 * ggc.h (gt_ggc_mx <T>): New.
8837 (gt_pch_nx_with_op <T>): New.
8838 (gt_pch_nx <T>): New.
8839 * passes.c (opt_pass::gt_ggc_mx): New.
8840 (opt_pass::gt_pch_nx): New.
8841 (opt_pass::gt_pch_nx_with_op): New.
8842 (pass_manager::gt_ggc_mx): New.
8843 (pass_manager::gt_pch_nx): New.
8844 (pass_manager::gt_pch_nx_with_op): New.
8845 (pass_manager::operator new): Use
8846 ggc_internal_cleared_alloc_stat rather than xcalloc.
8847 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8848 (pass_manager::gt_ggc_mx): New.
8849 (pass_manager::gt_pch_nx): New.
8850 (pass_manager::gt_pch_nx_with_op): New.
8851 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8852 (opt_pass::operator new): New.
8853 (opt_pass::gt_ggc_mx): New.
8854 (opt_pass::gt_pch_nx): New.
8855 (opt_pass::gt_pch_nx_with_op): New.
8856
8857 Revert r201864:
8858 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8859
8860 * Makefile.in (GTFILES): Add context.h.
8861 * context.c (gcc::context::operator new): New.
8862 (gcc::context::gt_ggc_mx): New.
8863 (gcc::context::gt_pch_nx): New.
8864 (gcc::context::gt_pch_nx): New.
8865 * context.h (gcc::context): Add GTY((user)) marking.
8866 (gcc::context::operator new): New.
8867 (gcc::context::gt_ggc_mx): New.
8868 (gcc::context::gt_pch_nx): New.
8869 (gcc::context::gt_pch_nx): New.
8870 (g): Add GTY marking.
8871 (gt_ggc_mx (gcc::context *)): New.
8872 (gt_pch_nx (gcc::context *)): New.
8873 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8874 void *cookie)): New.
8875 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8876
a046136a
AM		 88772013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8878
8879 * dce.c (fini_dce): Call df_analyze again just in case
8880 delete_unmarked_insns removed anything.
8881
67598720
TJ		 88822013-08-20 Teresa Johnson <tejohnson@google.com>
8883
8884 PR rtl-optimizations/57451
8885 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8886 from crossing split section boundaries.
8887
f8693faf
MGD		 88882013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8889
8890 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8891 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8892 with MULTLIB_DEFAULTS.
8893
7d0b9a9c
NC		 88942013-08-20 Nick Clifton <nickc@redhat.com>
8895
8896 * target.def (narrow_volatile_bitfield): Note that the default
8897 value is false, not !TARGET_STRICT_ALIGN.
8898 * doc/tm.texi: Regenerate.
8899
8edb8dc8
PC		 89002013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8901
8902 Fix LIB_SPEC for systems without libpthread.
8903
8904 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8905 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8906 for Android.
8907 * config/i386/linux-common.h: Likewise.
8908 * config/mips/linux-common.h: Likewise.
8909
12211b99 89102013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ		 8911
8912 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8913 checks.
8914
e42c64cb
DM		 89152013-08-20 David Malcolm <dmalcolm@redhat.com>
8916
8917 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8918 instances can own GC refs.
8919
8920 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8921 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8922 (gcc::context::gt_pch_nx): Likewise.
8923 (gcc::context::gt_pch_nx): Likewise.
8924 * ggc.h (gt_ggc_mx <T>): New.
8925 (gt_pch_nx_with_op <T>): New.
8926 (gt_pch_nx <T>): New.
8927 * passes.c (opt_pass::gt_ggc_mx): New.
8928 (opt_pass::gt_pch_nx): New.
8929 (opt_pass::gt_pch_nx_with_op): New.
8930 (pass_manager::gt_ggc_mx): New.
8931 (pass_manager::gt_pch_nx): New.
8932 (pass_manager::gt_pch_nx_with_op): New.
8933 (pass_manager::operator new): Use
8934 ggc_internal_cleared_alloc_stat rather than xcalloc.
8935 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8936 (pass_manager::gt_ggc_mx): New.
8937 (pass_manager::gt_pch_nx): New.
8938 (pass_manager::gt_pch_nx_with_op): New.
8939 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8940 (opt_pass::operator new): New.
8941 (opt_pass::gt_ggc_mx): New.
8942 (opt_pass::gt_pch_nx): New.
8943 (opt_pass::gt_pch_nx_with_op): New.
8944
910c02a0
DM		 89452013-08-20 David Malcolm <dmalcolm@redhat.com>
8946
8947 * Makefile.in (GTFILES): Add context.h.
8948 * context.c (gcc::context::operator new): New.
8949 (gcc::context::gt_ggc_mx): New.
8950 (gcc::context::gt_pch_nx): New.
8951 (gcc::context::gt_pch_nx): New.
8952 * context.h (gcc::context): Add GTY((user)) marking.
8953 (gcc::context::operator new): New.
8954 (gcc::context::gt_ggc_mx): New.
8955 (gcc::context::gt_pch_nx): New.
8956 (gcc::context::gt_pch_nx): New.
8957 (g): Add GTY marking.
8958 (gt_ggc_mx (gcc::context *)): New.
8959 (gt_pch_nx (gcc::context *)): New.
8960 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8961 void *cookie)): New.
8962 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8963
af4b4236
AM		 89642013-08-20 Alan Modra <amodra@gmail.com>
8965
8966 PR target/57865
8967 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8968 (rs6000_emit_epilogue): Likewise.
8969
12211b99 89702013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC		 8971
8972 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8973
e2323f5b
PB		 89742013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8975 Jakub Jelinek <jakub@redhat.com>
8976
8977 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8978 (BUILT_IN_FABSD64): Likewise.
8979 (BUILT_IN_FABSD128): Likewise.
8980 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8981 (fold_builtin_1): Likewise.
5339fc0c
UB		 8982 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8983 destination and source operands.
e2323f5b
PB		 8984 (*abstd2_fpr): Likewise.
8985 (*nabstd2_fpr): Likewise.
8986
cfbf3ee8
RS		 89872013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8988
8989 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8990 JUMP_P and INSN_P.
8991
d3136aeb
AH		 89922013-08-19 Aldy Hernandez <aldyh@redhat.com>
8993
8994 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8995 incomplete.
8996
d33d9e47
AI		 89972013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8998
8999 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9000 * builtins.c (default_libc_has_function): New.
9001 (gnu_libc_has_function): Ditto.
9002 (no_c99_libc_has_function): Ditto.
9003 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9004 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9005 (fold_builtin_sincos): Likewise.
9006 (fold_builtin_cexp): Likewise.
9007 * builtins.def (DEF_C94_BUILTIN): Likewise.
9008 (DEF_C99_BUILTIN): Likewise.
9009 (DEF_C99_C90RES_BUILTIN): Likewise.
9010 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9011 definitions to using this define.
9012 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 9013 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI		 9014 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9015 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9016 * config/darwin.h: Ditto.
9017 * config/elfos.h: Ditto.
9018 * config/freebsd.h: Ditto.
9019 * config/i386/cygming.h: Ditto.
9020 * config/i386/djgpp.h: Ditto.
9021 * config/i386/i386-interix.h: Ditto.
9022 * config/microblaze/microblaze.h: Ditto.
9023 * config/mmix/mmix.h: Ditto.
9024 * config/gnu-user.h: Ditto.
9025 * config/ia64/hpux.h: Ditto.
9026 * config/pa/pa-hpux.h: Ditto.
9027 * config/pdp11/pdp11.h: Ditto.
9028 * config/picochip/picochip.h: Ditto.
9029 * config/linux.h: Ditto.
9030 * config/netbsd.h: Ditto.
9031 * config/openbsd.h: Ditto.
9032 * config/rs6000/aix43.h: Ditto.
9033 * config/rs6000/aix51.h: Ditto.
9034 * config/rs6000/aix52.h: Ditto.
9035 * config/rs6000/aix53.h: Ditto.
9036 * config/rs6000/aix61.h: Ditto.
9037 * config/rs6000/darwin.h: Ditto.
9038 * config/rs6000/linux.h: Ditto.
9039 * config/rs6000/linux64.h: Ditto.
9040 * config/s390/tpf.h: Ditto.
9041 * config/sol2-10.h: Ditto.
9042 * config/sol2.h: Ditto.
9043 * config/vms/vms.h: Ditto.
9044 * config/vxworks.h: Ditto.
9045 * config/linux-android.c (linux_android_libc_has_function):
9046 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9047 * config/linux-protos.h (linux_android_libc_has_function):
9048 New declaration.
9049 * config/i386/i386.c (ix86_libc_has_function): New.
9050 * config/i386/i386-protos.h
9051 (ix86_libc_has_function): New declaration.
9052 * config/i386/i386.md
9053 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9054 ("isinf<mode>2): Likewise.
9055 * convert.c (convert_to_integer): Using new target hook
9056 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9057 TARGET_C99_FUNCTIONS.
9058 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9059 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9060 * coretypes.h (function_class): New enum for different
9061 classes of functions.
9062 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9063 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9064 (TARGET_HAS_SINCOS): Likewise.
9065 (TARGET_LIBC_HAS_FUNCTION): New.
9066 * doc/tm.texi: Regenerated.
9067 * targhooks.h (default_libc_has_function): New declaration.
9068 (no_c99_libc_has_function): Ditto.
9069 (gnu_libc_has_function): Ditto.
9070 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9071 and TARGET_HAS_SINCOS.
9072
eefe9a99
JH		 90732013-08-18 Jan Hubicka <jh@suse.cz>
9074
9075 * Makeifle-in (ipa-devirt.o): New.
9076 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9077 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9078 (analyze_functions): Look into possible targets of polymorphic call.
9079 * dumpfile.c (dump_files): Add type-inheritance dump.
9080 * dumpfile.h (TDI_inheritance): New.
9081 * ipa-devirt.c: New file.
9082 * ipa-utils.h (odr_type_d): Forward declare.
9083 (odr_type): New type.
9084 (build_type_inheritance_graph): Declare.
9085 (possible_polymorphic_call_targets): Declare and introduce inline
9086 variant when only edge is pased.
9087 (dump_possible_polymorphic_call_targets): Likewise.
9088 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9089 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9090 (types_same_for_odr): ... here.
9091 * tree.h (type_in_anonymous_namespace_p): Declare.
9092
29b89442
JJ		 90932013-08-18 Jakub Jelinek <jakub@redhat.com>
9094
9095 PR tree-optimization/58006
9096 * tree-parloops.c (take_address_of): Don't ICE if get_name
9097 returns NULL.
9098 (eliminate_local_variables_stmt): Remove clobber stmts.
9099
4f219961
EB		 91002013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9101
9102 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9103 error message is issued for an alias to undefined symbol.
9104
1d5755ef
JH		 91052013-08-18 Jan Hubicka <jh@suse.cz>
9106
9107 * cgraph.c (cgraph_create_indirect_edge): Discover
9108 polymorphic calls and record basic info into indirect_info.
9109 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9110 devirtualization, ignore objc function calls.
9111 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9112 call with no parm index info.
9113 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9114 * tree.c (virtual_method_call_p): New function.
9115 * tree.h (virtual_method_call_p): Declare.
9116
5945bebf
JH		 91172013-08-16 Jan Hubicka <jh@suse.cz>
9118
9119 PR middle-end/58179
9120 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9121
47fa5d23
DE		 91222013-08-16 David Edelsohn <dje.gcc@gmail.com>
9123
9124 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9125 attribute.
9126
c74f54a0
DM		 91272013-08-16 David Malcolm <dmalcolm@redhat.com>
9128
9129 * gengtype.c (type_for_name): Add special-case support for
9130 locating types within the "gcc::" namespace.
9131 (open_base_files): Emit a "using namespace gcc" directive.
9132
f758f299
MM		 91332013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9134
9135 PR target/58160
9136 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9137 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9138
9139 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9140 array instead of each individual operand as a separate argument.
9141 (emit_fusion_gpr_load): Likewise.
9142 (expand_fusion_gpr_load): Add new function declaration.
9143
9144 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9145 signature to have the operands passed as an array, instead of as
9146 separate arguments. Allow ZERO_EXTEND to be in the memory
9147 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9148 depend on the register live/dead flags when peepholes are run.
9149 (expand_fusion_gpr_load): New function to be called from the
9150 peephole2 pass, to change the register that addis sets to be the
9151 target register.
9152 (emit_fusion_gpr_load): Change the calling signature to have the
9153 operands passed as an array, instead of as separate arguments.
9154 Allow ZERO_EXTEND to be in the memory address, and also
9155 SIGN_EXTEND if -mpower8-fusion-sign.
9156
9157 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9158 unspec enumeration.
9159 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9160 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 9161 depend on the register live/dead state when the peephole pass is done.
f758f299 9162
158f4e4f
DM		 91632013-08-16 David Malcolm <dmalcolm@redhat.com>
9164
9165 * gengtype.c (create_user_defined_type): Ensure that the kind
9166 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9167 declaration is seen before the GTY((user)) marking.
9168
22f8cea5
BE		 91692013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9170
9171 PR target/58105
9172 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9173
c49bdb2e
JH		 91742013-08-16 Jan Hubicka <jh@suse.cz>
9175
9176 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9177 arugment expected_type.
9178 (gimple_fold_call): Use it.
9179 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9180 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 9181 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH		 9182 (try_make_edge_direct_virtual_call): Likewise.
9183 * tree.c (obj_type_ref_class): New.
9184 * tree.h (obj_type_ref_class): Use it.
9185
4042dca9
GDR		 91862013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9187
9188 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9189 (rtl_slim_pp): Likewise.
9190 (init_rtl_slim_pretty_print): Likewise.
9191 (dump_value_slim): Don't call it. Use local pretty printer.
9192 (dump_insn_slim): Likewise.
9193 (dump_rtl_slim): Likewise.
9194 (str_pattern_slim): Likewise.
9195 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9196 Simplify.
9197
7a460594
JJ		 91982013-08-16 Jakub Jelinek <jakub@redhat.com>
9199
639dc669
JJ		 9200 PR tree-optimization/58164
9201 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9202 walk gimple_goto_dest of GIMPLE_GOTO.
9203
7a460594
JJ		 9204 PR tree-optimization/58165
9205 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9206 bi_call must be the last stmt in a bb, don't split_block, instead
9207 use fallthru edge from it and give up if there is none.
9208 Release conds vector when returning early.
9209
e3c02cc7
XDL		 92102013-08-14 Xinliang David Li <davidxl@google.com>
9211
9212 * config/i386/i386.c (ix86_option_override_internal):
9213 Remove unused variable and field.
9214
6469da2c
BS		 92152013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9216
9217 PR target/57949
22f8cea5 9218 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS		 9219 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9220 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9221 and Linux, correct BLKmode alignment when 128-bit alignment is
9222 required and compatibility flag is not set.
22f8cea5
BE		 9223 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9224 for zero-size arguments when compatibility flag is not set.
6469da2c 9225
4ca890e2
JJ		 92262013-08-14 Jakub Jelinek <jakub@redhat.com>
9227
9228 PR tree-optimization/58145
9229 * tree-sra.c (build_ref_for_offset): If prev_base has
9230 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9231
daec1759 92322013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 9233
daec1759 9234 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 9235 Fix uninitialized variable error.
daec1759 9236
3ad20bd4
XDL		 92372013-08-14 Xinliang David Li <davidxl@google.com>
9238
9239 * config/i386/i386.opt: Define two new options.
9240 * config/i386/x86-tune.def: Add arch selector field in macros.
9241 * config/i386/i386.h: Adjust macro definition.
9242 * config/i386/i386.c (ix86_option_override_internal):
9243 Refactor the code.
9244 (parse_mtune_ctrl_str): New function.
9245 (set_ix86_tune_features): New function.
9246 (ix86_function_specific_restore): Call the new helper function.
9247
e839e2a9
AB		 92482013-08-14 Andrey Belevantsev <abel@ispras.ru>
9249
9250 PR rtl-optimization/57662
9251 * sel-sched.c (code_motion_process_successors): When the current insn
9252 is removed after the recursive traversal, break from the loop.
9253 Add comments and debug printouts.
9254
27ed665f
JJ		 92552013-08-14 Jakub Jelinek <jakub@redhat.com>
9256 Alexandre Oliva <aoliva@redhat.com>
9257
9258 PR target/58067
9259 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9260 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9261 there also UNSPEC_PLTOFF.
9262
61c1a609
MP		 92632013-08-14 Marek Polacek <polacek@redhat.com>
9264
9265 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9266 NUM_CONDITIONS bit positions.
9267
0e901c67
CC		 92682013-08-13 Cary Coutant <ccoutant@google.com>
9269
9270 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9271 (attr_checksum): Hash vector contents instead of pointer.
9272 (attr_checksum_ordered): Likewise.
9273
061eff6d
UB		 92742013-08-13 Uros Bizjak <ubizjak@gmail.com>
9275
9276 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9277 when Pmode != word_mode. Add length_address attribute.
9278 (sse3_monitor_<mode>): Merge from sse3_monitor and
9279 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9280 Pmode != word_mode. Update insn length attribute.
9281 * config/i386/i386.c (ix86_option_override_internal): Update
9282 ix86_gen_monitor selection for merged sse3_monitor insn.
9283
8ec9249b
JB		 92842013-08-13 Julian Brown <julian@codesourcery.com>
9285
9286 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9287 perform invalid legitimization on greater-than-word-size modes for
9288 TARGET_E500_DOUBLE.
9289
761a8eb7
VM		 92902013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9291
9292 * ira.c (setup_class_translate_array): Use aclass instead of cl
9293 for classes not fully covered by allocno classes.
9294
99c2bd54
JJ		 92952013-08-13 Jakub Jelinek <jakub@redhat.com>
9296
4029a5e0
JJ		 9297 PR tree-optimization/57661
9298 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9299 * tree-inline.c (tree_function_versioning): Initialize it.
9300 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9301 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9302 that is not being copied.
9303
99c2bd54
JJ		 9304 PR sanitizer/56417
9305 * asan.c (instrument_strlen_call): Fix typo in comment.
9306 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9307
b5f8f063
SE		 93082013-08-13 Steve Ellcey <sellcey@mips.com>
9309
9310 * config/mips/mips.md (prefetch): Use lw instead of ld on
9311 loongson in 32bit mode.
9312
4de80584
NC		 93132013-08-13 Nick Clifton <nickc@redhat.com>
9314
9315 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9316
09ce3660
JH		 93172013-08-13 Jan Hubicka <jh@suse.cz>
9318
9319 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9320 introduced edge; fix typo in sanity check.
9321 (cgraph_resolve_speculation): Export; improve diagnostic.
9322 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9323 speculation at type mismatch.
9324 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9325 (cgraph_resolve_speculation): Declare.
9326 (symtab_can_be_discarded): New function.
9327 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9328 * ipa-inline-transform.c (speculation_removed): New global var.
9329 (clone_inlined_nodes): See if speculation can be removed.
9330 (inline_call): If speculations was removed, we growths may not match.
9331 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9332 (speculation_useful_p): New function.
9333 (resolve_noninline_speculation): New function.
9334 (inline_small_functions): Resolve useless speculations.
9335 * ipa-inline.h (speculation_useful_p): Declare
9336 * ipa.c (can_replace_by_local_alias): Simplify.
9337 (ipa_profile): Produce speculative calls in non-lto, too;
9338 add simple cost model; produce local aliases.
9339
f971dc24
DM		 93402013-08-13 David Malcolm <dmalcolm@redhat.com>
9341
9342 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 9343 PASS_MANAGER_H.
f971dc24 9344
2e507b9a
PC		 93452013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9346
9347 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9348
c97fd4b8
UB		 93492013-08-12 Uros Bizjak <ubizjak@gmail.com>
9350
9351 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9352 iterator instead of X87MODEF.
9353
93542013-08-12 Perez Read <netfirewall@gmail.com>
9355
9356 PR target/58132
9357 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9358 operand 0 for intel asm alternative.
9359 (*movabs<mode>_2): Ditto for operand 1.
9360
f4f4204c
JG		 93612013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9362
9363 * config/aarch64/arm_none.h
9364 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9365
133b1a8e
NC		 93662013-08-12 Nick Clifton <nickc@redhat.com>
9367
9368 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9369
15dd8b3a
YR		 93702013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9371
9372 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9373 expand for QI/HImode operand to produce more effictive code for
9374 unsigned char(short) --> float(double) conversion.
9375
94109a6a
AM		 93762013-08-12 Alexander Monakov <amonakov@ispras.ru>
9377
9378 * doc/invoke.texi: Mention that -ftls-model does not force the final
9379 model.
9380
4e115102
MP		 93812013-08-12 Marek Polacek <polacek@redhat.com>
9382 Marc Glisse <marc.glisse@inria.fr>
9383
9384 PR tree-optimization/57980
9385 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9386 when creating -1 constant.
9387
748d71f3
JH		 93882013-08-10 Jan Hubicka <jh@suse.cz>
9389
1f97e193 9390 Workaround binutils PR14342.
748d71f3
JH		 9391 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9392 (gimple_init_edge_profiler): Likewise.
9393 (gimple_gen_ic_func_profiler): Likewise.
9394
5979aa54
JH		 93952013-08-09 Jan Hubicka <jh@suse.cz>
9396
9397 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9398
ad83025e
XDL		 93992013-08-09 Xinliang David Li <davidxl@google.com>
9400
9401 * config/i386/stringop.def: New file.
9402 * config/i386/stringop.opt: New file.
9403 * config/i386/i386-opts.h: Include stringopt.def.
9404 * config/i386/i386.opt: Include stringopt.opt.
9405 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 9406 Override default size based stringop inline strategies with options.
ad83025e
XDL		 9407 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9408 New function.
9409
e59df5fd
JH		 94102013-08-09 Jan Hubicka <jh@suse.cz>
9411
c97fd4b8 9412 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 9413
634ab819
JH		 94142013-08-09 Jan Hubicka <jh@suse.cz>
9415
9416 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9417 CGRAPH_FREQ_MAX.
9418 (dump_cgraph_node): Dump profile-id.
9419 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9420 and common_target_probability.
9421 * lto-cgraph.c (lto_output_edge): Stream common targets.
9422 (lto_output_node): Stream profile ids.
9423 (input_node): Stream profile ids.
9424 (input_edge): Stream common targets.
9425 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9426 * ipa.c: Include value-prof.h
9427 (ipa_profile_generate_summary): Turn indirect call statement histograms
9428 into common targets.
9429 (ipa_profile): Turn common targets into speculative edges.
9430
2fa3d31b
JH		 94312013-08-09 Jan Hubicka <jh@suse.cz>
9432
9433 * cgraph.h (cgraph_node): Add profile_id.
9434 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 9435 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH		 9436 (del_node_map): Update.
9437 (find_func_by_funcdef_no): Replace by ...
9438 (find_func_by_profile_id): ... this one.
9439 (gimple_ic_transform): Do not remove useful histograms when
9440 speculation is not done; dump info when indirect call removal
9441 can happen at LTO.
9442 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9443 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9444 (__gcov_indirect_call_profiler_v2): .. this one.
9445 * profile.h (init_node_map): Update.
9446 * coverage.c (coverage_compute_profile_id): New function.
9447 * coverage.h (coverage_compute_profile_id): Declare.
9448 * tree-profile.c (init_ic_make_global_vars): Make
9449 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9450 (gimple_init_edge_profiler): Update prototype of
9451 __gcov_indirect_call_profiler.
9452 (gimple_gen_ic_func_profiler): Simplify.
9453 (tree_profiling): Use init_node_map
9454
042ae7d2
JH		 94552013-08-09 Jan Hubicka <jh@suse.cz>
9456
c97fd4b8
UB		 9457 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9458 non-speculative refs.
042ae7d2
JH		 9459 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9460 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9461 (cgraph_set_call_stmt): Likewise.
9462 (cgraph_create_edge_1): Fix release checking compilatoin;
9463 clear lto_stmt_uid.
9464 (cgraph_free_edge): Free indirect info.
9465 (cgraph_turn_edge_to_speculative): New function.
9466 (cgraph_speculative_call_info): New function.
9467 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 9468 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH		 9469 (dump_cgraph_node): Dump speculation.
9470 (verify_edge_count_and_frequency): Accept speculative edges.
9471 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9472 (verify_cgraph_node): Handle speculation.
9473 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9474 (cgraph_set_call_stmt): Update prototype.
9475 (cgraph_make_edge_direct): Update prototype.
9476 (cgraph_speculative_call_info): Declare.
9477 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9478 to change; update call of ipa_find_references.
9479 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9480 lto_stmt_uid and speculative flags.
9481 (ipa_dump_references): Dump speculation.
9482 (ipa_clone_references): Clone speculative flag.
9483 (ipa_clone_referring): Likewise.
9484 (ipa_clone_ref): New function.
9485 (ipa_find_reference): Look into lto_stmt_uids
9486 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9487 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9488 (ipa_find_reference): Update declaration.
9489 (ipa_clone_ref): Declare.
9490 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9491 stream speculative flag.
9492 (lto_output_ref): Stream statements uids and speculation.
9493 (input_ref): Likewise.
9494 (input_edge): Stream speuclation.
9495 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9496 (cgraph_set_call_stmt_including_clones): Handle speculation.
9497 * ipa-inline.c (heap_edge_removal_hook): New function.
9498 (inline_small_functions): Register it.
9499 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9500 also initialize refs.
9501 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9502 edge to change.
9503 (try_make_edge_direct_simple_call): Likewise.
9504 (try_make_edge_direct_simple_call): Likewise.
9505 (update_indirect_edges_after_inlining): Likewise.
9506 (remove_described_reference): Look proper lto_stmt_uid.
9507 (propagate_controlled_uses): Likewise.
9508 (propagate_controlled_uses): Liekwise.
9509 * tree-inline.c (copy_bb): Copy speculative edges.
9510 (redirect_all_calls): New function.
c97fd4b8
UB		 9511 (copy_cfg_body): Do redirection after loop info is updated.
9512 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 9513
aa1e10cc
JH		 95142013-08-09 Jan Hubicka <jh@suse.cz>
9515
9516 * lto-streamer-out.c (output_function): Renumber PHIs.
9517 * lto-streamer-in.c (input_function): Likewise.
9518
66adb8eb
JG		 95192013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9520
9521 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9522 (get_lane_unsigned): Likewise.
9523 (dup_lane_scalar): Likewise.
9524 (get_lane): enable for VALL.
9525 * config/aarch64/aarch64-simd.md
9526 (aarch64_dup_lane_scalar<mode>): Remove.
9527 (aarch64_get_lane_signed<mode>): Likewise.
9528 (aarch64_get_lane_unsigned<mode>): Likewise.
9529 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9530 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9531 (aarch64_get_lane<mode>): Enable for all vector modes.
9532 (aarch64_get_lanedi): Remove misleading constraints.
9533 * config/aarch64/arm_neon.h
9534 (__aarch64_vget_lane_any): Define.
9535 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9536 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9537 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9538 * config/aarch64/iterators.md (VDQQH): New.
9539 (VDQQHS): Likewise.
9540 (vwcore): Likewise.
9541
1f65ae7a
EB		 95422013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9543
9544 * configure.ac: Add GAS check for LEON instructions on SPARC.
9545 * configure: Regenerate.
9546 * config.in: Likewise.
9547 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9548 sparc*-*-* block.
9549 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9550 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9551 for LEON or LEON3.
9552 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9553 (AS_LEON_FLAG): New macro.
9554 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9555 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9556 Deal with LEON and LEON3 for the memory model.
23a6cb78 9557 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB		 9558 (atomic_compare_and_swap<mode>_1): Likewise.
9559 (*atomic_compare_and_swap<mode>_1): Likewise.
9560
ff522f7f
ZC		 95612013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9562
9563 * config/arm/neon.md (vcond): Fix floating-point vector
9564 comparisons against 0.
9565
cb1cca12
VM		 95662013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9567
9568 * lra-constraints.c (emit_spill_move): Remove assert.
9569 (process_alt_operands): Add more debugging
9570 output. Increase reject for spilling into memory. Decrease
9571 reject for reloading scratch.
9572 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9573
a3719e31
SE		 95742013-08-08 Steve Ellcey <sellcey@mips.com>
9575
9576 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9577 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9578 micromips incompatible. Add nan2008.
9579 (MULTILIB_DIRNAMES): Add nan2008.
9580 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9581 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9582 and micromips incompatible. Add nan2008.
9583 (MULTILIB_DIRNAMES): Add nan2008.
9584 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9585
f40423e2
RS		 95862013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9587
9588 PR rtl-optimization/58079
9589 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9590 simplify_comparison has widened a comparison with an integer.
9591
869b9125
KT		 95922013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9593
9594 * config/arm/neon.md (movmisalign<mode>): Disable when we
9595 don't allow unaligned accesses.
9596 (*movmisalign<mode>_neon_store): Likewise.
9597 (*movmisalign<mode>_neon_load): Likewise.
9598 (*movmisalign<mode>_neon_store): Likewise.
9599 (*movmisalign<mode>_neon_load): Likewise.
9600
71cafea9
JH		 96012013-08-08 Jan Hubicka <jh@suse.cz>
9602
9603 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9604 (make_pass_rebuild_cgraph_edges): Also clear references.
9605 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9606 * ipa-inline-transform.c (inline_transform): Remove all references
9607 after inlining.
c97fd4b8
UB		 9608 * cgraphunit.c (expand_function): Remove all references after
9609 expansion.
71cafea9
JH		 9610 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9611 (ipa_find_reference): Rewrite to iterator.
9612 (remove_stmt_references): Likewise.
9613 (ipa_clear_stmts_in_references): New function.
9614 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB		 9615 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9616 clear references.
71cafea9
JH		 9617 * ipa-split.c (split_function): Remove references in split function.
9618
27d2e612
RE		 96192013-08-08 Richard Earnshaw <rearnsha@arm.com>
9620
9621 PR target/57431
1f65ae7a 9622 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE		 9623 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9624
5efc661c
RE		 96252013-08-08 Richard Earnshaw <rearnsha@arm.com>
9626
9627 PR target/56979
1f65ae7a 9628 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE		 9629 suggested mode for the assignment isn't compatible with the
9630 registers required.
9631
f276d31d
BE		 96322013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9633
9634 PR target/58065
9635 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9636
4b8bc035
XDL		 96372013-08-07 Xinliang David Li <davidxl@google.com>
9638
9639 * config/i386/i386.opt: New option -mtune-ctrl=.
9640 * config/i386/x86-tune.def: New file.
9641 * config/i386/i386.h: include x86-tune.def.
9642 * config/i386/i386.c (ix86_option_override_internal):
9643 Parsing -mtune-ctrl= option and set tune features.
9644
12211b99 96452013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE		 9646
9647 PR other/12081
03b0ee0a 9648 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE		 9649 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9650 to gen_2arg_fn_t.
9651
5cf6635b
EB		 96522013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9653
9654 * rtl.h (update_alignments): Declare.
9655 * final.c (grow_label_align): New function extracted from...
9656 (shorten_branches): ...here. Call it.
9657 (update_alignments): New function.
9658 * reorg.c (sibling_labels): New variable.
9659 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9660 the new label along with it onto the sibling_labels vector.
9661 (fill_simple_delay_slots): Adjust call to get_label_before.
9662 (fill_slots_from_thread): Likewise.
9663 (relax_delay_slots): Likewise.
9664 (make_return_insns): Likewise.
9665 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9666
3c8ca1ab
EB		 96672013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9668
9669 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9670 document its semantics.
9671 (diagnostic_report_diagnostic): Adjust accordingly.
9672
5ee5b32c
DM		 96732013-08-07 David Malcolm <dmalcolm@redhat.com>
9674
03b0ee0a 9675 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM		 9676 (sparc_option_override): ...and port to new C++ pass API.
9677 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9678
abd566fa
PB		 96792013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9680
9681 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9682
2077db1b
CT		 96832013-08-06 Caroline Tice <cmtice@google.com>
9684
9685 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9686 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9687 * tree-pass.h: Add pass_vtable_verify.
9688 * varasm.c (assemble_variable): Add code to properly set the comdat
9689 section and name for the .vtable_map_vars section.
9690 (assemble_vtyv_preinit_initializer): New function.
9691 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9692 LINK_ONCE flag.
9693 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9694 * vtable-verify.c: New file.
9695 * vtable-verify.h: New file.
9696 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9697 initialiation levels.
9698 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9699 * passes.def: Insert pass_vtable_verify.
9700 * aclocal.m4: Reorder includes.
9701 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9702 -fvtv-counts options.
9703 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9704 as appropriate, if -fvtable-verify=... is used.
9705 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9706 -fvtable-verify=... is used.
9707 * Makefile.in (OBJS): Add vtable-verify.o to list.
9708 (vtable-verify.o): Add new build rule.
9709 (GTFILES): Add vtable-verify.c to list.
9710 * common.opt (fvtable-verify=): New flag.
9711 (vtv_priority): Values for fvtable-verify= flag.
9712 (fvtv-counts): New flag.
9713 (fvtv-debug): New flag.
9714 * tree.h (save_vtable_map_decl): New extern function decl.
9715
03085d1c
DM		 97162013-08-07 David Malcolm <dmalcolm@redhat.com>
9717
9718 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9719 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9720 (pass_data_rl78_devirt): ...new pass_data instance and...
9721 (make_pass_rl78_devirt): ...new function.
9722 (rl78_asm_file_start): Port pass registration to new C++ API.
9723
05555c4a
DM		 97242013-08-07 David Malcolm <dmalcolm@redhat.com>
9725
9726 * coretypes.h (rtl_opt_pass): Add.
9727 (gcc::context): Add.
9728 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9729 (epiphany_init): Port to new C++ pass API.
9730 (epiphany_optimize_mode_switching): Likewise.
9731 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9732 (pass_manager::get_pass_mode_switching): New.
9733 (pass_manager::get_pass_peephole2): New.
9734 * mode-switching.c (pass_mode_switching): Add clone method.
9735 * recog.c (pass_peephole2): Add clone method.
9736 (pass_split_all_insns): Add clone method.
9737
c5a12e26
DM		 97382013-08-06 David Malcolm <dmalcolm@redhat.com>
9739
03b0ee0a
UB		 9740 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9741 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 9742
a2e2a668
JH		 97432013-08-06 Jan Hubicka <jh@suse.cz>
9744
9745 * cgraph.c (cgraph_get_body): New function based on lto.c
9746 implementation.
9747 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB		 9748 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9749 LTO paths.
a2e2a668
JH		 9750 * cgraphunit.c (expand_function): Get body prior expanding.
9751 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB		 9752 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9753 really need.
a2e2a668
JH		 9754 * passes.c (do_per_function_toporder): Get body.
9755 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9756 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9757
15682f24
MJ		 97582013-08-06 Martin Jambor <mjambor@suse.cz>
9759
9760 PR fortran/57987
9761 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9762 is not re-finalized. Rename second parameter to no_collect.
9763
78f6dd68
MJ		 97642013-08-06 Martin Jambor <mjambor@suse.cz>
9765
9766 PR middle-end/58041
9767 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9768 MEM_REF has proper alignment information.
9769
12211b99 97702013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE		 9771
9772 PR other/12081
9773 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9774 class insn_gen_fn.
9775 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9776 rtx (*) (rtx, ...) with insn_gen_fn.
9777 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9778 insn_gen_fn::stored_funcptr. Add initializer braces.
9779
8ac69a6c
DM		 97802013-08-05 David Malcolm <dmalcolm@redhat.com>
9781
9782 Rewrite how instances of passes are cloned to remove assumptions
9783 about their sizes (thus allowing pass subclasses to have
9784 additional data fields, albeit non-GC-managed ones at this point).
9785
9786 * passes.c (make_pass_instance): Now that passes have clone
9787 methods, rewrite this function to eliminate XNEW and memcpy
9788 calls that used hardcoded sizes. Since this function no longer
9789 creates pass instances, rename it to...
03b0ee0a
UB		 9790 (add_pass_instance): ...this. Document the old way that passes were
9791 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM		 9792 (next_pass_1): Add an initial_pass argument for use by
9793 add_pass_instance.
9794 (position_pass): When adding multiple instances of a pass, use
9795 the pass's clone method, rather than relying on the XNEW/memcpy
9796 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB		 9797 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9798 the initial instance of the current pass within the pass manager.
8ac69a6c 9799
27a4cd48
DM		 98002013-08-05 David Malcolm <dmalcolm@redhat.com>
9801
9802 This is the automated part of the conversion of passes from C
9803 structs to C++ classes.
9804
9805 Patch autogenerated by refactor_passes.py from
9806 https://github.com/davidmalcolm/gcc-refactoring-scripts
9807 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9808
9809 * asan.c (pass_asan): Convert from a global struct to a subclass of
9810 gimple_opt_pass along with...
9811 (pass_data_asan): ...new pass_data instance and...
9812 (make_pass_asan): ...new function.
9813 (pass_asan_O0): Convert from a global struct to a subclass of
9814 gimple_opt_pass along with...
9815 (pass_data_asan_O0): ...new pass_data instance and...
9816 (make_pass_asan_O0): ...new function.
9817 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9818 subclass of rtl_opt_pass along with...
9819 (pass_data_inc_dec): ...new pass_data instance and...
9820 (make_pass_inc_dec): ...new function.
9821 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9822 a subclass of rtl_opt_pass along with...
9823 (pass_data_reorder_blocks): ...new pass_data instance and...
9824 (make_pass_reorder_blocks): ...new function.
9825 (pass_duplicate_computed_gotos): Convert from a global struct to a
9826 subclass of rtl_opt_pass along with...
9827 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9828 (make_pass_duplicate_computed_gotos): ...new function.
9829 (pass_partition_blocks): Convert from a global struct to a subclass of
9830 rtl_opt_pass along with...
9831 (pass_data_partition_blocks): ...new pass_data instance and...
9832 (make_pass_partition_blocks): ...new function.
9833 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9834 struct to a subclass of rtl_opt_pass along with...
9835 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9836 and...
9837 (make_pass_branch_target_load_optimize1): ...new function.
9838 (pass_branch_target_load_optimize2): Convert from a global struct to a
9839 subclass of rtl_opt_pass along with...
9840 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9841 and...
9842 (make_pass_branch_target_load_optimize2): ...new function.
9843 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9844 of rtl_opt_pass along with...
9845 (pass_data_jump): ...new pass_data instance and...
9846 (make_pass_jump): ...new function.
9847 (pass_jump2): Convert from a global struct to a subclass of
9848 rtl_opt_pass along with...
9849 (pass_data_jump2): ...new pass_data instance and...
9850 (make_pass_jump2): ...new function.
9851 * cfgexpand.c (pass_expand): Convert from a global struct to a
9852 subclass of rtl_opt_pass along with...
9853 (pass_data_expand): ...new pass_data instance and...
9854 (make_pass_expand): ...new function.
9855 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9856 of rtl_opt_pass along with...
9857 (pass_data_free_cfg): ...new pass_data instance and...
9858 (make_pass_free_cfg): ...new function.
9859 (pass_into_cfg_layout_mode): Convert from a global struct to a
9860 subclass of rtl_opt_pass along with...
9861 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9862 (make_pass_into_cfg_layout_mode): ...new function.
9863 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9864 subclass of rtl_opt_pass along with...
9865 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9866 (make_pass_outof_cfg_layout_mode): ...new function.
9867 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9868 struct to a subclass of gimple_opt_pass along with...
9869 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9870 (make_pass_build_cgraph_edges): ...new function.
9871 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9872 subclass of gimple_opt_pass along with...
9873 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9874 (make_pass_rebuild_cgraph_edges): ...new function.
9875 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9876 subclass of gimple_opt_pass along with...
9877 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9878 and...
9879 (make_pass_remove_cgraph_callee_edges): ...new function.
9880 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9881 struct to a subclass of rtl_opt_pass along with...
9882 (pass_data_stack_adjustments): ...new pass_data instance and...
9883 (make_pass_stack_adjustments): ...new function.
9884 * combine.c (pass_combine): Convert from a global struct to a subclass
9885 of rtl_opt_pass along with...
9886 (pass_data_combine): ...new pass_data instance and...
9887 (make_pass_combine): ...new function.
9888 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9889 global struct to a subclass of rtl_opt_pass along with...
9890 (pass_data_compare_elim_after_reload): ...new pass_data instance
9891 and...
9892 (make_pass_compare_elim_after_reload): ...new function.
9893 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9894 of rtl_opt_pass along with...
9895 (pass_data_rtl_cprop): ...new pass_data instance and...
9896 (make_pass_rtl_cprop): ...new function.
9897 * cse.c (pass_cse): Convert from a global struct to a subclass of
9898 rtl_opt_pass along with...
9899 (pass_data_cse): ...new pass_data instance and...
9900 (make_pass_cse): ...new function.
9901 (pass_cse2): Convert from a global struct to a subclass of
9902 rtl_opt_pass along with...
9903 (pass_data_cse2): ...new pass_data instance and...
9904 (make_pass_cse2): ...new function.
9905 (pass_cse_after_global_opts): Convert from a global struct to a
9906 subclass of rtl_opt_pass along with...
9907 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9908 (make_pass_cse_after_global_opts): ...new function.
9909 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9910 of rtl_opt_pass along with...
9911 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9912 (make_pass_ud_rtl_dce): ...new function.
9913 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9914 rtl_opt_pass along with...
9915 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9916 (make_pass_fast_rtl_dce): ...new function.
9917 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9918 a subclass of rtl_opt_pass along with...
9919 (pass_data_df_initialize_opt): ...new pass_data instance and...
9920 (make_pass_df_initialize_opt): ...new function.
9921 (pass_df_initialize_no_opt): Convert from a global struct to a
9922 subclass of rtl_opt_pass along with...
9923 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9924 (make_pass_df_initialize_no_opt): ...new function.
9925 (pass_df_finish): Convert from a global struct to a subclass of
9926 rtl_opt_pass along with...
9927 (pass_data_df_finish): ...new pass_data instance and...
9928 (make_pass_df_finish): ...new function.
9929 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9930 rtl_opt_pass along with...
9931 (pass_data_rtl_dse1): ...new pass_data instance and...
9932 (make_pass_rtl_dse1): ...new function.
9933 (pass_rtl_dse2): Convert from a global struct to a subclass of
9934 rtl_opt_pass along with...
9935 (pass_data_rtl_dse2): ...new pass_data instance and...
9936 (make_pass_rtl_dse2): ...new function.
9937 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9938 subclass of rtl_opt_pass along with...
9939 (pass_data_dwarf2_frame): ...new pass_data instance and...
9940 (make_pass_dwarf2_frame): ...new function.
9941 * except.c (pass_set_nothrow_function_flags): Convert from a global
9942 struct to a subclass of rtl_opt_pass along with...
9943 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9944 and...
9945 (make_pass_set_nothrow_function_flags): ...new function.
9946 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9947 subclass of rtl_opt_pass along with...
9948 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9949 and...
9950 (make_pass_convert_to_eh_region_ranges): ...new function.
9951 * final.c (pass_compute_alignments): Convert from a global struct to a
9952 subclass of rtl_opt_pass along with...
9953 (pass_data_compute_alignments): ...new pass_data instance and...
9954 (make_pass_compute_alignments): ...new function.
9955 (pass_final): Convert from a global struct to a subclass of
9956 rtl_opt_pass along with...
9957 (pass_data_final): ...new pass_data instance and...
9958 (make_pass_final): ...new function.
9959 (pass_shorten_branches): Convert from a global struct to a subclass of
9960 rtl_opt_pass along with...
9961 (pass_data_shorten_branches): ...new pass_data instance and...
9962 (make_pass_shorten_branches): ...new function.
9963 (pass_clean_state): Convert from a global struct to a subclass of
9964 rtl_opt_pass along with...
9965 (pass_data_clean_state): ...new pass_data instance and...
9966 (make_pass_clean_state): ...new function.
9967 * function.c (pass_instantiate_virtual_regs): Convert from a global
9968 struct to a subclass of rtl_opt_pass along with...
9969 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9970 (make_pass_instantiate_virtual_regs): ...new function.
9971 (pass_leaf_regs): Convert from a global struct to a subclass of
9972 rtl_opt_pass along with...
9973 (pass_data_leaf_regs): ...new pass_data instance and...
9974 (make_pass_leaf_regs): ...new function.
9975 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9976 subclass of rtl_opt_pass along with...
9977 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9978 and...
9979 (make_pass_thread_prologue_and_epilogue): ...new function.
9980 (pass_match_asm_constraints): Convert from a global struct to a
9981 subclass of rtl_opt_pass along with...
9982 (pass_data_match_asm_constraints): ...new pass_data instance and...
9983 (make_pass_match_asm_constraints): ...new function.
9984 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9985 subclass of rtl_opt_pass along with...
9986 (pass_data_rtl_fwprop): ...new pass_data instance and...
9987 (make_pass_rtl_fwprop): ...new function.
9988 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9989 rtl_opt_pass along with...
9990 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9991 (make_pass_rtl_fwprop_addr): ...new function.
9992 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9993 rtl_opt_pass along with...
9994 (pass_data_rtl_pre): ...new pass_data instance and...
9995 (make_pass_rtl_pre): ...new function.
9996 (pass_rtl_hoist): Convert from a global struct to a subclass of
9997 rtl_opt_pass along with...
9998 (pass_data_rtl_hoist): ...new pass_data instance and...
9999 (make_pass_rtl_hoist): ...new function.
10000 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10001 subclass of gimple_opt_pass along with...
10002 (pass_data_lower_cf): ...new pass_data instance and...
10003 (make_pass_lower_cf): ...new function.
10004 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10005 from a global struct to a subclass of gimple_opt_pass along with...
10006 (pass_data_strength_reduction): ...new pass_data instance and...
10007 (make_pass_strength_reduction): ...new function.
10008 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10009 of rtl_opt_pass along with...
10010 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10011 (make_pass_rtl_ifcvt): ...new function.
10012 (pass_if_after_combine): Convert from a global struct to a subclass of
10013 rtl_opt_pass along with...
10014 (pass_data_if_after_combine): ...new pass_data instance and...
10015 (make_pass_if_after_combine): ...new function.
10016 (pass_if_after_reload): Convert from a global struct to a subclass of
10017 rtl_opt_pass along with...
10018 (pass_data_if_after_reload): ...new pass_data instance and...
10019 (make_pass_if_after_reload): ...new function.
10020 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10021 a subclass of rtl_opt_pass along with...
10022 (pass_data_initialize_regs): ...new pass_data instance and...
10023 (make_pass_initialize_regs): ...new function.
10024 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10025 of ipa_opt_pass_d along with...
10026 (pass_data_ipa_cp): ...new pass_data instance and...
10027 (make_pass_ipa_cp): ...new function.
10028 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10029 global struct to a subclass of gimple_opt_pass along with...
10030 (pass_data_inline_parameters): ...new pass_data instance and...
10031 (make_pass_inline_parameters): ...new function.
10032 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10033 subclass of gimple_opt_pass along with...
10034 (pass_data_early_inline): ...new pass_data instance and...
10035 (make_pass_early_inline): ...new function.
10036 (pass_ipa_inline): Convert from a global struct to a subclass of
10037 ipa_opt_pass_d along with...
10038 (pass_data_ipa_inline): ...new pass_data instance and...
10039 (make_pass_ipa_inline): ...new function.
10040 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10041 struct to a subclass of gimple_opt_pass along with...
10042 (pass_data_local_pure_const): ...new pass_data instance and...
10043 (make_pass_local_pure_const): ...new function.
10044 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10045 ipa_opt_pass_d along with...
10046 (pass_data_ipa_pure_const): ...new pass_data instance and...
10047 (make_pass_ipa_pure_const): ...new function.
10048 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10049 to a subclass of ipa_opt_pass_d along with...
10050 (pass_data_ipa_reference): ...new pass_data instance and...
10051 (make_pass_ipa_reference): ...new function.
10052 * ipa-split.c (pass_split_functions): Convert from a global struct to
10053 a subclass of gimple_opt_pass along with...
10054 (pass_data_split_functions): ...new pass_data instance and...
10055 (make_pass_split_functions): ...new function.
10056 (pass_feedback_split_functions): Convert from a global struct to a
10057 subclass of gimple_opt_pass along with...
10058 (pass_data_feedback_split_functions): ...new pass_data instance and...
10059 (make_pass_feedback_split_functions): ...new function.
10060 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10061 global struct to a subclass of simple_ipa_opt_pass along with...
10062 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10063 instance and...
10064 (make_pass_ipa_function_and_variable_visibility): ...new function.
10065 (pass_ipa_free_inline_summary): Convert from a global struct to a
10066 subclass of simple_ipa_opt_pass along with...
10067 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10068 (make_pass_ipa_free_inline_summary): ...new function.
10069 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10070 subclass of ipa_opt_pass_d along with...
10071 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10072 and...
10073 (make_pass_ipa_whole_program_visibility): ...new function.
10074 (pass_ipa_profile): Convert from a global struct to a subclass of
10075 ipa_opt_pass_d along with...
10076 (pass_data_ipa_profile): ...new pass_data instance and...
10077 (make_pass_ipa_profile): ...new function.
10078 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10079 ipa_opt_pass_d along with...
10080 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10081 (make_pass_ipa_cdtor_merge): ...new function.
10082 * ira.c (pass_ira): Convert from a global struct to a subclass of
10083 rtl_opt_pass along with...
10084 (pass_data_ira): ...new pass_data instance and...
10085 (make_pass_ira): ...new function.
10086 (pass_reload): Convert from a global struct to a subclass of
10087 rtl_opt_pass along with...
10088 (pass_data_reload): ...new pass_data instance and...
10089 (make_pass_reload): ...new function.
10090 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10091 subclass of rtl_opt_pass along with...
10092 (pass_data_cleanup_barriers): ...new pass_data instance and...
10093 (make_pass_cleanup_barriers): ...new function.
10094 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10095 of rtl_opt_pass along with...
10096 (pass_data_loop2): ...new pass_data instance and...
10097 (make_pass_loop2): ...new function.
10098 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10099 rtl_opt_pass along with...
10100 (pass_data_rtl_loop_init): ...new pass_data instance and...
10101 (make_pass_rtl_loop_init): ...new function.
10102 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10103 rtl_opt_pass along with...
10104 (pass_data_rtl_loop_done): ...new pass_data instance and...
10105 (make_pass_rtl_loop_done): ...new function.
10106 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10107 subclass of rtl_opt_pass along with...
10108 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10109 (make_pass_rtl_move_loop_invariants): ...new function.
10110 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10111 rtl_opt_pass along with...
10112 (pass_data_rtl_unswitch): ...new pass_data instance and...
10113 (make_pass_rtl_unswitch): ...new function.
10114 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10115 subclass of rtl_opt_pass along with...
10116 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10117 and...
10118 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10119 (pass_rtl_doloop): Convert from a global struct to a subclass of
10120 rtl_opt_pass along with...
10121 (pass_data_rtl_doloop): ...new pass_data instance and...
10122 (make_pass_rtl_doloop): ...new function.
10123 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10124 a subclass of rtl_opt_pass along with...
10125 (pass_data_lower_subreg): ...new pass_data instance and...
10126 (make_pass_lower_subreg): ...new function.
10127 (pass_lower_subreg2): Convert from a global struct to a subclass of
10128 rtl_opt_pass along with...
10129 (pass_data_lower_subreg2): ...new pass_data instance and...
10130 (make_pass_lower_subreg2): ...new function.
10131 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10132 struct to a subclass of ipa_opt_pass_d along with...
10133 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10134 (make_pass_ipa_lto_gimple_out): ...new function.
10135 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10136 of ipa_opt_pass_d along with...
10137 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10138 (make_pass_ipa_lto_finish_out): ...new function.
10139 * mode-switching.c (pass_mode_switching): Convert from a global struct
10140 to a subclass of rtl_opt_pass along with...
10141 (pass_data_mode_switching): ...new pass_data instance and...
10142 (make_pass_mode_switching): ...new function.
10143 * modulo-sched.c (pass_sms): Convert from a global struct to a
10144 subclass of rtl_opt_pass along with...
10145 (pass_data_sms): ...new pass_data instance and...
10146 (make_pass_sms): ...new function.
10147 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10148 subclass of gimple_opt_pass along with...
10149 (pass_data_expand_omp): ...new pass_data instance and...
10150 (make_pass_expand_omp): ...new function.
10151 (pass_lower_omp): Convert from a global struct to a subclass of
10152 gimple_opt_pass along with...
10153 (pass_data_lower_omp): ...new pass_data instance and...
10154 (make_pass_lower_omp): ...new function.
10155 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10156 of gimple_opt_pass along with...
10157 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10158 (make_pass_diagnose_omp_blocks): ...new function.
10159 * passes.c (pass_early_local_passes): Convert from a global struct to
10160 a subclass of simple_ipa_opt_pass along with...
10161 (pass_data_early_local_passes): ...new pass_data instance and...
10162 (make_pass_early_local_passes): ...new function.
10163 (pass_all_early_optimizations): Convert from a global struct to a
10164 subclass of gimple_opt_pass along with...
10165 (pass_data_all_early_optimizations): ...new pass_data instance and...
10166 (make_pass_all_early_optimizations): ...new function.
10167 (pass_all_optimizations): Convert from a global struct to a subclass
10168 of gimple_opt_pass along with...
10169 (pass_data_all_optimizations): ...new pass_data instance and...
10170 (make_pass_all_optimizations): ...new function.
10171 (pass_all_optimizations_g): Convert from a global struct to a subclass
10172 of gimple_opt_pass along with...
10173 (pass_data_all_optimizations_g): ...new pass_data instance and...
10174 (make_pass_all_optimizations_g): ...new function.
10175 (pass_rest_of_compilation): Convert from a global struct to a subclass
10176 of rtl_opt_pass along with...
10177 (pass_data_rest_of_compilation): ...new pass_data instance and...
10178 (make_pass_rest_of_compilation): ...new function.
10179 (pass_postreload): Convert from a global struct to a subclass of
10180 rtl_opt_pass along with...
10181 (pass_data_postreload): ...new pass_data instance and...
10182 (make_pass_postreload): ...new function.
10183 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10184 subclass of rtl_opt_pass along with...
10185 (pass_data_gcse2): ...new pass_data instance and...
10186 (make_pass_gcse2): ...new function.
10187 * postreload.c (pass_postreload_cse): Convert from a global struct to
10188 a subclass of rtl_opt_pass along with...
10189 (pass_data_postreload_cse): ...new pass_data instance and...
10190 (make_pass_postreload_cse): ...new function.
10191 * predict.c (pass_profile): Convert from a global struct to a subclass
10192 of gimple_opt_pass along with...
10193 (pass_data_profile): ...new pass_data instance and...
10194 (make_pass_profile): ...new function.
10195 (pass_strip_predict_hints): Convert from a global struct to a subclass
10196 of gimple_opt_pass along with...
10197 (pass_data_strip_predict_hints): ...new pass_data instance and...
10198 (make_pass_strip_predict_hints): ...new function.
10199 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10200 of rtl_opt_pass along with...
10201 (pass_data_peephole2): ...new pass_data instance and...
10202 (make_pass_peephole2): ...new function.
10203 (pass_split_all_insns): Convert from a global struct to a subclass of
10204 rtl_opt_pass along with...
10205 (pass_data_split_all_insns): ...new pass_data instance and...
10206 (make_pass_split_all_insns): ...new function.
10207 (pass_split_after_reload): Convert from a global struct to a subclass
10208 of rtl_opt_pass along with...
10209 (pass_data_split_after_reload): ...new pass_data instance and...
10210 (make_pass_split_after_reload): ...new function.
10211 (pass_split_before_regstack): Convert from a global struct to a
10212 subclass of rtl_opt_pass along with...
10213 (pass_data_split_before_regstack): ...new pass_data instance and...
10214 (make_pass_split_before_regstack): ...new function.
10215 (pass_split_before_sched2): Convert from a global struct to a subclass
10216 of rtl_opt_pass along with...
10217 (pass_data_split_before_sched2): ...new pass_data instance and...
10218 (make_pass_split_before_sched2): ...new function.
10219 (pass_split_for_shorten_branches): Convert from a global struct to a
10220 subclass of rtl_opt_pass along with...
10221 (pass_data_split_for_shorten_branches): ...new pass_data instance
10222 and...
10223 (make_pass_split_for_shorten_branches): ...new function.
10224 * ree.c (pass_ree): Convert from a global struct to a subclass of
10225 rtl_opt_pass along with...
10226 (pass_data_ree): ...new pass_data instance and...
10227 (make_pass_ree): ...new function.
10228 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10229 subclass of rtl_opt_pass along with...
10230 (pass_data_stack_regs): ...new pass_data instance and...
10231 (make_pass_stack_regs): ...new function.
10232 (pass_stack_regs_run): Convert from a global struct to a subclass of
10233 rtl_opt_pass along with...
10234 (pass_data_stack_regs_run): ...new pass_data instance and...
10235 (make_pass_stack_regs_run): ...new function.
10236 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10237 subclass of rtl_opt_pass along with...
10238 (pass_data_cprop_hardreg): ...new pass_data instance and...
10239 (make_pass_cprop_hardreg): ...new function.
10240 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10241 subclass of rtl_opt_pass along with...
10242 (pass_data_reginfo_init): ...new pass_data instance and...
10243 (make_pass_reginfo_init): ...new function.
10244 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10245 of rtl_opt_pass along with...
10246 (pass_data_regmove): ...new pass_data instance and...
10247 (make_pass_regmove): ...new function.
10248 * regrename.c (pass_regrename): Convert from a global struct to a
10249 subclass of rtl_opt_pass along with...
10250 (pass_data_regrename): ...new pass_data instance and...
10251 (make_pass_regrename): ...new function.
10252 * reorg.c (pass_delay_slots): Convert from a global struct to a
10253 subclass of rtl_opt_pass along with...
10254 (pass_data_delay_slots): ...new pass_data instance and...
10255 (make_pass_delay_slots): ...new function.
10256 (pass_machine_reorg): Convert from a global struct to a subclass of
10257 rtl_opt_pass along with...
10258 (pass_data_machine_reorg): ...new pass_data instance and...
10259 (make_pass_machine_reorg): ...new function.
10260 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10261 of rtl_opt_pass along with...
10262 (pass_data_sched): ...new pass_data instance and...
10263 (make_pass_sched): ...new function.
10264 (pass_sched2): Convert from a global struct to a subclass of
10265 rtl_opt_pass along with...
10266 (pass_data_sched2): ...new pass_data instance and...
10267 (make_pass_sched2): ...new function.
10268 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10269 to a subclass of rtl_opt_pass along with...
10270 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10271 (make_pass_stack_ptr_mod): ...new function.
10272 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10273 to a subclass of rtl_opt_pass along with...
10274 (pass_data_rtl_store_motion): ...new pass_data instance and...
10275 (make_pass_rtl_store_motion): ...new function.
10276 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10277 of gimple_opt_pass along with...
10278 (pass_data_tracer): ...new pass_data instance and...
10279 (make_pass_tracer): ...new function.
10280 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10281 to a subclass of gimple_opt_pass along with...
10282 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10283 (make_pass_diagnose_tm_blocks): ...new function.
10284 (pass_lower_tm): Convert from a global struct to a subclass of
10285 gimple_opt_pass along with...
10286 (pass_data_lower_tm): ...new pass_data instance and...
10287 (make_pass_lower_tm): ...new function.
10288 (pass_tm_init): Convert from a global struct to a subclass of
10289 gimple_opt_pass along with...
10290 (pass_data_tm_init): ...new pass_data instance and...
10291 (make_pass_tm_init): ...new function.
10292 (pass_tm_mark): Convert from a global struct to a subclass of
10293 gimple_opt_pass along with...
10294 (pass_data_tm_mark): ...new pass_data instance and...
10295 (make_pass_tm_mark): ...new function.
10296 (pass_tm_edges): Convert from a global struct to a subclass of
10297 gimple_opt_pass along with...
10298 (pass_data_tm_edges): ...new pass_data instance and...
10299 (make_pass_tm_edges): ...new function.
10300 (pass_tm_memopt): Convert from a global struct to a subclass of
10301 gimple_opt_pass along with...
10302 (pass_data_tm_memopt): ...new pass_data instance and...
10303 (make_pass_tm_memopt): ...new function.
10304 (pass_ipa_tm): Convert from a global struct to a subclass of
10305 simple_ipa_opt_pass along with...
10306 (pass_data_ipa_tm): ...new pass_data instance and...
10307 (make_pass_ipa_tm): ...new function.
10308 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10309 subclass of gimple_opt_pass along with...
10310 (pass_data_call_cdce): ...new pass_data instance and...
10311 (make_pass_call_cdce): ...new function.
10312 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10313 subclass of gimple_opt_pass along with...
10314 (pass_data_build_cfg): ...new pass_data instance and...
10315 (make_pass_build_cfg): ...new function.
10316 (pass_split_crit_edges): Convert from a global struct to a subclass of
10317 gimple_opt_pass along with...
10318 (pass_data_split_crit_edges): ...new pass_data instance and...
10319 (make_pass_split_crit_edges): ...new function.
10320 (pass_warn_function_return): Convert from a global struct to a
10321 subclass of gimple_opt_pass along with...
10322 (pass_data_warn_function_return): ...new pass_data instance and...
10323 (make_pass_warn_function_return): ...new function.
10324 (pass_warn_function_noreturn): Convert from a global struct to a
10325 subclass of gimple_opt_pass along with...
10326 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10327 (make_pass_warn_function_noreturn): ...new function.
10328 (pass_warn_unused_result): Convert from a global struct to a subclass
10329 of gimple_opt_pass along with...
10330 (pass_data_warn_unused_result): ...new pass_data instance and...
10331 (make_pass_warn_unused_result): ...new function.
10332 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10333 a subclass of gimple_opt_pass along with...
10334 (pass_data_merge_phi): ...new pass_data instance and...
10335 (make_pass_merge_phi): ...new function.
10336 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10337 a subclass of gimple_opt_pass along with...
10338 (pass_data_lower_complex): ...new pass_data instance and...
10339 (make_pass_lower_complex): ...new function.
10340 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10341 gimple_opt_pass along with...
10342 (pass_data_lower_complex_O0): ...new pass_data instance and...
10343 (make_pass_lower_complex_O0): ...new function.
10344 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10345 subclass of gimple_opt_pass along with...
10346 (pass_data_lower_eh): ...new pass_data instance and...
10347 (make_pass_lower_eh): ...new function.
10348 (pass_refactor_eh): Convert from a global struct to a subclass of
10349 gimple_opt_pass along with...
10350 (pass_data_refactor_eh): ...new pass_data instance and...
10351 (make_pass_refactor_eh): ...new function.
10352 (pass_lower_resx): Convert from a global struct to a subclass of
10353 gimple_opt_pass along with...
10354 (pass_data_lower_resx): ...new pass_data instance and...
10355 (make_pass_lower_resx): ...new function.
10356 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10357 of gimple_opt_pass along with...
10358 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10359 (make_pass_lower_eh_dispatch): ...new function.
10360 (pass_cleanup_eh): Convert from a global struct to a subclass of
10361 gimple_opt_pass along with...
10362 (pass_data_cleanup_eh): ...new pass_data instance and...
10363 (make_pass_cleanup_eh): ...new function.
10364 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10365 to a subclass of simple_ipa_opt_pass along with...
10366 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10367 (make_pass_ipa_lower_emutls): ...new function.
10368 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10369 a subclass of gimple_opt_pass along with...
10370 (pass_data_if_conversion): ...new pass_data instance and...
10371 (make_pass_if_conversion): ...new function.
10372 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10373 subclass of gimple_opt_pass along with...
10374 (pass_data_build_ssa): ...new pass_data instance and...
10375 (make_pass_build_ssa): ...new function.
10376 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10377 global struct to a subclass of gimple_opt_pass along with...
10378 (pass_data_loop_distribution): ...new pass_data instance and...
10379 (make_pass_loop_distribution): ...new function.
10380 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10381 subclass of gimple_opt_pass along with...
10382 (pass_data_mudflap_1): ...new pass_data instance and...
10383 (make_pass_mudflap_1): ...new function.
10384 (pass_mudflap_2): Convert from a global struct to a subclass of
10385 gimple_opt_pass along with...
10386 (pass_data_mudflap_2): ...new pass_data instance and...
10387 (make_pass_mudflap_2): ...new function.
10388 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10389 subclass of gimple_opt_pass along with...
10390 (pass_data_mudflap_1): ...new pass_data instance and...
10391 (make_pass_mudflap_1): ...new function.
10392 (pass_mudflap_2): Convert from a global struct to a subclass of
10393 gimple_opt_pass along with...
10394 (pass_data_mudflap_2): ...new pass_data instance and...
10395 (make_pass_mudflap_2): ...new function.
10396 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10397 gimple_opt_pass along with...
10398 (pass_data_nrv): ...new pass_data instance and...
10399 (make_pass_nrv): ...new function.
10400 (pass_return_slot): Convert from a global struct to a subclass of
10401 gimple_opt_pass along with...
10402 (pass_data_return_slot): ...new pass_data instance and...
10403 (make_pass_return_slot): ...new function.
10404 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10405 to a subclass of gimple_opt_pass along with...
10406 (pass_data_object_sizes): ...new pass_data instance and...
10407 (make_pass_object_sizes): ...new function.
10408 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10409 global struct to a subclass of gimple_opt_pass along with...
10410 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10411 and...
10412 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10413 (pass_fixup_cfg): Convert from a global struct to a subclass of
10414 gimple_opt_pass along with...
10415 (pass_data_fixup_cfg): ...new pass_data instance and...
10416 (make_pass_fixup_cfg): ...new function.
10417 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10418 (make_pass_mudflap_1): ...new function.
10419 (pass_mudflap_2): Replace declaration with that of...
10420 (make_pass_mudflap_2): ...new function.
10421 (pass_asan): Replace declaration with that of...
10422 (make_pass_asan): ...new function.
10423 (pass_asan_O0): Replace declaration with that of...
10424 (make_pass_asan_O0): ...new function.
10425 (pass_tsan): Replace declaration with that of...
10426 (make_pass_tsan): ...new function.
10427 (pass_tsan_O0): Replace declaration with that of...
10428 (make_pass_tsan_O0): ...new function.
10429 (pass_lower_cf): Replace declaration with that of...
10430 (make_pass_lower_cf): ...new function.
10431 (pass_refactor_eh): Replace declaration with that of...
10432 (make_pass_refactor_eh): ...new function.
10433 (pass_lower_eh): Replace declaration with that of...
10434 (make_pass_lower_eh): ...new function.
10435 (pass_lower_eh_dispatch): Replace declaration with that of...
10436 (make_pass_lower_eh_dispatch): ...new function.
10437 (pass_lower_resx): Replace declaration with that of...
10438 (make_pass_lower_resx): ...new function.
10439 (pass_build_cfg): Replace declaration with that of...
10440 (make_pass_build_cfg): ...new function.
10441 (pass_early_tree_profile): Replace declaration with that of...
10442 (make_pass_early_tree_profile): ...new function.
10443 (pass_cleanup_eh): Replace declaration with that of...
10444 (make_pass_cleanup_eh): ...new function.
10445 (pass_sra): Replace declaration with that of...
10446 (make_pass_sra): ...new function.
10447 (pass_sra_early): Replace declaration with that of...
10448 (make_pass_sra_early): ...new function.
10449 (pass_early_ipa_sra): Replace declaration with that of...
10450 (make_pass_early_ipa_sra): ...new function.
10451 (pass_tail_recursion): Replace declaration with that of...
10452 (make_pass_tail_recursion): ...new function.
10453 (pass_tail_calls): Replace declaration with that of...
10454 (make_pass_tail_calls): ...new function.
10455 (pass_tree_loop): Replace declaration with that of...
10456 (make_pass_tree_loop): ...new function.
10457 (pass_tree_loop_init): Replace declaration with that of...
10458 (make_pass_tree_loop_init): ...new function.
10459 (pass_lim): Replace declaration with that of...
10460 (make_pass_lim): ...new function.
10461 (pass_tree_unswitch): Replace declaration with that of...
10462 (make_pass_tree_unswitch): ...new function.
10463 (pass_predcom): Replace declaration with that of...
10464 (make_pass_predcom): ...new function.
10465 (pass_iv_canon): Replace declaration with that of...
10466 (make_pass_iv_canon): ...new function.
10467 (pass_scev_cprop): Replace declaration with that of...
10468 (make_pass_scev_cprop): ...new function.
10469 (pass_empty_loop): Replace declaration with that of...
10470 (make_pass_empty_loop): ...new function.
10471 (pass_record_bounds): Replace declaration with that of...
10472 (make_pass_record_bounds): ...new function.
10473 (pass_graphite): Replace declaration with that of...
10474 (make_pass_graphite): ...new function.
10475 (pass_graphite_transforms): Replace declaration with that of...
10476 (make_pass_graphite_transforms): ...new function.
10477 (pass_if_conversion): Replace declaration with that of...
10478 (make_pass_if_conversion): ...new function.
10479 (pass_loop_distribution): Replace declaration with that of...
10480 (make_pass_loop_distribution): ...new function.
10481 (pass_vectorize): Replace declaration with that of...
10482 (make_pass_vectorize): ...new function.
10483 (pass_slp_vectorize): Replace declaration with that of...
10484 (make_pass_slp_vectorize): ...new function.
10485 (pass_complete_unroll): Replace declaration with that of...
10486 (make_pass_complete_unroll): ...new function.
10487 (pass_complete_unrolli): Replace declaration with that of...
10488 (make_pass_complete_unrolli): ...new function.
10489 (pass_parallelize_loops): Replace declaration with that of...
10490 (make_pass_parallelize_loops): ...new function.
10491 (pass_loop_prefetch): Replace declaration with that of...
10492 (make_pass_loop_prefetch): ...new function.
10493 (pass_iv_optimize): Replace declaration with that of...
10494 (make_pass_iv_optimize): ...new function.
10495 (pass_tree_loop_done): Replace declaration with that of...
10496 (make_pass_tree_loop_done): ...new function.
10497 (pass_ch): Replace declaration with that of...
10498 (make_pass_ch): ...new function.
10499 (pass_ccp): Replace declaration with that of...
10500 (make_pass_ccp): ...new function.
10501 (pass_phi_only_cprop): Replace declaration with that of...
10502 (make_pass_phi_only_cprop): ...new function.
10503 (pass_build_ssa): Replace declaration with that of...
10504 (make_pass_build_ssa): ...new function.
10505 (pass_build_alias): Replace declaration with that of...
10506 (make_pass_build_alias): ...new function.
10507 (pass_build_ealias): Replace declaration with that of...
10508 (make_pass_build_ealias): ...new function.
10509 (pass_dominator): Replace declaration with that of...
10510 (make_pass_dominator): ...new function.
10511 (pass_dce): Replace declaration with that of...
10512 (make_pass_dce): ...new function.
10513 (pass_dce_loop): Replace declaration with that of...
10514 (make_pass_dce_loop): ...new function.
10515 (pass_cd_dce): Replace declaration with that of...
10516 (make_pass_cd_dce): ...new function.
10517 (pass_call_cdce): Replace declaration with that of...
10518 (make_pass_call_cdce): ...new function.
10519 (pass_merge_phi): Replace declaration with that of...
10520 (make_pass_merge_phi): ...new function.
10521 (pass_split_crit_edges): Replace declaration with that of...
10522 (make_pass_split_crit_edges): ...new function.
10523 (pass_pre): Replace declaration with that of...
10524 (make_pass_pre): ...new function.
10525 (pass_profile): Replace declaration with that of...
10526 (make_pass_profile): ...new function.
10527 (pass_strip_predict_hints): Replace declaration with that of...
10528 (make_pass_strip_predict_hints): ...new function.
10529 (pass_lower_complex_O0): Replace declaration with that of...
10530 (make_pass_lower_complex_O0): ...new function.
10531 (pass_lower_complex): Replace declaration with that of...
10532 (make_pass_lower_complex): ...new function.
10533 (pass_lower_vector): Replace declaration with that of...
10534 (make_pass_lower_vector): ...new function.
10535 (pass_lower_vector_ssa): Replace declaration with that of...
10536 (make_pass_lower_vector_ssa): ...new function.
10537 (pass_lower_omp): Replace declaration with that of...
10538 (make_pass_lower_omp): ...new function.
10539 (pass_diagnose_omp_blocks): Replace declaration with that of...
10540 (make_pass_diagnose_omp_blocks): ...new function.
10541 (pass_expand_omp): Replace declaration with that of...
10542 (make_pass_expand_omp): ...new function.
10543 (pass_expand_omp_ssa): Replace declaration with that of...
10544 (make_pass_expand_omp_ssa): ...new function.
10545 (pass_object_sizes): Replace declaration with that of...
10546 (make_pass_object_sizes): ...new function.
10547 (pass_strlen): Replace declaration with that of...
10548 (make_pass_strlen): ...new function.
10549 (pass_fold_builtins): Replace declaration with that of...
10550 (make_pass_fold_builtins): ...new function.
10551 (pass_stdarg): Replace declaration with that of...
10552 (make_pass_stdarg): ...new function.
10553 (pass_early_warn_uninitialized): Replace declaration with that of...
10554 (make_pass_early_warn_uninitialized): ...new function.
10555 (pass_late_warn_uninitialized): Replace declaration with that of...
10556 (make_pass_late_warn_uninitialized): ...new function.
10557 (pass_cse_reciprocals): Replace declaration with that of...
10558 (make_pass_cse_reciprocals): ...new function.
10559 (pass_cse_sincos): Replace declaration with that of...
10560 (make_pass_cse_sincos): ...new function.
10561 (pass_optimize_bswap): Replace declaration with that of...
10562 (make_pass_optimize_bswap): ...new function.
10563 (pass_optimize_widening_mul): Replace declaration with that of...
10564 (make_pass_optimize_widening_mul): ...new function.
10565 (pass_warn_function_return): Replace declaration with that of...
10566 (make_pass_warn_function_return): ...new function.
10567 (pass_warn_function_noreturn): Replace declaration with that of...
10568 (make_pass_warn_function_noreturn): ...new function.
10569 (pass_cselim): Replace declaration with that of...
10570 (make_pass_cselim): ...new function.
10571 (pass_phiopt): Replace declaration with that of...
10572 (make_pass_phiopt): ...new function.
10573 (pass_forwprop): Replace declaration with that of...
10574 (make_pass_forwprop): ...new function.
10575 (pass_phiprop): Replace declaration with that of...
10576 (make_pass_phiprop): ...new function.
10577 (pass_tree_ifcombine): Replace declaration with that of...
10578 (make_pass_tree_ifcombine): ...new function.
10579 (pass_dse): Replace declaration with that of...
10580 (make_pass_dse): ...new function.
10581 (pass_nrv): Replace declaration with that of...
10582 (make_pass_nrv): ...new function.
10583 (pass_rename_ssa_copies): Replace declaration with that of...
10584 (make_pass_rename_ssa_copies): ...new function.
10585 (pass_sink_code): Replace declaration with that of...
10586 (make_pass_sink_code): ...new function.
10587 (pass_fre): Replace declaration with that of...
10588 (make_pass_fre): ...new function.
10589 (pass_check_data_deps): Replace declaration with that of...
10590 (make_pass_check_data_deps): ...new function.
10591 (pass_copy_prop): Replace declaration with that of...
10592 (make_pass_copy_prop): ...new function.
10593 (pass_vrp): Replace declaration with that of...
10594 (make_pass_vrp): ...new function.
10595 (pass_uncprop): Replace declaration with that of...
10596 (make_pass_uncprop): ...new function.
10597 (pass_return_slot): Replace declaration with that of...
10598 (make_pass_return_slot): ...new function.
10599 (pass_reassoc): Replace declaration with that of...
10600 (make_pass_reassoc): ...new function.
10601 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10602 (make_pass_rebuild_cgraph_edges): ...new function.
10603 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10604 (make_pass_remove_cgraph_callee_edges): ...new function.
10605 (pass_build_cgraph_edges): Replace declaration with that of...
10606 (make_pass_build_cgraph_edges): ...new function.
10607 (pass_local_pure_const): Replace declaration with that of...
10608 (make_pass_local_pure_const): ...new function.
10609 (pass_tracer): Replace declaration with that of...
10610 (make_pass_tracer): ...new function.
10611 (pass_warn_unused_result): Replace declaration with that of...
10612 (make_pass_warn_unused_result): ...new function.
10613 (pass_diagnose_tm_blocks): Replace declaration with that of...
10614 (make_pass_diagnose_tm_blocks): ...new function.
10615 (pass_lower_tm): Replace declaration with that of...
10616 (make_pass_lower_tm): ...new function.
10617 (pass_tm_init): Replace declaration with that of...
10618 (make_pass_tm_init): ...new function.
10619 (pass_tm_mark): Replace declaration with that of...
10620 (make_pass_tm_mark): ...new function.
10621 (pass_tm_memopt): Replace declaration with that of...
10622 (make_pass_tm_memopt): ...new function.
10623 (pass_tm_edges): Replace declaration with that of...
10624 (make_pass_tm_edges): ...new function.
10625 (pass_split_functions): Replace declaration with that of...
10626 (make_pass_split_functions): ...new function.
10627 (pass_feedback_split_functions): Replace declaration with that of...
10628 (make_pass_feedback_split_functions): ...new function.
10629 (pass_strength_reduction): Replace declaration with that of...
10630 (make_pass_strength_reduction): ...new function.
10631 (pass_ipa_lower_emutls): Replace declaration with that of...
10632 (make_pass_ipa_lower_emutls): ...new function.
10633 (pass_ipa_function_and_variable_visibility): Replace declaration with
10634 that of...
10635 (make_pass_ipa_function_and_variable_visibility): ...new function.
10636 (pass_ipa_tree_profile): Replace declaration with that of...
10637 (make_pass_ipa_tree_profile): ...new function.
10638 (pass_early_local_passes): Replace declaration with that of...
10639 (make_pass_early_local_passes): ...new function.
10640 (pass_ipa_whole_program_visibility): Replace declaration with that
10641 of...
10642 (make_pass_ipa_whole_program_visibility): ...new function.
10643 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10644 (make_pass_ipa_lto_gimple_out): ...new function.
10645 (pass_ipa_increase_alignment): Replace declaration with that of...
10646 (make_pass_ipa_increase_alignment): ...new function.
10647 (pass_ipa_inline): Replace declaration with that of...
10648 (make_pass_ipa_inline): ...new function.
10649 (pass_ipa_free_lang_data): Replace declaration with that of...
10650 (make_pass_ipa_free_lang_data): ...new function.
10651 (pass_ipa_free_inline_summary): Replace declaration with that of...
10652 (make_pass_ipa_free_inline_summary): ...new function.
10653 (pass_ipa_cp): Replace declaration with that of...
10654 (make_pass_ipa_cp): ...new function.
10655 (pass_ipa_reference): Replace declaration with that of...
10656 (make_pass_ipa_reference): ...new function.
10657 (pass_ipa_pure_const): Replace declaration with that of...
10658 (make_pass_ipa_pure_const): ...new function.
10659 (pass_ipa_pta): Replace declaration with that of...
10660 (make_pass_ipa_pta): ...new function.
10661 (pass_ipa_lto_finish_out): Replace declaration with that of...
10662 (make_pass_ipa_lto_finish_out): ...new function.
10663 (pass_ipa_tm): Replace declaration with that of...
10664 (make_pass_ipa_tm): ...new function.
10665 (pass_ipa_profile): Replace declaration with that of...
10666 (make_pass_ipa_profile): ...new function.
10667 (pass_ipa_cdtor_merge): Replace declaration with that of...
10668 (make_pass_ipa_cdtor_merge): ...new function.
10669 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10670 of...
10671 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10672 (pass_init_datastructures): Replace declaration with that of...
10673 (make_pass_init_datastructures): ...new function.
10674 (pass_fixup_cfg): Replace declaration with that of...
10675 (make_pass_fixup_cfg): ...new function.
10676 (pass_expand): Replace declaration with that of...
10677 (make_pass_expand): ...new function.
10678 (pass_instantiate_virtual_regs): Replace declaration with that of...
10679 (make_pass_instantiate_virtual_regs): ...new function.
10680 (pass_rtl_fwprop): Replace declaration with that of...
10681 (make_pass_rtl_fwprop): ...new function.
10682 (pass_rtl_fwprop_addr): Replace declaration with that of...
10683 (make_pass_rtl_fwprop_addr): ...new function.
10684 (pass_jump): Replace declaration with that of...
10685 (make_pass_jump): ...new function.
10686 (pass_jump2): Replace declaration with that of...
10687 (make_pass_jump2): ...new function.
10688 (pass_lower_subreg): Replace declaration with that of...
10689 (make_pass_lower_subreg): ...new function.
10690 (pass_cse): Replace declaration with that of...
10691 (make_pass_cse): ...new function.
10692 (pass_fast_rtl_dce): Replace declaration with that of...
10693 (make_pass_fast_rtl_dce): ...new function.
10694 (pass_ud_rtl_dce): Replace declaration with that of...
10695 (make_pass_ud_rtl_dce): ...new function.
10696 (pass_rtl_dce): Replace declaration with that of...
10697 (make_pass_rtl_dce): ...new function.
10698 (pass_rtl_dse1): Replace declaration with that of...
10699 (make_pass_rtl_dse1): ...new function.
10700 (pass_rtl_dse2): Replace declaration with that of...
10701 (make_pass_rtl_dse2): ...new function.
10702 (pass_rtl_dse3): Replace declaration with that of...
10703 (make_pass_rtl_dse3): ...new function.
10704 (pass_rtl_cprop): Replace declaration with that of...
10705 (make_pass_rtl_cprop): ...new function.
10706 (pass_rtl_pre): Replace declaration with that of...
10707 (make_pass_rtl_pre): ...new function.
10708 (pass_rtl_hoist): Replace declaration with that of...
10709 (make_pass_rtl_hoist): ...new function.
10710 (pass_rtl_store_motion): Replace declaration with that of...
10711 (make_pass_rtl_store_motion): ...new function.
10712 (pass_cse_after_global_opts): Replace declaration with that of...
10713 (make_pass_cse_after_global_opts): ...new function.
10714 (pass_rtl_ifcvt): Replace declaration with that of...
10715 (make_pass_rtl_ifcvt): ...new function.
10716 (pass_into_cfg_layout_mode): Replace declaration with that of...
10717 (make_pass_into_cfg_layout_mode): ...new function.
10718 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10719 (make_pass_outof_cfg_layout_mode): ...new function.
10720 (pass_loop2): Replace declaration with that of...
10721 (make_pass_loop2): ...new function.
10722 (pass_rtl_loop_init): Replace declaration with that of...
10723 (make_pass_rtl_loop_init): ...new function.
10724 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10725 (make_pass_rtl_move_loop_invariants): ...new function.
10726 (pass_rtl_unswitch): Replace declaration with that of...
10727 (make_pass_rtl_unswitch): ...new function.
10728 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10729 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10730 (pass_rtl_doloop): Replace declaration with that of...
10731 (make_pass_rtl_doloop): ...new function.
10732 (pass_rtl_loop_done): Replace declaration with that of...
10733 (make_pass_rtl_loop_done): ...new function.
10734 (pass_web): Replace declaration with that of...
10735 (make_pass_web): ...new function.
10736 (pass_cse2): Replace declaration with that of...
10737 (make_pass_cse2): ...new function.
10738 (pass_df_initialize_opt): Replace declaration with that of...
10739 (make_pass_df_initialize_opt): ...new function.
10740 (pass_df_initialize_no_opt): Replace declaration with that of...
10741 (make_pass_df_initialize_no_opt): ...new function.
10742 (pass_reginfo_init): Replace declaration with that of...
10743 (make_pass_reginfo_init): ...new function.
10744 (pass_inc_dec): Replace declaration with that of...
10745 (make_pass_inc_dec): ...new function.
10746 (pass_stack_ptr_mod): Replace declaration with that of...
10747 (make_pass_stack_ptr_mod): ...new function.
10748 (pass_initialize_regs): Replace declaration with that of...
10749 (make_pass_initialize_regs): ...new function.
10750 (pass_combine): Replace declaration with that of...
10751 (make_pass_combine): ...new function.
10752 (pass_if_after_combine): Replace declaration with that of...
10753 (make_pass_if_after_combine): ...new function.
10754 (pass_ree): Replace declaration with that of...
10755 (make_pass_ree): ...new function.
10756 (pass_partition_blocks): Replace declaration with that of...
10757 (make_pass_partition_blocks): ...new function.
10758 (pass_match_asm_constraints): Replace declaration with that of...
10759 (make_pass_match_asm_constraints): ...new function.
10760 (pass_regmove): Replace declaration with that of...
10761 (make_pass_regmove): ...new function.
10762 (pass_split_all_insns): Replace declaration with that of...
10763 (make_pass_split_all_insns): ...new function.
10764 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10765 (make_pass_fast_rtl_byte_dce): ...new function.
10766 (pass_lower_subreg2): Replace declaration with that of...
10767 (make_pass_lower_subreg2): ...new function.
10768 (pass_mode_switching): Replace declaration with that of...
10769 (make_pass_mode_switching): ...new function.
10770 (pass_sms): Replace declaration with that of...
10771 (make_pass_sms): ...new function.
10772 (pass_sched): Replace declaration with that of...
10773 (make_pass_sched): ...new function.
10774 (pass_ira): Replace declaration with that of...
10775 (make_pass_ira): ...new function.
10776 (pass_reload): Replace declaration with that of...
10777 (make_pass_reload): ...new function.
10778 (pass_clean_state): Replace declaration with that of...
10779 (make_pass_clean_state): ...new function.
10780 (pass_branch_prob): Replace declaration with that of...
10781 (make_pass_branch_prob): ...new function.
10782 (pass_value_profile_transformations): Replace declaration with that
10783 of...
10784 (make_pass_value_profile_transformations): ...new function.
10785 (pass_postreload_cse): Replace declaration with that of...
10786 (make_pass_postreload_cse): ...new function.
10787 (pass_gcse2): Replace declaration with that of...
10788 (make_pass_gcse2): ...new function.
10789 (pass_split_after_reload): Replace declaration with that of...
10790 (make_pass_split_after_reload): ...new function.
10791 (pass_branch_target_load_optimize1): Replace declaration with that
10792 of...
10793 (make_pass_branch_target_load_optimize1): ...new function.
10794 (pass_thread_prologue_and_epilogue): Replace declaration with that
10795 of...
10796 (make_pass_thread_prologue_and_epilogue): ...new function.
10797 (pass_stack_adjustments): Replace declaration with that of...
10798 (make_pass_stack_adjustments): ...new function.
10799 (pass_peephole2): Replace declaration with that of...
10800 (make_pass_peephole2): ...new function.
10801 (pass_if_after_reload): Replace declaration with that of...
10802 (make_pass_if_after_reload): ...new function.
10803 (pass_regrename): Replace declaration with that of...
10804 (make_pass_regrename): ...new function.
10805 (pass_cprop_hardreg): Replace declaration with that of...
10806 (make_pass_cprop_hardreg): ...new function.
10807 (pass_reorder_blocks): Replace declaration with that of...
10808 (make_pass_reorder_blocks): ...new function.
10809 (pass_branch_target_load_optimize2): Replace declaration with that
10810 of...
10811 (make_pass_branch_target_load_optimize2): ...new function.
10812 (pass_leaf_regs): Replace declaration with that of...
10813 (make_pass_leaf_regs): ...new function.
10814 (pass_split_before_sched2): Replace declaration with that of...
10815 (make_pass_split_before_sched2): ...new function.
10816 (pass_compare_elim_after_reload): Replace declaration with that of...
10817 (make_pass_compare_elim_after_reload): ...new function.
10818 (pass_sched2): Replace declaration with that of...
10819 (make_pass_sched2): ...new function.
10820 (pass_stack_regs): Replace declaration with that of...
10821 (make_pass_stack_regs): ...new function.
10822 (pass_stack_regs_run): Replace declaration with that of...
10823 (make_pass_stack_regs_run): ...new function.
10824 (pass_df_finish): Replace declaration with that of...
10825 (make_pass_df_finish): ...new function.
10826 (pass_compute_alignments): Replace declaration with that of...
10827 (make_pass_compute_alignments): ...new function.
10828 (pass_duplicate_computed_gotos): Replace declaration with that of...
10829 (make_pass_duplicate_computed_gotos): ...new function.
10830 (pass_variable_tracking): Replace declaration with that of...
10831 (make_pass_variable_tracking): ...new function.
10832 (pass_free_cfg): Replace declaration with that of...
10833 (make_pass_free_cfg): ...new function.
10834 (pass_machine_reorg): Replace declaration with that of...
10835 (make_pass_machine_reorg): ...new function.
10836 (pass_cleanup_barriers): Replace declaration with that of...
10837 (make_pass_cleanup_barriers): ...new function.
10838 (pass_delay_slots): Replace declaration with that of...
10839 (make_pass_delay_slots): ...new function.
10840 (pass_split_for_shorten_branches): Replace declaration with that of...
10841 (make_pass_split_for_shorten_branches): ...new function.
10842 (pass_split_before_regstack): Replace declaration with that of...
10843 (make_pass_split_before_regstack): ...new function.
10844 (pass_convert_to_eh_region_ranges): Replace declaration with that
10845 of...
10846 (make_pass_convert_to_eh_region_ranges): ...new function.
10847 (pass_shorten_branches): Replace declaration with that of...
10848 (make_pass_shorten_branches): ...new function.
10849 (pass_set_nothrow_function_flags): Replace declaration with that of...
10850 (make_pass_set_nothrow_function_flags): ...new function.
10851 (pass_dwarf2_frame): Replace declaration with that of...
10852 (make_pass_dwarf2_frame): ...new function.
10853 (pass_final): Replace declaration with that of...
10854 (make_pass_final): ...new function.
10855 (pass_rtl_seqabstr): Replace declaration with that of...
10856 (make_pass_rtl_seqabstr): ...new function.
10857 (pass_release_ssa_names): Replace declaration with that of...
10858 (make_pass_release_ssa_names): ...new function.
10859 (pass_early_inline): Replace declaration with that of...
10860 (make_pass_early_inline): ...new function.
10861 (pass_inline_parameters): Replace declaration with that of...
10862 (make_pass_inline_parameters): ...new function.
10863 (pass_update_address_taken): Replace declaration with that of...
10864 (make_pass_update_address_taken): ...new function.
10865 (pass_convert_switch): Replace declaration with that of...
10866 (make_pass_convert_switch): ...new function.
10867 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10868 to a subclass of simple_ipa_opt_pass along with...
10869 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10870 (make_pass_ipa_tree_profile): ...new function.
10871 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10872 subclass of gimple_opt_pass along with...
10873 (pass_data_sra_early): ...new pass_data instance and...
10874 (make_pass_sra_early): ...new function.
10875 (pass_sra): Convert from a global struct to a subclass of
10876 gimple_opt_pass along with...
10877 (pass_data_sra): ...new pass_data instance and...
10878 (make_pass_sra): ...new function.
10879 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10880 gimple_opt_pass along with...
10881 (pass_data_early_ipa_sra): ...new pass_data instance and...
10882 (make_pass_early_ipa_sra): ...new function.
10883 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10884 subclass of gimple_opt_pass along with...
10885 (pass_data_ccp): ...new pass_data instance and...
10886 (make_pass_ccp): ...new function.
10887 (pass_fold_builtins): Convert from a global struct to a subclass of
10888 gimple_opt_pass along with...
10889 (pass_data_fold_builtins): ...new pass_data instance and...
10890 (make_pass_fold_builtins): ...new function.
10891 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10892 subclass of gimple_opt_pass along with...
10893 (pass_data_copy_prop): ...new pass_data instance and...
10894 (make_pass_copy_prop): ...new function.
10895 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10896 global struct to a subclass of gimple_opt_pass along with...
10897 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10898 (make_pass_rename_ssa_copies): ...new function.
10899 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10900 subclass of gimple_opt_pass along with...
10901 (pass_data_dce): ...new pass_data instance and...
10902 (make_pass_dce): ...new function.
10903 (pass_dce_loop): Convert from a global struct to a subclass of
10904 gimple_opt_pass along with...
10905 (pass_data_dce_loop): ...new pass_data instance and...
10906 (make_pass_dce_loop): ...new function.
10907 (pass_cd_dce): Convert from a global struct to a subclass of
10908 gimple_opt_pass along with...
10909 (pass_data_cd_dce): ...new pass_data instance and...
10910 (make_pass_cd_dce): ...new function.
10911 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10912 subclass of gimple_opt_pass along with...
10913 (pass_data_dominator): ...new pass_data instance and...
10914 (make_pass_dominator): ...new function.
10915 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10916 gimple_opt_pass along with...
10917 (pass_data_phi_only_cprop): ...new pass_data instance and...
10918 (make_pass_phi_only_cprop): ...new function.
10919 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10920 subclass of gimple_opt_pass along with...
10921 (pass_data_dse): ...new pass_data instance and...
10922 (make_pass_dse): ...new function.
10923 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10924 a subclass of gimple_opt_pass along with...
10925 (pass_data_forwprop): ...new pass_data instance and...
10926 (make_pass_forwprop): ...new function.
10927 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10928 struct to a subclass of gimple_opt_pass along with...
10929 (pass_data_tree_ifcombine): ...new pass_data instance and...
10930 (make_pass_tree_ifcombine): ...new function.
10931 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10932 subclass of gimple_opt_pass along with...
10933 (pass_data_ch): ...new pass_data instance and...
10934 (make_pass_ch): ...new function.
10935 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10936 subclass of gimple_opt_pass along with...
10937 (pass_data_tree_loop): ...new pass_data instance and...
10938 (make_pass_tree_loop): ...new function.
10939 (pass_tree_loop_init): Convert from a global struct to a subclass of
10940 gimple_opt_pass along with...
10941 (pass_data_tree_loop_init): ...new pass_data instance and...
10942 (make_pass_tree_loop_init): ...new function.
10943 (pass_lim): Convert from a global struct to a subclass of
10944 gimple_opt_pass along with...
10945 (pass_data_lim): ...new pass_data instance and...
10946 (make_pass_lim): ...new function.
10947 (pass_tree_unswitch): Convert from a global struct to a subclass of
10948 gimple_opt_pass along with...
10949 (pass_data_tree_unswitch): ...new pass_data instance and...
10950 (make_pass_tree_unswitch): ...new function.
10951 (pass_predcom): Convert from a global struct to a subclass of
10952 gimple_opt_pass along with...
10953 (pass_data_predcom): ...new pass_data instance and...
10954 (make_pass_predcom): ...new function.
10955 (pass_vectorize): Convert from a global struct to a subclass of
10956 gimple_opt_pass along with...
10957 (pass_data_vectorize): ...new pass_data instance and...
10958 (make_pass_vectorize): ...new function.
10959 (pass_graphite): Convert from a global struct to a subclass of
10960 gimple_opt_pass along with...
10961 (pass_data_graphite): ...new pass_data instance and...
10962 (make_pass_graphite): ...new function.
10963 (pass_graphite_transforms): Convert from a global struct to a subclass
10964 of gimple_opt_pass along with...
10965 (pass_data_graphite_transforms): ...new pass_data instance and...
10966 (make_pass_graphite_transforms): ...new function.
10967 (pass_check_data_deps): Convert from a global struct to a subclass of
10968 gimple_opt_pass along with...
10969 (pass_data_check_data_deps): ...new pass_data instance and...
10970 (make_pass_check_data_deps): ...new function.
10971 (pass_iv_canon): Convert from a global struct to a subclass of
10972 gimple_opt_pass along with...
10973 (pass_data_iv_canon): ...new pass_data instance and...
10974 (make_pass_iv_canon): ...new function.
10975 (pass_scev_cprop): Convert from a global struct to a subclass of
10976 gimple_opt_pass along with...
10977 (pass_data_scev_cprop): ...new pass_data instance and...
10978 (make_pass_scev_cprop): ...new function.
10979 (pass_record_bounds): Convert from a global struct to a subclass of
10980 gimple_opt_pass along with...
10981 (pass_data_record_bounds): ...new pass_data instance and...
10982 (make_pass_record_bounds): ...new function.
10983 (pass_complete_unroll): Convert from a global struct to a subclass of
10984 gimple_opt_pass along with...
10985 (pass_data_complete_unroll): ...new pass_data instance and...
10986 (make_pass_complete_unroll): ...new function.
10987 (pass_complete_unrolli): Convert from a global struct to a subclass of
10988 gimple_opt_pass along with...
10989 (pass_data_complete_unrolli): ...new pass_data instance and...
10990 (make_pass_complete_unrolli): ...new function.
10991 (pass_parallelize_loops): Convert from a global struct to a subclass
10992 of gimple_opt_pass along with...
10993 (pass_data_parallelize_loops): ...new pass_data instance and...
10994 (make_pass_parallelize_loops): ...new function.
10995 (pass_loop_prefetch): Convert from a global struct to a subclass of
10996 gimple_opt_pass along with...
10997 (pass_data_loop_prefetch): ...new pass_data instance and...
10998 (make_pass_loop_prefetch): ...new function.
10999 (pass_iv_optimize): Convert from a global struct to a subclass of
11000 gimple_opt_pass along with...
11001 (pass_data_iv_optimize): ...new pass_data instance and...
11002 (make_pass_iv_optimize): ...new function.
11003 (pass_tree_loop_done): Convert from a global struct to a subclass of
11004 gimple_opt_pass along with...
11005 (pass_data_tree_loop_done): ...new pass_data instance and...
11006 (make_pass_tree_loop_done): ...new function.
11007 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11008 struct to a subclass of gimple_opt_pass along with...
11009 (pass_data_cse_reciprocals): ...new pass_data instance and...
11010 (make_pass_cse_reciprocals): ...new function.
11011 (pass_cse_sincos): Convert from a global struct to a subclass of
11012 gimple_opt_pass along with...
11013 (pass_data_cse_sincos): ...new pass_data instance and...
11014 (make_pass_cse_sincos): ...new function.
11015 (pass_optimize_bswap): Convert from a global struct to a subclass of
11016 gimple_opt_pass along with...
11017 (pass_data_optimize_bswap): ...new pass_data instance and...
11018 (make_pass_optimize_bswap): ...new function.
11019 (pass_optimize_widening_mul): Convert from a global struct to a
11020 subclass of gimple_opt_pass along with...
11021 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11022 (make_pass_optimize_widening_mul): ...new function.
11023 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11024 subclass of gimple_opt_pass along with...
11025 (pass_data_phiopt): ...new pass_data instance and...
11026 (make_pass_phiopt): ...new function.
11027 (pass_cselim): Convert from a global struct to a subclass of
11028 gimple_opt_pass along with...
11029 (pass_data_cselim): ...new pass_data instance and...
11030 (make_pass_cselim): ...new function.
11031 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11032 subclass of gimple_opt_pass along with...
11033 (pass_data_phiprop): ...new pass_data instance and...
11034 (make_pass_phiprop): ...new function.
11035 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11036 subclass of gimple_opt_pass along with...
11037 (pass_data_pre): ...new pass_data instance and...
11038 (make_pass_pre): ...new function.
11039 (pass_fre): Convert from a global struct to a subclass of
11040 gimple_opt_pass along with...
11041 (pass_data_fre): ...new pass_data instance and...
11042 (make_pass_fre): ...new function.
11043 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11044 subclass of gimple_opt_pass along with...
11045 (pass_data_reassoc): ...new pass_data instance and...
11046 (make_pass_reassoc): ...new function.
11047 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11048 subclass of gimple_opt_pass along with...
11049 (pass_data_sink_code): ...new pass_data instance and...
11050 (make_pass_sink_code): ...new function.
11051 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11052 subclass of gimple_opt_pass along with...
11053 (pass_data_strlen): ...new pass_data instance and...
11054 (make_pass_strlen): ...new function.
11055 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11056 struct to a subclass of gimple_opt_pass along with...
11057 (pass_data_build_alias): ...new pass_data instance and...
11058 (make_pass_build_alias): ...new function.
11059 (pass_build_ealias): Convert from a global struct to a subclass of
11060 gimple_opt_pass along with...
11061 (pass_data_build_ealias): ...new pass_data instance and...
11062 (make_pass_build_ealias): ...new function.
11063 (pass_ipa_pta): Convert from a global struct to a subclass of
11064 simple_ipa_opt_pass along with...
11065 (pass_data_ipa_pta): ...new pass_data instance and...
11066 (make_pass_ipa_pta): ...new function.
11067 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11068 subclass of gimple_opt_pass along with...
11069 (pass_data_uncprop): ...new pass_data instance and...
11070 (make_pass_uncprop): ...new function.
11071 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11072 global struct to a subclass of gimple_opt_pass along with...
11073 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11074 (make_pass_late_warn_uninitialized): ...new function.
11075 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11076 to a subclass of gimple_opt_pass along with...
11077 (pass_data_init_datastructures): ...new pass_data instance and...
11078 (make_pass_init_datastructures): ...new function.
11079 (pass_early_warn_uninitialized): Convert from a global struct to a
11080 subclass of gimple_opt_pass along with...
11081 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11082 (make_pass_early_warn_uninitialized): ...new function.
11083 (pass_update_address_taken): Convert from a global struct to a
11084 subclass of gimple_opt_pass along with...
11085 (pass_data_update_address_taken): ...new pass_data instance and...
11086 (make_pass_update_address_taken): ...new function.
11087 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11088 struct to a subclass of gimple_opt_pass along with...
11089 (pass_data_release_ssa_names): ...new pass_data instance and...
11090 (make_pass_release_ssa_names): ...new function.
11091 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11092 subclass of gimple_opt_pass along with...
11093 (pass_data_stdarg): ...new pass_data instance and...
11094 (make_pass_stdarg): ...new function.
11095 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11096 global struct to a subclass of gimple_opt_pass along with...
11097 (pass_data_convert_switch): ...new pass_data instance and...
11098 (make_pass_convert_switch): ...new function.
11099 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11100 to a subclass of gimple_opt_pass along with...
11101 (pass_data_tail_recursion): ...new pass_data instance and...
11102 (make_pass_tail_recursion): ...new function.
11103 (pass_tail_calls): Convert from a global struct to a subclass of
11104 gimple_opt_pass along with...
11105 (pass_data_tail_calls): ...new pass_data instance and...
11106 (make_pass_tail_calls): ...new function.
11107 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11108 struct to a subclass of gimple_opt_pass along with...
11109 (pass_data_lower_vector): ...new pass_data instance and...
11110 (make_pass_lower_vector): ...new function.
11111 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11112 gimple_opt_pass along with...
11113 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11114 (make_pass_lower_vector_ssa): ...new function.
11115 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11116 to a subclass of gimple_opt_pass along with...
11117 (pass_data_slp_vectorize): ...new pass_data instance and...
11118 (make_pass_slp_vectorize): ...new function.
11119 (pass_ipa_increase_alignment): Convert from a global struct to a
11120 subclass of simple_ipa_opt_pass along with...
11121 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11122 (make_pass_ipa_increase_alignment): ...new function.
11123 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11124 gimple_opt_pass along with...
11125 (pass_data_vrp): ...new pass_data instance and...
11126 (make_pass_vrp): ...new function.
11127 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11128 subclass of simple_ipa_opt_pass along with...
11129 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11130 (make_pass_ipa_free_lang_data): ...new function.
11131 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11132 gimple_opt_pass along with...
11133 (pass_data_tsan): ...new pass_data instance and...
11134 (make_pass_tsan): ...new function.
11135 (pass_tsan_O0): Convert from a global struct to a subclass of
11136 gimple_opt_pass along with...
11137 (pass_data_tsan_O0): ...new pass_data instance and...
11138 (make_pass_tsan_O0): ...new function.
11139 * var-tracking.c (pass_variable_tracking): Convert from a global
11140 struct to a subclass of rtl_opt_pass along with...
11141 (pass_data_variable_tracking): ...new pass_data instance and...
11142 (make_pass_variable_tracking): ...new function.
11143 * web.c (pass_web): Convert from a global struct to a subclass of
11144 rtl_opt_pass along with...
11145 (pass_data_web): ...new pass_data instance and...
11146 (make_pass_web): ...new function.
11147 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11148 declaration with that of...
11149 (make_pass_mode_switch_use): ...new function.
11150 (pass_resolve_sw_modes): Replace declaration with that of...
11151 (make_pass_resolve_sw_modes): ...new function.
11152 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11153 from a global struct to a subclass of rtl_opt_pass along with...
11154 (pass_data_mode_switch_use): ...new pass_data instance and...
11155 (make_pass_mode_switch_use): ...new function.
11156 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11157 from a global struct to a subclass of rtl_opt_pass along with...
11158 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11159 (make_pass_resolve_sw_modes): ...new function.
11160 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11161 struct to a subclass of rtl_opt_pass along with...
11162 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11163 (make_pass_insert_vzeroupper): ...new function.
11164 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11165 global struct to a subclass of rtl_opt_pass along with...
11166 (pass_data_work_around_errata): ...new pass_data instance and...
11167 (make_pass_work_around_errata): ...new function.
11168 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11169 struct to a subclass of rtl_opt_pass along with...
11170 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11171 (make_pass_mips_machine_reorg2): ...new function.
11172
6a389ed5
DM		 111732013-08-05 David Malcolm <dmalcolm@redhat.com>
11174
11175 * passes.c (pass_manager::operator new): New.
11176
f7695dbf
DM		 111772013-08-05 David Malcolm <dmalcolm@redhat.com>
11178
11179 Handwritten part of conversion of passes to C++ classes.
11180
11181 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11182 (toplev.o): Add dep on PASS_MANAGER_H.
11183 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11184 of early local pases to reflect this moving from a global to a
11185 member of gcc::pass_manager.
11186 (cgraph_add_new_function): Likewise.
11187 * lto-cgraph.c (lto_output_node): Update for conversion of
11188 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11189 * passes.c (opt_pass::clone): New.
11190 (opt_pass::gate): New.
11191 (opt_pass::execute): New.
11192 (opt_pass::opt_pass): New.
11193 (pass_manager::execute_early_local_passes): New.
11194 (pass_manager::execute_pass_mode_switching): new.
11195 (finish_optimization_passes): Convert to...
11196 (pass_manager::finish_optimization_passes): ...this.
11197 (finish_optimization_passes): Update for conversion of passes to
11198 C++ classes.
11199 (register_dump_files_1): Use has_gate since we cannot portably
11200 check a vtable entry against NULL.
11201 (dump_one_pass): Likewise.
11202 (ipa_write_summaries_2): Likewise.
11203 (ipa_write_optimization_summaries_1): Likewise.
11204 (ipa_read_summaries_1): Likewise.
11205 (ipa_read_optimization_summaries_1): Likewise.
11206 (execute_ipa_stmt_fixups): Likewise.
11207 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11208 pass-creation functions rather than wiring up globals, and
11209 storing the results in fields of pass_manager generated using
11210 pass-instances.def.
11211 (pass_manager::dump_profile_report): Update for conversion of
11212 passes to C++ classes.
11213 (pass_manager::execute_ipa_summary_passes): Likewise.
11214 (execute_one_ipa_transform_pass): Likewise.
11215 (execute_one_pass): Use has_gate and has_execute since we cannot
11216 portably check a vtable entry against NULL.
11217 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11218 (pass_manager): Use pass-instances.def to add fields for the
11219 various pass instances.
11220 * toplev.c (finalize): Update for move of
11221 finish_optimization_passes to a method of gcc::pass_manager.
11222 * toplev.h (finish_optimization_passes): Move to method of class
11223 pass_manager.
11224 * tree-pass.h (struct pass_data): New.
03b0ee0a 11225 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM		 11226 (opt_pass::gate): Convert to virtual function.
11227 (opt_pass::~opt_pass): New.
11228 (opt_pass::clone): New.
11229 (opt_pass::execute): Convert to virtual function.
11230 (opt_pass::opt_pass): New.
11231 (opt_pass::ctxt_): new.
11232 (gimple_opt_pass): Convert to subclass of opt_pass.
11233 (gimple_opt_pass::gimple_opt_pass): New.
11234 (rtl_opt_pass): Convert to subclass of opt_pass.
11235 (rtl_opt_pass::rtl_opt_pass): New.
11236 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11237 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11238 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11239 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11240 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11241 invocation of pass_mode_switching to reflect this moving from a
11242 global to a member of gcc::pass_manager.
11243 (ix86_option_override): Rework how pass_insert_vzeroupper is
11244 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 11245 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 11246
d0b2f831
RE		 112472013-08-05 Richard Earnshaw <rearnsha@arm.com>
11248
11249 PR rtl-optimization/57708
11250 * recog.c (peep2_find_free_register): Validate all regs in a
11251 multi-reg mode.
11252
51a5c0c2
JH		 112532013-08-05 Jan Hubicka <jh@suse.cz>
11254
11255 PR lto/57602
03b0ee0a
UB		 11256 * cgraph.c (verify_cgraph_node): Accept local flags from other
11257 partitions.
51a5c0c2
JH		 11258 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11259 (function_and_variable_visibility): Likewise.
11260 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11261
65f0a120
GDR		 112622013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11263
11264 * graph.c (init_graph_slim_pretty_print): Remove.
11265 (print_graph_cfg): Do not call it. Use local pretty printer.
11266 (start_graph_dump): Likewise.
11267
b3f80694
GDR		 112682013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11269
11270 * gimple-pretty-print.c (buffer): Remove.
11271 (initialized): Likewise.
11272 (maybe_init_pretty_print): Likewise.
11273 (print_gimple_stmt): Do not call it. Use non-static local
11274 pretty_printer variable.
11275 (print_gimple_expr): Likewise.
11276 (print_gimple_seq): Likewise.
11277 (gimple_dump_bb): Likewise.
11278
11a877b3
GDR		 112792013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11280
11281 * asan.c (asan_pp): Remove.
11282 (asan_pp_initialized): Likewise.
11283 (asan_pp_initialize): Likewise.
11284 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11285 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11286 (asan_add_global): Likewise.
11287
b066401f
GDR		 112882013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11289
11290 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11291 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11292 * pretty-print.c (pp_formatted_text_data): Likewise.
11293 (pp_indent): Rename from pp_base_indent.
11294 (pp_format): Rename from pp_base_format.
11295 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11296 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11297 (pp_flush): Rename from pp_base_flush.
11298 (pp_set_line_maximum_length): Rename from
11299 pp_base_set_line_maximum_length.
11300 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11301 (pp_set_prefix): Rename from pp_base_set_prefix.
11302 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11303 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11304 (pp_append_text): Rename from pp_base_append_text.
11305 (pp_formatted_text): Rename from pp_base_formatted_text.
11306 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11307 (pp_remaining_character_count_for_line): Rename from
11308 pp_base_remaining_character_count_for_line.
11309 (pp_newline): Rename from pp_base_newline.
11310 (pp_character): Rename from pp_base_character.
11311 (pp_string): Rename from pp_base_string.
11312 (pp_maybe_space): Rename from pp_base_maybe_space.
11313 * asan.c (asan_pp_string): Adjust.
11314 (asan_emit_stack_protection): Likewise.
11315 (asan_add_global): Likewise.
11316 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11317 * tree-mudflap.c (mf_varname_tree): Likewise.
11318 * tree-pretty-print.c (pp_tree_identifier): Rename from
11319 pp_base_tree_identifier.
11320 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11321 Declare as function.
11322
137a1a27
GDR		 113232013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11324
11325 * pretty-print.h (pp_bar_bar): New.
11326 (pp_ampersand_ampersand): Likewise.
11327 (pp_less_equal): Likewise.
11328 (pp_greater_equal): Likewise.
11329 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11330 printer functions instead of pp_string or operators and punctuators.
11331 (dump_gimple_call): Likewise.
11332 (dump_gimple_omp_for): Likewise.
11333 (dump_gimple_transaction): Likewise.
11334 (dump_gimple_phi): Likewise.
11335 (pp_gimple_stmt_1): Likewise.
11336 * sched-vis.c (print_insn): Likewise.
11337 * tree-mudflap.c (mf_varname_tree): Likewise.
11338 * tree-pretty-print.c (dump_block_node): Likewise.
11339 (dump_generic_node): Likewise.
11340
815effe1
JH		 113412013-08-02 Jan Hubicka <jh@suse.cz>
11342
11343 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11344 boundaries.
11345 * lto-streamer-out.c (tree_is_indexable): Results decls and
11346 parm decls are not indexable.
11347 (DFS_write_tree_body): Do not follow args and results.
11348 (hash_tree): Likewise.
11349 (output_functions): Rearrange so struct function is needed
11350 only when real body is output; be able to also ouptut abstract
11351 functions; output DECL_ARGUMENTS and DECL_RESULT.
11352 (lto_output): When not in WPA, ale store abstract functions.
11353 (write_symbol): Do not care about RESULT_DECL.
11354 (output_symbol_p): Handle correctly sbtract decls.
11355 * lto-streamer-in.c (input_function): Rearrange so struct
11356 function can be NULL at entry; allow streaming of
11357 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11358 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11359 sanity check during LTO.
11360 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11361 RESULT_DECl and DECL_ARGUMENTS.
11362 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11363 Likewise.
11364
07838b13
GDR		 113652013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11366
11367 * pretty-print.h (pp_underscore): New.
11368 (pp_comma): Tidy.
11369 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11370 printer functions instead of pp_character.
11371 (dump_binary_rhs): Likewise.
11372 (dump_ternary_rhs): Likewise.
11373 (dump_gimple_call_args): Likewise.
11374 (pp_points_to_solution): Likewise.
11375 (dump_gimple_call): Likewise.
11376 (dump_gimple_switch): Likewise.
11377 (dump_gimple_cond): Likewise.
11378 (dump_gimple_bind): Likewise.
11379 (dump_gimple_try): Likewise.
11380 (dump_gimple_omp_for): Likewise.
11381 (dump_gimple_omp_continue): Likewise.
11382 (dump_gimple_omp_single): Likewise.
11383 (dump_gimple_omp_sections): Likewise.
11384 (dump_gimple_omp_block): Likewise.
11385 (dump_gimple_omp_critical): Likewise.
11386 (dump_gimple_transaction): Likewise.
11387 (dump_gimple_asm): Likewise.
11388 (dump_gimple_phi): Likewise.
11389 (dump_gimple_omp_parallel): Likewise.
11390 (dump_gimple_omp_task): Likewise.
11391 (dump_gimple_omp_atomic_load): Likewise.
11392 (dump_gimple_omp_atomic_store): Likewise.
11393 (dump_gimple_mem_ops): Likewise.
11394 (pp_gimple_stmt_1): Likewise.
11395 (pp_cfg_jump): Likewise.
11396 (dump_implicit_edges): Likewise.
11397 (gimple_dump_bb_for_graph): Likewise.
11398 * graph.c (draw_cfg_node): Likewise.
11399 * langhooks.c (lhd_print_error_function): Likewise.
11400 * sched-vis.c (print_exp): Likewise.
11401 (print_value): Likewise.
11402 (print_pattern): Likewise.
11403 (print_insn): Likewise.
11404 (rtl_dump_bb_for_graph): Likewise.
11405 * tree-pretty-print.c (dump_function_declaration): Likewise.
11406 (dump_array_domain): Likewise.
11407 (dump_omp_clause): Likewise.
11408 (dump_location): Likewise.
11409 (dump_generic_node): Likewise.
11410 (print_struct_decl): Likewise.
11411 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11412
bb0d2039
BS		 114132013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11414
11415 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11416 candidate table when replacing a candidate statement.
11417 (replace_rhs_if_not_dup): Likewise.
11418 (replace_one_candidate): Likewise.
11419
bc0ec027 114202013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 11421 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH		 11422
11423 * cgraphunit.c (add_new_function): Fix logic when adding from
11424 late IPA pass.
11425 (assemble_thunk): Rename to ...
03b0ee0a
UB		 11426 (expand_thunk); .. this one; export; get it working with
11427 general functions; make produced gimple valid.
bc0ec027
JH		 11428 * cgraph.h (expand_thunk): Declare.
11429
0e8853ee
JH		 114302013-08-02 Jan Hubicka <jh@suse.cz>
11431
03b0ee0a
UB		 11432 * ipa-cp.c (gather_context_independent_values): Use
11433 ipa_get_param_move_cost.
11434 (get_replacement_map): Remove PARAM; move parameter folding
11435 into tree-inline.c
0e8853ee
JH		 11436 (create_specialized_node): Update.
11437 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11438 assert that we have gimple body; update move_cost.
11439 (count_formal_params): Assert that we have gimple body.
11440 (ipa_dump_param): New function.
11441 (ipa_alloc_node_params): Break out from ...
11442 (ipa_initialize_node_params): ... here.
11443 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11444 (ipa_write_node_info): Stream move costs.
11445 (ipa_read_node_info): Read move costs.
11446 (ipa_update_after_lto_read): Do not recompute node params.
11447 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11448 (ipa_get_param): Check we are not in WPA.
11449 (ipa_get_param_move_cost): New.
11450 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11451 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11452 parm numbers to be present.
11453
ecee672b
VM		 114542013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11455
11456 PR rtl-optimization/58048
11457 * lra-constraints.c (process_alt_operands): Don't check asm
11458 operand on register.
11459
13ec0527
EB		 114602013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11461
11462 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11463 the implied StoreLoad barrier for atomic operations if before.
11464
af15184a
JH		 114652013-08-02 Jan Hubicka <jh@suse.cz>
11466 Martin Liska <marxin.liska@gmail.com>
11467
03b0ee0a
UB		 11468 * cgraph.c (cgraph_function_body_availability): Do not check
11469 cgraph flags.
af15184a
JH		 11470 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11471 symtab_node_availability): Declare.
11472 * ipa.c (can_replace_by_local_alias): New.
11473 (function_and_variable_visibility): Use it.
03b0ee0a
UB		 11474 * symtab.c (symtab_for_node_and_aliases,
11475 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 11476
01e54ef8
VM		 114772013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11478
11479 PR rtl-optimization/57963
03b0ee0a 11480 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM		 11481 (lra_constraints): Use them.
11482
e2fd7ca7
SN		 114832013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11484
03b0ee0a
UB		 11485 * config/arm/types.md (define_attr "type"): Add "load_acq"
11486 and "store_rel".
e2fd7ca7
SN		 11487 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11488 changes.
11489 (cortex_a53_store1): Likewise.
11490
0dcc26c3
JH		 114912013-08-01 Jan Hubicka <jh@suse.cz>
11492
03b0ee0a
UB		 11493 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11494 partitions are not needed.
0dcc26c3 11495
0bd72901
UB		 114962013-08-01 Uros Bizjak <ubizjak@gmail.com>
11497
11498 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11499 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11500 MAYBE_NON_Q_CLASS_P where appropriate.
11501
12123452
JH		 115022013-08-01 Jan Hubicka <jh@suse.cz>
11503
11504 * cgraph.h (release_function_body): Declare.
11505 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11506 of unused delcarations.
11507
ce852f9c
KT		 115082013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11509
03b0ee0a
UB		 11510 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11511 RTL form when subtracting a constant.
ce852f9c 11512
c743b246
KT		 115132013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11514
11515 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11516 Generate canonical plus rtx with negated immediate instead of minus
11517 where appropriate.
11518 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11519
c0c123ef
JH		 115202013-08-01 Jan Hubicka <jh@suse.cz>
11521
11522 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11523 (cgraph_release_function_body): Likewise.
11524 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11525 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11526 to used_as_abstract_origin.
11527 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11528 symbols used as abstract origins.
11529 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB		 11530 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11531 used_as_abstract_origin.
c0c123ef 11532 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB		 11533 used_as_abstract_origin; be ready for DECL_RESULT and
11534 DECL_ARGUMENTS to be NULL.
c0c123ef 11535
0bd72901
UB		 11536 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11537 for abstract functions.
11538 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11539 real symbols.
c0c123ef 11540
1f029433
JH		 115412013-08-01 Jan Hubicka <jh@suse.cz>
11542
11543 * profile.c (compute_value_histograms): Fix thinko.
11544
c451f4d6
SN		 115452013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11546
11547 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11548 aarch-common-protos.h to extra_headers.
11549 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11550 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11551 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11552
34b512f5
SN		 115532013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11554
11555 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11556 Include "../arm/types.md". Define "type" attribute for all patterns.
11557 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11558 attribute changes.
11559
d86e633a
MM		 115602013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11561
11562 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11563 to support power8 load fusion.
11564 (fusion_gpr_mem_load): Likewise.
11565
11566 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11567
11568 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11569 declarations for power8 load fusion.
11570 (emit_fusion_gpr_load): Likewise.
11571
11572 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11573 tuning for power8, turn on fusion mode by default. Turn on sign
11574 extending fusion mode if normal fusion mode is on, and we are at
11575 -O2 or -O3.
11576 (fusion_gpr_load_p): New function, return true if we can fuse an
11577 addis instruction with a dependent load to a GPR.
11578 (emit_fusion_gpr_load): Emit the instructions for power8 load
11579 fusion to GPRs.
11580
0bd72901 11581 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM		 11582 (VSX load fusion peepholes): New peepholes to fuse together an
11583 addi instruction with a VSX load instruction.
11584
11585 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11586 peepholes to fuse an addis instruction with a load to a GPR base
11587 register. If we are supporting sign extending fusions, convert
11588 sign extending loads to zero extending loads and add an explicit
11589 sign extension.
11590
c3f35647
SN		 115912013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11592
11593 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11594 aarch-common-protos.h to extra_headers.
11595 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11596 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11597 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11598 (arm_no_early_alu_shift_dep): Likewise.
11599 (arm_no_early_alu_shift_value_dep): Likewise.
11600 (arm_no_early_mul_dep): Likewise.
11601 (arm_no_early_store_addr_dep): Likewise.
11602 (arm_mac_accumulator_is_mul_result): Likewise.
11603 (arm_mac_accumulator_is_result): Likewise.
c3f35647 11604 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB		 11605 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11606 here to ...
c3f35647 11607 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11608 (arm_no_early_alu_shift_dep): Likewise.
11609 (arm_no_early_alu_shift_value_dep): Likewise.
11610 (arm_no_early_mul_dep): Likewise.
11611 (arm_no_early_store_addr_dep): Likewise.
11612 (arm_mac_accumulator_is_mul_result): Likewise.
11613 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN		 11614 * config/arm/aarch-common-protos.h: ... here. New file.
11615 * config/arm/t-arm (aarch-common.o): Define.
11616
116172013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11618
11619 * config/arm/arm.md: Include new file "types.md".
11620 (define_attr "type"): Move from here to ...
11621 (define_attr "mul32"): Likewise.
11622 (define_attr "mul64"): Likewise.
11623 * config/arm/types.md: ... here. New file.
11624
228c1313
SH		 116252013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11626
11627 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11628 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11629
d6e9046d 116302013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG		 11631
11632 * gen-pass-instances.awk: Fix offset of substr().
11633
a167b052
DM		 116342013-07-31 David Malcolm <dmalcolm@redhat.com>
11635
11636 * Makefile.in (pass-instances.def): New.
11637 (passes.o): Replace dependency on passes.def with one on
11638 pass-instances.def
11639
11640 * gen-pass-instances.awk: New.
11641
11642 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11643 rather than passes.def, updating local definition of NEXT_PASS
11644 macro to add an extra NUM parameter (currently unused).
11645
315f8c0e
DM		 116462013-07-30 David Malcolm <dmalcolm@redhat.com>
11647
11648 * Makefile.in (PASS_MANAGER_H): New.
11649 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11650 (passes.o): Likewise.
11651 (statistics.o): Likewise.
11652 (cgraphunit.o): Likewise.
11653 (context.o): Depend on PASS_MANAGER_H.
11654
11655 * pass_manager.h: New.
11656
11657 * cgraphunit.c (cgraph_add_new_function): Update for moves
11658 of globals to fields of pass_manager.
11659 (analyze_function): Likewise.
11660 (expand_function): Likewise.
11661 (ipa_passes): Likewise.
11662 (compile): Likewise.
11663
11664 * context.c (context::context): New.
11665 * context.h (context::context): New.
11666 (context::get_passes): New.
11667 (context::passes_): New.
11668
11669 * lto-cgraph.c (input_node): Update for moves of globals to
11670 fields of pass_manager.
11671
11672 * passes.c (all_passes): Remove, in favor of a field of the
11673 same name within the new class pass_manager.
11674 (all_small_ipa_passes): Likewise.
11675 (all_lowering_passes): Likewise.
11676 (all_regular_ipa_passes): Likewise.
11677 (all_late_ipa_passes): Likewise.
11678 (all_lto_gen_passes): Likewise.
11679 (passes_by_id): Likewise.
11680 (passes_by_id_size): Likewise.
11681 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11682 the new class pass_manager.
11683 (set_pass_for_id): Convert to...
11684 (pass_manager::set_pass_for_id): ...method.
11685 (get_pass_for_id): Convert to...
11686 (pass_manager::get_pass_for_id): ...method.
11687 (register_one_dump_file): Move body of implementation into...
11688 (pass_manager::register_one_dump_file): ...here.
11689 (register_dump_files_1): Convert to...
11690 (pass_manager::register_dump_files_1): ...method.
11691 (register_dump_files): Convert to...
11692 (pass_manager::register_dump_files): ...method.
11693 (create_pass_tab): Update for moves of globals to fields of
11694 pass_manager.
11695 (dump_passes): Move body of implementation into...
11696 (pass_manager::dump_passes): ...here.
11697 (register_pass): Move body of implementation into...
11698 (pass_manager::register_pass): ...here.
11699 (init_optimization_passes): Convert into...
11700 (pass_manager::pass_manager): ...constructor for new
11701 pass_manager class, and initialize the pass_lists array.
11702 (check_profile_consistency): Update for moves of globals to
11703 fields of pass_manager.
11704 (dump_profile_report): Move body of implementation into...
11705 (pass_manager::dump_profile_report): ...here.
11706 (ipa_write_summaries_1): Update for moves of pass lists from
11707 being globals to fields of pass_manager.
11708 (ipa_write_optimization_summaries): Likewise.
11709 (ipa_read_summaries): Likewise.
11710 (ipa_read_optimization_summaries): Likewise.
11711 (execute_all_ipa_stmt_fixups): Likewise.
11712
11713 * statistics.c (statistics_fini): Update for moves of globals to
11714 fields of pass_manager.
11715
11716 * toplev.c (general_init): Replace call to
11717 init_optimization_passes with construction of the pass_manager
11718 instance.
11719
11720 * tree-pass.h (all_passes): Remove, in favor of a field of the
11721 same name within the new class pass_manager.
11722 (all_small_ipa_passes): Likewise.
11723 (all_lowering_passes): Likewise.
11724 (all_regular_ipa_passes): Likewise.
11725 (all_lto_gen_passes): Likewise.
11726 (all_late_ipa_passes): Likewise.
11727 (passes_by_id): Likewise.
11728 (passes_by_id_size): Likewise.
11729 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11730 the new class pass_manager.
11731 (get_pass_for_id): Remove.
11732
02cab1c1
RE		 117332013-07-30 Richard Earnshaw <rearnsha@arm.com>
11734
11735 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11736 configs.
11737
b2d3f886
RE		 117382013-07-30 Richard Earnshaw <rearnsha@arm.com>
11739
11740 * arm.md (mulhi3): New expand pattern.
11741
605e86fa
JH		 117422013-07-30 Jan Hubicka <jh@suse.cz>
11743 Martin Liska <marxin.liska@gmail.com>
11744
11745 * profile.c (compute_value_histograms): Do not ICE when
11746 there is mismatch only on some counters.
11747
14a87636
ZC		 117482013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11749
11750 PR rtl-optimization/57637
11751 * function.c (move_insn_for_shrink_wrap): Also check the
11752 GEN set of the LIVE problem for the liveness analysis
11753 if it exists, otherwise give up.
11754
0100cd3f
BS		 117552013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11756
11757 PR tree-optimization/57993
11758 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11759 replaced statement in the candidate table.
11760 (phi_add_costs): Return infinite cost when the hidden basis does
11761 not dominate all phis on which the candidate is dependent.
11762 (replace_one_candidate): Record replaced statement in the
11763 candidate table.
11764
f0d811f7
JR		 117652013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11766
11767 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11768 (ashlv2si3): New expander.
11769 (*ashlv2si3_i): New define_insn_and_split.
11770 * predicates.md (float_operation): Allow patterns with three
11771 basic sub-patterns.
11772
bba33211
JR		 11773 PR rtl-optimization/58021
11774 * mode-switching.c (create_pre_exit): Always split off preceding
11775 insns if we are not at the basic block head.
11776
ff3f3951
MR		 117772013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11778
11779 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11780 (UCLIBC_DYNAMIC_LINKER): New macro.
11781 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11782 `-mnan=2008'.
11783 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11784 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11785 `-mnan=2008'.
11786 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11787 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11788 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11789 for SF and DF modes. Use ieee_quad_format for TF mode.
11790 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11791 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11792 (mips_option_override): Handle `-mnan=legacy'.
11793 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11794 `-mabs=2008' and `-mnan=2008'.
11795 (OPTION_DEFAULT_SPECS): Add "nan" default.
11796 (ASM_SPEC): Handle `-mnan='.
11797 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11798 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11799 comment accordingly.
11800 (neg<mode>2): Likewise.
11801 * config/mips/mips.opt (mabs, mnan): New options.
11802 * doc/install.texi (Configuration): Document `--with-nan=' option.
11803 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11804 `-mnan=' options.
11805 (MIPS Options): Document them.
11806 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11807 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11808 * configure: Regenerate.
11809 * config.in: Regenerate.
11810
576e2f37
UB		 118112013-07-29 Uros Bizjak <ubizjak@gmail.com>
11812
11813 * config/i386/i386.md (float post-reload splitters): Do not check
11814 for subregs of SSE registers.
11815
118162013-07-29 Uros Bizjak <ubizjak@gmail.com>
11817 H.J. Lu <hongjiu.lu@intel.com>
11818
11819 PR target/57954
11820 PR target/57988
11821 * config/i386/i386.md (post-reload splitter
11822 to avoid partial SSE reg dependency stalls): New pattern.
11823
a71f0749
DV		 118242013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11825
11826 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11827
4c97f1cc
JR		 118282013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11829
11830 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11831 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11832
b0c733d8
CC		 118332013-07-26 Cary Coutant <ccoutant@google.com>
11834
11835 * dwarf2out.c (die_checksum_ordered): Don't include template
11836 instantiations in signature.
11837 (is_template_parameter): New function.
11838 (is_template_instantiation): New function.
11839 (generate_skeleton_bottom_up): Don't include template instantiations
11840 in type unit DIE.
11841 (generate_skeleton): Likewise.
11842 (break_out_comdat_types): Move recursive call to break out nested
11843 types earlier.
11844 (prune_unused_types_mark_generic_parms_dies): Call
11845 is_template_parameter.
11846
040d8a1c
IB		 118472013-07-26 Ian Bolton <ian.bolton@arm.com>
11848
11849 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11850 uses vector registers.
89fdc743 11851 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 11852
85bd4ac6 118532013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 11854 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT		 11855
11856 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 11857 where appropriate.
85bd4ac6
KT		 11858 * config/arm/arm.md (movcond_addsi): New splitter.
11859
b6051207
SA		 118602013-07-25 Sterling Augustine <saugustine@google.com>
11861
11862 * dwarf2out.c (size_of_pubnames): Move code to...
11863 (include_pubname_in_output): ...here. New.
11864 (want_pubnames): Rearrange.
11865 (output_pubnames): Call include_pubname_in_output. Move assertion.
11866
41a7c215
CM		 118672013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11868
11869 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11870
118712013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11872
11873 PR target/38836
11874 * doc/extend.texi: Remove obsolete builtins. Fix
11875 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11876
49bde175
JH		 118772013-07-25 Jan Hubicka <jh@suse.cz>
11878
11879 * cgraph.c (release_function_body): Break out from ...
11880 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11881 and DECL_ARGUMENTS.
11882 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11883 old_tree in the map.
11884 (create_specialized_node): Update.
11885 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11886 into index.
41a7c215
CM		 11887 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11888 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH		 11889 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11890 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11891 DECL_RESULT.
11892
4870352d
KT		 118932013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11894
11895 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11896 addsi3_carryin_alt2_<optab>): Correct output template.
11897
1d6aee1c
KT		 118982013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11899
11900 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11901 Adjust for arm_restrict_it.
11902 Remove trailing whitespace.
11903
c7cafd75
MK		 119042013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11905
017d38f5
MK		 11906 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11907 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11908
c7cafd75
MK		 11909 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11910
00b7527b
VM		 119112013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11912
11913 PR rtl-optimization/57960
11914 * lra-constraints.c (process_alt_operands): Use the right mode
11915 when checking strict_low.
11916
4f63dfc6
JH		 119172013-07-25 Jan Hubicka <jh@suse.cz>
11918
11919 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11920 * cgraph.c (cgraph_remove_node): Do not release function body
11921 when in cgraph streaming.
11922 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM		 11923 in other partitions are not considered reachable; fix handling of
11924 clones.
4f63dfc6 11925
e6dcfa18
RR		 119262013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11927
11928 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11929
0b93d3b6
RR		 119302013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11931
11932 PR target/19599
11933 PR target/57731
bb80c2eb 11934 PR target/57837
e6dcfa18 11935 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 11936 Ss with US. Adjust output for v5 and v4t.
41a7c215 11937 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 11938
0b93d3b6
RR		 11939 * config/arm/constraints.md ("Ss"): Rename to US.
11940
690688b3
TG		 119412013-07-25 Terry Guo <terry.guo@arm.com>
11942
11943 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11944 shift_add/shift_sub0/shift_sub1 RTXs.
11945
a4ad093b 119462013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 11947 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS		 11948
11949 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11950 (altivec_vpks<VI_char>ss): Likewise.
11951 (altivec_vpks<VI_char>us): Likewise.
11952 (altivec_vpku<VI_char>us): Likewise.
11953 (altivec_vpku<VI_char>um): Likewise.
11954
8c5005ce
DM		 119552013-07-24 David Malcolm <dmalcolm@redhat.com>
11956
11957 Introduce context class.
11958
11959 * Makefile.in (CONTEXT_H): New.
11960 (OBJS): Add context.o.
11961 (toplev.o): Add CONTEXT_H to dependencies.
11962 (context.o): New.
11963
41a7c215 11964 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM		 11965
11966 * context.c: New.
11967
11968 * context.h: New.
11969
ce4a9422
JR		 119702013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11971
11972 PR rtl-optimization/57968
11973 * mode-switching.c (create_pre_exit): Allow instructions that
11974 don't set a return register to need a non-exit mode.
11975
493f4c9e 119762013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11977 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 11978
b5860fd3
WS		 11979 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11980 operands to vperm for little endian.
11981 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11982 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 11983
73792b92 119842013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11985 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS		 11986
11987 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11988 two operands for little-endian.
11989
14c21302
SE		 119902013-07-23 Steve Ellcey <sellcey@mips.com>
11991
11992 * config/mips/mips.c (mips_case_values_threshold): New.
11993 (TARGET_CASE_VALUES_THRESHOLD): Define.
11994
a6056198 119952013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11996 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS		 11997
11998 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11999 selection of field for vector splat in little endian mode.
12000
dd7a40e1
MM		 120012013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12002
12003 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12004 expanders to rs6000.md.
12005 (ior<mode>3): Likewise.
12006 (and<mode>3): Likewise.
12007 (one_cmpl<mode>2): Likewise.
12008 (nor<mode>3): Likewise.
12009 (andc<mode>3): Likewise.
12010 (eqv<mode>3): Likewise.
12011 (nand<mode>3): Likewise.
12012 (orc<mode>3): Likewise.
12013
12014 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12015 declaration.
12016
12017 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12018 to split multi-word logical operations.
12019 (rs6000_split_logical_di): Likewise.
12020 (rs6000_split_logical): Likewise.
12021
12022 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12023 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12024 and allow TImode operations in 32-bit.
12025 (vsx_and<mode>3_64bit): Likewise.
12026 (vsx_ior<mode>3_32bit): Likewise.
12027 (vsx_ior<mode>3_64bit): Likewise.
12028 (vsx_xor<mode>3_32bit): Likewise.
12029 (vsx_xor<mode>3_64bit): Likewise.
12030 (vsx_one_cmpl<mode>2_32bit): Likewise.
12031 (vsx_one_cmpl<mode>2_64bit): Likewise.
12032 (vsx_nor<mode>3_32bit): Likewise.
12033 (vsx_nor<mode>3_64bit): Likewise.
12034 (vsx_andc<mode>3_32bit): Likewise.
12035 (vsx_andc<mode>3_64bit): Likewise.
12036 (vsx_eqv<mode>3_32bit): Likewise.
12037 (vsx_eqv<mode>3_64bit): Likewise.
12038 (vsx_nand<mode>3_32bit): Likewise.
12039 (vsx_nand<mode>3_64bit): Likewise.
12040 (vsx_orc<mode>3_32bit): Likewise.
12041 (vsx_orc<mode>3_64bit): Likewise.
12042
12043 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12044 logical types in GPRs.
12045
12046 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12047 logical insns to rs6000.md, and allow TImode operations in
12048 32-bit.
12049 (altivec_ior<mode>3): Likewise.
12050 (altivec_xor<mode>3): Likewise.
12051 (altivec_one_cmpl<mode>2): Likewise.
12052 (altivec_nor<mode>3): Likewise.
12053 (altivec_andc<mode>3): Likewise.
12054
12055 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12056 attributes for moving the 128-bit logical operations into
12057 rs6000.md.
12058 (BOOL_REGS_OUTPUT): Likewise.
12059 (BOOL_REGS_OP1): Likewise.
12060 (BOOL_REGS_OP2): Likewise.
12061 (BOOL_REGS_UNARY): Likewise.
12062 (BOOL_REGS_AND_CR0): Likewise.
12063 (one_cmpl<mode>2): Add support for DI logical operations on
12064 32-bit, splitting the operations to 32-bit.
12065 (anddi3): Likewise.
12066 (iordi3): Likewise.
12067 (xordi3): Likewise.
12068 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12069 changes to combine the 32/64-bit code, allow logical operations on
12070 TI mode in 32-bit, and to use similar match_operator patterns like
12071 scalar mode uses. Combine the Altivec and VSX code for logical
12072 operations, and move it here.
12073 (ior<mode>3, 128-bit types): Likewise.
12074 (xor<mode>3, 128-bit types): Likewise.
12075 (one_cmpl<mode>3, 128-bit types): Likewise.
12076 (nor<mode>3, 128-bit types): Likewise.
12077 (andc<mode>3, 128-bit types): Likewise.
12078 (eqv<mode>3, 128-bit types): Likewise.
12079 (nand<mode>3, 128-bit types): Likewise.
12080 (orc<mode>3, 128-bit types): Likewise.
12081 (and<mode>3_internal): Likewise.
12082 (bool<mode>3_internal): Likewise.
12083 (boolc<mode>3_internal1): Likewise.
12084 (boolc<mode>3_internal2): Likewise.
12085 (boolcc<mode>3_internal1): Likewise.
12086 (boolcc<mode>3_internal2): Likewise.
12087 (eqv<mode>3_internal1): Likewise.
12088 (eqv<mode>3_internal2): Likewise.
12089 (one_cmpl1<mode>3_internal): Likewise.
12090
12211b99 120912013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH		 12092
12093 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12094 Rename flag_stack_usage to flag_stack_usage_info.
12095
12211b99 120962013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 12097
a5f257fd
ME		 12098 * config/microblaze/sync.md: New file.
12099 * config/microblaze/microblaze.md: Include sync.md
12100 * config/microblaze/microblaze.c: Add print_operand 'y'.
12101 * config/microblaze/constraints.md: Add memory_contraint
12102 'Q' which is a single register.
450b0ebb 12103
49fba14f
EB		 121042013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12105
12106 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12107
c7b3b99f
PCC		 121082013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12109
12110 * reload.c (find_reloads): Exit loop once we find this operand
12111 cannot be reloaded somehow for this alternative.
12112
12113 * reload.c (find_reloads): Exit loop once we find a hard register.
12114
12115 * rtlanal.c (computed_jump_p): Exit loop once we find label
12116 reference is used.
12117
12118 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12119
12120 * cfgloopmanip.c (remove_path): Exit loop after setting
12121 irred_invalidated.
12122
12123 * gensupport.c (subst_dup): Avoid loop if code is not
12124 MATCH_DUP nor MATCH_OP_DUP.
12125
c8fbf1fa
NBJ		 121262013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12127
12128 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12129
9259db42
YZ		 121302013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12131
12132 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12133 true for SP_REGNUM if mode == ptr_mode.
12134 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12135 with value R0_REGNUM + 31.
12136
00edcfbe
YZ		 121372013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12138
12139 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12140 pad pointer-typed argument downward.
12141
43be9a95
YZ		 121422013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12143
12144 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12145 and __ILP32__ when the ILP32 model is in use.
12146
28514dda
YZ		 121472013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12148
12149 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12150 (aarch64_load_symref_appropriately): In the case of
12151 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12152 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12153 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12154 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12155 if 'mode' doesn't equal to 'ptr_mode'.
12156 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12157 'vcall_offset'; change to call aarch64_emit_move differently depending
12158 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12159 to calculate the upper bound of 'vcall_offset'.
12160 (aarch64_cannot_force_const_mem): Change to also return true if
12161 mode != ptr_mode.
12162 (aarch64_legitimize_reload_address): In the case of large
12163 displacements, add new local variable 'xmode' and an assertion
12164 based on it; change to use 'xmode' to generate the new rtx and
12165 reload.
12166 (aarch64_asm_trampoline_template): Change to generate the template
12167 differently depending on TARGET_ILP32 or not; change to use
12168 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12169 (aarch64_trampoline_size): Removed.
12170 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12171 and replace immediate literals with it. Change to use 'ptr_mode'
12172 instead of 'DImode' and call convert_memory_address if the mode
12173 of 'fnaddr' doesn't equal to 'ptr_mode'.
12174 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12175 to output symbol.
12176 (aarch64_elf_asm_destructor): Likewise.
12177 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12178 on TARGET_ILP32 instead of aarch64_trampoline_size.
12179 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12180 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12181 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12182 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12183 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12184 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12185 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12186 depending on the value of 'mode'.
12187 (add_losym_<mode>): New.
12188 (ldr_got_small_<mode>): New, based on ldr_got_small.
12189 (ldr_got_small): Remove.
12190 (ldr_got_small_sidi): New.
12191 * config/aarch64/iterators.md (P): New.
12192 (PTR): Change to 'ptr_mode' in the condition.
12193
17a819cb
YZ		 121942013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12195
12196 * config.gcc (aarch64*-*-*): Support --with-abi.
12197 (aarch64*-*-elf): Support --with-multilib-list.
12198 (aarch64*-*-linux*): Likewise.
12199 (supported_defaults): Add abi to aarch64*-*-*.
12200 * configure.ac: Mention AArch64 for --with-multilib-list.
12201 * configure: Re-generated.
12202 * config/aarch64/biarchilp32.h: New file.
12203 * config/aarch64/biarchlp64.h: New file.
12204 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12205 (ABI_SPEC): Ditto.
12206 (MULTILIB_DEFAULTS): Ditto.
12207 (DRIVER_SELF_SPECS): Ditto.
12208 (ASM_SPEC): Update to also substitute -mabi.
12209 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12210 file whose name depends on -mabi= and -mbig-endian.
12211 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12212 TARGET_ILP32.
12213 (POINTER_SIZE): New define.
12214 (POINTERS_EXTEND_UNSIGNED): Ditto.
12215 (enum aarch64_abi_type): New enumeration tag.
12216 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12217 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12218 (TARGET_ILP32): New define.
12219 * config/aarch64/aarch64.opt (mabi): New.
12220 (aarch64_abi): New.
12221 (ilp32, lp64): New values for -mabi.
12222 * config/aarch64/t-aarch64 (comma): New define.
12223 (MULTILIB_OPTIONS): Ditto.
12224 (MULTILIB_DIRNAMES): Ditto.
12225 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12226 * doc/invoke.texi: Document -mabi for AArch64.
12227
9f9cbdce
GJL		 122282013-07-23 Georg-Johann Lay <avr@gjlay.de>
12229
12230 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12231
7061977b 122322013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 12233 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS		 12234
12235 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12236 endianness when selecting field to splat.
12237
75a41878
EC		 122382013-07-22 Eric Christopher <echristo@gmail.com>
12239
12240 * dwarf2out.c (die_odr_checksum): New function to use
12241 CHECKSUM_ macros and ULEB128 for DIE tag.
12242 (generate_type_signature): Use.
41a7c215 12243
38ae58ca
EB		 122442013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12245
12246 * config.gcc (sparc*-*-*): Accept leon3 processor.
12247 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12248 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12249 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12250 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12251 (mfix-ut699): Adjust comment.
12252 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12253 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12254 (CPP_CPU_SPEC): Likewise.
12255 (ASM_CPU_SPEC): Likewise.
12256 * config/sparc/sparc.c (leon3_cost): New constant.
12257 (sparc_option_override): Add leon3 support.
12258 (mem_ref): New function.
12259 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12260 (sparc_do_work_around_errata): Look into the instruction in the delay
12261 slot and adjust accordingly. Add fix for the data cache nullify issues
12262 of the UT699. Change insertion position for the NOP.
12263 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12264 (leon3_load): New reservation.
12265 (leon_store): Bump latency to 2.
12266 (grfpu): New automaton.
12267 (grfpu_alu): New unit.
12268 (grfpu_ds): Likewise.
12269 (leon_fp_alu): Adjust.
12270 (leon_fp_mult): Delete.
12271 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12272 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12273 * config/sparc/sparc.md (cpu): Add leon3.
12274 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12275 (swapsi): Likewise.
12276 (atomic_test_and_set): Likewise.
12277 (ldstub): Likewise.
12278
04dfc6df
JU		 122792013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12280
12281 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12282 default for R5900 targets.
12283 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12284 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12285 * config/mips/mips.c (mips_option_override): Report an error for
12286 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12287 for -march=r5900 -mhard-float.
12288
c7b3b99f 122892013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC		 12290
12291 * df-problems.c (can_move_insns_across): Exit loop once we
12292 find a non-fixed, non-global register.
12293
12294 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12295 setting can_throw.
12296
12297 * omega.c (omega_eliminate_red): Break after setting red_found.
12298 (omega_problem_has_red_equations): Similarly after setting found.
12299 (omega_query_variable): Similarly after setting coupled.
12300
9b6e6981
MP		 123012013-07-22 Marek Polacek <polacek@redhat.com>
12302
12303 * gimplify.c: Don't include gimple.h twice.
12304
5e5f7673
KT		 123052013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12306
12307 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12308 instead of TARGET_THUMB1.
12309 (Pz): New constraint.
12310 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12311 encodings.
12312 (compare_negsi_si): Likewise.
12313 (compare_addsi2_op0): Likewise.
12314 (compare_addsi2_op1): Likewise.
12315 (addsi3_carryin_<optab>): Likewise.
12316 (addsi3_carryin_alt2_<optab>): Likewise.
12317 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12318 for arm_restrict_it.
12319 (subsi3_carryin): Likewise.
12320 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12321 (minmax_arithsi): Disable for arm_restrict_it.
12322 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12323 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12324 (satsi_<SAT:code>_shift): Likewise.
12325 (arm_shiftsi3): Add alternative for 16-bit encoding.
12326 (arm32_movhf): Disable for arm_restrict_it.
12327 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12328 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12329
859abddd
SN		 123302013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12331
12332 * config/arm/arm.md (attribute "insn"): Delete.
12333 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12334 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12335 (not_shiftsi): Update for attribute change.
12336 (not_shiftsi_compare0): Likewise.
12337 (not_shiftsi_compare0_scratch): Likewise.
12338 (arm_one_cmplsi2): Likewise.
12339 (thumb1_one_cmplsi2): Likewise.
12340 (notsi_compare0): Likewise.
12341 (notsi_compare0_scratch): Likewise.
12342 (thumb1_movdi_insn): Likewise.
12343 (arm_movsi_insn): Likewise.
12344 (movhi_insn_arch4): Likewise.
12345 (movhi_bytes): Likewise.
12346 (arm_movqi_insn): Likewise.
12347 (thumb1_movqi_insn): Likewise.
12348 (arm32_movhf): Likewise.
12349 (thumb1_movhf): Likewise.
12350 (arm_movsf_soft_insn): Likewise.
12351 (thumb1_movsf_insn): Likewise.
12352 (thumb_movdf_insn): Likewise.
12353 (movsicc_insn): Likewise.
12354 (movsfcc_soft_insn): Likewise.
12355 (and_scc): Likewise.
12356 (cond_move): Likewise.
12357 (if_move_not): Likewise.
12358 (if_not_move): Likewise.
12359 (if_shift_move): Likewise.
12360 (if_move_shift): Likewise.
12361 (if_shift_shift): Likewise.
12362 (if_not_arith): Likewise.
12363 (if_arith_not): Likewise.
12364 (cond_move_not): Likewise.
12365 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12366 (neon_mov<mode>): Likewise.
12367 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12368 (thumb2_movsi_vfp): Likewise.
12369 (movsf_vfp): Likewise.
12370 (thumb2_movsf_vfp): Likewise.
371e77e3
UB		 12371 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12372 change.
859abddd
SN		 12373 (cortexa7_older_only): Likewise.
12374 (cortexa7_younger): Likewise.
12375 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12376 (1020alu_shift_op): Likewise.
12377 (1020alu_shift_reg_op): Likewise.
12378 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12379 (alu_shift_op): Likewise.
12380 (alu_shift_reg_op): Likewise.
12381 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12382 (11_alu_shift_op): Likewise.
12383 (11_alu_shift_reg_op): Likewise.
12384 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12385 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12386 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12387 change.
859abddd
SN		 12388 (cortex_a15_alu_shift): Likewise.
12389 (cortex_a15_alu_shift_reg): Likewise.
12390 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12391 (cortex_a5_alu_shift): Likewise.
371e77e3
UB		 12392 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12393 change.
859abddd 12394 (cortex_a53_alu_shift): Likewise.
371e77e3
UB		 12395 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12396 change.
859abddd
SN		 12397 (cortex_a7_alu_reg): Likewise.
12398 (cortex_a7_alu_shift): Likewise.
12399 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12400 (cortex_a8_alu_shift): Likewise.
12401 (cortex_a8_alu_shift_reg): Likewise.
12402 (cortex_a8_mov): Likewise.
12403 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12404 (cortex_a9_dp_shift): Likewise.
12405 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12406 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12407 (cortex_r4_mov): Likewise.
12408 (cortex_r4_alu_shift): Likewise.
12409 (cortex_r4_alu_shift_reg): Likewise.
12410 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12411 (526_alu_shift_op): Likewise.
12412 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12413 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12414 (626te_alu_shift_op): Likewise.
12415 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12416 (726te_alu_op): Likewise.
12417 (726te_alu_shift_op): Likewise.
12418 (726te_alu_shift_reg_op): Likewise.
12419 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12420 (mp626_alu_shift_op): Likewise.
12421 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12422 (pj4_alu_e1_conds): Likewise.
12423 (pj4_alu): Likewise.
12424 (pj4_alu_conds): Likewise.
12425 (pj4_shift): Likewise.
12426 (pj4_shift_conds): Likewise.
12427 (pj4_alu_shift): Likewise.
12428 (pj4_alu_shift_conds): Likewise.
12429
95b97fac
KT		 124302013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12431
12432 * config/arm/predicates.md (shiftable_operator_strict_it):
12433 New predicate.
12434 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12435 Disable cond_exec version for arm_restrict_it.
12436 (thumb2_smaxsi3): Convert to generate cond_exec.
12437 (thumb2_sminsi3): Likewise.
12438 (thumb32_umaxsi3): Likewise.
12439 (thumb2_uminsi3): Likewise.
12440 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12441 (thumb2_neg_abssi2): Likewise.
12442 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12443 (thumb2_movsicc_insn): Adjust alternatives.
12444 (thumb2_mov_negscc): Disable for arm_restrict_it.
12445 (thumb2_mov_negscc_strict_it): New pattern.
12446 (thumb2_mov_notscc_strict_it): New pattern.
12447 (thumb2_mov_notscc): Disable for arm_restrict_it.
12448 (thumb2_ior_scc): Likewise.
12449 (thumb2_ior_scc_strict_it): New pattern.
12450 (thumb2_cond_move): Adjust for arm_restrict_it.
12451 (thumb2_cond_arith): Disable for arm_restrict_it.
12452 (thumb2_cond_arith_strict_it): New pattern.
12453 (thumb2_cond_sub): Adjust for arm_restrict_it.
12454 (thumb2_movcond): Likewise.
12455 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12456 (thumb2_zero_extendhisi2_v6): Likewise.
12457 (thumb2_zero_extendqisi2_v6): Likewise.
12458 (orsi_notsi_si): Likewise.
12459 (orsi_not_shiftsi_si): Likewise.
12460
d5a6ef82
GJL		 124612013-07-22 Georg-Johann Lay <avr@gjlay.de>
12462
12463 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12464 instruction sequence is 1 byte shorter.
12465
d9e71154
UB		 124662013-07-22 Uros Bizjak <ubizjak@gmail.com>
12467
12468 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12469 it is not needed after split.
12470
8c6381ec
IS		 124712013-07-20 Iain Sandoe <iain@codesourcery.com>
12472
12473 PR target/51784
12474 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12475 second label for nonlocal goto receivers. Don't output pic base labels
12476 unless we're producing PIC; mark that action unreachable().
12477 (ix86_save_reg): If the function contains a nonlocal label, save the
12478 PIC base reg.
12479 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 12480 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS		 12481 (update_pic_label_number_if_needed): New.
12482 (machopic_output_function_base_name): Adjust for nonlocal receiver
12483 case.
12484 (machopic_should_output_picbase_label): New.
12485 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12486 (nonlocal_goto_receiver): New insn and split.
12487
d05d0709
JG		 124882013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12489
12490 * config/aarch64/aarch64-builtins.c
12491 (aarch64_fold_builtin): Fold abs in all modes.
12492 * config/aarch64/aarch64-simd-builtins.def
12493 (abs): Enable for all modes.
12494 * config/aarch64/arm_neon.h
12495 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12496 (vabs_f64): Add missing intrinsic.
12497
ba590e63
IB		 124982013-07-19 Ian Bolton <ian.bolton@arm.com>
12499
12500 * config/aarch64/arm_neon.h (vabs_s64): New function
12501
02371798
GJL		 125022013-07-19 Georg-Johann Lay <avr@gjlay.de>
12503
12504 PR target/57516
12505 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12506 * config/avr/avr.md (adjust_len): Add `round'.
12507 * config/avr/avr-protos.h (avr_out_round): New prototype.
12508 (avr_out_plus): Add `out_label' argument.
12509 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12510 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12511 Handle the case where `insn' is just a pattern.
12512 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12513 (avr_out_round): New function.
12514 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12515
12211b99 125162013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH		 12517
12518 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12519 Add check for flag_stack_usage to handle -fstack-usage support
12520
12211b99 125212013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH		 12522
12523 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12524 interaction for new Power8 flags and VSX.
12525
df9dda2d
ST		 125262013-07-18 Sriraman Tallam <tmsriram@google.com>
12527
12528 PR middle-end/57698
12529 * tree-inline.c (expand_call_inline): Emit errors during
12530 early_inlining only if optimization is not turned on.
12531
0f9cdef4
DM		 125322013-07-18 David Malcolm <dmalcolm@redhat.com>
12533
12534 * passes.def: New.
12535
12536 * passes.c (init_optimization_passes): Move the construction of
12537 the pass hierarchy into a new passes.def file.
12538
12539 * Makefile.in (passes.o): Add dependency on passes.def.
12540
2efa4087
DM		 125412013-07-18 David Malcolm <dmalcolm@redhat.com>
12542
12543 * passes.c (init_optimization_passes): Introduce macros for
12544 constructing the tree of passes (INSERT_PASSES_AFTER,
12545 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12546 TERMINATE_PASS_LIST).
12547
bc404e1b
VM		 125482013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12549 Wei Mi <wmi@google.com>
12550
12551 PR rtl-optimization/57878
12552 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12553 top.
12554 (reload_pseudo_compare_func): Check nregs first for reload
12555 pseudos.
12556
d675843f
DM		 125572013-07-18 David Malcolm <dmalcolm@redhat.com>
12558
12559 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12560
c7b3b99f 125612013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC		 12562
12563 * read-rtl.c (validate_const_int): Once an invalid character is
12564 seen, quit the loop.
12565
12566 * gengtype.c (write_roots): Similarly once we find the "deletable"
12567 or "if_marked" option.
12568
52fceb44
SN		 125692013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12570
12571 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12572 "xtab" and "sat". Move value "clz" from here to ...
12573 (attriubte "type"): ... here.
12574 (satsi_<SAT:code>): Delete "insn" attribute.
12575 (satsi_<SAT:code>_shift): Likewise.
12576 (arm_zero_extendqisi2addsi): Likewise.
12577 (arm_extendqisi2addsi): Likewise.
12578 (clzsi2): Update for attribute changes.
12579 (rbitsi2): Likewise.
371e77e3
UB		 12580 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12581 attribute.
52fceb44
SN		 12582 (arm_usatsihi): Likewise.
12583 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12584
006bd006
SN		 125852013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12586
12587 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB		 12588 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12589 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN		 12590 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12591 in alphabetical order.
12592 (attribute "core_cycles"): Update for attribute changes.
12593 (arm_addsi3): Likewise.
12594 (addsi3_compare0): Likewise.
12595 (addsi3_compare0_scratch): Likewise.
12596 (addsi3_compare_op1): Likewise.
12597 (addsi3_compare_op2): Likewise.
12598 (compare_addsi2_op0): Likewise.
12599 (compare_addsi2_op1): Likewise.
12600 (addsi3_carryin_shift_<optab>): Likewise.
12601 (subsi3_carryin_shift): Likewise.
12602 (rsbsi3_carryin_shift): Likewise.
12603 (arm_subsi3_insn): Likewise.
12604 (subsi3_compare0): Likewise.
12605 (subsi3_compare): Likewise.
12606 (arm_andsi3_insn): Likewise.
12607 (thumb1_andsi3_insn): Likewise.
12608 (andsi3_compare0): Likewise.
12609 (andsi3_compare0_scratch): Likewise.
12610 (zeroextractsi_compare0_scratch
12611 (andsi_not_shiftsi_si): Likewise.
12612 (iorsi3_insn): Likewise.
12613 (iorsi3_compare0): Likewise.
12614 (iorsi3_compare0_scratch): Likewise.
12615 (arm_xorsi3): Likewise.
12616 (thumb1_xorsi3_insn): Likewise.
12617 (xorsi3_compare0): Likewise.
12618 (xorsi3_compare0_scratch): Likewise.
12619 (satsi_<SAT:code>_shift): Likewise.
12620 (rrx): Likewise.
12621 (arm_shiftsi3): Likewise.
12622 (shiftsi3_compare0): Likewise.
12623 (not_shiftsi): Likewise.
12624 (not_shiftsi_compare0): Likewise.
12625 (not_shiftsi_compare0_scratch): Likewise.
12626 (arm_one_cmplsi2): Likewise.
12627 (thumb_one_complsi2): Likewise.
12628 (notsi_compare0): Likewise.
12629 (notsi_compare0_scratch): Likewise.
12630 (thumb1_zero_extendhisi2): Likewise.
12631 (arm_zero_extendhisi2): Likewise.
12632 (arm_zero_extendhisi2_v6): Likewise.
12633 (arm_zero_extendhisi2addsi): Likewise.
12634 (thumb1_zero_extendqisi2): Likewise.
12635 (thumb1_zero_extendqisi2_v6): Likewise.
12636 (arm_zero_extendqisi2): Likewise.
12637 (arm_zero_extendqisi2_v6): Likewise.
12638 (arm_zero_extendqisi2addsi): Likewise.
12639 (thumb1_extendhisi2): Likewise.
12640 (arm_extendhisi2): Likewise.
12641 (arm_extendhisi2_v6): Likewise.
12642 (arm_extendqisi): Likewise.
12643 (arm_extendqisi_v6): Likewise.
12644 (arm_extendqisi2addsi): Likewise.
12645 (thumb1_extendqisi2): Likewise.
12646 (thumb1_movdi_insn): Likewise.
12647 (arm_movsi_insn): Likewise.
12648 (movsi_compare0): Likewise.
12649 (movhi_insn_arch4): Likewise.
12650 (movhi_bytes): Likewise.
12651 (arm_movqi_insn): Likewise.
12652 (thumb1_movqi_insn): Likewise.
12653 (arm32_movhf): Likewise.
12654 (thumb1_movhf): Likewise.
12655 (arm_movsf_soft_insn): Likewise.
12656 (thumb1_movsf_insn): Likewise.
12657 (movdf_soft_insn): Likewise.
12658 (thumb_movdf_insn): Likewise.
12659 (arm_cmpsi_insn): Likewise.
12660 (cmpsi_shiftsi): Likewise.
12661 (cmpsi_shiftsi_swp): Likewise.
12662 (arm_cmpsi_negshiftsi_si): Likewise.
12663 (movsicc_insn): Likewise.
12664 (movsfcc_soft_insn): Likewise.
12665 (arith_shiftsi): Likewise.
12666 (arith_shiftsi_compare0
12667 (arith_shiftsi_compare0_scratch
12668 (sub_shiftsi): Likewise.
12669 (sub_shiftsi_compare0
12670 (sub_shiftsi_compare0_scratch
12671 (and_scc): Likewise.
12672 (cond_move): Likewise.
12673 (if_plus_move): Likewise.
12674 (if_move_plus): Likewise.
12675 (if_move_not): Likewise.
12676 (if_not_move): Likewise.
12677 (if_shift_move): Likewise.
12678 (if_move_shift): Likewise.
12679 (if_shift_shift): Likewise.
12680 (if_not_arith): Likewise.
12681 (if_arith_not): Likewise.
12682 (cond_move_not): Likewise.
12683 (thumb1_ashlsi3): Set type attribute.
12684 (thumb1_ashrsi3): Likewise.
12685 (thumb1_lshrsi3): Likewise.
12686 (thumb1_rotrsi3): Likewise.
12687 (shiftsi3_compare0_scratch): Likewise.
12688 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12689 (neon_mov<mode>): Likewise.
371e77e3
UB		 12690 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12691 attribute changes.
006bd006
SN		 12692 (thumb2_movsi_insn): Likewise.
12693 (thumb2_cmpsi_neg_shiftsi): Likewise.
12694 (thumb2_extendqisi_v6): Likewise.
12695 (thumb2_zero_extendhisi2_v6): Likewise.
12696 (thumb2_zero_extendqisi2_v6): Likewise.
12697 (thumb2_shiftsi3_short): Likewise.
12698 (thumb2_addsi3_compare0_scratch): Likewise.
12699 (orsi_not_shiftsi_si): Likewise.
12700 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12701 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12702 changes.
12703 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12704 (1020alu_shift_op): Likewise.
12705 (1020alu_shift_reg_op): Likewise.
12706 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12707 (alu_shift_op): Likewise.
12708 (alu_shift_reg_op): Likewise.
12709 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12710 (11_alu_shift_op): Likewise.
12711 (11_alu_shift_reg_op): Likewise.
12712 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12713 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12714 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12715 changes.
006bd006
SN		 12716 (cortex_a15_alu_shift): Likewise.
12717 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB		 12718 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12719 changes.
006bd006
SN		 12720 (cortex_a5_alu_shift): Likewise.
12721 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12722 changes.
12723 (cortex_a53_alu_shift): Likewise.
12724 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12725 changes.
12726 (cortex_a7_alu_reg): Likewise.
12727 (cortex_a7_alu_shift): Likewise.
371e77e3
UB		 12728 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12729 changes.
006bd006
SN		 12730 (cortex_a8_alu_shift): Likewise.
12731 (cortex_a8_alu_shift_reg): Likewise.
12732 (cortex_a8_mov): Likewise.
12733 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12734 (cortex_a9_dp_shift): Likewise.
371e77e3
UB		 12735 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12736 changes.
12737 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12738 changes.
006bd006
SN		 12739 (cortex_r4_mov): Likewise.
12740 (cortex_r4_alu_shift): Likewise.
12741 (cortex_r4_alu_shift_reg): Likewise.
12742 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12743 (526_alu_shift_op): Likewise.
12744 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12745 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12746 (626te_alu_shift_op): Likewise.
12747 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12748 (726te_alu_op): Likewise.
12749 (726te_alu_shift_op): Likewise.
12750 (726te_alu_shift_reg_op): Likewise.
12751 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12752 (mp626_alu_shift_op): Likewise.
12753 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12754 (pj4_alu_e1_conds): Likewise.
12755 (pj4_alu): Likewise.
12756 (pj4_alu_conds): Likewise.
12757 (pj4_shift): Likewise.
12758 (pj4_shift_conds): Likewise.
12759 (pj4_alu_shift): Likewise.
12760 (pj4_alu_shift_conds): Likewise.
371e77e3
UB		 12761 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12762 changes.
006bd006
SN		 12763 (cortexa7_older_only): Likewise.
12764 (cortexa7_younger): Likewise.
12765
651df1b2
DM		 127662013-07-18 David Malcolm <dmalcolm@redhat.com>
12767
12768 * ipa-pure-const.c (generate_summary): Rename to...
12769 (pure_const_generate_summary): ... this.
12770
ea717bd7
IS		 127712013-07-17 Iain Sandoe <iain@codesourcery.com>
12772
12773 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12774
a1259a13
YR		 127752013-07-17 Yvan Roux <yvan.roux@linaro.org>
12776
12777 PR target/57909
12778 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12779 usage in HI mode.
12780
167f68ed
AK		 127812013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12782
12783 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12784 enabled without -march=zEC12.
12785 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12786 flags to be set.
12787
f900a982
MR		 127882013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12789
12790 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12791 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12792 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12793 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12794 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12795 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12796 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12797 ISA_HAS_FP4.
12798 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12799 and ISA_HAS_NMADD3_NMSUB3.
12800 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12801 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12802 (nmsub4<mode>, nmsub3<mode>): Likewise.
12803 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12804
d4ed27eb
MR		 128052013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12806
12807 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12808 TARGET_MIPS5400 checking.
12809
7a5add18
PB		 128102013-07-16 Jakub Jelinek <jakub@redhat.com>
12811 Peter Bergner <bergner@vnet.ibm.com>
12812
12813 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12814 registers in the comment.
12815 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12816 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12817 rather than FIRST_PSEUDO_REGISTERS.
12818
128192013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB		 12820
12821 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12822 enable extra ISA flags with TARGET_HTM.
12823
d2ab0929
MR		 128242013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12825
12826 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12827 Fix comment typos.
12828
839c74bc
CH		 128292013-07-15 Cong Hou <congh@google.com>
12830
12831 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12832 in compare function for sorting.
12833
7a5add18 128342013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB		 12835
12836 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12837 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12838 * config/rs6000/rs6000.opt: Add -mhtm option.
12839 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12840 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12841 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12842 __HTM__ if the HTM instructions are available.
12843 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12844 htm_spr_reg_operand): New define_predicates.
12845 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12846 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12847 Include htm.md.
12848 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12849 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12850 HTM builtin functions.
12851 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12852 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12853 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12854 (rs6000_builtin_mask_calculate): Likewise.
12855 (rs6000_option_override_internal): Likewise.
12856 (bdesc_htm): Add new HTM builtin support.
12857 (htm_spr_num): New function.
12858 (htm_spr_regno): Likewise.
12859 (rs6000_htm_spr_icode): Likewise.
12860 (htm_expand_builtin): Likewise.
12861 (htm_init_builtins): Likewise.
12862 (rs6000_expand_builtin): Add support for HTM builtin functions.
12863 (rs6000_init_builtins): Likewise.
371e77e3
UB		 12864 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12865 option.
0258b6e4
PB		 12866 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12867 (TARGET_HTM, MASK_HTM): Define macros.
12868 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12869 (FIXED_REGISTERS): Likewise.
12870 (CALL_USED_REGISTERS): Likewise.
12871 (CALL_REALLY_USED_REGISTERS): Likewise.
12872 (REG_ALLOC_ORDER): Likewise.
12873 (enum reg_class): Likewise.
12874 (REG_CLASS_NAMES): Likewise.
12875 (REG_CLASS_CONTENTS): Likewise.
12876 (REGISTER_NAMES): Likewise.
12877 (ADDITIONAL_REGISTER_NAMES): Likewise.
12878 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12879 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12880 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12881 * config/rs6000/htm.md: New file.
12882 * config/rs6000/htmintrin.h: New file.
12883 * config/rs6000/htmxlintrin.h: New file.
12884
87dd8ab0
MS		 128852013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12886
12887 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12888 Define SYMBOL_TINY_GOT, update comment.
12889 * config/aarch64/aarch64.c
12890 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12891 (aarch64_expand_mov_immediate): Likewise.
12892 (aarch64_print_operand): Likewise.
12893 (aarch64_classify_symbol): Likewise.
12894 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12895 (ldr_got_tiny): Define.
12896
52d676b6
TG		 128972013-07-13 Tobias Grosser <tobias@grosser.es>
12898
12899 PR tree-optimization/54094
12900 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12901 scheduling dimension for the parallelism check from the polyhedral
12902 information in the AST.
12903 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12904 in 2D + 1 form.
12905
286e8fc1
JM		 129062013-07-13 Jason Merrill <jason@redhat.com>
12907
12908 * print-tree.c (debug_vec_tree): Use debug_raw.
12909 (debug_raw (vec<tree, va_gc> &)): New.
12910 (debug_raw (vec<tree, va_gc> *)): New.
12911 * tree.h: Declare them.
12912
e43257e8
BC		 129132013-07-13 Bin Cheng <bin.cheng@arm.com>
12914
12915 * ifcvt.c (ifcvt_after_combine): New static variable.
12916 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12917 for size.
12918 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12919 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12920 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12921
da582d46
MR		 129222013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12923
12924 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12925
18b0ea8f
MM		 129262013-07-12 Michael Matz <matz@suse.de>
12927
12928 PR middle-end/55771
12929 * convert.c (convert_to_real): Reject non-float inner types.
12930
e4f0f84d
TB		 129312013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12932
12933 * config/aarch64/aarch64-protos.h
12934 (aarch64_simd_immediate_valid_for_move): Remove.
12935 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12936 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12937 cases.
12938 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12939
76500022
SE		 129402013-07-11 Steve Ellcey <sellcey@mips.com>
12941
12942 * config/mips/mips.c (mips_conditional_register_usage): Do not
12943 use t[0-7] registers in MIPS16 mode when optimizing for size.
12944
78039734
ST		 129452013-07-11 Sriraman Tallam <tmsriram@google.com>
12946
12947 * config/i386/i386.c (dispatch_function_versions): Fix array
12948 indexing of function_version_info to match actual_versions.
12949
7b24b675
TJ		 129502013-07-11 Teresa Johnson <tejohnson@google.com>
12951
12952 * vec.h (struct va_gc): Move release out-of-line.
12953 (va_gc::release): Call ggc_free on released vec.
12954
85d4cbb8
UW		 129552013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12956
12957 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12958 Require GOT register as additional operand in UNSPEC.
12959 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12960 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12961 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12962 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12963 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12964 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12965 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12966
52b3f9e2
GJL		 129672013-07-11 Georg-Johann Lay <avr@gjlay.de>
12968
12969 PR target/57631
12970 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12971 name seen by assembler/linker rather if available.
12972
3f3e266e
AS		 129732013-07-11 Andreas Schwab <schwab@suse.de>
12974
12975 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12976
21231ca6
VM		 129772013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12978
371e77e3 12979 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 12980
d331d374
JM		 129812013-07-10 Joseph Myers <joseph@codesourcery.com>
12982
673c2f63
JM		 12983 * doc/tm.texi.in: Move hook documentation to ....
12984 * target.def: ... here.
12985
d331d374
JM		 12986 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12987 text on @hook line.
12988 * doc/tm.texi: Regenerate.
12989
dfeadaa0
PC		 129902013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12991
12992 PR c++/57869
12993 * doc/invoke.texi: Document -Wconditionally-supported.
12994
bb6f2bac
GJL		 129952013-07-10 Georg-Johann Lay <avr@gjlay.de>
12996
12997 PR target/57844
12998 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12999 of my_fp.
13000
5219b232
GJL		 130012013-07-10 Georg-Johann Lay <avr@gjlay.de>
13002
13003 PR target/57506
13004 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13005 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13006 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13007 Remove duplicate devices.
13008 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13009 * config/avr/t-multilib: Regenerate.
13010 * config/avr/avr-tables.opt: Regenerate.
13011 * doc/avr-mmcu.texi: Regenerate.
13012
283045c0
GJL		 130132013-07-10 Georg-Johann Lay <avr@gjlay.de>
13014
13015 PR target/56987
13016 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13017
ceb2d59f 130182013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 13019
ceb2d59f
GS		 13020 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13021 the cost of MULT when optimizing for size.
13022
055e0a99
JBG		 130232013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13024
13025 * config/cr16/cr16-protos.h: Don't include target.h.
13026
ef43e868
JM		 130272013-07-09 Joseph Myers <joseph@codesourcery.com>
13028
13029 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13030 adjust register size for TDmode and TFmode for VSX registers.
13031
4a283fdf
KT		 130322013-07-08 Kai Tietz <ktietz@redhat.com>
13033
13034 PR target/56892
13035 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13036 hook_bool_const_tree_true.
13037
2cf4c39e
AK		 130382013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13039
13040 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13041 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13042 * config/s390/s390.md: Define FPR*_REGNUM constants.
13043 Fix FPR2_REGNUM constant (18 -> 17).
13044 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13045 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13046 ("trunc<BFP:mode><DFP_ALL:mode>2")
13047 ("trunc<DFP_ALL:mode><BFP:mode>2")
13048 ("*extend<BFP:mode><DFP_ALL:mode>2")
13049 ("*extend<DFP_ALL:mode><BFP:mode>2")
13050 ("extend<BFP:mode><DFP_ALL:mode>2")
13051 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13052 FPR4_REGNUM.
13053
e6ac0270 130542013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 13055
6a2fa4b2 13056 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 13057
b89b22fc
AK		 130582013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13059
13060 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13061 and cfun_fpr_bit_p to cfun_fpr_save_p.
13062 (s390_frame_area, s390_register_info, s390_frame_info)
13063 (s390_emit_prologue, s390_emit_epilogue)
13064 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13065 register numbers.
13066 * config/s390/s390.h: Define *_REGNUM macros for floating point
13067 register numbers.
13068
07f398aa
EB		 130692013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13070
13071 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13072
75fefa91
PCC		 130732013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13074
13075 PR rtl-optimization/57786
13076 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13077 and break out of the loop when it is set to false.
13078
046f1eee
JJ		 130792013-07-08 Jakub Jelinek <jakub@redhat.com>
13080
8140c065
JJ		 13081 PR target/57819
13082 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13083 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13084 (const_int 63)) 0)).
13085 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13086 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13087 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13088
046f1eee
JJ		 13089 PR rtl-optimization/57829
13090 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13091 mask bits outside of mode are just sign-extension from mode to HWI.
13092
930b700b
MZ		 130932013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13094
13095 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13096 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13097 adjust_address instead of change_address to keep info about alignment.
13098 (emit_strmov): Remove.
13099 (emit_memmov): New function.
13100 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13101 (expand_movmem_epilogue): Likewise and return updated rtx for
13102 destination.
13103 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13104 destination and source.
13105 (decide_alignment): Refactor, handle vector_loop.
13106 (ix86_expand_movmem): Likewise.
13107 (ix86_expand_setmem): Likewise.
13108 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13109
7b9d1bd8
UB		 131102013-07-07 Uros Bizjak <ubizjak@gmail.com>
13111
13112 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13113 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13114
eabb5f48
UB		 131152013-07-06 Uros Bizjak <ubizjak@gmail.com>
13116
13117 * config/i386/sse.md (sse_movlhps): Change alternative 3
13118 of operand 2 to "m".
13119
131202013-07-06 Uros Bizjak <ubizjak@gmail.com>
13121
13122 PR target/57807
13123 * config/i386/sse.md (iptr): New mode attribute.
13124 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13125 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13126 (<sse>_vmmul<mode>3): Ditto.
13127 (<sse>_vmdiv<mode>3): Ditto.
13128 (sse_vmrcpv4sf2): Ditto.
13129 (<sse>_vmsqrt<mode>2): Ditto.
13130 (sse_vmrsqrtv4sf2): Ditto.
13131 (<sse>_vm<code><mode>3): Ditto.
13132 (avx_vmcmp<mode>3): Ditto.
13133 (<sse>_vmmaskcmp<mode>3): Ditto.
13134 (<sse>_comi): Ditto.
13135 (<sse>_ucomi): Ditto.
13136 (*xop_vmfrcz_<mode>): Ditto.
13137 (*fmai_fmadd_<mode>): Ditto.
13138 (*fmai_fmsub_<mode>): Ditto.
13139 (*fmai_fnmadd_<mode>): Ditto.
13140 (*fmai_fnmsub_<mode>): Ditto.
13141 (*fma4i_vmfmadd_<mode>): Ditto.
13142 (*fma4i_vmfmsub_<mode>): Ditto.
13143 (*fma4i_vmfnmadd_<mode>): Ditto.
13144 (*fma4i_vmfnmsub_<mode>): Ditto.
13145 (*xop_vmfrcz_<mode>): Ditto.
13146 (sse_cvtps2pi): Ditto.
13147 (sse_cvttps2pi): Ditto.
13148 (sse_cvtss2si): Ditto.
13149 (sse_cvtss2si_2): Ditto.
13150 (sse_cvtss2siq_2): Ditto.
13151 (sse_cvttss2si): Ditto.
13152 (sse_cvttss2siq): Ditto.
13153 (sse_cvtsd2si): Ditto.
13154 (sse_cvtsd2si_2): Ditto.
13155 (sse_cvtsd2siq_2): Ditto.
13156 (sse_cvttsd2si): Ditto.
13157 (sse_cvttsd2siq): Ditto.
13158 (sse_cvtsd2ss): Ditto.
13159 (sse_cvtss2sd): Ditto.
13160 (avx2_pbroadcast<mode>): Ditto.
13161 (avx2_pbroadcast<mode>_1): Ditto.
13162 (*avx_vperm_broadcast_v4sf): Ditto.
13163
13164 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13165 (sse_movlhps): Ditto.
13166 (sse_storehps): Ditto.
13167 (sse_loadhps): Ditto.
13168 (sse_storelps): Ditto.
13169 (sse_loadlps): Ditto.
13170 (*vec_concatv4sf): Ditto.
13171 (*vec_interleave_highv2df): Ditto.
13172 (*vec_interleave_lowv2df): Ditto.
13173 (*vec_extractv2df_1_sse): Ditto.
13174 (*vec_extractv2df_0_sse): Ditto.
13175 (sse2_storelpd): Ditto.
13176 (sse2_loadlpd): Ditto.
13177 (sse2_movsd): Ditto.
13178 (*vec_concatv4si): Ditto.
13179 (vec_concatv2di): Ditto.
13180
13181 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13182 for Intel asm dialect.
13183 (mmx_punpcklwd): Ditto.
13184 (mmx_punpckldq): Ditto.
13185
13186 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13187 for intel assembler dialect.
13188
1f6eac90
JJ		 131892013-07-06 Jakub Jelinek <jakub@redhat.com>
13190
13191 PR target/29776
13192 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13193 for BUILT_IN_C{LZ,LRSB}*.
13194 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13195 * tree-vrp.c (extract_range_basic): Handle
13196 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13197 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13198 fall thru to code calling set_value*.
13199 * builtins.c (expand_builtin): Remove *IMAX cases.
13200 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13201 if width is bigger than 2*HWI.
13202
2b778c9d
VM		 132032013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13204
13205 PR rtl-optimization/55342
13206 * lra-int.h (lra_subreg_reload_pseudos): New.
13207 * lra.c: Add undoing optional reloads to the block diagram.
13208 (lra_subreg_reload_pseudos): New.
13209 (lra_optional_reload_pseudos): Change comments.
13210 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13211 lra_optional_reload_pseudos after undo transformations.
13212 * lra-assigns.c (pseudo_prefix_title): New.
13213 (lra_setup_reg_renumber): Use it.
13214 (spill_for): Ditto. Check subreg reload pseudos too.
13215 (assign_by_spills): Consider subreg reload pseudos too.
13216 * lra-constraints.c (simplify_operand_subreg): Use
13217 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13218 (curr_insn_transform): Recognize and do optional reloads.
13219 (undo_optional_reloads): New.
13220 (lra_undo_inheritance): Call undo_optional_reloads.
13221
207156e4
TQ		 132222013-07-05 Thomas Quinot <quinot@adacore.com>
13223
13224 * tree-complex.c (expand_complex_operations_1): Fix typo.
13225
8990e73a
TB		 132262013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13227
13228 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13229 (tune_params): New member 'const vec_costs'.
13230 * config/aarch64/aarch64.c (generic_vector_cost): New.
13231 (generic_tunings): New member 'generic_vector_cost'.
13232 (aarch64_builtin_vectorization_cost): New.
13233 (aarch64_add_stmt_cost): New.
13234 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13235 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13236
f7bc421d
JJ		 132372013-07-03 Jakub Jelinek <jakub@redhat.com>
13238
13239 PR target/57777
13240 * config/i386/predicates.md (vsib_address_operand): Disallow
13241 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13242
e90d1568
HPN		 132432013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13244
13245 PR middle-end/55030
13246 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13247 expand_builtin_setjmp_receiver.
13248 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13249 with NULL for the label parameter.
13250 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13251 the frame-pointer. Adjust comments.
13252 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13253 only if LABEL is non-NULL.
13254
cbd7413b
YZ		 132552013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13256
13257 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13258 (ARM_ABI_AAPCS64): Ditto.
13259 (arm_abi): Ditto.
13260 (ARM_DEFAULT_ABI): Ditto.
13261
dec11868
JG		 132622013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13263
13264 * config/aarch64/aarch64-builtins.c
13265 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13266 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13267 (st1): Likewise.
13268 * config/aarch64/aarch64-simd.md
13269 (aarch64_ld1<VALL:mode>): New.
13270 (aarch64_st1<VALL:mode>): Likewise.
13271 * config/aarch64/arm_neon.h
13272 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13273
786fb9b6
ST		 132742013-07-02 Sriraman Tallam <tmsriram@google.com>
13275
0ba6aff3
UB		 13276 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13277 target ISA is AVX.
786fb9b6
ST		 13278 (ix86_option_override_internal):Turn on all -mavx target flags by
13279 default as they are dependent on AVX anyway.
13280
4b043553
CC		 132812013-07-02 Cary Coutant <ccoutant@google.com>
13282
13283 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13284 deterministic hash.
13285 (loc_checksum_ordered): Likewise.
13286 (hash_loc_operands): Remove inline keyword.
13287
c4e87a13
JJ		 132882013-07-02 Jakub Jelinek <jakub@redhat.com>
13289
13290 PR tree-optimization/57741
13291 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13292 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13293 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13294 Allow REAL_CST step_exprs if flag_associative_math.
13295 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13296
26366d28
IB		 132972013-07-02 Ian Bolton <ian.bolton@arm.com>
13298
0ba6aff3 13299 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 13300
2879bb2b
IB		 133012013-07-02 Ian Bolton <ian.bolton@arm.com>
13302
13303 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13304
afc5e8a6
KT		 133052013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13306
13307 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13308 encoding.
13309 (iorsi3_insn): Likewise.
13310 (arm_xorsi3): Likewise.
13311
31f8442b
SN		 133122013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13313
13314 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13315 to ...
13316 (attribute "type"): ... here, and prefix with "wmmx_".
13317 (attribute "core_cycles"): Update for attribute changes.
13318 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13319 (tbcstv4hi): Likewise.
13320 (tbcstv2si): Likewise.
13321 (iwmmxt_iordi3): Likewise.
13322 (iwmmxt_xordi3): Likewise.
13323 (iwmmxt_anddi3): Likewise.
13324 (iwmmxt_nanddi3): Likewise.
13325 (iwmmxt_arm_movdi): Likewise.
13326 (iwmmxt_movsi_insn): Likewise.
13327 (mov<mode>_internal): Likewise.
13328 (and<mode>3_iwmmxt): Likewise.
13329 (ior<mode>3_iwmmxt): Likewise.
13330 (xor<mode>3_iwmmxt): Likewise.
13331 (add<mode>3_iwmmxt): Likewise.
13332 (ssaddv8qi3): Likewise.
13333 (ssaddv4hi3): Likewise.
13334 (ssaddv2si3): Likewise.
13335 (usaddv8qi3): Likewise.
13336 (usaddv4hi3): Likewise.
13337 (usaddv2si3): Likewise.
13338 (sub<mode>3_iwmmxt): Likewise.
13339 (sssubv8qi3): Likewise.
13340 (sssubv4hi3): Likewise.
13341 (sssubv2si3): Likewise.
13342 (ussubv8qi3): Likewise.
13343 (ussubv4hi3): Likewise.
13344 (ussubv2si3): Likewise.
13345 (mulv4hi3_iwmmxt): Likewise.
13346 (smulv4hi3_highpart): Likewise.
13347 (umulv4hi3_highpart): Likewise.
13348 (iwmmxt_wmacs): Likewise.
13349 (iwmmxt_wmacsz): Likewise.
13350 (iwmmxt_wmacu): Likewise.
13351 (iwmmxt_wmacuz): Likewise.
13352 (iwmmxt_clrdi): Likewise.
13353 (iwmmxt_clrv8qi): Likewise.
13354 (iwmmxt_clr4hi): Likewise.
13355 (iwmmxt_clr2si): Likewise.
13356 (iwmmxt_uavgrndv8qi3): Likewise.
13357 (iwmmxt_uavgrndv4hi3): Likewise.
13358 (iwmmxt_uavgv8qi3): Likewise.
13359 (iwmmxt_uavgv4hi3): Likewise.
13360 (iwmmxt_tinsrb): Likewise.
13361 (iwmmxt_tinsrh): Likewise.
13362 (iwmmxt_tinsrw): Likewise.
13363 (iwmmxt_textrmub): Likewise.
13364 (iwmmxt_textrmsb): Likewise.
13365 (iwmmxt_textrmuh): Likewise.
13366 (iwmmxt_textrmsh): Likewise.
13367 (iwmmxt_textrmw): Likewise.
13368 (iwmxxt_wshufh): Likewise.
13369 (eqv8qi3): Likewise.
13370 (eqv4hi3): Likewise.
13371 (eqv2si3): Likewise.
13372 (gtuv8qi3): Likewise.
13373 (gtuv4hi3): Likewise.
13374 (gtuv2si3): Likewise.
13375 (gtv8qi3): Likewise.
13376 (gtv4hi3): Likewise.
13377 (gtv2si3): Likewise.
13378 (smax<mode>3_iwmmxt): Likewise.
13379 (umax<mode>3_iwmmxt): Likewise.
13380 (smin<mode>3_iwmmxt): Likewise.
13381 (umin<mode>3_iwmmxt): Likewise.
13382 (iwmmxt_wpackhss): Likewise.
13383 (iwmmxt_wpackwss): Likewise.
13384 (iwmmxt_wpackdss): Likewise.
13385 (iwmmxt_wpackhus): Likewise.
13386 (iwmmxt_wpackwus): Likewise.
13387 (iwmmxt_wpackdus): Likewise.
13388 (iwmmxt_wunpckihb): Likewise.
13389 (iwmmxt_wunpckihh): Likewise.
13390 (iwmmxt_wunpckihw): Likewise.
13391 (iwmmxt_wunpckilb): Likewise.
13392 (iwmmxt_wunpckilh): Likewise.
13393 (iwmmxt_wunpckilw): Likewise.
13394 (iwmmxt_wunpckehub): Likewise.
13395 (iwmmxt_wunpckehuh): Likewise.
13396 (iwmmxt_wunpckehuw): Likewise.
13397 (iwmmxt_wunpckehsb): Likewise.
13398 (iwmmxt_wunpckehsh): Likewise.
13399 (iwmmxt_wunpckehsw): Likewise.
13400 (iwmmxt_wunpckelub): Likewise.
13401 (iwmmxt_wunpckeluh): Likewise.
13402 (iwmmxt_wunpckeluw): Likewise.
13403 (iwmmxt_wunpckelsb): Likewise.
13404 (iwmmxt_wunpckelsh): Likewise.
13405 (iwmmxt_wunpckelsw): Likewise.
13406 (ror<mode>3): Likewise.
13407 (ashr<mode>3_iwmmxt): Likewise.
13408 (lshr<mode>3_iwmmxt): Likewise.
13409 (ashl<mode>3_iwmmxt): Likewise.
13410 (ror<mode>3_di): Likewise.
13411 (ashr<mode>3_di): Likewise.
13412 (lshr<mode>3_di): Likewise.
13413 (ashl<mode>3_di): Likewise.
13414 (iwmmxt_wmadds): Likewise.
13415 (iwmmxt_wmaddu): Likewise.
13416 (iwmmxt_tmia): Likewise.
13417 (iwmmxt_tmiaph): Likewise.
13418 (iwmmxt_tmiabb): Likewise.
13419 (iwmmxt_tmiatb): Likewise.
13420 (iwmmxt_tmiabt): Likewise.
13421 (iwmmxt_tmiatt): Likewise.
13422 (iwmmxt_tmovmskb): Likewise.
13423 (iwmmxt_tmovmskh): Likewise.
13424 (iwmmxt_tmovmskw): Likewise.
13425 (iwmmxt_waccb): Likewise.
13426 (iwmmxt_wacch): Likewise.
13427 (iwmmxt_waccw): Likewise.
13428 (iwmmxt_waligni): Likewise.
13429 (iwmmxt_walignr): Likewise.
13430 (iwmmxt_walignr0): Likewise.
13431 (iwmmxt_walignr1): Likewise.
13432 (iwmmxt_walignr2): Likewise.
13433 (iwmmxt_walignr3): Likewise.
13434 (iwmmxt_wsadb): Likewise.
13435 (iwmmxt_wsadh): Likewise.
13436 (iwmmxt_wsadbz): Likewise.
13437 (iwmmxt_wsadhz): Likewise.
13438 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13439 (iwmmxt_wabsdiffb): Likewise.
13440 (iwmmxt_wabsdiffh): Likewise.
13441 (iwmmxt_wabsdiffw): Likewise.
13442 (iwmmxt_waddsubhx): Likewise
13443 (iwmmxt_wsubaddhx): Likewise.
13444 (addc<mode>3): Likewise.
13445 (iwmmxt_avg4): Likewise.
13446 (iwmmxt_avg4r): Likewise.
13447 (iwmmxt_wmaddsx): Likewise.
13448 (iwmmxt_wmaddux): Likewise.
13449 (iwmmxt_wmaddsn): Likewise.
13450 (iwmmxt_wmaddun): Likewise.
13451 (iwmmxt_wmulwsm): Likewise.
13452 (iwmmxt_wmulwum): Likewise.
13453 (iwmmxt_wmulsmr): Likewise.
13454 (iwmmxt_wmulumr): Likewise.
13455 (iwmmxt_wmulwsmr): Likewise.
13456 (iwmmxt_wmulwumr): Likewise.
13457 (iwmmxt_wmulwl): Likewise.
13458 (iwmmxt_wqmulm): Likewise.
13459 (iwmmxt_wqmulwm): Likewise.
13460 (iwmmxt_wqmulmr): Likewise.
13461 (iwmmxt_wqmulwmr): Likewise.
13462 (iwmmxt_waddbhusm): Likewise.
13463 (iwmmxt_waddbhusl): Likewise.
13464 (iwmmxt_wqmiabb): Likewise.
13465 (iwmmxt_wqmiabt): Likewise.
13466 (iwmmxt_wqmiatb): Likewise.
13467 (iwmmxt_wqmiatt): Likewise.
13468 (iwmmxt_wqmiabbn): Likewise.
13469 (iwmmxt_wqmiabtn): Likewise.
13470 (iwmmxt_wqmiatbn): Likewise.
13471 (iwmmxt_wqmiattn): Likewise.
13472 (iwmmxt_wmiabb): Likewise.
13473 (iwmmxt_wmiabt): Likewise.
13474 (iwmmxt_wmiatb): Likewise.
13475 (iwmmxt_wmiatt): Likewise.
13476 (iwmmxt_wmiabbn): Likewise.
13477 (iwmmxt_wmiabtn): Likewise.
13478 (iwmmxt_wmiatbn): Likewise.
13479 (iwmmxt_wmiattn): Likewise.
13480 (iwmmxt_wmiawbb): Likewise.
13481 (iwmmxt_wmiawbt): Likewise.
13482 (iwmmxt_wmiawtb): Likewise.
13483 (iwmmxt_wmiawtt): Likewise.
13484 (iwmmxt_wmiawbbn): Likewise.
13485 (iwmmxt_wmiawbtn): Likewise.
13486 (iwmmxt_wmiawtbn): Likewise.
13487 (iwmmxt_wmiawttn): Likewise.
13488 (iwmmxt_wmerge): Likewise.
13489 (iwmmxt_tandc<mode>3): Likewise.
13490 (iwmmxt_torc<mode>3): Likewise.
13491 (iwmmxt_torvsc<mode>3): Likewise.
13492 (iwmmxt_textrc<mode>3): Likewise.
13493 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13494 (wmmxt_pack): Likewise.
13495 (wmmxt_mult_c1): Likewise.
13496 (wmmxt_mult_c2): Likewise.
13497 (wmmxt_alu_c1): Likewise.
13498 (wmmxt_alu_c2): Likewise.
13499 (wmmxt_alu_c3): Likewise.
13500 (wmmxt_transfer_c1): Likewise.
13501 (wmmxt_transfer_c2): Likewise.
13502 (wmmxt_transfer_c3): Likewise.
13503 (marvell_f_iwmmxt_wstr): Likewise.
13504 (marvell_f_iwmmxt_wldr): Likewise.
13505
bdb7bf8a
YZ		 135062013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13507
0ba6aff3 13508 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 13509
e6631d38
VM		 135102013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13511
13512 Revert:
13513 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13514 * lra-constraints.c (need_for_split_p): Check call used hard regs
13515 living through calls.
13516
13517 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13518 call used regs for call insn.
13519
53c1275b
JJ		 135202013-06-28 Jakub Jelinek <jakub@redhat.com>
13521
13522 PR target/57736
0ba6aff3
UB		 13523 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13524 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 13525
1141ed3f
BI		 135262013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13527
13528 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 13529
e32e4c4a
VM		 135302013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13531
13532 * lra-constraints.c (need_for_split_p): Check call used hard regs
13533 living through calls.
13534
f161bfd3
MM		 135352013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13536
13537 PR target/57744
13538 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13539 to tie with any other modes. Eliminate Altivec vector mode tests,
13540 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13541 code, to return 0 if testing MODE2 for a condition, if we've
13542 already tested MODE1 for the same condition.
13543
7eda14e1
MS		 135442013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13545
13546 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13547 layout.
13548
c822f852
MS		 135492013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13550
13551 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13552 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13553
da4f13a4
MS		 135542013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13555
0ba6aff3
UB		 13556 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13557 Define.
da4f13a4
MS		 13558 (aarch64_symbolic_constant_p): Remove.
13559 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13560 static. Fix line length and white space.
13561 (aarch64_symbolic_constant_p): Remove.
13562 * config/aarch64/predicates.md (aarch64_valid_symref):
13563 Use aarch64_classify_symbol_expression.
13564
956a95a5
KT		 135652013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13566
13567 * config/arm/constraints.md (Ts): New constraint.
13568 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13569 16-bit encodings.
13570 (compare_scc): Use "Ts" constraint for operand 0.
13571 (ior_scc_scc): Likewise.
13572 (and_scc_scc): Likewise.
13573 (and_scc_scc_nodom): Likewise.
13574 (ior_scc_scc_cmp): Likewise for operand 7.
13575 (and_scc_scc_cmp): Likewise.
13576 * config/arm/thumb2.md (thumb2_movsi_insn):
13577 Add alternatives for 16-bit encodings.
13578 (thumb2_movhi_insn): Likewise.
13579 (thumb2_movsicc_insn): Likewise.
13580 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13581 (thumb2_negscc): Use "Ts" constraint.
13582 Move mvn instruction outside cond_exec block.
13583 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13584 for 16-bit encodings.
13585
31dfce10
KT		 135862013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13587
13588 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13589 encoding.
13590 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13591 (mulsi3subsi): Likewise.
13592 (mulsidi3adddi): Likewise.
13593 (mulsidi3_v6): Likewise.
13594 (umulsidi3_v6): Likewise.
13595 (umulsidi3adddi_v6): Likewise.
13596 (smulsi3_highpart_v6): Likewise.
13597 (umulsi3_highpart_v6): Likewise.
13598 (mulhisi3tb): Likewise.
13599 (mulhisi3bt): Likewise.
13600 (mulhisi3tt): Likewise.
13601 (maddhisi4): Likewise.
13602 (maddhisi4tb): Likewise.
13603 (maddhisi4tt): Likewise.
13604 (maddhidi4): Likewise.
13605 (maddhidi4tb): Likewise.
13606 (maddhidi4tt): Likewise.
13607 (zeroextractsi_compare0_scratch): Likewise.
13608 (insv_zero): Likewise.
13609 (insv_t2): Likewise.
13610 (anddi_notzesidi_di): Likewise.
13611 (anddi_notsesidi_di): Likewise.
13612 (andsi_notsi_si): Likewise.
13613 (iordi_zesidi_di): Likewise.
13614 (xordi_zesidi_di): Likewise.
13615 (andsi_iorsi3_notsi): Likewise.
13616 (smax_0): Likewise.
13617 (smax_m1): Likewise.
13618 (smin_0): Likewise.
13619 (not_shiftsi): Likewise.
13620 (unaligned_loadsi): Likewise.
13621 (unaligned_loadhis): Likewise.
13622 (unaligned_loadhiu): Likewise.
13623 (unaligned_storesi): Likewise.
13624 (unaligned_storehi): Likewise.
13625 (extv_reg): Likewise.
13626 (extzv_t2): Likewise.
13627 (divsi3): Likewise.
13628 (udivsi3): Likewise.
13629 (arm_zero_extendhisi2addsi): Likewise.
13630 (arm_zero_extendqisi2addsi): Likewise.
13631 (compareqi_eq0): Likewise.
13632 (arm_extendhisi2_v6): Likewise.
13633 (arm_extendqisi2addsi): Likewise.
13634 (arm_movt): Likewise.
13635 (thumb2_ldrd): Likewise.
13636 (thumb2_ldrd_base): Likewise.
13637 (thumb2_ldrd_base_neg): Likewise.
13638 (thumb2_strd): Likewise.
13639 (thumb2_strd_base): Likewise.
13640 (thumb2_strd_base_neg): Likewise.
13641 (arm_negsi2): Add alternative for 16-bit encoding.
13642 (arm_one_cmplsi2): Likewise.
13643
1572e697
KT		 136442013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13645
13646 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13647 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13648 (movdfcc): Likewise.
13649 * config/arm/vfp.md (*thumb2_movsf_vfp):
13650 Disable predication for arm_restrict_it.
13651 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13652 (*thumb2_movdfcc_vfp): Likewise.
13653 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13654 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13655 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13656 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13657 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13658 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13659 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13660 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13661 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13662 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13663 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13664 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13665 Disable predication for arm_restrict_it.
13666
ebf8f0ea
KY		 136672013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13668
13669 * config/i386/bmiintrin.h (_bextr_u32): New.
13670 (_bextr_u64): Ditto.
13671
a290fcda
RS		 136722013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13673
13674 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13675 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13676 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13677 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13678 * config/mips/n32-elf.h: ...this new file.
13679
ea169996
MG		 136802013-06-27 Marc Glisse <marc.glisse@inria.fr>
13681
13682 PR target/57224
13683 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13684 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13685
12211b99 136862013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM		 13687
13688 * config/mips/mips-tables.opt: Regenerate.
13689 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 13690 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 13691 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 13692
2444244e
JJ		 136932013-06-27 Jakub Jelinek <jakub@redhat.com>
13694
cd06d2a2
JJ		 13695 PR target/57623
13696 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13697 constraints of operand 1 and 2.
13698
2444244e
JJ		 13699 PR target/57623
13700 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13701 to match RTL canonicalization. Swap predicates and
13702 constraints of operand 1 and 2.
13703
2f259720
VM		 137042013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13705
0ba6aff3
UB		 13706 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13707 Process OP_INOUT regs for splitting too.
2f259720 13708
fef4d2b3
JJ		 137092013-06-27 Jakub Jelinek <jakub@redhat.com>
13710
13711 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13712 decl before the loop, initialize to NULL.
13713 (vectorizable_load): Initialize ptr_incr to NULL.
13714
79ee9826
MJ		 137152013-06-27 Martin Jambor <mjambor@suse.cz>
13716
13717 PR lto/57208
13718 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13719 * ipa-ref.c (ipa_maybe_record_reference): New function.
13720 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13721 * ipa-cp.c (create_specialized_node): Record potential references from
13722 aggvals.
13723 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13724
e18b4a81
YZ		 137252013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13726
13727 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13728 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13729 'mode' to force_reg.
13730 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13731 (aarch64_expand_mov_immediate): Likewise.
13732
9c023bf0
YZ		 137332013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13734
13735 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13736 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13737
5a3fe9b6
AK		 137382013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13739
13740 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13741 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13742 (struct machine_function): Add tbegin_p.
13743 (s390_canonicalize_comparison): Fold CC mode compares to
13744 conditional jump if possible.
13745 (s390_emit_jump): Return the emitted jump.
13746 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13747 Handle CCRAWmode compares.
13748 (s390_option_override): Default to -mhtm if available.
13749 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13750 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13751 FPRs instead of df_regs_ever_live_p.
13752 (s390_optimize_nonescaping_tx): New function.
13753 (s390_init_frame_layout): Extend clobbered_regs array to cover
13754 FPRs as well.
13755 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13756 (s390_expand_tbegin): New function.
13757 (enum s390_builtin): New enum definition.
13758 (code_for_builtin): New array definition.
13759 (s390_init_builtins): New function.
13760 (s390_expand_builtin): New function.
13761 (TARGET_INIT_BUILTINS): Define.
13762 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 13763 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK		 13764 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13765 (s390_alc_comparison): Likewise.
13766 * config/s390/s390-modes.def: Add CCRAWmode.
13767 * config/s390/s390.h (processor_flags): Add PF_TX.
13768 (TARGET_CPU_HTM): Define macro.
13769 (TARGET_HTM): Define macro.
13770 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13771 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13772 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13773 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 13774 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK		 13775 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13776 ("*cc_to_int"): Move up.
13777 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13778 constants other than 0.
13779 ("*ccraw_to_int"): New insn and splitter definition.
13780 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13781 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13782 ("tx_assist"): New expander.
13783 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13784 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13785 * config/s390/s390.opt: Add -mhtm option.
13786 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13787 * config/s390/htmxlintrin.h: New file.
13788 * config/s390/htmintrin.h: New file.
13789 * config/s390/s390intrin.h: New file.
13790 * doc/extend.texi: Document htm builtins.
13791 * config.gcc: Add the new header files to extra_headers.
13792
9e216629
TS		 137932013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13794
13795 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 13796 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 13797
fd8c65e7 137982013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 13799 Pat Haugen <pthaugen@us.ibm.com>
13800 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM		 13801
13802 * config/rs6000/power8.md: New.
13803 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13804 setting for power8 entry.
13805 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13806 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13807 test for Power4/Power5 only.
13808 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13809 support.
13810 (force_new_group): Adjust comment.
13811 * config/rs6000/rs6000.md: Include power8.md.
13812
b24a2ce5
GY		 138132013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13814
13815 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13816 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13817 declaration.
13818 (tune_params): Update comment.
13819 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13820 (arm_max_conditional_execute): New function.
13821 (thumb2_final_prescan_insn): Use max_insn_skipped and
13822 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13823
f6fdeb62
JJ		 138242013-06-25 Jakub Jelinek <jakub@redhat.com>
13825
13826 PR tree-optimization/57705
13827 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13828 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 13829 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ		 13830 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13831
c13bc3d9
MJ		 138322013-06-25 Martin Jambor <mjambor@suse.cz>
13833
13834 PR middle-end/57670
13835 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13836 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13837 calls in the dump.
13838 (ipa_note_param_call): Initialize member_ptr flag.
13839 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13840 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13841 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13842 (ipa_read_indirect_edge_info): Likewise.
13843
8a845901
RB		 138442013-06-25 Richard Biener <rguenther@suse.de>
13845
13846 PR middle-end/56977
13847 * passes.c (init_optimization_passes): Move pass_fold_builtins
13848 and pass_dce earlier with -Og.
13849
33be0bec
EB		 138502013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13851
13852 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13853 <BIT_FIELD_REF>: Remove trailing TAB.
13854 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13855 remove blank line.
13856
5fe8e757
MJ		 138572013-06-24 Martin Jambor <mjambor@suse.cz>
13858
13859 PR tree-optimization/57358
13860 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13861 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13862 (ipa_analyze_params_uses): Generate pessimistic info when true.
13863
44a60244
MJ		 138642013-06-24 Martin Jambor <mjambor@suse.cz>
13865
13866 PR tree-optimization/57539
13867 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13868 global.inlined_to of the new node to it. All callers changed.
13869 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13870 inlining_into, pass it to cgraph_clone_node.
13871 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13872 ipa_free_edge_args_substructures.
13873 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13874 rdesc linked list. Do not assert rdesc edges have inlined caller.
13875 Assert we have found an rdesc in the rdesc list.
13876
7c5848b8
RB		 138772013-06-24 Richard Biener <rguenther@suse.de>
13878
0ba6aff3 13879 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB		 13880 (pointer_set_lookup): Declare.
13881 (class pointer_map): New template class implementing a
13882 generic pointer to T map.
13883 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13884 pointer_map<T>::contains, pointer_map<T>::insert,
13885 pointer_map<T>::traverse): New functions.
13886 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13887 (pointer_set_lookup): New function.
13888 (pointer_set_contains): Use pointer_set_lookup.
13889 (pointer_set_insert): Likewise.
13890 (insert_aux): Remove.
13891 (struct pointer_map_t): Embed a pointer_set_t.
13892 (pointer_map_create): Adjust.
13893 (pointer_map_destroy): Likewise.
13894 (pointer_map_contains): Likewise.
13895 (pointer_map_insert): Likewise.
13896 (pointer_map_traverse): Likewise.
13897 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13898 pointer_map<unsigned> instead of a pointer_map_t.
13899 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13900 (streamer_tree_cache_lookup): Likewise.
13901 (streamer_tree_cache_create): Likewise.
13902 (streamer_tree_cache_delete): Likewise.
13903 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13904 pointer_map<unsigned> instead of a pointer_map_t.
13905 (lto_init_tree_ref_encoder): Adjust.
13906 (lto_destroy_tree_ref_encoder): Likewise.
13907 * lto-section-out.c (lto_output_decl_index): Likewise.
13908 (lto_record_function_out_decl_state): Likewise.
13909 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13910
3dbc97a9
RB		 139112013-06-24 Richard Biener <rguenther@suse.de>
13912
13913 PR tree-optimization/57488
13914 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13915
8345965a
AM		 139162013-06-24 Alan Modra <amodra@gmail.com>
13917
13918 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13919 (gen_easy_altivec_constant): Likewise.
13920 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13921 easy_vector_constant_msb): Likewise.
13922
b7df379f
JJ		 139232013-06-23 Jakub Jelinek <jakub@redhat.com>
13924
13925 PR target/57688
13926 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13927 add missing return true.
13928
12211b99 139292013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE		 13930
13931 PR target/52483
13932 * config/sh/predicates.md (general_extend_operand): Invoke
13933 general_movsrc_operand for memory operands.
13934 (general_movsrc_operand): Allow reg+reg addressing, do not use
13935 general_operand for memory operands.
13936
97db2bf7
ST		 139372013-06-23 Sriraman Tallam <tmsriram@google.com>
13938
13939 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13940 when current target options does not apply.
13941 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13942 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB		 13943 * config/i386/bmiintrin.h: Pass appropriate target
13944 attributes to header.
97db2bf7
ST		 13945 * config/i386/mmintrin.h: Ditto.
13946 * config/i386/nmmintrin.h: Ditto.
13947 * config/i386/avx2intrin.h: Ditto.
13948 * config/i386/fxsrintrin.h: Ditto.
13949 * config/i386/tbmintrin.h: Ditto.
13950 * config/i386/xsaveintrin.h: Ditto.
13951 * config/i386/f16cintrin.h: Ditto.
13952 * config/i386/xtestintrin.h: Ditto.
13953 * config/i386/xsaveoptintrin.h: Ditto.
13954 * config/i386/bmi2intrin.h: Ditto.
13955 * config/i386/lzcntintrin.h: Ditto.
13956 * config/i386/smmintrin.h: Ditto.
13957 * config/i386/wmmintrin.h: Ditto.
13958 * config/i386/x86intrin.h: Remove all header include guards.
13959 * config/i386/prfchwintrin.h: Ditto.
13960 * config/i386/pmmintrin.h: Ditto.
13961 * config/i386/tmmintrin.h: Ditto.
13962 * config/i386/xmmintrin.h: Ditto.
13963 * config/i386/popcntintrin.h: Ditto.
13964 * config/i386/rdseedintrin.h: Ditto.
13965 * config/i386/ammintrin.h: Ditto.
13966 * config/i386/emmintrin.h: Ditto.
13967 * config/i386/immintrin.h: Remove all header include guards.
13968 * config/i386/fma4intrin.h: Ditto.
13969 * config/i386/lwpintrin.h: Ditto.
13970 * config/i386/xopintrin.h: Ditto.
13971 * config/i386/ia32intrin.h: Ditto.
13972 * config/i386/avxintrin.h: Ditto.
13973 * config/i386/rtmintrin.h: Ditto.
13974 * config/i386/fmaintrin.h: Ditto.
13975 * config/i386/mm3dnow.h: Ditto.
13976
495e6879
ST		 139772013-06-22 Sriraman Tallam <tmsriram@google.com>
13978
13979 * common/config/i386/i386-common.c: Handle LZCNT.
13980
11e8f687
AK		 139812013-06-22 Andi Kleen <ak@linux.intel.com>
13982
13983 * doc/extend.texi: Use __atomic_store_n instead of
13984 __atomic_store in HLE example.
13985
12211b99 139862013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE		 13987
13988 * config/sh/sh.c: Remove <cstdlib> workaround.
13989
927f908b
AK		 139902013-06-21 Andi Kleen <ak@linux.intel.com>
13991
0ba6aff3 13992 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 13993
cca410b6
AK		 139942013-06-21 Andi Kleen <ak@linux.intel.com>
13995
13996 * doc/extend.texi: Document that __atomic_clear and
13997 __atomic_test_and_set should only be used with bool.
13998
d84db1ed
JH		 139992013-06-20 Jan Hubicka <jh@suse.cz>
14000
14001 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14002 types_same_for_odr.
14003 * tree.c (decls_same_for_odr): New function.
14004 (same_for_edr): New function.
14005 (types_same_for_odr): New function.
14006 (get_binfo_at_offset): Use it.
14007 * tree.h (types_same_for_odr): Declare.
14008
12211b99 140092013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE		 14010 Jason Merrill <jason@redhat.com>
14011
14012 * system.h: Include <cstdlib> as well as <stdlib.h>.
14013
49b4ff63
UB		 140142013-06-20 Uros Bizjak <ubizjak@gmail.com>
14015
14016 PR target/57655
14017 * config/i386/i386.c (construct_container): Report error if
14018 long double is used with disabled x87 float returns.
14019
7a89ad00
JH		 140202013-06-20 Jan Hubicka <jh@suse.cz>
14021
14022 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14023
12211b99 140242013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR		 14025
14026 PR rtl-optimization/57425
14027 PR rtl-optimization/57569
14028 * alias.c (write_dependence_p): Remove parameters mem_mode and
14029 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14030 Changed all callers.
14031 (canon_anti_dependence): Get comments and semantics in sync.
14032 Add parameter mem_canonicalized. Changed all callers.
14033 * rtl.h (canon_anti_dependence): Update prototype.
14034
d16e9a99
RB		 140352013-06-20 Richard Biener <rguenther@suse.de>
14036
14037 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14038 byte case, inline streamer_read_uchar and defer section
14039 overrun check.
14040
c3a9b91b
RB		 140412013-06-20 Richard Biener <rguenther@suse.de>
14042
14043 PR tree-optimization/57584
14044 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14045 SSA names into the expanded expression that take part in
14046 abnormal coalescing.
14047
371e77e3 140482013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS		 14049
14050 * gcov.c (print_usage): Handle new option.
14051 (process_args): Ditto.
14052 (get_gcov_intermediate_filename): New function.
14053 (output_intermediate_file): New function.
14054 (output_gcov_file): New function
14055 (generate_results): Handle new option.
14056 (release_function): Relase demangled name.
14057 (read_graph_file): Handle demangled name.
14058 (output_lines): Ditto.
14059 * doc/gcov.texi: Document gcov intermediate format.
14060
11f5b71a
VM		 140612013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14062
14063 PR bootstrap/57604
14064 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14065 (lra_emit_add): Use the functions. Add comment about Y as an
14066 address segment.
14067
e1f9340b
DE		 140682013-06-19 David Edelsohn <dje.gcc@gmail.com>
14069
14070 PR driver/57652
14071 * collect2.c (collect_atexit): New.
14072 (collect_exit): Delete.
14073 (main): Register collect_atexit with atexit.
14074 (collect_wait): Change collect_exit to exit.
14075 (do_wait): Same.
14076 * collect2.h (collect_exit): Delete.
14077 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14078
e3f9e0ac
WM		 140792013-06-19 Wei Mi <wmi@google.com>
14080
14081 PR rtl-optimization/57518
14082 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14083 if regno is used in paradoxical subreg.
14084 (update_equiv_regs): Check pdx_subregs[regno] before
14085 set a reg to be equivalent with a mem.
14086
a09f1a76
MK		 140872013-06-19 Matthias Klose <doko@ubuntu.com>
14088
14089 PR driver/57651
14090 * file-find.h (find_a_file): Add a mode parameter.
14091 * file-find.c (find_a_file): Likewise.
14092 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14093 with X_OK for the executables.
14094 * collect2.c (main): Call find_a_file with X_OK.
14095
7d18b0ad
SE		 140962013-06-19 Steve Ellcey <sellcey@mips.com>
14097
14098 PR target/56942
14099 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14100 Use NEXT_INSN instead of next_real_insn.
14101
6a6dac52
JH		 141022013-06-19 Jan Hubicka <jh@suse.cz>
14103
14104 * cgraph.h (const_value_known_p): Replace by ...
14105 (ctor_for_folding): .. this one.
14106 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14107 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14108 * expr.c (expand_expr_real_1): Likewise.
14109 (string_constant): Likewise.
14110 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14111 * ipa.c (process_references): Likewise.
14112 (symtab_remove_unreachable_nodes): Likewise.
14113 * ipa-inline-analysis.c (param_change_prob): Likewise.
14114 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14115 (get_base_constructor): Likwise.
14116 * varpool.c (varpool_remove_node): Likewise.
14117 (varpool_remove_initializer): LIkewise.
14118 (dump_varpool_node): LIkwise.
14119 (const_value_known_p): Rewrite to ...
14120 (ctor_for_folding): ... this one.
14121
216c12ab
JJ		 141222013-06-19 Jakub Jelinek <jakub@redhat.com>
14123
14124 PR driver/57651
14125 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14126 PERSONALITY in $PATH derived prefixes.
14127
d53e2f99
JL		 141282013-06-19 Jeff Law <law@redhat.com>
14129
ecdbf306
JL		 14130 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14131 in comment.
14132
d53e2f99
JL		 14133 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14134 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14135 booleans.
14136
89e7fc04
SN		 141372013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14138
14139 * config/arm/vfp.md: Move VFP instruction classification documentation
14140 to ...
14141 * config/arm/arm.md: ... here. Update instruction classification
14142 documentation.
14143
5e7f6aaa
RE		 141442013-06-19 Richard Earnshaw <rearnsha@arm.com>
14145
14146 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14147 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14148 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14149
5fdc1e5d
SB		 141502013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14151
14152 PR target/57609
14153 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14154 with NEXT_INSN. Use tablejump_p to check for jump table data
14155 insns.
14156
e952afba
PC		 141572013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14158
14159 PR c++/56544
14160 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14161 that now in C++ the value is correct per the C++ standards.
14162
3fd0303a
RB		 141632013-06-19 Richard Biener <rguenther@suse.de>
14164
14165 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14166 for global context.
14167
b04e4b08
AK		 141682013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14169
14170 Revert:
14171 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14172
14173 PR target/57609
14174 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14175 with next_active_insn.
14176
bef8491a
ST		 141772013-06-18 Sriraman Tallam <tmsriram@google.com>
14178
14179 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14180 functions are inlined during failures to flag an error.
14181 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14182 in early inline pass.
22fc7d3d 14183
c21b257d
L		 141842013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14185
14186 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14187 in comments.
14188
33255ae3
JB		 141892013-06-18 Julian Brown <julian@codesourcery.com>
14190
14191 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14192 Permit virtual register pre-reload if !strict.
14193 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14194 change.
14195 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14196 prototype.
14197 * config/arm/neon.md (movmisalign<mode>): Use
14198 neon_perm_struct_or_reg_operand instead of
14199 neon_struct_or_register_operand.
14200 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14201 neon_permissive_struct_operand instead of neon_struct_operand.
14202 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14203 neon_vector_mem_operand.
14204 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14205 neon_vector_mem_operand.
14206 (neon_permissive_struct_operand): New.
14207 (neon_struct_or_register_operand): Rename to...
14208 (neon_perm_struct_or_reg_operand): This. Adjust call to
14209 neon_vector_mem_operand.
14210
d579fcda
RB		 142112013-06-18 Richard Biener <rguenther@suse.de>
14212
14213 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14214 * lto-streamer.h: Include pointer-set.h.
14215 (struct lto_decl_slot): Remove.
14216 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14217 Remove next_index entry.
14218 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14219 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14220 (lto_init_tree_ref_encoder): Adjust.
14221 (lto_destroy_tree_ref_encoder): Likewise.
14222 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14223 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14224 (lto_output_decl_index): Adjust.
14225 (lto_new_out_decl_state): Likewise.
14226 (lto_record_function_out_decl_state): Likewise.
14227 * lto-streamer-out.c (copy_function): Likewise.
14228
31ee20ba
RB		 142292013-06-18 Richard Biener <rguenther@suse.de>
14230
14231 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14232 * cgraphunit.c: Include cfgloop.h.
14233 (init_lowered_empty_function): Initialize the loop tree.
14234 (assemble_thunk): Insert new BBs into loops.
14235
ac0511f2
RB		 142362013-06-18 Richard Biener <rguenther@suse.de>
14237
14238 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14239 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14240 the map from cache entry to cache index optional.
14241 (streamer_tree_cache_replace_tree): Adjust accordingly.
14242 (streamer_tree_cache_append): Likewise.
14243 (streamer_tree_cache_delete): Likewise.
14244 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14245 streamer cache map from cache entry to cache index.
14246 * lto-streamer-out.c (create_output_block): Adjust.
14247
09485a08
SN		 142482013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14249
22fc7d3d
UB		 14250 * config/arm/arm.md (attribute "insn"): Move multiplication and
14251 division attributes to...
09485a08
SN		 14252 (attribute "type"): ... here. Remove mult.
14253 (attribute "mul32"): New attribute.
14254 (attribute "mul64"): Add umaal.
14255 (*arm_mulsi3): Update attributes.
14256 (*arm_mulsi3_v6): Likewise.
14257 (*thumb_mulsi3): Likewise.
14258 (*thumb_mulsi3_v6): Likewise.
14259 (*mulsi3_compare0): Likewise.
14260 (*mulsi3_compare0_v6): Likewise.
14261 (*mulsi_compare0_scratch): Likewise.
14262 (*mulsi_compare0_scratch_v6): Likewise.
14263 (*mulsi3addsi): Likewise.
14264 (*mulsi3addsi_v6): Likewise.
14265 (*mulsi3addsi_compare0): Likewise.
14266 (*mulsi3addsi_compare0_v6): Likewise.
14267 (*mulsi3addsi_compare0_scratch): Likewise.
14268 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14269 (*mulsi3subsi): Likewise.
14270 (*mulsidi3adddi): Likewise.
14271 (*mulsi3addsi_v6): Likewise.
14272 (*mulsidi3adddi_v6): Likewise.
14273 (*mulsidi3_nov6): Likewise.
14274 (*mulsidi3_v6): Likewise.
14275 (*umulsidi3_nov6): Likewise.
14276 (*umulsidi3_v6): Likewise.
14277 (*umulsidi3adddi): Likewise.
14278 (*umulsidi3adddi_v6): Likewise.
14279 (*smulsi3_highpart_nov6): Likewise.
14280 (*smulsi3_highpart_v6): Likewise.
14281 (*umulsi3_highpart_nov6): Likewise.
14282 (*umulsi3_highpart_v6): Likewise.
14283 (mulhisi3): Likewise.
14284 (*mulhisi3tb): Likewise.
14285 (*mulhisi3bt): Likewise.
14286 (*mulhisi3tt): Likewise.
14287 (maddhisi4): Likewise.
14288 (*maddhisi4tb): Likewise.
14289 (*maddhisi4tt): Likewise.
14290 (maddhidi4): Likewise.
14291 (*maddhidi4tb): Likewise.
14292 (*maddhidi4tt): Likewise.
14293 (divsi3): Likewise.
14294 (udivsi3): Likewise.
14295 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14296 (thumb2_mulsi_short_compare0): Likewise.
14297 (thumb2_mulsi_short_compare0_scratch): Likewise.
14298 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14299 (1020mult2): Likewise.
14300 (1020mult3): Likewise.
14301 (1020mult4): Likewise.
14302 (1020mult5): Likewise.
14303 (1020mult6): Likewise.
22fc7d3d
UB		 14304 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14305 change.
09485a08
SN		 14306 (cortex_a15_mult64): Likewise.
14307 (cortex_a15_sdiv): Likewise.
14308 (cortex_a15_udiv): Likewise.
14309 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14310 (mult2): Likewise.
14311 (mult3): Likewise.
14312 (mult4): Likewise.
14313 (mult5): Likewise.
14314 (mult6): Likewise.
14315 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14316 (pj4_ir_div): Likewise.
14317 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14318 (11_mult2): Likewise.
14319 (11_mult3): Likewise.
14320 (11_mult4): Likewise.
14321 (11_mult5): Likewise.
14322 (11_mult6): Likewise.
14323 (11_mult7): Likewise.
14324 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14325 (cortex_a8_mla): Likewise.
14326 (cortex_a8_mull): Likewise.
14327 (cortex_a8_smulwy): Likewise.
14328 (cortex_a8_smlald): Likewise.
14329 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14330 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14331 (cortex_r4_mul_3): Likewise.
14332 (cortex_r4_mla_4): Likewise.
14333 (cortex_r4_mla_3): Likewise.
14334 (cortex_r4_smlald): Likewise.
14335 (cortex_r4_mull): Likewise.
14336 (cortex_r4_sdiv): Likewise.
14337 (cortex_r4_udiv): Likewise.
14338 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14339 (cortex_a7_idiv): Likewise.
14340 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14341 (9_mult2): Likewise.
14342 (9_mult3): Likewise.
14343 (9_mult4): Likewise.
14344 (9_mult5): Likewise.
14345 (9_mult6): Likewise.
14346 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14347 (cortex_a53_sdiv): Likewise.
14348 (cortex_a53_udiv): Likewise.
14349 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14350 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14351 (mp626_mult2): Likewise.
14352 (mp626_mult3): Likewise.
14353 (mp626_mult4): Likewise.
14354 * config/arm/fa526.md (526_mult1): Update attribute change.
14355 (526_mult2): Likewise.
14356 * config/arm/arm-generic.md (mult): Update attribute change.
14357 (mult_ldsched_strongarm): Likewise.
14358 (mult_ldsched): Likewise.
14359 (multi_cycle): Likewise.
14360 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14361 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14362 (606te_mult2): Likewise.
14363 (606te_mult3): Likewise.
14364 (606te_mult4): Likewise.
14365 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14366 (cortex_a9_mac16): Likewise.
14367 (cortex_a9_multiply): Likewise.
14368 (cortex_a9_mac): Likewise.
14369 (cortex_a9_multiply_long): Likewise.
14370 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14371 (626te_mult2): Likewise.
14372 (626te_mult3): Likewise.
14373 (626te_mult4): Likewise.
14374
8951f345
RB		 143752013-06-18 Richard Biener <rguenther@suse.de>
14376
14377 PR lto/57334
14378 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14379
d8101d05
AK		 143802013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14381
14382 PR target/57609
14383 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14384 with next_active_insn.
14385
69eff9da
AM		 143862013-06-18 Alan Modra <amodra@gmail.com>
14387
14388 * config/rs6000/rs6000.h (enum data_align): New.
14389 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14390 (DATA_ABI_ALIGNMENT): Define.
14391 (CONSTANT_ALIGNMENT): Correct comment.
14392 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14393 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14394
8ed00d76
DM		 143952013-06-17 David Malcolm <dmalcolm@redhat.com>
14396
14397 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14398 ATTRIBUTE_UNUSED marking.
14399
3d33d06b
SN		 144002013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14401
14402 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14403 alternative and update.
14404 (aarch64_dup_lanedi): Delete.
14405 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14406 * config/aarch64/aarch64-simd-builtins.def: Update.
14407
ee03e71d
RB		 144082013-06-17 Richard Biener <rguenther@suse.de>
14409
14410 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14411 (lto_input_scc): Declare.
14412 (lto_input_tree_1): Likewise.
14413 (struct lto_stats_d): Add num_tree_bodies_output and
14414 num_pickle_refs_output.
14415 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14416 (lto_read_tree_1): Split out from ...
14417 (lto_read_tree): ... this.
14418 (lto_input_scc): New function.
14419 (lto_input_tree_1): Split out from ...
14420 (lto_input_tree): ... this. Handle LTO_tree_scc.
14421 (lto_data_in_create): Create the streamer cache without hashes.
14422 * lto-streamer-out.c (create_output_block): Create the streamer
14423 cache with hashes when not doing WPA.
14424 (lto_write_tree_1): Split out from ...
14425 (lto_write_tree): ... this.
14426 (get_symbol_initial_value): New function.
14427 (lto_output_tree_1): Split out from ...
14428 (lto_output_tree): ... this. Write trees as series of SCCs
14429 using a DFS walk via DFS_write_tree.
14430 (struct sccs, struct scc_entry): New types.
14431 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14432 (DFS_write_tree_body): New function.
14433 (DFS_write_tree): Likewise.
14434 (hash_tree): Likewise.
14435 (scc_entry_compare): Likewise.
14436 (hash_scc): Likewise.
14437 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14438 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14439 TREE_CHAIN as regular reference.
14440 (streamer_read_integer_cst): Remove.
14441 (streamer_get_pickled_tree): Adjust.
14442 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14443 of DECL_EXTERNALs in BLOCK_VARS for now.
14444 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14445 reference.
14446 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14447 Add hash value argument and record that if hashes are recorded
14448 in the cache.
14449 (streamer_tree_cache_insert_1): Adjust.
14450 (streamer_tree_cache_insert): Likewise.
14451 (streamer_tree_cache_insert_at): Rename to ...
14452 (streamer_tree_cache_replace_tree): ... this and adjust.
14453 (streamer_tree_cache_append): Adjust.
14454 (record_common_node): Likewise.
14455 (streamer_tree_cache_create): Add argument whether to
14456 record hash values together with trees.
14457 (streamer_tree_cache_delete): Adjust.
14458 * tree-streamer.h (struct streamer_tree_cache_d): Add
14459 vector of hashes.
14460 (streamer_read_integer_cst): Remove.
14461 (streamer_tree_cache_insert): Adjust.
14462 (streamer_tree_cache_append): Likewise.
14463 (streamer_tree_cache_insert_at): Rename to ...
14464 (streamer_tree_cache_replace_tree): ... this and adjust.
14465 (streamer_tree_cache_create): Add argument whether to record hashes.
14466 (streamer_tree_cache_get): Rename to ...
14467 (streamer_tree_cache_get_tree): ... this.
14468 (streamer_tree_cache_get_hash): New function.
14469 * tree.c (cache_integer_cst): New function.
14470 * tree.h (cache_integer_cst): Declare.
14471 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14472 * lto-symtab.c (lto_varpool_replace_node): Only release
14473 DECL_INITIAL of non-prevailing decls.
14474 * varpool.c (varpool_remove_initializer): Do not release
14475 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14476
0de86a92
JU		 144772013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14478
14479 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14480 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14481 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14482 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14483 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14484 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14485 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14486 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14487 instead of TARGET_64BIT.
14488 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14489 Require ISA_HAS_<D>DIV.
14490
74240413
RS		 144912013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14492
14493 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14494 (mips*-*-linux*): Move default with_llsc setting to where other
14495 defaults are set.
14496 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14497 with_arch block.
14498 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14499 Likewise. Remove default with_tune setting. Move default float
14500 setting to its own block. Handle with_llsc in the same block as above.
14501
12211b99 145022013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR		 14503
14504 PR rtl-optimization/57425
14505 PR rtl-optimization/57569
299a5f6a 14506 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR		 14507 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14508 Changed all callers.
14509 (canon_anti_dependence): New function.
14510 * cse.c (check_dependence): Use canon_anti_dependence.
14511 * cselib.c (cselib_invalidate_mem): Likewise.
14512 * rtl.h (canon_anti_dependence): Declare.
14513
b259d352
JU		 145142013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14515
14516 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14517 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14518 ".set mips3" for 64-bit targets.
14519
4de09b85
DC		 145202013-06-15 Dehao Chen <dehao@google.com>
14521
14522 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14523 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14524 (gimple_check_call_args): Likewise.
14525 * value-prof.c (check_ic_target): Likewise.
14526 * ipa-inline.c (early_inliner): Likewise.
14527 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14528 * cgraph.c (cgraph_create_edge_1): Likewise.
14529 (cgraph_make_edge_direct): Likewise.
14530
d1aa4a76
MM		 145312013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14532
14533 PR target/57615
14534 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14535 rs6000_output_move_128bit to handle emitting quad memory
14536 operations. Set attribute length to 8 bytes.
14537
12211b99 145382013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP		 14539
14540 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14541 New pattern.
14542 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14543 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14544 (aarch64_<su>mlsl<mode>): Likewise.
14545
f5273574
MS		 145462013-06-14 Mike Stump <mikestump@comcast.net>
14547
14548 * Makefile.in (TARGET_H): Add insn-codes.h.
14549
ef7e7ab2
AM		 145502013-06-14 Alan Modra <amodra@gmail.com>
14551
14552 PR middle-end/57134
14553 PR middle-end/57586
14554 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14555 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14556 bitfield expansion when EXPAND_MEMORY.
14557 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14558
135faab6
MM		 145592013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14560
14561 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14562 test for clearing quad memory on 32-bit later.
14563
948a1fd9
MG		 145642013-06-13 Marc Glisse <marc.glisse@inria.fr>
14565
14566 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14567 (fold_negate_expr): Likewise.
14568 (fold_real_zero_addition_p): Handle vectors.
14569 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14570
0597fb9c
AM		 145712013-06-14 Alan Modra <amodra@gmail.com>
14572
14573 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14574
e5b962d0
JH		 145752013-06-13 Jan Hubicka <jh@suse.cz>
14576
14577 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14578 Local comdats are not externally visible.
14579 * symtab.c (dump_symtab_base): Dump externally visible.
14580 (verify_symtab_base): Verify back links in the symtab hash.
14581
a8355e51
BC		 145822013-06-13 Bin Cheng <bin.cheng@arm.com>
14583
14584 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14585 CONVERT_EXPR as equal nodes.
14586
8f7e6e33
BC		 145872013-06-13 Bin Cheng <bin.cheng@arm.com>
14588
22fc7d3d 14589 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 14590
5d418483
MG		 145912013-06-13 Marc Glisse <marc.glisse@inria.fr>
14592
14593 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14594 Generalize to complex and vector.
14595 * tree.c (build_all_ones_cst): New function.
14596 * tree.h (build_all_ones_cst): Declare it.
14597
50751417
AM		 145982013-06-13 Alan Modra <amodra@gmail.com>
14599
14600 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14601 * config/rs6000/rs6000.md (signbittf2): New insn.
14602 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14603 (abstf2_internal, cmptf_internal2): Likewise.
14604 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14605
4b02c962 146062013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14607 Pat Haugen <pthaugen@us.ibm.com>
14608 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM		 14609
14610 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14611 power8 byte, half-word, and quad-word atomic instructions.
14612 (emit_store_conditional): Likewise.
14613 (rs6000_expand_atomic_compare_and_swap): Likewise.
14614 (rs6000_expand_atomic_op): Likewise.
14615
14616 * config/rs6000/sync.md (larx): Add new modes for power8.
14617 (stcx): Likewise.
14618 (AINT): New mode iterator to include TImode as well as normal
14619 integer modes on power8.
14620 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14621 that VSX registers are not considered. Use AINT mode iterator
14622 instead of INT1 to allow inclusion of quad word atomic operations
14623 on power8.
14624 (load_locked<mode>): Likewise.
14625 (store_conditional<mode>): Likewise.
14626 (atomic_compare_and_swap<mode>): Likewise.
14627 (atomic_exchange<mode>): Likewise.
14628 (atomic_nand<mode>): Likewise.
14629 (atomic_fetch_<fetchop_name><mode>): Likewise.
14630 (atomic_nand_fetch<mode>): Likewise.
14631 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14632 each type.
14633 (ATOMIC): On power8, add QImode, HImode modes.
14634 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14635 modes that promote to SImode.
14636 (load_lockedti): Convert TImode arguments to PTImode, so that we
14637 get a guaranteed even/odd register pair.
14638 (load_lockedpti): Likewise.
14639 (store_conditionalti): Likewise.
14640 (store_conditionalpti): Likewise.
14641
14642 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14643 atomic load/store instructions.
14644 (HSI): Likewise.
14645
710b6cc1
RS		 146462013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14647
14648 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14649 loads.
14650 (insn_count): New attribute, with most cases extracted from...
14651 (length): ...here. Redefine most cases in terms of insn_count.
14652 (single_insn): Delete.
14653 (can_delay): Use insn_count to check for single instructions.
14654 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14655 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14656 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14657 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14658 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14659 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14660 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14661 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14662 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14663 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14664 rather than "length".
14665 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14666 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14667 Use "insn_count" rather than "length".
14668 * config/mips/mips-dsp.md
14669 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14670 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14671 length attributes.
14672
1951f101
MG		 146732013-06-12 Marc Glisse <marc.glisse@inria.fr>
14674
14675 PR tree-optimization/57361
14676 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14677
8b033a8a
SN		 146782013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14679
22fc7d3d
UB		 14680 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14681 to split.
8b033a8a
SN		 14682 (aarch64_simd_combine<mode>): New instruction expansion.
14683 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14684 function prototype.
14685 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14686 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14687
ca0f62a8
JH		 146882013-06-12 Jan Hubicka <jh@suse.cz>
14689
14690 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14691 decl has when in streaming stage.
14692 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14693 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14694
12211b99 146952013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS		 14696
14697 PR target/57578
14698 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14699
bfdeda2c
JJ		 147002013-06-12 Jakub Jelinek <jakub@redhat.com>
14701
14702 PR tree-optimization/57537
14703 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14704 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14705
a4fa02d1
RB		 147062013-06-12 Richard Biener <rguenther@suse.de>
14707
14708 * data-streamer.h (streamer_write_char_stream): CSE
14709 obs->current_pointer.
14710 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14711 streamer_write_char_stream manually and optimize the resulting loop.
14712 (streamer_write_hwi_stream): Likewise.
14713
bbf9ad07
JH		 147142013-06-12 Jan Hubicka <jh@suse.cz>
14715
14716 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14717 * cgraph.h (varpool_create_empty_node): Declare.
14718 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14719 duplicated nodes.
14720 * symtab.c (symtab_unregister_node): Be lax about missin entries
14721 in node hash.
14722 (symtab_get_node): Update comment.
14723 * varpool.c (varpool_create_empty_node): Break out from ...
14724 (varpool_node_for_decl): ... here.
14725 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14726
a53f90ad
EB		 147272013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14728
14729 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14730 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14731 part. Use straight-line flow at the end.
14732 <COMPONENT_REF>: Remove superfluous else.
14733 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14734
946f9306
JJ		 147352013-06-12 Jakub Jelinek <jakub@redhat.com>
14736
14737 PR target/56564
14738 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14739 target hook even for !TREE_PUBLIC decls. If no resolution info
14740 is available, return false for common and external decls.
14741
a7fca6f0
KP		 147422013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14743
14744 * config/rl78/constraints.md (U): New constraint.
14745 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14746 valloc attribute.
14747
c7d8f446
MM		 147482013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14749
14750 PR target/57589
14751 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14752 to allow returning address to AT_PLATFORM name.
14753
08346abd
JH		 147542013-06-11 Jan Hubicka <jh@suse.cz>
14755
14756 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14757 * cgraph.h (symtab_node_base): Add weakref flag.
14758 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14759 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14760 (output_weakrefs): Use weakref flag.
14761 * fold-const.c (simple_operand_p): Handle WEAK.
14762 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14763 * ipa.c (varpool_externally_visible_p): Drop weakref.
14764 (function_and_variable_visibility): Update comment; fix weakref
14765 sanity checks; do not clear DECL_WEAK on them.
14766 * lto-cgraph.c (lto_output_node): update.
14767 (lto_output_varpool_node): Update.
14768 (input_overwrite_node): Update.
14769 (input_node): Update.
14770 (input_varpool_node): Update.
14771 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14772 (lto_symtab_merge_symbols): Add sanity check.
14773 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14774 * passes.c (rest_of_decl_compilation): Set static flag, too.
14775 * symtab.c (dump_symtab_base): Dump weakref.
14776 (verify_symtab_base): Sanity check weakrefs.
14777 (symtab_make_decl_local): Remove duplicated code.
14778 (symtab_alias_ultimate_target): Simplify.
14779 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 14780
2e6a2f88
TV		 147812013-06-11 Tom de Vries <tom@codesourcery.com>
14782
14783 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14784 sequence_vect == NULL.
14785
d9af3652
DD		 147862013-06-11 DJ Delorie <dj@redhat.com>
14787
14788 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14789 (rl78_unwind_word_mode): New.
14790
ffa66012
DM		 147912013-06-11 David Malcolm <dmalcolm@redhat.com>
14792
14793 * final.c (debug_prefix_maps): Make static.
14794
783f2d35
DM		 147952013-06-11 David Malcolm <dmalcolm@redhat.com>
14796
14797 * function.c (initial_trampoline): Remove stray copy.
14798
7af26def
SN		 147992013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14800
14801 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14802
97756c0e
MJ		 148032013-06-11 Martin Jambor <mjambor@suse.cz>
14804
14805 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14806 within bounds at the beginning of the function.
14807
2c7b8bf6
AM		 148082013-06-11 Alan Modra <amodra@gmail.com>
14809
14810 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14811 reporting.
14812 (get_named_section): Don't NULL !DECL_P decl.
14813
77cea46e
IZ		 148142013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14815
14816 * doc/invoke.texi (core-avx2): Document.
14817 (slm): Likewise.
14818 (atom): Updated with MOVBE.
14819
888f0920
RB		 148202013-06-11 Richard Biener <rguenther@suse.de>
14821
22fc7d3d 14822 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 14823
6b438d58
AB		 148242013-06-11 Anton Blanchard <anton@samba.org>
14825
14826 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14827 correct shift value in little-endian mode.
14828
3bc66938
JJ		 148292013-06-11 Jakub Jelinek <jakub@redhat.com>
14830
14831 PR target/56564
14832 * varasm.c (get_variable_align): Move #endif to the right place.
14833
a5dc7f88
CC		 148342013-06-10 Cary Coutant <ccoutant@google.com>
14835
14836 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14837 for hash so that hash table traversal order is deterministic.
14838
19be72ab 148392013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14840 Pat Haugen <pthaugen@us.ibm.com>
14841 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM		 14842
14843 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14844 of vectors in GPRS if they are direct moves or quad word load or
14845 store moves.
14846
14847 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14848 declaration.
14849 (direct_move_p): Likewise.
14850 (quad_load_store_p): Likewise.
14851
14852 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14853 classes into bins based on the physical register type.
14854 (reg_class_to_reg_type): Likewise.
14855 (IS_STD_REG_TYPE): Likewise.
14856 (IS_FP_VECT_REG_TYPE): Likewise.
14857 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14858 use direct move instructions.
14859 (reload_gpr_vsx): Likewise.
14860 (reload_vsx_gpr): Likewise.
14861 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14862 information that is a simplification of register classes. Also
14863 precalculate direct move reload helpers.
14864 (direct_move_p): New function to return true if the operation can
14865 be done as a direct move instruciton.
14866 (quad_load_store_p): New function to return true if the operation
14867 is a quad memory operation.
14868 (rs6000_legitimize_address): If quad memory, only allow register
14869 indirect for TImode addresses.
14870 (rs6000_legitimate_address_p): Likewise.
14871 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14872 (rs6000_reload_register_type): Likewise.
14873 (register_to_reg_type): Return register type.
14874 (rs6000_secondary_reload_simple_move): New helper function for
14875 secondary reload and secondary memory needed to identify anything
14876 that is a simple move, and does not need reloading.
14877 (rs6000_secondary_reload_direct_move): New helper function for
14878 secondary reload to identify cases that can be done with several
14879 instructions via the direct move instructions.
14880 (rs6000_secondary_reload_move): New helper function for secondary
14881 reload to identify moves between register types that can be done.
14882 (rs6000_secondary_reload): Add support for quad memory operations
14883 and for direct move.
14884 (rs6000_secondary_memory_needed): Likewise.
14885 (rs6000_debug_secondary_memory_needed): Change argument names.
14886 (rs6000_output_move_128bit): New function to return the move to
14887 use for 128-bit moves, including knowing about the various
14888 limitations of quad memory operations.
14889
14890 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14891 memory operations. call rs6000_output_move_128bit for the actual
14892 instruciton(s) to generate.
14893 (vsx_movti_64bit): Likewise.
14894
14895 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14896 (UNSPEC_P8V_MTVSRWZ): Likewise.
14897 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14898 (UNSPEC_P8V_MTVSRD): Likewise.
14899 (UNSPEC_P8V_XXPERMDI): Likewise.
14900 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14901 (UNSPEC_FUSION_GPR): Likewise.
14902 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 14903 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM		 14904 (f32_sv): Likewise.
14905 (f32_dm): Likewise.
14906 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14907 loads and direct move instructions.
14908 (zero_extendsidi2_lfiwzx): Likewise.
14909 (extendsidi2_lfiwax): Likewise.
14910 (extendsidi2_nocell): Likewise.
14911 (floatsi<mode>2_lfiwax): Likewise.
14912 (lfiwax): Likewise.
14913 (floatunssi<mode>2_lfiwzx): Likewise.
14914 (lfiwzx): Likewise.
14915 (fix_trunc<mode>_stfiwx): Likewise.
14916 (fixuns_trunc<mode>_stfiwx): Likewise.
14917 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14918 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14919 (parity<mode>2_cmpb): Set length/type attr.
14920 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14921 for 'mr.' to fast_compare.
14922 (bpermd_<mode>): Change type attr to popcnt.
14923 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14924 (p8_mtvsrwz_1): Likewise.
14925 (p8_mtvsrwz_2): Likewise.
14926 (reload_fpr_from_gpr<mode>): Likewise.
14927 (p8_mtvsrd_1): Likewise.
14928 (p8_mtvsrd_2): Likewise.
14929 (p8_xxpermdi_<mode>): Likewise.
14930 (reload_vsx_from_gpr<mode>): Likewise.
14931 (reload_vsx_from_gprsf): Likewise.
14932 (p8_mfvsrd_3_<mode>): LIkewise.
14933 (reload_gpr_from_vsx<mode>): Likewise.
14934 (reload_gpr_from_vsxsf): Likewise.
14935 (p8_mfvsrd_4_disf): Likewise.
14936 (multi-word GPR splits): Do not split direct moves or quad memory
14937 operations.
14938
16876bdc
DM		 149392013-06-10 David Malcolm <dmalcolm@redhat.com>
14940
14941 * tree-into-ssa.c (interesting_blocks): Make static.
14942
df8a1d28
JJ		 149432013-06-10 Jakub Jelinek <jakub@redhat.com>
14944
14945 PR target/56564
14946 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14947 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14948 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14949 (get_variable_align): New function.
14950 (get_variable_section, emit_bss, emit_common,
14951 assemble_variable_contents, place_block_symbol): Use
14952 get_variable_align instead of DECL_ALIGN.
14953 (assemble_noswitch_variable): Add align argument, use it
14954 instead of DECL_ALIGN.
14955 (assemble_variable): Adjust caller. Use get_variable_align
14956 instead of DECL_ALIGN.
14957 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14958 caller.
14959 (DATA_ABI_ALIGNMENT): Define.
14960 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14961 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14962 opt is false, only return the psABI mandated alignment increase.
14963 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14964 (DATA_ABI_ALIGNMENT): ... this.
14965 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14966 (DATA_ABI_ALIGNMENT): ... this.
14967 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14968 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14969 (DATA_ABI_ALIGNMENT): ... this.
14970 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14971 * doc/tm.texi: Regenerated.
14972
57b29ca6
UB		 149732013-06-10 Uros Bizjak <ubizjak@gmail.com>
14974
14975 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14976 cmp_code to construct REG_EQUAL note.
14977
3e56ed50
JJ		 149782013-06-09 Jakub Jelinek <jakub@redhat.com>
14979
14980 PR target/57568
14981 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14982 that operands[2] doesn't overlap with operands[0].
14983
74fe2a1d
DE		 149842013-06-09 David Edelsohn <dje.gcc@gmail.com>
14985 Jan Hubicka <jh@suse.cz>
14986
14987 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14988 hack to mark symbols as used.
14989
ed52a84e
VM		 149902013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14991
14992 PR rtl-optimization/57559
14993 * lra-constraints.c (process_alt_operands): Don't discourage
14994 memory with known offset for offsetable memory constraint.
14995 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14996
299316ed
EB		 149972013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14998
14999 * varasm.c (struct oc_local_state): Reorder fields.
15000 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15001 and adjust accordingly.
15002 (output_constructor): Reorder initialization code and adjust call to
15003 output_constructor_bitfield.
15004
88ac60d3
JH		 150052013-06-07 Jan Hubicka <jh@suse.cz>
15006
15007 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15008
4cdce1a8
DM		 150092013-06-07 David Malcolm <dmalcolm@redhat.com>
15010
15011 * tree-object-size.c (unknown): Make const.
15012
b5e0425c
AK		 150132013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15014
15015 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15016 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15017 for last alternative in the cpu_facility attribute.
15018
79678d04
KT		 150192013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15020
15021 PR target/56315
15022 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15023 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15024 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15025 * config/arm/constraints.md (Dg): New constraint.
15026 * config/arm/neon.md (xordi3_neon): Remove.
15027 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15028 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15029
b31ddbdb
KT		 150302013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15031
15032 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15033 Clean up alternatives.
15034
8f90b7d4
AM		 150352013-06-07 Alan Modra <amodra@gmail.com>
15036
15037 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15038 va_list_gpr_size.
15039
961ce119
AM		 150402013-06-07 Alan Modra <amodra@gmail.com>
15041
15042 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15043
33e49835
KT		 150442013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15045
15046 * config/arm/constraints.md (Df): New constraint.
15047 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15048 Correct length attribute for last two alternatives.
15049
acd17ae6
AM		 150502013-06-07 Alan Modra <amodra@gmail.com>
15051
15052 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15053 override user -mfp-in-toc.
15054 (offsettable_ok_by_alignment): Consider just the current access
15055 rather than the whole object, unless BLKmode. Handle
15056 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15057 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15058 for -mcmodel=medium.
15059 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15060 override user -mfp-in-toc or -msum-in-toc. Default to
15061 -mno-fp-in-toc for -mcmodel=medium.
15062
73310b0e
DD		 150632013-06-06 DJ Delorie <dj@redhat.com>
15064
15065 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15066 TARGET_VALID_POINTER_MODE.
15067
0bd62dca 150682013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15069 Pat Haugen <pthaugen@us.ibm.com>
15070 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM		 15071
15072 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15073 Document new power8 builtins.
15074
15075 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15076 condition code register, to allow 128-bit logical operations to be
15077 done in the VSX or GPR registers.
15078 (nor<mode>3): Use the canonical form for nor.
15079 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15080 vclz*, and vpopcnt* vector instructions.
15081 (nand<mode>3): Likewise.
15082 (orc<mode>3): Likewise.
15083 (clz<mode>2): LIkewise.
15084 (popcount<mode>2): Likewise.
15085
15086 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15087 that only the GPRs are recognized.
15088
15089 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15090 support for new power8 builtins.
15091
15092 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15093 builtin functions.
15094 (xscvdpspn): Likewise.
15095 (vclz): Likewise.
15096 (vclzb): Likewise.
15097 (vclzh): Likewise.
15098 (vclzw): Likewise.
15099 (vclzd): Likewise.
15100 (vpopcnt): Likewise.
15101 (vpopcntb): Likewise.
15102 (vpopcnth): Likewise.
15103 (vpopcntw): Likewise.
15104 (vpopcntd): Likewise.
15105 (vgbbd): Likewise.
15106 (vmrgew): Likewise.
15107 (vmrgow): Likewise.
15108 (eqv): Likewise.
15109 (eqv_v16qi3): Likewise.
15110 (eqv_v8hi3): Likewise.
15111 (eqv_v4si3): Likewise.
15112 (eqv_v2di3): Likewise.
15113 (eqv_v4sf3): Likewise.
15114 (eqv_v2df3): Likewise.
15115 (nand): Likewise.
15116 (nand_v16qi3): Likewise.
15117 (nand_v8hi3): Likewise.
15118 (nand_v4si3): Likewise.
15119 (nand_v2di3): Likewise.
15120 (nand_v4sf3): Likewise.
15121 (nand_v2df3): Likewise.
15122 (orc): Likewise.
15123 (orc_v16qi3): Likewise.
15124 (orc_v8hi3): Likewise.
15125 (orc_v4si3): Likewise.
15126 (orc_v2di3): Likewise.
15127 (orc_v4sf3): Likewise.
15128 (orc_v2df3): Likewise.
15129
15130 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15131 allow power8 quad mode in 64-bit.
15132 (rs6000_builtin_vectorized_function): Add support to vectorize
15133 ISA 2.07 count leading zeros, population count builtins.
15134 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15135 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15136 (builtin_function_type): Add vgbbd builtin function which takes an
15137 unsigned argument.
15138 (altivec_expand_vec_perm_const): Add support for new power8 merge
15139 instructions.
15140
15141 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15142 that does not include TImdoe for use with 32-bit.
15143 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15144 instructions.
15145 (UNSPEC_VSX_CVDPSPN): Likewise.
15146 (vsx_xscvdpspn): Likewise.
15147 (vsx_xscvspdpn): Likewise.
15148 (vsx_xscvdpspn_scalar): Likewise.
15149 (vsx_xscvspdpn_directmove): Likewise.
15150 (vsx_and<mode>3): Split logical operations into 32-bit and
15151 64-bit. Add support to do logical operations on TImode as well as
15152 VSX vector types. Allow logical operations to be done in either
15153 VSX registers or in general purpose registers in 64-bit mode. Add
15154 splitters if GPRs were used. For AND, add clobber of CCmode to
15155 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15156 encoding.
15157 (vsx_and<mode>3_32bit): Likewise.
15158 (vsx_and<mode>3_64bit): Likewise.
15159 (vsx_ior<mode>3): Likewise.
15160 (vsx_ior<mode>3_32bit): Likewise.
15161 (vsx_ior<mode>3_64bit): Likewise.
15162 (vsx_xor<mode>3): Likewise.
15163 (vsx_xor<mode>3_32bit): Likewise.
15164 (vsx_xor<mode>3_64bit): Likewise.
15165 (vsx_one_cmpl<mode>2): Likewise.
15166 (vsx_one_cmpl<mode>2_32bit): Likewise.
15167 (vsx_one_cmpl<mode>2_64bit): Likewise.
15168 (vsx_nor<mode>3): Likewise.
15169 (vsx_nor<mode>3_32bit): Likewise.
15170 (vsx_nor<mode>3_64bit): Likewise.
15171 (vsx_andc<mode>3): Likewise.
15172 (vsx_andc<mode>3_32bit): Likewise.
15173 (vsx_andc<mode>3_64bit): Likewise.
15174 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15175 and xxlorc instructions.
15176 (vsx_eqv<mode>3_64bit): Likewise.
15177 (vsx_nand<mode>3_32bit): Likewise.
15178 (vsx_nand<mode>3_64bit): Likewise.
15179 (vsx_orc<mode>3_32bit): Likewise.
15180 (vsx_orc<mode>3_64bit): Likewise.
15181
15182 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15183
15184 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15185 instruction.
15186 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15187 (p8_vmrgow): Likewise.
15188 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15189 GPRs to be split under VSX.
15190 (p8v_clz<mode>2): Add power8 count leading zero support.
15191 (p8v_popcount<mode>2): Add power8 population count support.
15192 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15193 support.
15194
15195 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15196 instruction.
15197
15198 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15199 builtin functions.
15200 (vec_nand): Likewise.
15201 (vec_vclz): Likewise.
15202 (vec_vclzb): Likewise.
15203 (vec_vclzd): Likewise.
15204 (vec_vclzh): Likewise.
15205 (vec_vclzw): Likewise.
15206 (vec_vgbbd): Likewise.
15207 (vec_vmrgew): Likewise.
15208 (vec_vmrgow): Likewise.
15209 (vec_vpopcnt): Likewise.
15210 (vec_vpopcntb): Likewise.
15211 (vec_vpopcntd): Likewise.
15212 (vec_vpopcnth): Likewise.
15213 (vec_vpopcntw): Likewise.
15214
37684c46
VM		 152152013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15216
15217 PR rtl-optimization/57468
15218 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15219 spilled pseudos.
15220
d89ae27c
VM		 152212013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15222
15223 PR rtl-optimization/57459
15224 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15225 type when setting live regs.
15226
3597e113
VM		 152272013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15228
15229 * config/s390/s390.opt (mlra): New option.
15230 * config/s390/s390.c (s390_decompose_address): Check displacement
15231 for all registers for LRA.
15232 (s390_secondary_reload): Don't used secondary reloads for LRA.
15233 (s390_lra_p): New function.
15234 (TARGET_LRA_P): Define.
15235 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15236 of attribute cpu_facility to zarch for the last alternative.
15237 (*cmpmem_short): Ditto.
15238
01037aeb
EB		 152392013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15240
15241 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15242 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15243 (arm_expand_prologue): Likewise.
15244
3371a64f
TJ		 152452013-06-06 Teresa Johnson <tejohnson@google.com>
15246
15247 PR c++/53743
15248 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15249 as this is now done by redirect_edge_and_branch_force.
15250 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15251 barriers, and fix interaction with splitting.
15252 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15253 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15254 reflect changes made in the routine.
15255 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15256 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15257 since this is called in cfglayout mode, and replace partition fixup
15258 with assert as that is now done by force_nonfallthru_and_redirect.
15259 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15260 already be marked with region crossing note.
15261 (insert_section_boundary_note): Make non-static, gate on flag
15262 has_bb_partition, rewrite to also check for multiple partitions.
15263 (rest_of_handle_reorder_blocks): Remove call to
15264 insert_section_boundary_note, now done later during free_cfg.
15265 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15266 * bb-reorder.h (insert_section_boundary_note): Declare.
15267 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15268 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15269 invoke insert_section_boundary_note.
15270 (try_redirect_by_replacing_jump): Remove unnecessary
15271 check for region crossing note.
15272 (fixup_partition_crossing): New function.
15273 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15274 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15275 in non-cfglayout mode.
15276 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15277 remove old code that tried to do this. Emit barrier correctly
15278 when we are in cfglayout mode.
15279 (last_bb_in_partition): New function.
15280 (rtl_split_edge): Correctly fixup partition boundaries.
15281 (commit_one_edge_insertion): Remove old code that tried to
15282 fixup region crossing edge since this is now handled in
15283 split_block, and set up insertion point correctly since
15284 block may now end in a jump.
15285 (verify_hot_cold_block_grouping): Guard against checking when not in
15286 linearized RTL mode.
15287 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15288 notes.
15289 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15290 rtl_verify_flow_info, so not called in cfglayout mode.
15291 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15292 (fixup_reorder_chain): Remove old code that attempted to fixup region
15293 crossing note as this is now handled in force_nonfallthru_and_redirect.
15294 (duplicate_insn_chain): Don't duplicate switch section notes.
15295 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15296 note.
15297 * basic-block.h (emit_barrier_after_bb): Declare.
15298
66071e10
KT		 152992013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15300
15301 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15302 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15303 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15304
0e26bf3d
KT		 153052013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15306
15307 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15308 where appropriate.
15309 * config/arm/ldmstm.md: Regenerate.
15310
12b4e7ef
KT		 153112013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15312
15313 * config/arm/sync.md (atomic_loaddi_1):
15314 Disable predication for arm_restrict_it.
15315 (arm_load_exclusive<mode>): Likewise.
15316 (arm_load_exclusivesi): Likewise.
15317 (arm_load_exclusivedi): Likewise.
15318 (arm_load_acquire_exclusive<mode>): Likewise.
15319 (arm_load_acquire_exclusivesi): Likewise.
15320 (arm_load_acquire_exclusivedi): Likewise.
15321 (arm_store_exclusive<mode>): Likewise.
15322 (arm_store_exclusive<mode>): Likewise.
15323 (arm_store_release_exclusivedi): Likewise.
15324 (arm_store_release_exclusive<mode>): Likewise.
15325
15d16c8a
RB		 153262013-06-06 Richard Biener <rguenther@suse.de>
15327
15328 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15329 after LTO_null.
15330 (lto_tag_is_tree_code_p): Adjust.
15331 (lto_tag_is_gimple_code_p): Likewise.
15332 (lto_gimple_code_to_tag): Likewise.
15333 (lto_tag_to_gimple_code): Likewise.
15334 (lto_tree_code_to_tag): Likewise.
15335 (lto_tag_to_tree_code): Likewise.
15336 * data-streamer.h (streamer_write_hwi_in_range): Use
15337 uhwi streaming to stream the normalized range.
15338 (streamer_read_hwi_in_range): Likewise.
15339
17a7fc37
KT		 153402013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15341
15342 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15343 (predicable_short_it): Likewise.
15344 (predicated): Likewise.
15345 (enabled): Handle above.
15346 (define_cond_exec): Set predicated attribute to yes.
15347
b57ca59b
MS		 153482013-06-05 Mike Stump <mikestump@comcast.net>
15349
15350 * gdbinit.in (__FUNCTION__): Add.
15351
c1e183a9
UB		 153522013-06-05 Uros Bizjak <ubizjak@gmail.com>
15353
15354 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15355 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15356
0378bddb
JH		 153572013-06-05 Jan Hubicka <jh@suse.cz>
15358
15359 * varasm.c (mark_decl_referenced): Revert the removal until targets
15360 are fixed.
15361
8f8a2057
DE		 153622013-06-05 David Edelsohn <dje.gcc@gmail.com>
15363
15364 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15365 instead of mark_decl_referenced.
15366
edb983b2
JH		 153672013-06-05 Jan Hubicka <jh@suse.cz>
15368
15369 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15370 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15371 and symtab_used_from_object_file_p.
15372 (cgraph_make_node_local_1): Clear forced_by_abi.
15373 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15374 * cgraph.h (symtab_node_base): Add forced_by_abi.
15375 (decide_is_variable_needed): Remove.
15376 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15377 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15378 (decide_is_symbol_needed): ... this one; handle symbols in general;
15379 always analyze virtuals; honnor forced_by_abi.
15380 (cgraph_finalize_function): Update.
15381 (varpool_finalize_decl): Update.
15382 (symbol_defined_and_needed): Remove.
15383 (analyze_functions): Update.
15384 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15385 output_refs, input_overwrite_node): Handle forced_by_abi.
15386 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15387 (address_taken_from_non_vtable_p): ... this one.
15388 (comdat_can_be_unshared_p_1): New function.
15389 (cgraph_comdat_can_be_unshared_p): Rename to ...
15390 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15391 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15392 (function_and_variable_visibility): Clear forced_by_abi as needed.
15393 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15394 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH		 15395 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15396 * varpool.c (decide_is_variable_needed): Remove.
15397
9912dbe5
KT		 153982013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15399
15400 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15401 (arm_option_override): Override arm_restrict_it where appropriate.
15402 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15403 * config/arm/arm.opt (mrestrict-it): New command-line option.
15404 * doc/invoke.texi: Document -mrestrict-it.
15405
34c136b6
DM		 154062013-06-05 David Malcolm <dmalcolm@redhat.com>
15407
15408 * tsan.c (tsan_atomic_table): Make const.
15409
a367df53
RB		 154102013-06-05 Richard Biener <rguenther@suse.de>
15411
15412 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15413 index associated with the tree we are supposed to replace.
15414 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15415 TREE_ASM_WRITTEN as zero for everything but SSA names.
15416
70d51a19
DM		 154172013-06-05 David Malcolm <dmalcolm@redhat.com>
15418
15419 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15420
b9bd2075
JH		 154212013-06-04 Jan Hubicka <jh@suse.cz>
15422
15423 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15424 (input_node, input_varpool_node): Handle correctly external same
15425 body aliases.
15426 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15427 nodes at ltrans stage.
15428
aaae719d
JH		 154292013-06-04 Jan Hubicka <jh@suse.cz>
15430
15431 * ipa-inline.c (update_caller_keys): Fix availability test.
15432 (update_callee_keys): Likewise.
15433 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15434 to follow ELF standard.
15435
107eea2c
JU		 154362013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15437
15438 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15439 (mips64r5900el-*-elf*): New configurations.
15440 * config/mips/mips-cpus.def (r5900): New processor.
15441 * config/mips/mips-tables.opt: Regenerate.
15442 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15443 (mips_issue_rate): Handle PROCESSOR_R5900.
15444 (mips_reorg_process_insns): Force reorder mode for the R5900.
15445 * config/mips/mips.h (TARGET_MIPS5900): Define.
15446 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15447 TARGET_MIPS5900.
15448 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15449 TARGET_MIPS5900.
15450 * config/mips/mips.md (processor): Add r5900.
15451 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15452
b7342d25
IB		 154532013-06-04 Ian Bolton <ian.bolton@arm.com>
15454
15455 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15456 into function to generate MOVI instruction.
c1e183a9 15457 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB		 15458 (aarch64_preferred_simd_mode): Turn into wrapper.
15459 (aarch64_output_scalar_simd_mov_immediate): New function.
15460 * config/aarch64/aarch64-protos.h: Add prototype for above.
15461
81c2dfb9
IB		 154622013-06-04 Ian Bolton <ian.bolton@arm.com>
15463
15464 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15465 element_char member.
15466 (sizetochar): Return signed char.
15467 (aarch64_simd_valid_immediate): Remove elchar and other
15468 unnecessary variables.
15469 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15470 Calculate element_char as required.
15471 * config/aarch64/aarch64-protos.h: Update and move prototype
15472 for aarch64_output_simd_mov_immediate.
15473 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15474 Update arguments.
15475
48063b9d
IB		 154762013-06-04 Ian Bolton <ian.bolton@arm.com>
15477
15478 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15479 information completed by aarch64_simd_valid_immediate.
15480 (aarch64_legitimate_constant_p): Update arguments.
15481 (aarch64_simd_valid_immediate): Work with struct rather than many
15482 pointers.
15483 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15484 (aarch64_simd_make_constant): Update arguments.
15485 (aarch64_output_simd_mov_immediate): Work with struct rather than
15486 many pointers. Output immediate directly rather than as operand.
15487 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15488 Update prototype.
15489 * config/aarch64/constraints.md (Dn): Update arguments.
15490
3ea63f60
IB		 154912013-06-04 Ian Bolton <ian.bolton@arm.com>
15492
15493 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15494 longer static.
15495 (aarch64_simd_immediate_valid_for_move): Remove.
15496 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15497 (aarch64_simd_make_constant): Update call.
15498 (aarch64_output_simd_mov_immediate): Update call.
15499 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15500 Add prototype.
15501 * config/aarch64/constraints.md (Dn): Update call.
15502
d8edd899
IB		 155032013-06-04 Ian Bolton <ian.bolton@arm.com>
15504
15505 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15506 return type to bool for prototype.
15507 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15508 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15509 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15510
44b20bb8
CM		 155112013-06-04 Catherine Moore <clm@codesourcery.com>
15512
15513 * config/mips/mips.opt (meva): New.
c1e183a9 15514 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM		 15515 (ASM_SPEC): Handle -meva.
15516 * doc/invoke.texi (meva): Document.
15517
52befbd8
AM		 155182013-06-04 Alan Modra <amodra@gmail.com>
15519
15520 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15521 constant output.
15522
aadaf24e
KT		 155232013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15524
15525 * rtl.def: Add extra fourth optional field to define_cond_exec.
15526 * gensupport.c (process_one_cond_exec): Process attributes from
15527 define_cond_exec.
15528 * doc/md.texi: Document fourth field in define_cond_exec.
15529
0cd9e9ee
EB		 155302013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15531
15532 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15533 out the processing order as in store_bit_field_1.
15534
cb0f665d
JH		 155352013-06-04 Jan Hubicka <jh@suse.cz>
15536
15537 PR middle-end/57500
15538 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15539 non-VAR_DECL node if it does not exist yet.
15540
53984b9b
RS		 155412013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15542
15543 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15544 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15545 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15546 target_cpu_default setting.
15547
5d318fd4
TJ		 155482013-06-03 Teresa Johnson <tejohnson@google.com>
15549
15550 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15551 default to -fopt-info=optimized instead of all.
15552 * doc/invoke.texi: Ditto.
15553 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15554 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15555 (execute_vect_slp): Emit BB vectorization success under
15556 MSG_OPTIMIZED_LOCATIONS.
15557 * tree-vect-slp.c (vect_slp_transform_bb): Change
15558 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15559 * tree-vect-loop.c (vect_transform_loop): Ditto.
15560
ec3af349
JM		 155612013-06-03 Jason Merrill <jason@redhat.com>
15562
15563 PR c++/57415
15564 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15565 Use TARGET_EXPR for C++.
15566
1baf9159
JJ		 155672013-06-03 Jakub Jelinek <jakub@redhat.com>
15568
15569 PR rtl-optimization/57268
0cd9e9ee 15570 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ		 15571 if DEBUG_INSN_P (insn).
15572
15573 Reapply
15574 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15575
15576 PR rtl-optimization/57268
15577 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15578 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15579
0f1d3965
YR		 155802013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15581
15582 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15583 (ix86_avoid_lea_for_addr): Likewise.
15584 (exact_dependency_1): Likewise.
15585 (ix86_adjust_cost): Likewise.
15586 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15587 removed.
15588 (do_reorder_for_imul): Fix typo, formatting and
15589 !reload_completed check removed.
15590 (ix86_sched_reorder): Fix typo and formatting.
15591 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15592 list.
15593
3bc30db4
SN		 155942013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15595
15596 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15597
19cc6fac
EB		 155982013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15599
15600 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15601 <STRING_CST>: Likewise.
15602 <VECTOR_CST>: Likewise.
15603
c68a6e08
JW		 156042013-06-01 Janus Weil <janus@gcc.gnu.org>
15605 Mikael Morin <mikael@gcc.gnu.org>
15606
15607 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15608 * config.in: Regenerated.
15609 * configure: Regenerated.
15610
38e55e5c
JH		 156112013-06-01 Jan Hubicka <jh@suse.cz>
15612
15613 PR middle-end/57366
15614 * cgraphunit.c (compile): When weakref is not supported,
15615 set up transparent aliases before final output pass.
15616 * varasm.c (assemble_alias): Do not try to do it here.
15617
eb51d2ff
JH		 156182013-06-01 Jan Hubicka <jh@suse.cz>
15619
15620 PR middle-end/57467
15621 * passes.c (for_per_function): Skip unanalyzed functions.
15622
40a7fe1e
JH		 156232013-06-01 Jan Hubicka <jh@suse.cz>
15624
15625 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15626 (lto_symtab_merge_symbols_1): ... this one.
15627 (lto_symtab_merge_cgraph_nodes): Rename to ...
15628 (lto_symtab_merge_symbols): ... this one; simplify.
15629 * cgraph.c (same_body_aliases_done): Rename to ...
15630 (cpp_implicit_aliases_done): ... this one.
15631 (cgraph_create_function_alias): Update.
15632 (cgraph_same_body_alias): Update.
c1e183a9 15633 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 15634 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 15635 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH		 15636 (cgraph_node): Remove same_body_alias.
15637 (varpool_node): Remove alias_of and extra_name_alias.
15638 (same_body_aliases_done): Rename to ..
15639 (cpp_implicit_aliases_done): ... this one.
15640 (symtab_alias_ultimate_target): Add default parameter.
15641 (symtab_resolve_alias): New function.
15642 (fixup_same_cpp_alias_visibility): Declare.
15643 (cgraph_function_node): Add default parameter.
15644 (cgraph_node_asm_name): Likewise.
15645 (cgraph_function_or_thunk_node): Add default parameter; do
15646 not ICE when it is NULL.
15647 (varpool_variable_node): Likewise.
15648 * tree-emutls.c (create_emultls_var): Update.
15649 (ipa_lower_emutls): Update.
15650 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15651 (cgraph_reset_node): Reset alias info.
15652 (cgraph_finalize_function): Update.
15653 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15654 (analyze_function): Simplify.
15655 (cgraph_process_same_body_aliases): Simplify.
15656 (analyze_functions): Fixup same body aliases.
15657 (handle_alias_pairs): Simplify.
15658 (assemble_thunk): Update.
15659 (assemble_thunks_and_aliases): Update.
15660 (output_weakrefs): Rewrite.
15661 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15662 (lto_output_varpool_node): Likewise.
15663 (compute_ltrans_boundary): Remve assert.
15664 (get_alias_symbol): New functoin.
15665 (input_node): Rewrite alias handling.
15666 (input_varpool_node): Likewise.
15667 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15668 * ipa.c (process_references): Handle weakrefs correctly.
15669 (symtab_remove_unreachable_nodes): Likewise.
15670 * trans-mem.c (get_cg_data): Update.
15671 (ipa_tm_create_version_alias): Update.
15672 (ipa_tm_execute): Update.
15673 * symtab.c (dump_symtab_base): Dump aliases.
15674 (verify_symtab_base): Verify aliases.
15675 (symtab_node_availability): New function.
15676 (symtab_alias_ultimate_target): Simplify.
15677 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15678 handle all the fixup cases.
15679 (symtab_resolve_alias): New function.
15680 * passes.c (ipa_write_summaries): Handle weakrefs.
15681 * varpool.c (varpool_analyze_node): Simplify.
15682 (assemble_aliases): Update.
15683 (varpool_create_variable_alias): Simplify.
15684 (varpool_extra_name_alias): Simplify.
15685 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15686 (lto_symtab_merge_symbols): ... this one.
15687
21759881
DT		 156882013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15689
15690 Revert
15691 PR rtl-optimization/57268
15692 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15693 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15694
daf4e940
TB		 156952013-06-01 Tobias Burnus <burnus@net-b.de>
15696
15697 Partially reverted:
15698 2013-05-31 Tobias Burnus <burnus@net-b.de>
15699
15700 PR middle-end/57073
15701 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15702 further up.
15703
c359d8d0
DT		 157042013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15705
15706 PR rtl-optimization/57268
15707 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15708 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15709
4172215d
EB		 157102013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15711
15712 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15713 unordered comparison operators when -fno-trapping-math is in effect
15714 on the e500.
15715 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15716 and implement unordered comparison operators properly on the e500.
15717
a8c50132
EB		 157182013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15719
15720 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15721 for constant scalar integers.
15722 (simplify_relational_operation_1): Likewise.
15723
44626634
SB		 157242013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15725
15726 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15727 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15728 Fix comment.
15729
55805e54
YR		 157302013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15731 Igor Zamyatin <igor.zamyatin@intel.com>
15732
15733 Silvermont (SLM) architecture performance tuning.
15734 * config/i386/i386.h (enum ix86_tune_indices): Add
15735 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15736 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15737
15738 * config/i386/i386.c (initial_ix86_tune_features)
15739 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15740 (ix86_lea_outperforms): Handle Silvermont tuning.
15741 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15742 call.
15743 (ix86_use_lea_for_mov): Likewise.
15744 (ix86_avoid_lea_for_addr): Likewise.
15745 (ix86_lea_for_add_ok): Likewise.
15746 (exact_dependency_1): New function.
15747 (exact_store_load_dependency): Likewise.
15748 (ix86_adjust_cost): Handle Silvermont tuning.
15749 (do_reoder_for_imul): Likewise.
15750 (swap_top_of_ready_list): New function.
15751 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15752
15753 * config/i386/i386.md (peepholes that split memory operand in fp
15754 converts): New.
15755
e19c9de2
MS		 157562013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15757
15758 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15759 Remove un-necessary braces.
15760
38e6c9a6
MS		 157612013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15762
15763 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15764 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15765
e3530904
TB		 157662013-05-31 Tobias Burnus <burnus@net-b.de>
15767
15768 PR middle-end/57073
c1e183a9 15769 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 15770
b6af05a9
KT		 157712013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15772
15773 PR target/56315
15774 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15775 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15776 * config/arm/neon.md (iordi3_neon): Remove.
15777 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15778 * config/arm/predicates.md (imm_for_neon_logic_operand):
15779 Move to earlier in the file.
15780 (neon_logic_op2): Likewise.
15781 (arm_iordi_operand_neon): New predicate.
15782
f800c166
RB		 157832013-05-31 Richard Biener <rguenther@suse.de>
15784
15785 PR tree-optimization/57478
15786 PR tree-optimization/57453
15787 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15788 are life as well.
15789
e571fa59
KP		 157902013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15791
15792 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15793 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15794
0fa6e0ef
TB		 157952013-05-30 Tobias Burnus <burnus@net-b.de>
15796 Thomas Koenig <tkoenig@gcc.gnu.org>
15797
15798 PR middle-end/57073
15799 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15800 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15801
4688ddf5
SB		 158022013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15803
0fa6e0ef 15804 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 15805
3b9ceb4b
VM		 158062013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15807
15808 * target.def (register_usage_leveling_p): New hook.
15809 * targhooks.c (default_register_usage_leveling_p): New.
15810 * targhooks.h (default_register_usage_leveling_p): New prototype.
15811 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15812 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15813 * doc/tm.texi: Update.
15814 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15815
65f9e789
IB		 158162013-05-30 Ian Bolton <ian.bolton@arm.com>
15817
15818 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15819 (*insv_reg<mode>): New define_insn.
15820
12211b99 158212013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR		 15822
15823 PR rtl-optimization/57439
15824 * postreload.c (move2add_valid_value_p): Check that we have
15825 a zero subreg_regno_offset when accessing the register in
15826 the requested mode.
15827
0b871ccf
YR		 158282013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15829 Igor Zamyatin <igor.zamyatin@intel.com>
15830
15831 Silvermont (SLM) architecture pipeline model, tuning and
15832 insn selection.
15833 * config.gcc: Add slm config options and target.
15834
15835 * config/i386/slm.md: New.
15836
15837 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15838
f43245d1 15839 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR		 15840 PROCESSOR_SLM.
15841 (ix86_target_macros_internal): Likewise.
15842
f43245d1 15843 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR		 15844 (m_SLM): New macro flag.
15845 (initial_ix86_tune_features): Set m_SLM.
15846 (x86_accumulate_outgoing_args): Likewise.
15847 (x86_arch_always_fancy_math_387): Likewise.
15848 (processor_target_table): Add slm cost.
15849 (cpu_names): Add slm cpu name.
15850 (x86_option_override_internal): Set SLM ISA.
15851 (ix86_issue_rate): New case PROCESSOR_SLM.
15852 (ia32_multipass_dfa_lookahead): Likewise.
15853 (fold_builtin_cpu): Add slm.
15854
15855 * config/i386/i386.h (TARGET_SLM): New target macro.
15856 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15857 (processor_type): Add PROCESSOR_SLM.
15858
15859 * config/i386/i386.md (cpu): Add new value "slm".
15860 (slm.md): Include slm.md.
15861
24d5b097
XG		 158622013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15863 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15864
15865 * config/arm/arm-protos.h: Add and update function protos.
15866 * config/arm/arm.c (use_simple_return_p): New added.
15867 (thumb2_expand_return): Check simple_return flag.
15868 * config/arm/arm.md: Add simple_return and conditional simple_return.
15869 * config/arm/iterators.md: Add iterator for return and simple_return.
15870
c1cccc15
ZC		 158712013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15872
15873 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15874 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15875 (arm_emit_vfp_multi_reg_pop): Likewise.
15876 (thumb2_emit_ldrd_pop): Likewise.
15877 (arm_expand_epilogue): Add misc REG_CFA notes.
15878 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15879
3a4f280b
LC		 158802013-05-29 Lawrence Crowl <crowl@google.com>
15881
15882 * config/arm/t-arm: Update for below.
15883
15884 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15885 Change type to hash_table. Update dependent calls and types.
15886
15887 * config/i386/t-cygming: Update for below.
15888
15889 * config/i386/t-interix: Update for below.
15890
15891 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15892 Change type to hash_table. Update dependent calls and types.
15893 (i386_find_on_wrapper_list::wrappers): Likewise.
15894
15895 * config/ia64/t-ia64: Update for below.
15896
15897 * config/ia64/ia64.c (bundle_state_table):
15898 Change type to hash_table. Update dependent calls and types.
15899
15900 * config/mips/mips.c (mips_reorg_process_insns::htab):
15901 Change type to hash_table. Update dependent calls and types.
15902
15903 * config/sol2.c (solaris_comdat_htab):
15904 Change type to hash_table. Update dependent calls and types.
15905
15906 * config/t-sol2: Update for above.
15907
1388a0e3
TJ		 159082013-05-29 Teresa Johnson <tejohnson@google.com>
15909
15910 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15911 functions are not yet marked as defined.
15912
a5965b52 159132013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15914 Pat Haugen <pthaugen@us.ibm.com>
15915 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM		 15916
15917 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15918 instructions.
15919 (VEC_A): Likewise.
15920 (VEC_C): Likewise.
15921 (vrotl<mode>3): Likewise.
15922 (vashl<mode>3): Likewise.
15923 (vlshr<mode>3): Likewise.
15924 (vashr<mode>3): Likewise.
15925
15926 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15927 support for power8 V2DI builtins.
15928
15929 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15930 power8 V2DI builtins.
15931 (vupkhsw): Likewise.
15932 (vupklsw): Likewise.
15933 (vaddudm): Likewise.
15934 (vminsd): Likewise.
15935 (vmaxsd): Likewise.
15936 (vminud): Likewise.
15937 (vmaxud): Likewise.
15938 (vpkudum): Likewise.
15939 (vpksdss): Likewise.
15940 (vpkudus): Likewise.
15941 (vpksdus): Likewise.
15942 (vrld): Likewise.
15943 (vsld): Likewise.
15944 (vsrd): Likewise.
15945 (vsrad): Likewise.
15946 (vsubudm): Likewise.
15947 (vcmpequd): Likewise.
15948 (vcmpgtsd): Likewise.
15949 (vcmpgtud): Likewise.
15950 (vcmpequd_p): Likewise.
15951 (vcmpgtsd_p): Likewise.
15952 (vcmpgtud_p): Likewise.
15953 (vupkhsw): Likewise.
15954 (vupklsw): Likewise.
15955 (vaddudm): Likewise.
15956 (vmaxsd): Likewise.
15957 (vmaxud): Likewise.
15958 (vminsd): Likewise.
15959 (vminud): Likewise.
15960 (vpksdss): Likewise.
15961 (vpksdus): Likewise.
15962 (vpkudum): Likewise.
15963 (vpkudus): Likewise.
15964 (vrld): Likewise.
15965 (vsld): Likewise.
15966 (vsrad): Likewise.
15967 (vsrd): Likewise.
15968 (vsubudm): Likewise.
15969
15970 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15971 support for power8 V2DI instructions.
15972
15973 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15974 power8 V2DI instructions. Combine pack and unpack insns to use an
15975 iterator for each mode. Check whether a particular mode supports
15976 Altivec instructions instead of just checking TARGET_ALTIVEC.
15977 (UNSPEC_VPKUWUM): Likewise.
15978 (UNSPEC_VPKSHSS): Likewise.
15979 (UNSPEC_VPKSWSS): Likewise.
15980 (UNSPEC_VPKUHUS): Likewise.
15981 (UNSPEC_VPKSHUS): Likewise.
15982 (UNSPEC_VPKUWUS): Likewise.
15983 (UNSPEC_VPKSWUS): Likewise.
15984 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15985 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15986 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15987 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15988 (UNSPEC_VUPKHSB): Likewise.
15989 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15990 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15991 (UNSPEC_VUPKHSH): Likewise.
15992 (UNSPEC_VUPKLSB): Likewise.
15993 (UNSPEC_VUPKLSH): Likewise.
15994 (VI2): Likewise.
15995 (VI_char): Likewise.
15996 (VI_scalar): Likewise.
15997 (VI_unit): Likewise.
15998 (VP): Likewise.
15999 (VP_small): Likewise.
16000 (VP_small_lc): Likewise.
16001 (VU_char): Likewise.
16002 (add<mode>3): Likewise.
16003 (altivec_vaddcuw): Likewise.
16004 (altivec_vaddu<VI_char>s): Likewise.
16005 (altivec_vadds<VI_char>s): Likewise.
16006 (sub<mode>3): Likewise.
16007 (altivec_vsubcuw): Likewise.
16008 (altivec_vsubu<VI_char>s): Likewise.
16009 (altivec_vsubs<VI_char>s): Likewise.
16010 (altivec_vavgs<VI_char>): Likewise.
16011 (altivec_vcmpbfp): Likewise.
16012 (altivec_eq<mode>): Likewise.
16013 (altivec_gt<mode>): Likewise.
16014 (altivec_gtu<mode>): Likewise.
16015 (umax<mode>3): Likewise.
16016 (smax<mode>3): Likewise.
16017 (umin<mode>3): Likewise.
16018 (smin<mode>3): Likewise.
16019 (altivec_vpkuhum): Likewise.
16020 (altivec_vpkuwum): Likewise.
16021 (altivec_vpkshss): Likewise.
16022 (altivec_vpkswss): Likewise.
16023 (altivec_vpkuhus): Likewise.
16024 (altivec_vpkshus): Likewise.
16025 (altivec_vpkuwus): Likewise.
16026 (altivec_vpkswus): Likewise.
16027 (altivec_vpks<VI_char>ss): Likewise.
16028 (altivec_vpks<VI_char>us): Likewise.
16029 (altivec_vpku<VI_char>us): Likewise.
16030 (altivec_vpku<VI_char>um): Likewise.
16031 (altivec_vrl<VI_char>): Likewise.
16032 (altivec_vsl<VI_char>): Likewise.
16033 (altivec_vsr<VI_char>): Likewise.
16034 (altivec_vsra<VI_char>): Likewise.
16035 (altivec_vsldoi_<mode>): Likewise.
16036 (altivec_vupkhsb): Likewise.
16037 (altivec_vupkhs<VU_char>): Likewise.
16038 (altivec_vupkls<VU_char>): Likewise.
16039 (altivec_vupkhsh): Likewise.
16040 (altivec_vupklsb): Likewise.
16041 (altivec_vupklsh): Likewise.
16042 (altivec_vcmpequ<VI_char>_p): Likewise.
16043 (altivec_vcmpgts<VI_char>_p): Likewise.
16044 (altivec_vcmpgtu<VI_char>_p): Likewise.
16045 (abs<mode>2): Likewise.
16046 (vec_unpacks_hi_v16qi): Likewise.
16047 (vec_unpacks_hi_v8hi): Likewise.
16048 (vec_unpacks_lo_v16qi): Likewise.
16049 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16050 (vec_unpacks_lo_v8hi): Likewise.
16051 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16052 (vec_pack_trunc_v8h): Likewise.
16053 (vec_pack_trunc_v4si): Likewise.
16054 (vec_pack_trunc_<mode>): Likewise.
16055
16056 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16057 V2DI builtins.
16058 (vec_vmaxsd): Likewise.
16059 (vec_vmaxud): Likewise.
16060 (vec_vminsd): Likewise.
16061 (vec_vminud): Likewise.
16062 (vec_vpksdss): Likewise.
16063 (vec_vpksdus): Likewise.
16064 (vec_vpkudum): Likewise.
16065 (vec_vpkudus): Likewise.
16066 (vec_vrld): Likewise.
16067 (vec_vsld): Likewise.
16068 (vec_vsrad): Likewise.
16069 (vec_vsrd): Likewise.
16070 (vec_vsubudm): Likewise.
16071 (vec_vupkhsw): Likewise.
16072 (vec_vupklsw): Likewise.
16073
e70670cf
JH		 160742013-05-29 Jan Hubicka <jh@suse.cz>
16075
16076 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16077 flags; reorder rest of fields in more consistent way.
16078 (varpool_node): Remove analyzed, finalized and alias.
16079 (cgraph_ndoe): Likewise.
16080 (symtab_alias_ultimate_target): New function.
16081 (cgraph_function_node): Move offline.
16082 (cgraph_reset_node): Declare.
16083 (cgraph_comdat_can_be_unshared_p): Remove.
16084 (varpool_remove_initializer): Declare.
16085 (varpool_first_defined_variable, varpool_next_defined_variable
16086 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16087 (cgraph_function_with_gimple_body_p): Update.
16088 (varpool_all_refs_explicit_p): Update.
16089 (symtab_alias_target): New function.
16090 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16091 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB		 16092 (cgraph_function_or_thunk_node): Simplify using
16093 symtab_alias_ultimate_target.
e70670cf
JH		 16094 (varpool_variable_node): Likewise.
16095 * cgraph.c (cgraph_create_function_alias): Update.
16096 (cgraph_add_thunk): Update.
16097 (cgraph_remove_node): Update.
16098 (dump_cgraph_node): Do not dump removed flags.
16099 (cgraph_function_body_availability): Update.
16100 (cgraph_propagate_frequency): Update.
16101 (verify_cgraph_node): Check sanity of local flag.
16102 (cgraph_function_node): Move here from cgraph.h; revamp for
16103 cgraph_function_or_thunk_node.
16104 * lto-symtab.c (lto_varpool_replace_node): Update.
16105 (lto_symtab_resolve_can_prevail_p): Update.
16106 (lto_symtab_merge_cgraph_nodes): Update.
16107 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16108 propagate_constants_accross_call, devirtualization_time_bonus,
16109 ipcp_propagate_stage): Update.
16110 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB		 16111 * ipa-inline-transform.c (clone_inlined_nodes,
16112 preserve_function_body_p): Update.
e70670cf
JH		 16113 * ipa-reference.c (propagate): Update.
16114 (write_node_summary_p): Update.
16115 * toplev.c (wrapup_global_declaration_2): Update.
16116 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16117 (analyze_function) ... this one.
16118 (cgraph_process_new_functions): Update.
16119 (cgraph_reset_node): Export.
16120 (cgraph_finalize_function): Update.
16121 (cgraph_add_new_function): Update.
16122 (process_function_and_variable_attributes): Update.
16123 (varpool_finalize_decl): Update.
16124 (symbol_finalized): Remove.
16125 (symbol_finalized_and_needed): Rename to ...
16126 (symbol_defined_and_needed): ... update.
16127 (cgraph_analyze_functions): Update.
16128 (handle_alias_pairs): Update.
16129 (mark_functions_to_output): Update.
16130 (assemble_thunk): Update.
16131 (output_in_order): Update.
16132 (output_weakrefs): Update.
16133 (finalize_compilation_unit): Update.
16134 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16135 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16136 input_node, input_varpool_node): Update.
16137 * dbxout.c (dbxout_expand_expr): Update.
16138 * cgraphclones.c (cgraph_clone_node): Update.
16139 (cgraph_copy_node_for_versioning): Update.
16140 (cgraph_materialize_clone): Update.
16141 (cgraph_materialize_all_clones): Update.
16142 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16143 propagate_pure_const, propagate_nothrow): Update.
16144 * lto-streamer-out.c (lto_output, write_symbol): Update.
16145 * ipa-utils.c (ipa_reverse_postorder): Update.
16146 * ipa-inline.c (can_inline_edge_p): Update.
16147 (update_caller_keys, ipa_inline): Update.
16148 * dwarf2out.c (reference_to_unused,
16149 premark_types_used_by_global_vars_helper): Update.
16150 * tree-eh.c (tree_could_trap_p): Update.
16151 * ipa-split.c (consider_split, execute_split_functions): Update.
16152 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB		 16153 has_addr_references_p): Update; move ahead in file for better
16154 readability.
e70670cf
JH		 16155 (process_references): Simplify.
16156 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16157 bodies are removed.
16158 (cgraph_comdat_can_be_unshared_p): Make static.
16159 (cgraph_externally_visible_p): Update.
16160 (varpool_externally_visible_p): Update.
16161 (function_and_variable_visibility): Update.
16162 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16163 ipa_tm_mark_force_output_node): Update.
16164 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16165 estimate_edge_devirt_benefit, inline_generate_summary,
16166 inline_write_summary): Update.
16167 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16168 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB		 16169 (ipa_print_node_params, ipa_prop_read_section,
16170 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH		 16171 * varasm.c (mark_decl_referenced): Update.
16172 (assemble_alias, dump_tm_clone_pairs): Update.
16173 * tree-inline.c (copy_bb): Update.
16174 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16175 Update.
16176 * symtab.c (dump_symtab_base): Print new flags.
16177 (verify_symtab_base): Verify new flags.
16178 (symtab_alias_ultimate_target): New function.
16179 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16180 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16181 Update.
c1e183a9
UB		 16182 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16183 Update.
e70670cf
JH		 16184 * i386.c (ix86_get_function_versions_dispatcher,
16185 ix86_generate_version_dispatcher_body): Update.
16186 (fold_builtin_cpu): Use varpool_add_new_variable.
16187 * varpool.c (varpool_remove_initializer): Break out from ...
16188 (varpool_remove_node): ... this one.
16189 (dump_varpool_node, varpool_node_for_asm,
16190 cgraph_variable_initializer_availability, varpool_analyze_node,
16191 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB		 16192 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16193 Update.
e70670cf 16194
182802ad
JH		 161952013-05-29 Jan Hubicka <jh@suse.cz>
16196
16197 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16198
8190b609
ER		 161992013-05-29 Easwaran Raman <eraman@google.com>
16200
16201 PR tree-optimization/57442
16202 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16203 when control exits the main loop.
16204
69f5aa9b
SKS		 162052013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16206
16207 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16208 and RX600.
c1e183a9 16209 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS		 16210 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16211 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16212 option.
16213
4b847da9
BS		 162142013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16215
16216 PR tree-optimization/57441
16217 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16218 Don't limit size of incr_vec to number of candidates.
16219
4fc43c39
SE		 162202013-05-29 Steve Ellcey <sellcey@imgtec.com>
16221
16222 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16223 and mips16 directories.
c1e183a9 16224 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE		 16225 (MULTILIB_DIRNAMES): Ditto.
16226 (MULTILIB_EXCEPTIONS): Add new exceptions.
16227 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16228 (MULTILIB_DIRNAMES): Ditto.
16229 (MULTILIB_EXCEPTIONS): Add new exceptions.
16230
12211b99 162312012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS		 16232 Marcus Shawcroft <marcus.shawcroft@arm.com>
16233
16234 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16235 SYMBOL_TINY_ABSOLUTE.
16236 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16237 SYMBOL_TINY_ABSOLUTE.
16238 (aarch64_expand_mov_immediate): Likewise.
16239 (aarch64_classify_symbol): Likewise.
16240 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16241 Permit SYMBOL_TINY_ABSOLUTE.
16242 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16243
12211b99 162442013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS		 16245 Marcus Shawcroft <marcus.shawcroft@arm.com>
16246
16247 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16248 Refactor if/switch. Replace gcc_assert with if.
16249
c0186656
GG		 162502013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16251
16252 * config/i386/i386.c (initial_ix86_tune_features): Enable
16253 FP Reassociation for AMD bdver1 and bdver2.
16254
d20188f3
MJ		 162552013-05-29 Martin Jambor <mjambor@suse.cz>
16256
16257 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16258 and IMAGPART_EXPR do not occur within other handled_components.
16259
292cba13
RB		 162602013-05-29 Richard Biener <rguenther@suse.de>
16261
16262 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16263 access on whether the use is in the BB we currently try to
16264 vectorize.
16265 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16266 vectorize to vect_bb_slp_scalar_cost.
16267
6eddf228
RB		 162682013-05-29 Richard Biener <rguenther@suse.de>
16269
16270 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16271 computing scalar cost offsetted by stmts that are kept live
16272 by scalar uses.
16273 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16274 for computation of scalar cost.
16275
7df36117
SE		 162762013-05-28 Steve Ellcey <sellcey@mips.com>
16277
16278 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16279
36536d79
BI		 162802013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16281
16282 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16283 array notation built-in reduction functions.
16284 * doc/passes.texi (Passes): Added documentation about changes done
16285 for Cilk Plus.
16286 * doc/invoke.texi (C Dialect Options): Added documentation about
16287 the -fcilkplus flag.
16288 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16289 (BUILTINS_DEF): Depend on cilkplus.def.
16290 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16291 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16292 * cilkplus.def: New file.
16293
12211b99 162942013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR		 16295
16296 PR rtl-optimization/57439
16297 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16298
9daf6dbb
ER		 162992013-05-28 Easwaran Raman <eraman@google.com>
16300
16301 PR tree-optimization/57337
16302 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16303 (find_insert_point): Correctly identify the insertion point
16304 when two statements with the same UID is compared.
16305
fbd7e877
RB		 163062013-05-28 Richard Biener <rguenther@suse.de>
16307
16308 PR tree-optimization/56787
16309 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16310 from the list of data references.
16311 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16312 clobbers.
16313 (vect_analyze_loop_operations): Likewise.
16314 (vect_transform_loop): Remove clobbers.
16315
bbba1117
MJ		 163162013-05-28 Martin Jambor <mjambor@suse.cz>
16317
16318 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16319 and REALPART_EXPRs have scalar type.
16320
bd388c2a
RB		 163212013-05-28 Richard Biener <rguenther@suse.de>
16322
16323 PR tree-optimization/57411
16324 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16325 virtual operands.
16326 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16327 virtual operand propagation.
16328
2f56a311
EB		 163292013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16330
16331 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16332 destination register for bmasksi_vis.
16333 (vector_init_bshuffle): Likewise.
16334 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16335
5c3eacbb
EB		 163362013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16337
16338 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16339 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16340 mode if the instruction isn't available in the original mode.
16341 * config/sparc/sparc.opt (mfix-ut699): New option.
16342 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16343 (divdf3): Turn into expander.
16344 (divdf3_nofix): New insn.
16345 (divdf3_fix): Likewise.
16346 (divsf3): Disable if -mfix-ut699.
16347 (sqrtdf2): Turn into expander.
16348 (sqrtdf2_nofix): New insn.
16349 (sqrtdf2_fix): Likewise.
16350 (sqrtsf2): Disable if -mfix-ut699.
16351
a1756c0a
RB		 163522013-05-27 Richard Biener <rguenther@suse.de>
16353
16354 PR middle-end/57412
16355 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16356 block for the new loop.
16357
5a892248
RB		 163582013-05-27 Richard Biener <rguenther@suse.de>
16359
16360 PR tree-optimization/57343
16361 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16362 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16363 (number_of_iterations_cond): Do not build the folded tree.
16364
d1de852b
RB		 163652013-05-27 Richard Biener <rguenther@suse.de>
16366
16367 Revert
16368 PR middle-end/57381
16369 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16370 OEP_CONSTANT_ADDRESS_OF retained.
16371
16372 PR tree-optimization/57417
16373 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16374 for unchanged base.
16375 (set_ssa_val_to): Compare addresses using
16376 get_addr_base_and_unit_offset.
16377
12211b99 163782013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR		 16379
16380 PR rtl-optimization/56833
16381 * postreload.c (move2add_record_mode): New function.
16382 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16383 (move2add_use_add2_insn): Use move2add_record_sym_value.
16384 (move2add_use_add3_insn): Likewise.
16385 (reload_cse_move2add): Use move2add_valid_value_p and
16386 move2add_record_mode. Invalidate call-clobbered and REG_INC
16387 affected regs by setting reg_mode to VOIDmode.
16388 (move2add_note_store): Don't pretend the inside of a SUBREG is
16389 the actual destination. Invalidate single/leading registers by
16390 setting reg_mode to VOIDmode.
16391 Use move2add_record_sym_value, move2add_valid_value_p and
16392 move2add_record_mode.
16393
b03be25f
RB		 163942013-05-27 Richard Biener <rguenther@suse.de>
16395
16396 PR tree-optimization/57396
16397 * tree-affine.c (double_int_constant_multiple_p): Properly
16398 return false for val == 0 and div != 0.
16399
44e88db2
RS		 164002013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16401
16402 * config/mips/mips.h: Use #elif in preprocessor conditions.
16403
3b859704
RS		 164042013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16405
16406 PR target/53916
16407 * config/mips/constraints.md (kl): New constraint.
16408 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16409 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16410 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16411 from LO for MIPS16.
16412 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16413
c979d5f5
RS		 164142013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16415
16416 PR target/55777
16417 * config/mips/mips.c (mips_can_inline_p): New function.
16418 (TARGET_CAN_INLINE_P): Define.
16419
8e90de43
SB		 164202013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16421
16422 * sched-int.h (ds_t, dw_t): Make unsigned int.
16423 Fix documentation that describes how all the ds_t bits are used.
16424 Reserve the last bit for delayed-branch scheduling.
16425 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16426 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16427 (gen_dep_weak_1): Remove prototype.
16428 * sched-deps.c (get_dep_weak_1): Make static.
16429 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16430 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16431 * doc/tm.texi: Regenerate.
16432 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16433
cb5cb194
SB		 164342013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16435
16436 PR debug/56950
16437 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16438
5bd93ff6
NS		 164392013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16440 Sandra Loosemore <sandra@codesourcery.com>
16441
16442 * config.gcc (powerpc-*): Allow native for with-cpu.
16443
2343af65
JL		 164442013-05-24 Jeff Law <law@redhat.com>
16445
16446 PR tree-optimization/57124
16447 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16448 conversion feeding a condition if the range has an overflow
16449 if -fstrict-overflow. Add warnings for when we do make the
16450 transformation.
16451
3ad695b9
DC		 164522013-05-24 Dehao Chen <dehao@google.com>
16453
04960246 16454 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC		 16455 (locus_discrim_hasher): Likewise.
16456 (locus_discrim_hasher::hash): Likewise.
16457 (locus_discrim_hasher::equal): Likewise.
16458
82338059
MJ		 164592013-05-24 Martin Jambor <mjambor@suse.cz>
16460
16461 PR tree-optimization/57294
16462 * cgraph.h (ipa_record_stmt_references): Declare.
16463 * cgraphbuild.c (ipa_record_stmt_references): New function.
16464 (build_cgraph_edges): Use ipa_record_stmt_references.
16465 (rebuild_cgraph_edges): Likewise.
16466 (cgraph_rebuild_references): Likewise.
16467 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16468 associated with the old statement and build references from the
16469 newly built statements.
16470 * ipa-ref.c (ipa_remove_stmt_references): New function.
16471 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16472
1ccd4874
VM		 164732013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16474
55805e54 16475 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 16476 mem-mem moves.
55805e54 16477 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 16478 too.
55805e54 16479 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM		 16480 secondary memory moves.
16481 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16482 reg set up in the current insn.
16483
25e25c73
DC		 164842013-05-24 Dehao Chen <dehao@google.com>
16485
04960246 16486 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC		 16487 hash function.
16488 (locus_descrim_hasher::equal): Likewise.
04960246 16489 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC		 16490 (make_edges): Likewise.
16491 (next_discriminator_for_locus): Likewise.
16492 (same_line_p): Likewise.
16493 (assign_discriminators): Likewise.
16494 (make_cond_expr_edges): Likewise.
16495 (make_gimple_switch_edges): Likewise.
16496 (make_goto_expr_edges): Likewise.
16497 (make_gimple_asm_edges): Likewise.
16498
50d38551
IB		 164992013-05-24 Ian Bolton <ian.bolton@arm.com>
16500
16501 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16502 X format specifier to only display bottom 16 bits.
16503 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16504 immediate to match for operand 2, since it will be masked.
16505
aea0101d
RB		 165062013-05-24 Richard Biener <rguenther@suse.de>
16507
16508 PR tree-optimization/57287
16509 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16510 all SSA names that occur in abnormal PHIs.
16511
634e03d3
AI		 165122013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16513
16514 PR tree-ssa/57385
16515 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16516 that index is not negative.
16517
b17c024f
EB		 165182013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16519
16520 PR rtl-optimization/55177
16521 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16522 (simplify_byte_swapping_operation): New.
16523 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16524 (simplify_relational_operation_1): Deal with BSWAP.
16525
46aeac1b
RH		 165262013-05-23 Richard Henderson <rth@redhat.com>
16527
16528 PR target/56742
16529 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16530 (ix86_reorg): Call it.
16531
70cc1536
UB		 165322013-05-23 Uros Bizjak <ubizjak@gmail.com>
16533
16534 PR target/57379
16535 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16536 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16537 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16538
b3851501
CB		 165392013-05-23 Christian Bruel <christian.bruel@st.com>
16540
16541 PR debug/57351
16542 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16543
12211b99 165442013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS		 16545 Marcus Shawcroft <marcus.shawcroft@arm.com>
16546
16547 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16548 * config/aarch64/constraints.md (Usa): Remove.
16549 * doc/md.texi (AArch64 Usa): Remove.
16550
12211b99 165512013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS		 16552 Marcus Shawcroft <marcus.shawcroft@arm.com>
16553
16554 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16555 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16556 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16557 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16558
12211b99 165592013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP		 16560
16561 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16562 instruction (AdvSIMD).
16563 * config/aarch64/aarch64-builtins.c
16564 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16565 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16566
76c36cb1 165672013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ		 16568
16569 PR middle-end/57347
16570 * tree.h (contains_bitfld_component_ref_p): Declare.
16571 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB		 16572 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16573 caller.
7d2fb524
MJ		 16574 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16575 not access a bit-field. Assert all final offsets are byte-aligned.
16576
ce521ae6
RB		 165772013-05-23 Richard Biener <rguenther@suse.de>
16578
16579 PR tree-optimization/57380
16580 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16581 least one invariant or re-used load.
16582 * passes.c (init_optimization_passes): Move pass_phiprop before
16583 pass_forwprop.
16584
75c7257f
JG		 165852013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16586
16587 * config/aarch64/aarch64-simd.md
16588 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16589
af6d44b5
RB		 165902013-05-23 Richard Biener <rguenther@suse.de>
16591
16592 PR middle-end/57381
16593 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16594 OEP_CONSTANT_ADDRESS_OF retained.
16595
bd3647bf
JJ		 165962013-05-23 Jakub Jelinek <jakub@redhat.com>
16597
16598 PR middle-end/57344
70cc1536
UB		 16599 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16600 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 16601
a55757ea
RB		 166022013-05-23 Richard Biener <rguenther@suse.de>
16603
16604 PR rtl-optimization/57341
16605 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16606 instead of true_dependence.
16607
cfb00b41
DM		 166082013-05-22 David Malcolm <dmalcolm@redhat.com>
16609
dd1c676f
DM		 16610 * bb-reorder.c (branch_threshold): Make const.
16611 (exec_threshold): Ditto.
cfb00b41 16612
f62511da 166132013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16614 Pat Haugen <pthaugen@us.ibm.com>
16615 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16616
16617 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16618 documentation for the power8 crypto builtins.
16619
16620 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16621
16622 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16623 macros for defining power8 builtin functions.
16624 (BU_P8V_AV_2): Likewise.
16625 (BU_P8V_AV_P): Likewise.
16626 (BU_P8V_VSX_1): Likewise.
16627 (BU_P8V_OVERLOAD_1): Likewise.
16628 (BU_P8V_OVERLOAD_2): Likewise.
16629 (BU_CRYPTO_1): Likewise.
16630 (BU_CRYPTO_2): Likewise.
16631 (BU_CRYPTO_3): Likewise.
16632 (BU_CRYPTO_OVERLOAD_1): Likewise.
16633 (BU_CRYPTO_OVERLOAD_2): Likewise.
16634 (XSCVSPDP): Fix typo, point to the correct instruction.
16635 (VCIPHER): Add power8 crypto builtins.
16636 (VCIPHERLAST): Likewise.
16637 (VNCIPHER): Likewise.
16638 (VNCIPHERLAST): Likewise.
16639 (VPMSUMB): Likewise.
16640 (VPMSUMH): Likewise.
16641 (VPMSUMW): Likewise.
16642 (VPERMXOR_V2DI): Likewise.
16643 (VPERMXOR_V4SI: Likewise.
16644 (VPERMXOR_V8HI: Likewise.
16645 (VPERMXOR_V16QI: Likewise.
16646 (VSHASIGMAW): Likewise.
16647 (VSHASIGMAD): Likewise.
16648 (VPMSUM): Likewise.
16649 (VPERMXOR): Likewise.
16650 (VSHASIGMA): Likewise.
16651
16652 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16653 __CRYPTO__ if the crypto instructions are available.
16654 (altivec_overloaded_builtins): Add support for overloaded power8
16655 builtins.
16656
16657 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16658 support for power8 crypto builtins.
16659 (builtin_function_type): Likewise.
16660 (altivec_init_builtins): Add support for builtins that take vector
16661 long long (V2DI) arguments.
16662
16663 * config/rs6000/crypto.md: New file, define power8 crypto
16664 instructions.
16665
166662013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16667 Pat Haugen <pthaugen@us.ibm.com>
16668 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16669
16670 * doc/invoke.texi (Option Summary): Add power8 options.
16671 (RS/6000 and PowerPC Options): Likewise.
16672
16673 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16674 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16675 wm, wn, wr documentation.
16676
f43245d1 16677 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM		 16678 registers if direct move instructions are enabled.
16679 (wn): New constraint for no registers.
16680 (wq): New constraint for quad word even GPR registers.
16681 (wr): New constraint if 64-bit instructions are enabled.
16682 (wv): New constraint if power8 vector instructions are enabled.
16683 (wQ): New constraint for quad word memory locations.
16684
f43245d1 16685 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM		 16686 constraint for 0..15 for crypto instructions.
16687 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16688 as GPR and floating point registers.
16689 (int_reg_operand): New predicate to match only GPR registers.
16690 (base_reg_operand): New predicate to match base registers.
16691 (quad_int_reg_operand): New predicate to match even GPR registers
16692 for quad memory operations.
16693 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16694 operations in both GPR and VSX registers.
16695 (quad_memory_operand): New predicate for quad memory operations.
16696 (reg_or_indexed_operand): New predicate for direct move support.
16697
f43245d1 16698 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM		 16699 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16700 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16701 (POWERPC_MASKS): Add power8 options.
16702 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16703 various options.
16704
f43245d1 16705 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM		 16706 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16707
f43245d1 16708 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM		 16709 (-mpower8-fusion): New power8 options.
16710 (-mpower8-fusion-sign): Likewise.
16711 (-mpower8-vector): Likewise.
16712 (-mcrypto): Likewise.
16713 (-mdirect-move): Likewise.
16714 (-mquad-memory): Likewise.
16715
f43245d1 16716 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM		 16717 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16718 registers.
70cc1536 16719 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM		 16720 (rs6000_debug_vector_unit): Add p8_vector.
16721 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16722 definitions. Also print fusion state.
16723 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16724 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16725 (rs6000_option_override_internal): Add support for power8.
16726 (rs6000_common_init_builtins): Add debugging for skipped builtins
16727 if -mdebug=builtin.
16728 (rs6000_adjust_cost): Add power8 support.
16729 (rs6000_issue_rate): Likewise.
16730 (insn_must_be_first_in_group): Likewise.
16731 (insn_must_be_last_in_group): Likewise.
16732 (force_new_group): Likewise.
16733 (rs6000_register_move_cost): Likewise.
16734 (rs6000_opt_masks): Likewise.
16735
16736 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16737 power8 capable assembler, default to power7 options.
16738 (TARGET_DIRECT_MOVE): Likewise.
16739 (TARGET_CRYPTO): Likewise.
16740 (TARGET_P8_VECTOR): Likewise.
16741 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16742 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16743 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16744 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16745 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16746 (TARGET_XSCVDPSPN): Likewise.
16747 (TARGET_XSCVSPDPN): Likewsie.
16748 (TARGET_SYNC_HI_QI): Likewise.
16749 (TARGET_SYNC_TI): Likewise.
16750 (MASK_CRYPTO): Likewise.
16751 (MASK_DIRECT_MOVE): Likewise.
16752 (MASK_P8_FUSION): Likewise.
16753 (MASK_P8_VECTOR): Likewise.
70cc1536
UB		 16754 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16755 temporary used by some of the direct move instructions to get two FP
16756 temporary registers does not force creation of a stack frame.
f62511da
MM		 16757 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16758 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16759 that any VSX registers are tieable, even if they are also an
16760 Altivec vector mode.
16761 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16762 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16763 (RS6000_BTM_CRYPTO): Likewise.
16764 (RS6000_BTM_COMMON): Likewise.
16765
16766 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16767 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16768 (enum rs6000_vector): Add power8 vector support.
16769
73a1a707
RR		 167702013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16771
16772 PR target/19599
16773 PR target/57340
16774 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16775 (any_sibcall_could_use_r3): this and handle indirect calls.
16776 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16777
d28073d4
BS		 167782013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16779
16780 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16781
bee0b10c
RB		 167822013-05-22 Richard Biener <rguenther@suse.de>
16783
16784 PR middle-end/57349
16785 * profile.c (branch_prob): Do not split blocks that are
16786 abnormally receiving from ECF_RETURNS_TWICE functions.
16787
98409b51
RS		 167882013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16789
16790 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16791 address mode. Move pointer mode initialization to the same place.
16792
c0602ab8
MZ		 167932013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16794
16795 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16796 while it has any effect.
16797
4f7a634e
ER		 167982013-05-21 Easwaran Raman <eraman@google.com>
16799
16800 PR tree-optimization/57322
9daf6dbb
ER		 16801 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16802 UID of the statement added to the BB to be 1.
4f7a634e 16803
a4ce1258
JJ		 168042013-05-21 Jakub Jelinek <jakub@redhat.com>
16805
16806 PR tree-optimization/57331
70cc1536
UB		 16807 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16808 of conversion from pointer type to integral type with integer.
a4ce1258 16809
1b14621a
MJ		 168102013-05-21 Martin Jambor <mjambor@suse.cz>
16811
16812 PR lto/57289
16813 * ipa-prop.c (ipa_read_node_info): Process param_used and
16814 controlled_uses in the same order as when writing.
16815
e60661f0
MG		 168162013-05-21 Magnus Granberg <baldrick@free.fr>
16817
16818 PR plugins/56754
ee49aa34 16819 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 16820
4a61be9a
RB		 168212013-05-21 Richard Biener <rguenther@suse.de>
16822
16823 PR tree-optimization/57318
16824 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16825 estimate stmts with side-effects as likely eliminated.
16826
c52da5f7
RB		 168272013-05-21 Richard Biener <rguenther@suse.de>
16828
16829 PR tree-optimization/57330
16830 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16831 preserve the call stmts fntype.
16832
7ec67e2a
RB		 168332013-05-21 Richard Biener <rguenther@suse.de>
16834
16835 PR tree-optimization/57303
16836 * tree-ssa-sink.c (statement_sink_location): Improve killing
16837 stmt detection and properly handle self-assignments.
16838
b112d513
CB		 168392013-05-21 Christian Bruel <christian.bruel@st.com>
16840
55805e54
YR		 16841 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16842 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB		 16843 registers. Set register size out of the PARALLEL loop.
16844
14c2ec26
OE		 168452013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16846
16847 PR target/56547
16848 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16849 (*fmasf4, *fmasf4_media): New insns.
16850
da734fa1
RS		 168512013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16852
16853 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16854 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16855 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16856 (mips_idiv_insns): Update the comments to say that the returned
16857 instruction counts are in units of BASE_INSN_LENGTH.
16858 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16859 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16860 using 2 rather than 4 as the length of indirect MIPS16 and
16861 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16862 length of a NOP. Don't divide MIPS16 lengths by 2.
16863 (mips16_split_long_branches): Assume a branch is long if the
16864 length is greater than 4 rather than 8.
16865 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16866 rather than multiplying them by 2. Multiply instruction counts
16867 by BASE_INSN_LENGTH rather than 4.
16868 (*jump_mips16, tls_get_tp_mips16_<mode>)
16869 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16870
13719e8b
RS		 168712013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16872
16873 * config/mips/mips.md (extended_mips16): Remove branch case.
16874 (length): Remove duplicated extended_mips16 test.
16875
c3850d14
RS		 168762013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16877
16878 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16879
8da2e059
RS		 168802013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16881
16882 * recog.h (Recog_data): Rename to...
16883 (recog_data_d): ...this.
16884 (recog_data): Update accordingly.
16885 * recog.c (recog_data): Likewise.
16886 * reload.c (save_recog_data): Likewise.
16887 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16888 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16889
2df013f3
JB		 168902013-05-17 Julian Brown <julian@codesourcery.com>
16891
16892 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16893 found in a REG_EQUAL note, invalidate it.
16894
371e77e3 168952013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER		 16896
16897 * tree-ssa-reassoc.c (find_insert_point): New function.
16898 (insert_stmt_after): Likewise.
16899 (get_def_stmt): Likewise.
16900 (ensure_ops_are_available): Likewise.
16901 (not_dominated_by): Likewise.
16902 (rewrite_expr_tree): Do not move statements beyond what is
16903 necessary. Remove call to swap_ops_for_binary_stmt...
16904 (reassociate_bb): ... and move it here.
16905 (build_and_add_sum): Assign UIDs for new statements.
16906 (linearize_expr): Likewise.
16907 (do_reassoc): Renumber gimple statement UIDs.
16908
e01c7cca
JH		 169092013-05-17 Jan Hubicka <jh@suse.cz>
16910
16911 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16912 weakrefs.
16913 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB		 16914 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16915 weakrefs.
e01c7cca
JH		 16916 (output_weakrefs): Update.
16917
c3272a92
PCC		 169182013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16919 Martin Jambor <mjambor@suse.cz>
16920
16921 PR middle-end/57276
16922 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16923 value that corresponds to the given aggval is found in values vector.
16924
11c2aa39
UB		 169252013-05-17 Uros Bizjak <ubizjak@gmail.com>
16926
16927 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16928 sse, sse2, sse3, ssse3 and sse4a flags to options.
16929
abecc8c6
DM		 169302013-05-17 David Malcolm <dmalcolm@redhat.com>
16931
16932 * gengtype-state.c: (s_expr_writer): New class, to handle
16933 prettifying of output layout of s-expressions.
16934 (state_writer): New class, to write out gtype.state.
16935 (state_written_type_count): Move this variable into member data of
16936 state_writer.
16937 (s_expr_writer::s_expr_writer): New code: constructor for new class
16938 (state_writer::state_writer(): ditto
16939 (s_expr_writer::write_new_line): New function
16940 (s_expr_writer::write_any_indent): ditto
16941 (s_expr_writer::begin_s_expr): ditto
16942 (s_expr_writer::end_s_expr): ditto
16943 (write_state_fileloc): convert to method of state_writer...
16944 (state_writer:: write_state_fileloc): ...and use methods of
16945 s_expr_writer to write indentation into the gtype.state output file
16946 to visually represent the hierarchical structure of the list
16947 structures
16948 (write_state_fields): ditto, renaming to...
16949 (state_writer::write_state_fields)
16950 (write_state_a_string): ditto, renaming to...
16951 (state_writer::write_state_a_string)
16952 (write_state_string_option): ditto, renaming to...
16953 (state_writer::write_state_string_option)
16954 (write_state_type_option): ditto, renaming to...
16955 (state_writer::write_state_type_option)
16956 (write_state_nested_option): ditto, renaming to...
16957 (state_writer::write_state_nested_option)
16958 (write_state_option): ditto, renaming to...
16959 (state_writer::write_state_option)
16960 (write_state_options): ditto, renaming to...
16961 (state_writer::write_state_options)
16962 (write_state_lang_bitmap): ditto, renaming to...
16963 (state_writer::write_state_lang_bitmap)
16964 (write_state_version): ditto, renaming to...
16965 (state_writer::write_state_version)
16966 (write_state_scalar_type): ditto, renaming to...
16967 (state_writer::write_state_scalar_type)
16968 (write_state_string_type): ditto, renaming to...
16969 (state_writer::write_state_string_type)
16970 (write_state_undefined_type): ditto, renaming to...
16971 (state_writer::write_state_undefined_type)
16972 (write_state_struct_union_type): ditto, renaming to...
16973 (state_writer::write_state_struct_union_type)
16974 (write_state_struct_type): ditto, renaming to...
16975 (state_writer::write_state_struct_type)
16976 (write_state_user_struct_type): ditto, renaming to...
16977 (state_writer::write_state_user_struct_type)
16978 (write_state_lang_struct_type): ditto, renaming to...
16979 (state_writer::write_state_lang_struct_type)
16980 (write_state_param_struct_type): ditto, renaming to...
16981 (state_writer::write_state_param_struct_type)
16982 (write_state_pointer_type): ditto, renaming to...
16983 (state_writer::write_state_pointer_type)
16984 (write_state_array_type): ditto, renaming to...
16985 (state_writer::write_state_array_type)
16986 (write_state_gc_used): ditto, renaming to...
16987 (state_writer::write_state_gc_used)
16988 (write_state_common_type_content): ditto, renaming to...
16989 (state_writer::write_state_common_type_content)
16990 (write_state_type): ditto, renaming to...
16991 (state_writer::write_state_type)
16992 (write_state_pair_list): ditto, renaming to...
16993 (state_writer::write_state_pair_list)
16994 (write_state_pair): ditto, renaming to...
16995 (state_writer::write_state_pair)
16996 (write_state_typedefs): ditto, renaming to...
16997 (state_writer::write_state_typedefs)
16998 (write_state_structures): ditto, renaming to...
16999 (state_writer::write_state_structures)
17000 (write_state_param_structs): ditto, renaming to...
17001 (state_writer::write_state_param_structs)
17002 (write_state_variables): ditto, renaming to...
17003 (state_writer::write_state_variables)
17004 (write_state_srcdir): ditto, renaming to...
17005 (state_writer::write_state_srcdir)
17006 (write_state_files_list): ditto, renaming to...
17007 (state_writer::write_state_files_list)
17008 (write_state_languages): ditto, renaming to...
17009 (state_writer::write_state_languages)
17010 (write_state): create a state_writer instance and use it when
17011 writing out the state file
17012
d6545f29
MS		 170132013-05-17 Mike Stump <mikestump@comcast.net>
17014
816a3f73 17015 PR rtl-optimization/57304
d6545f29
MS		 17016 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17017 accessing DF_REF_REAL_LOC.
17018
38c821cf
JJ		 170192013-05-17 Jakub Jelinek <jakub@redhat.com>
17020
17021 PR rtl-optimization/57281
17022 PR rtl-optimization/57300
17023 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17024 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17025 what the other splitter did if the registers are dead.
17026
2a293391
RB		 170272013-05-17 Richard Biener <rguenther@suse.de>
17028
17029 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17030 MEM_REF offsets.
17031
17042d2b
JJ		 170322013-05-17 Jakub Jelinek <jakub@redhat.com>
17033
17034 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17035 linking.
17036
5b115c1f
MP		 170372013-05-17 Marek Polacek <polacek@redhat.com>
17038
11c2aa39
UB		 17039 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17040 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 17041
52d84413
JJ		 170422013-05-17 Jakub Jelinek <jakub@redhat.com>
17043
68119618
JJ		 17044 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17045 vect_external_def oprnd1 with loop_vinfo, try to emit
17046 optional cast, negation and and stmts on the loop preheader
17047 edge instead of into the pattern def seq.
17048
52d84413
JJ		 17049 PR tree-optimization/57051
17050 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17051 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17052
17b962bd
NC		 170532013-05-16 Nick Clifton <nickc@redhat.com>
17054
17055 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17056 (rl78_is_naked_func): New function.
56aefbf7
UB		 17057 (rl78_expand_prologue): Skip prologue generation for naked functions.
17058 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC		 17059 * doc/extend.texi (naked): Add RL78 to the list of processors
17060 that supports this attribute.
17061
b1a0f84e
JL		 170622013-05-16 Jeff Law <law@redhat.com>
17063
17064 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17065
af0e415b
UB		 170662013-05-16 Uros Bizjak <ubizjak@gmail.com>
17067
17068 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17069 cache parameters using detect_caches_amd also for CYRIX,
17070 NSC and TM2 signatures.
17071
19db293a
UB		 170722013-05-16 Uros Bizjak <ubizjak@gmail.com>
17073 Dzianis Kahanovich <mahatma@eu.by>
17074
17075 PR target/45359
17076 PR target/46396
17077 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17078 VIA/Centaur processors and determine their cache parameters
17079 using detect_caches_amd.
17080
251a41b9
TJ		 170812013-05-16 Teresa Johnson <tejohnson@google.com>
17082
17083 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17084 (rtl_verify_edges): New function.
17085 (rtl_verify_bb_insns): Ditto.
17086 (rtl_verify_bb_pointers): Ditto.
17087 (rtl_verify_bb_insn_chain): Ditto.
17088 (rtl_verify_fallthru): Ditto.
17089 (rtl_verify_bb_layout): Ditto.
17090 (rtl_verify_flow_info_1): Outline checks into new functions.
17091 (rtl_verify_flow_info): Ditto.
17092
f14540b6
SE		 170932013-05-16 Steve Ellcey <sellcey@imgtec.com>
17094
17095 * cfghooks.c (copy_bbs): Add update_dominance argument.
17096 * cfghooks.h (copy_bbs): Update prototype.
17097 * tree-cfg.c (gimple_duplicate_sese_region):
17098 Add update_dominance argument.
17099 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17100 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17101 gimple_duplicate_sese_region call.
17102 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17103 Update copy_bbs call.
17104 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17105 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17106
7e9a3abb
JJ		 171072013-05-16 Jakub Jelinek <jakub@redhat.com>
17108
17109 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17110 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17111 vect_recog_rotate_pattern.
17112 (vect_recog_rotate_pattern): New function.
17113
427b248d
JM		 171142013-05-16 Jason Merrill <jason@redhat.com>
17115
17116 * Makefile.in (LLINKER): New variable.
17117 (mostlyclean): Remove link mutex.
17118 * configure.ac: Handle --enable-link-mutex.
17119 * lock-and-run.sh: New script.
17120
b871e3d2
RR		 171212013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17122
17123 PR target/19599
17124 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17125 for NULL decl.
17126
ce858126
RO		 171272013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17128
17129 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17130
798d3d04
GY		 171312013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17132
17133 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17134 * config/arm/arm.c (next_consecutive_mem): New function.
17135 (gen_movmem_ldrd_strd): Likewise.
17136 * config/arm/arm.md (movmemqi): Update condition and code.
17137 (unaligned_loaddi, unaligned_storedi): New patterns.
17138
0baddc45
RO		 171392013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17140
17141 * config.gcc: Obsolete *-*-solaris2.9*.
17142 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17143
aa9d5bcf
RB		 171442013-05-16 Richard Biener <rguenther@suse.de>
17145
17146 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17147 earlier, after GRAPHITE transforms and IV canonicalization.
17148
5a0f4dd3
JJ		 171492013-05-16 Jakub Jelinek <jakub@redhat.com>
17150
17151 * omp-low.c (extract_omp_for_data): For collapsed loops,
17152 if at least one of the loops is known at compile time to
17153 iterate zero times, set count to 0.
17154 (expand_omp_regimplify_p): New function.
17155 (expand_omp_for_generic): For collapsed loops, if at least
17156 one of the loops isn't known to iterate at least once,
17157 add runtime check with setting count to 0.
17158 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17159 For unsigned types if it isn't known at compile time that
17160 the loop will iterate at least once, add runtime check to bypass
17161 the whole loop if initial condition isn't true.
17162
e3753785
NS		 171632013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17164
17165 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17166
43bb4dd1
MG		 171672013-05-16 Marc Glisse <marc.glisse@inria.fr>
17168
17169 PR middle-end/57286
17170 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17171 transformations to avoid an infinite loop.
17172
3571dde6
MP		 171732013-05-16 Marek Polacek <polacek@redhat.com>
17174
17175 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17176
b8b3f0ca
LE		 171772013-05-15 Leif Ekblad <leif@rdos.net>
17178
17179 * config/i386/i386.c (ix86_decompose_address): Use
17180 DEFAULT_TLS_SEG_REG to access TLS segment register.
17181 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17182 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17183 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17184
e299a383
RS		 171852013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17186
17187 PR target/57260
17188 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17189 sibling calls to functions that would normally be lazily bound,
17190 unless $gp is call-clobbered.
17191
e7208ea3
UB		 171922013-05-15 Uros Bizjak <ubizjak@gmail.com>
17193
19db293a 17194 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 17195 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB		 17196 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17197 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB		 17198 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17199 of TARGET_3DNOW.
17200 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17201
e7413f3d
AS		 172022013-05-15 Andreas Schwab <schwab@suse.de>
17203
17204 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17205 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17206 third operand.
17207
ad4db775
TJ		 172082013-05-15 Teresa Johnson <tejohnson@google.com>
17209
17210 * loop-unroll.c (report_unroll_peel): Check decision before
17211 emitting unroll/peel message.
17212
af205f67
TJ		 172132013-05-15 Teresa Johnson <tejohnson@google.com>
17214
17215 * function.h (has_bb_partition): New rtl_data flag.
17216 (bb_reorder_complete): Ditto.
17217 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17218 instead of flag_reorder_blocks_and_partition.
17219 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17220 with some enhancements.
17221 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17222 * bb-reorder.c (connect_traces): Check for has_bb_partition
17223 instead of flag_reorder_blocks_and_partition.
17224 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17225 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17226 verify_hot_cold_block_grouping.
17227 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17228
9adcfa3c
RR		 172292013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17230
17231 PR target/19599
17232 * config/arm/predicates.md (call_insn_operand): New predicate.
17233 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17234 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17235 if insn is not a tail call.
17236 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17237 registers.
17238 * config/arm/arm.h (enum reg_class): New caller save register class.
17239 (REG_CLASS_NAMES): Likewise.
17240 (REG_CLASS_CONTENTS): Likewise.
17241 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17242 without decls.
17243
ccb3ad87
RB		 172442013-05-15 Richard Biener <rguenther@suse.de>
17245
17246 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17247 of MSG_OPTIMIZED_LOCATIONS.
17248 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17249 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17250 message.
17251 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17252 of MSG_OPTIMIZED_LOCATIONS.
17253 (execute_vect_slp): Likewise.
17254 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17255 (vect_create_cond_for_alias_checks): Likewise.
17256 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17257 (vect_recog_widen_mult_pattern): Likewise.
17258 (vect_recog_widen_sum_pattern): Likewise.
17259 (vect_recog_over_widening_pattern): Likewise.
17260 (vect_recog_widen_shift_pattern): Likewise.
17261 (vect_recog_vector_vector_shift_pattern): Likewise.
17262 (vect_recog_divmod_pattern): Likewise.
17263 (vect_recog_mixed_size_cond_pattern): Likewise.
17264 (vect_recog_bool_pattern): Likewise.
17265 (vect_pattern_recog_1): Likewise.
17266
48b1474e
MJ		 172672013-05-15 Martin Jambor <mjambor@suse.cz>
17268
17269 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17270 non-functions to builtin_unreachable.
17271 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17272 correct when new direct edges were discovered.
17273
9de04252
MJ		 172742013-05-15 Martin Jambor <mjambor@suse.cz>
17275
17276 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17277 header, print symbol order instead of node uid, print more information
17278 about indirect edge targets.
17279 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17280 uids.
17281 (ipa_make_edge_direct_to_target): Likewise.
17282 (remove_described_reference): Likewise.
17283 (propagate_controlled_uses): Likewise.
17284 (ipa_print_node_params): Also print symbol order.
17285 (ipcp_transform_function): Print symbol order instead of node uids.
17286 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17287 (cgraph_get_create_real_symbol_node): Likewise.
17288 * ipa-cp.c (print_lattice): Likewise.
17289 (print_all_lattices): Likewise.
17290 (determine_versionability): Likewise.
17291 (initialize_node_lattices): Likewise.
17292 (estimate_local_effects): Likewise.
17293 (update_profiling_info): Likewise.
17294 (create_specialized_node): Likewise.
17295 (perhaps_add_new_callers): Likewise.
17296 (decide_about_value): Likewise.
17297 (decide_whether_version_node): Likewise.
17298 (identify_dead_nodes): Likewise.
17299 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17300 (dump_inline_summary): Likewise.
17301 (estimate_node_size_and_time): Likewise.
17302 (inline_analyze_function): Likewise.
17303 * ipa-inline.c (report_inline_failed_reason): Likewise.
17304 (want_early_inline_function_p): Likewise.
17305 (edge_badness): Likewise.
17306 (update_edge_key): Likewise.
17307 (inline_small_functions): Likewise. Add dumping of order to two other
17308 dumps.
17309 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17310 instead of node uids.
17311 (propagate_pure_const): Likewise.
17312 (propagate_pure_const): Likewise.
17313 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17314 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17315 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17316 of node uids.
17317 * tree-pretty-print.c (dump_function_header): Likewise.
17318 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17319 Print symbol order instead of node uids.
17320
1dd03b91
AK		 173212013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17322
17323 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17324 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17325
fdf6a7b9
RB		 173262013-05-15 Richard Biener <rguenther@suse.de>
17327
17328 PR tree-optimization/57275
17329 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17330 return value for fail to do runtime alias checks for gather loads.
17331
2d6e4603
JH		 173322013-05-15 Jan Hubicka <jh@suse.cz>
17333
17334 PR lto/57038
17335 PR lto/47375
e7208ea3
UB		 17336 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17337 weakrefs are not external.
17338 (lto_symtab_merge_decls): Fix thinko when dealing with
17339 non-lto_symtab decls.
2d6e4603
JH		 17340 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17341 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17342 * varpool.c (dump_varpool_node): Dump more flags.
17343
83f44b39
GG		 173442013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17345
17346 * config/i386/i386.c (processor_alias_table): Add instruction
17347 FSGSBASE for AMD bdver3 architecture.
17348
2e55d062
JJ		 173492013-05-14 Jakub Jelinek <jakub@redhat.com>
17350
17351 * tree.c (warn_deprecated_use): Print file:line using locus color.
17352 * diagnostic.c (diagnostic_report_current_module): Print file:line
17353 and file:line:column using locus color.
17354
0bfdb81e
MS		 173552013-05-14 Mike Stump <mikestump@comcast.net>
17356
17357 * gdbinit.in: Add __null.
17358
a508ef22
MS		 173592013-05-14 Mike Stump <mikestump@comcast.net>
17360
17361 * recog.h: Rename struct recog_data to Recog_data.
17362 * recog.c: Likewise.
17363 * reload.c (can_reload_into): Likewise.
17364 * config/picochip/picochip.c: Likewise.
17365
e7180acb
MS		 173662013-05-14 Mike Stump <mikestump@comcast.net>
17367
17368 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17369
61216c26
SB		 173702013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17371
fd6beed4
SB		 17372 * resource.h (struct resources): Remove unch_memory member.
17373 (CLEAR_RESOURCE): Don't clear unch_memory.
17374 * resource.c (mark_referenced_resources): Don't set it.
17375 (mark_set_resources): Likewise.
17376 (mark_target_live_regs): Don't clear it.
17377 (init_resource_info): Likewise.
17378 * reorg.c (resource_conflicts_p): Don't compare it.
17379 (redundant_insn): Don't set it.
17380
61216c26
SB		 17381 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17382 Remove prototypes.
17383 * emit-rtl.c (next_label): Remove unused function.
17384 (skip_consecutive_labels, link_cc0_insns): Move to ...
17385 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17386 only place where these functions are used, and make them static.
17387
418d1b87
MG		 173882013-05-14 Marc Glisse <marc.glisse@inria.fr>
17389
17390 * fold-const.c (fold_negate_expr): Handle vectors.
17391 (fold_truth_not_expr): Make it static.
17392 (fold_invert_truthvalue): New static function.
17393 (invert_truthvalue_loc): Handle vectors. Do not call
17394 fold_truth_not_expr directly.
17395 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17396 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17397 (fold_comparison): Handle vector constants.
17398 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17399 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17400 * tree.h (fold_truth_not_expr): Remove declaration.
17401
fc21784d
JG		 174022013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17403
17404 * config/aarch64/aarch64-simd.md
17405 (aarch64_vcond_internal<mode>): Rename to...
17406 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17407 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17408 float modes. Clarify all iterator modes.
17409 (vcond<mode><mode>): Use new name for vcond expanders.
17410 (vcond<v_cmp_result><mode>): Likewise.
17411 (vcondu<mode><mode>: Likewise.
17412 * config/aarch64/iterators.md (VDQF_COND): New.
17413
d4c52634
MG		 174142013-05-14 Marc Glisse <marc.glisse@inria.fr>
17415
17416 PR bootstrap/57266
17417 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17418 variable for the shift amount. Check that we shift by non-negative
17419 amounts.
17420
2b261262
CLT		 174212013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17422
17423 PR target/42017
17424 * config/arm/arm.h (EPILOGUE_USES): Only return true
17425 for LR_REGNUM after epilogue_completed.
17426
12211b99 174272013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR		 17428
17429 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 17430 is error_mark_node.
6e022d7b 17431
56cf7859
RO		 174322013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17433
17434 PR target/57261
17435 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17436 and Solaris 11+/x86 with gld.
17437 * configure: Regenerate.
17438
75776c6d
JJ		 174392013-05-14 Jakub Jelinek <jakub@redhat.com>
17440
17441 * expmed.c (expand_shift_1): Canonicalize rotates by
17442 constant bitsize / 2 to bitsize - 1.
7f998021 17443 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ		 17444 case ROTATERT>: Likewise.
17445
17446 Revert:
17447 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17448
17449 * config/i386/i386.md (rotateinv): New code attr.
17450 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17451 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17452 roll $31, %eax, etc.
17453
df35498a
RB		 174542013-05-14 Richard Biener <rguenther@suse.de>
17455
17456 PR middle-end/57235
17457 * tree-eh.c (sink_clobbers): Give up for successors with
17458 multiple predecessors and no virtual uses.
17459
cc6e7ece
EB		 174602013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17461
17462 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17463 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17464
f778dd4d
JJ		 174652013-05-14 Jakub Jelinek <jakub@redhat.com>
17466
17467 PR middle-end/57251
17468 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17469 the case when both op0 and op1 have VOIDmode.
17470
bad4df9b
KP		 174712013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17472
17473 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17474 in multiply-accumulate mode.
17475
56f3e9ac
GW		 174762013-05-13 Guozhi Wei <carrot@google.com>
17477
17478 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17479
a3d7ab92
KT		 174802013-05-13 Kai Tietz <ktietz@redhat.com>
17481
17482 PR target/56975
17483 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 17484 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 17485 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 17486 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT		 17487 (ix86_expand_prologue): Likewise.
17488 (ix86_expand_split_stack_prologue): Likewise.
17489 (legitimate_pic_address_disp_p): Likewise.
17490 (legitimize_pic_address): Likewise.
17491 (legitimize_tls_address): Likewise.
17492 (legitimize_pe_coff_symbol): Likewise.
17493 (output_pic_addr_const): Likewise.
17494 (construct_plt_address): Likewise.
17495 (ix86_expand_call): Likewise.
17496 (x86_output_mi_thunk): Likewise.
17497 (x86_function_profiler): Likewise.
17498
c59b7e28
SN		 174992013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17500
17501 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17502 similar switch cases.
17503 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17504 (aarch64_simd_mov_to_<mode>low): Delete.
17505 (aarch64_simd_mov_to_<mode>high): Delete.
17506 (move_lo_quad_<mode>): Add w<-r alternative.
17507 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17508 (aarch64_simd_mov_from_*): Update type attribute.
17509 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17510 statement.
17511
5f28524a
JH		 175122013-05-13 Jan Hubicka <jh@suse.cz>
17513
17514 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17515 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB		 17516 ix86_expand_epilogue, emit_i387_cw_initialization,
17517 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17518 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 17519
ae6fa899
JJ		 175202013-05-13 Jakub Jelinek <jakub@redhat.com>
17521
17522 PR tree-optimization/45216
17523 PR tree-optimization/57157
17524 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17525 the (-Y) & (B - 1) variant if OP is |.
17526 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17527 return shifted. Use (-op1) & (prec - 1) as other_amount
17528 instead of prec - op1.
17529
4502fe8d
MJ		 175302013-05-13 Martin Jambor <mjambor@suse.cz>
17531
17532 PR middle-end/42371
17533 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17534 (ipa_constant_data): New type.
17535 (ipa_jump_func): Use ipa_constant_data to hold information about
17536 constant jump functions.
17537 (ipa_get_jf_constant): Adjust to jump function type changes.
17538 (ipa_get_jf_constant_rdesc): New function.
17539 (ipa_param_descriptor): New field controlled_uses.
17540 (ipa_get_controlled_uses): New function.
17541 (ipa_set_controlled_uses): Likewise.
17542 * ipa-ref.h (ipa_find_reference): Declare.
17543 * ipa-prop.c (ipa_cst_ref_desc): New type.
17544 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17545 changes.
17546 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17547 New parameter cs. Adjust all callers.
17548 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17549 (remove_described_reference): New function.
17550 (jfunc_rdesc_usable): Likewise.
17551 (try_make_edge_direct_simple_call): Decrement controlled use count,
17552 attempt to remove reference if it hits zero.
17553 (combine_controlled_uses_counters): New function.
17554 (propagate_controlled_uses): Likewise.
17555 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17556 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17557 (ipa_print_node_params): Print described use counter.
17558 (ipa_write_jump_function): Adjust to jump function type changes.
17559 (ipa_read_jump_function): New parameter CS, pass it to
17560 ipa_set_jf_constant. Adjust caller.
17561 (ipa_write_node_info): Stream controlled use count
17562 (ipa_read_node_info): Likewise.
17563 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17564 asserting.
17565 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17566 count. Remove cloning-added reference if it reaches zero.
17567 * ipa-ref.c (ipa_find_reference): New function.
17568
0864bfc2
GG		 175692013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17570
e7208ea3 17571 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG		 17572 alignment values for AMD BD and BT architectures.
17573
640bfeb2
MG		 175742013-05-13 Marc Glisse <marc.glisse@inria.fr>
17575
17576 * tree-vect-generic.c (uniform_vector_p): Move ...
17577 * tree.c (uniform_vector_p): ... here.
17578 * tree.h (uniform_vector_p): Declare it.
17579 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17580 into a scalar.
17581
3a60f32b
JJ		 175822013-05-13 Jakub Jelinek <jakub@redhat.com>
17583
198fe1bf
JJ		 17584 PR tree-optimization/57230
17585 * tree-ssa-strlen.c (handle_char_store): Record length for
17586 array store from STRING_CST.
17587
3a60f32b
JJ		 17588 PR tree-optimization/57230
17589 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17590 check.
17591
566be57c
JR		 175922013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17593
17594 * config/epiphany/epiphany.c (epiphany_init): Check size of
17595 NUM_MODES_FOR_MODE_SWITCHING.
17596 (epiphany_expand_prologue):
17597 Remove CONFIG_REGNUM initial value handling code.
17598 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17599 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 17600 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR		 17601 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17602 Don't return 1 for FP_MODE_NONE.
17603 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17604 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17605 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17606 * config/epiphany/epiphany.md (save_config): New pattern.
17607
0f2c2331
UB		 176082013-05-12 Uros Bizjak <ubizjak@gmail.com>
17609
17610 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17611
5b3f93c7
UB		 176122013-05-10 Uros Bizjak <ubizjak@gmail.com>
17613
17614 * config/i386/i386.md (memory): Handle sseishft1.
17615 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17616 (*vec_extractv2di_1): Ditto.
17617
1f873f0e
VM		 176182013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17619
17620 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17621 saved registers.
17622
341427fa 176232013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH		 17624
17625 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17626 Add mthumb/march=armv7-a multilib.
17627 Add mthumb/march=armv7-r multilib.
17628 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17629
9e69bdde
RC		 176302013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17631
17632 * config/v850/t-rtems: Add more multilibs.
17633
9ff09a22
RB		 176342013-05-10 Richard Biener <rguenther@suse.de>
17635
17636 PR tree-optimization/57214
17637 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17638 not propagate from SSA names that occur in abnormal PHI nodes.
17639
a5e0cd1d
MG		 176402013-05-10 Marc Glisse <marc.glisse@inria.fr>
17641
17642 * stor-layout.c (element_precision): New function.
17643 * machmode.h (element_precision): Declare it.
17644 * tree.c (build_minus_one_cst): New function.
17645 (element_precision): Likewise.
17646 * tree.h (build_minus_one_cst): Declare new function.
17647 (element_precision): Likewise.
17648 * fold-const.c (operand_equal_p): Use element_precision.
17649 (fold_binary_loc): Handle vector types.
17650 * convert.c (convert_to_integer): Use element_precision.
17651 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17652 separately.
17653
cb2558bc
RS		 176542013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17655
17656 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17657 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17658 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17659 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17660 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17661 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17662 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17663 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17664 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17665 (Uuw8): New constraints.
17666 (Usb4): Move into alphabetical order.
17667 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17668 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17669 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17670 previously unnamed patterns.
17671 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17672 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17673 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17674 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17675 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17676 of specific lengths.
17677
cb3b8d33
JJ		 176782013-05-10 Jakub Jelinek <jakub@redhat.com>
17679
6f93c008
JJ		 17680 * config/i386/i386.md (rotateinv): New code attr.
17681 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17682 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17683 roll $31, %eax, etc.
17684
cb3b8d33
JJ		 17685 PR tree-optimization/45216
17686 PR tree-optimization/57157
17687 * tree-ssa-forwprop.c (simplify_rotate): New function.
17688 (ssa_forward_propagate_and_combine): Call it.
17689
afb119be
RB		 176902013-05-10 Richard Biener <rguenther@suse.de>
17691
17692 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17693 disable peeling when we version for aliasing.
17694 (vector_alignment_reachable_p): Honor explicit user alignment.
17695 (vect_supportable_dr_alignment): Likewise.
17696 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17697 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17698 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17699 then peeling to arrange for the cost-model check to come first.
17700
01ae4861
AM		 177012013-05-10 Alan Modra <amodra@gmail.com>
17702
17703 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17704 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17705 * configure: Regenerate.
17706
ebc9a431
AM		 177072013-05-10 Alan Modra <amodra@gmail.com>
17708
17709 PR target/55033
17710 * varasm.c (default_elf_select_section): Move !DECL_P check..
17711 (get_named_section): ..to here before calling get_section_name.
17712 Adjust assertion.
17713 (default_section_type_flags): Add DECL_P check.
17714 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17715 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17716
d4bca93c
JR		 177172013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17718
17719 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17720 When using gen_stack_adjust_str with a register offset, add a
17721 REG_FRAME_RELATED_EXPR note.
17722
60ca9a65
UB		 177232013-05-09 Uros Bizjak <ubizjak@gmail.com>
17724
17725 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17726 (*vec_extractv4si_zext_mem): Ditto.
17727 (*vec_extractv2di): Add 0->x and x->x alternatives.
17728 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17729 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17730
a3409c02
JM		 177312013-05-09 Jason Merrill <jason@redhat.com>
17732
0138d6b2
JM		 17733 N3639 C++1y VLA support
17734 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17735 DECL_VALUE_EXPR.
17736
a3409c02
JM		 17737 * tree.c (build_constructor_va): New.
17738 * tree.h: Declare it.
17739
66e6b990
MJ		 177402013-05-09 Martin Jambor <mjambor@suse.cz>
17741
17742 PR lto/57084
17743 * gimple-fold.c (canonicalize_constructor_val): Call
17744 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17745
64cfa6c0
JH		 177462013-05-09 Jan Hubicka <jh@suse.cz>
17747 Richard Biener <rguenther@suse.de>
17748
17749 PR lto/54095
17750 * symtab.c (symtab_make_decl_local): Do not add private names.
17751
c3167b00
JH		 177522013-05-09 Jan Hubicka <jh@suse.cz>
17753
17754 PR lto/54095
17755 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17756 (unlink_from_assembler_name_hash): Likewise.
17757 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17758 symtab_unregister_node, symtab_initialize_asm_name_hash,
17759 change_decl_assembler_name): Update.
17760
12dc6974
SN		 177612013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17762
17763 * config/aarch64/aarch64.md: New movtf split.
17764 (*movtf_aarch64): Update.
17765 (aarch64_movdi_tilow): Handle TF modes and rename to
17766 aarch64_movdi_<mode>low.
17767 (aarch64_movdi_tihigh): Handle TF modes and rename to
17768 aarch64_movdi_<mode>high
17769 (aarch64_movtihigh_di): Handle TF modes and rename to
17770 aarch64_mov<mode>high_di
17771 (aarch64_movtilow_di): Handle TF modes and rename to
17772 aarch64_mov<mode>low_di
17773 (aarch64_movtilow_tilow): Remove spurious whitespace.
17774 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17775 splits.
17776 (aarch64_print_operand): Update.
17777
227eb343
AM		 177782013-05-09 Alan Modra <amodra@gmail.com>
17779
17780 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17781 powerpc64le.
17782 * configure: Regenerate.
17783
0b013847
UB		 177842013-05-08 Uros Bizjak <ubizjak@gmail.com>
17785
17786 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17787 splitter preparation statements.
17788 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17789 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17790 adjust_address_nv.
17791
1dc3d6e9
BS		 177922013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17793
17794 * gimple-ssa-strength-reduction.c (count_candidates): Change
17795 return value to int.
17796 (analyze_candidates_and_replace): Change type of length to int.
17797
e61e7d28
UB		 177982013-05-08 Uros Bizjak <ubizjak@gmail.com>
17799
17800 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17801 (*vec_extract<mode>): Use VI12_128 mode iterator.
17802 (*vec_extract<mode>_mem): Ditto.
17803 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17804 attribute.
17805
4fbfcf44
DN		 178062013-05-08 Diego Novillo <dnovillo@google.com>
17807
17808 PR bootstrap/54659
17809
17810 Revert:
4fbfcf44
DN		 17811 2012-08-17 Diego Novillo <dnovillo@google.com>
17812
e61e7d28
UB		 17813 PR bootstrap/54281
17814 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17815 * config.in: Regenerate.
17816 * configure: Regenerate.
17817 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 17818
702d8703
JH		 178192013-05-08 Jan Hubicka <jh@suse.cz>
17820
17821 PR lto/54095
17822 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17823 * cgraph.h (symtab_node_base): Add unique_name.
17824 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17825 input_overwrite_node, input_varpool_node): Stream unique_name.
17826 * cgraphclones.c (cgraph_create_virtual_clone,
17827 cgraph_function_versioning): Set unique_name.
17828 * ipa.c (function_and_variable_visibility): Set unique_name.
17829
8b28cf47
BS		 178302013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17831
17832 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17833 (alloc_cand_and_find_basis): Restrict conditional candidate
17834 processing to CAND_MULTs.
17835
e86074fd
JH		 178362013-05-08 Jan Hubicka <jh@suse.cz>
17837
17838 PR lto/54095
17839 lto-symtab.c (lto_symtab_symbol_p): New function.
17840 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17841 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17842 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17843 Skip static symbols.
17844
44398cbe
PC		 178452013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17846
17847 PR tree-optimization/57200
17848 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17849 Only call inform if the preceding warning_at returns true.
17850
f6bc1c4a
HS		 178512013-05-07 Han Shen <shenhan@google.com>
17852
17853 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17854 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17855 * common.opt (fstack-protector-strong): New option.
17856 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17857 * doc/invoke.texi (Optimization Options): Document
17858 "-fstack-protector-strong".
17859 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17860
4ffecb1f
SB		 178612013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17862
17863 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17864
5a107a0f
VM		 178652013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17866
17867 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17868 it is not lhs of the insn.
17869 (setup_reg_equiv): Remove insn having equiv memory which it is not
17870 lhs of the insn.
17871 * lra-constraints.c (process_address): Try to improve generation
17872 code for address base + disp.
17873 (lra_constraints): Make correct the code for checking insn setting
17874 up backward equivalence. Remove insn only if it is in the init
17875 insn list.
17876 * lra-eliminations.c (update_reg_eliminate): Change return value.
17877 (lra_eliminate): Use the result.
17878
3f5783ea
UB		 178792013-05-07 Uros Bizjak <ubizjak@gmail.com>
17880
17881 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17882 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17883 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17884 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17885 register target operands.
17886 (*vec_extractv8hi_sse2): New pattern.
17887 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17888 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17889 (*vec_extract<mode>_mem): New insn and split pattern.
17890
8a5800b8
CL		 178912013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17892
17893 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17894 (TARGET_ASAN_SHADOW_OFFSET): Define.
17895 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17896 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17897
7bf55a70
BS		 178982013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17899
17900 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17901 (incr_vec_index): Return -1 if increment not found.
17902 (create_add_on_incoming_edge): Assert if increment not found.
17903 (record_increment): Limit number of increments recorded.
17904 (all_phi_incrs_profitable): Return false if an increment not found.
17905 (replace_profitable_candidates): Don't process increments that were
17906 not recorded.
17907 (analyze_candidates_and_replace): Limit size of incr_vec.
17908
3f8825c0
RB		 179092013-05-07 Richard Biener <rguenther@suse.de>
17910
17911 * calls.c (special_function_p): setjmp-like functions are leaf.
17912 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17913 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17914
fd4842cd
SN		 179152013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17916
17917 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17918 (aarch64_simd_mov<mode>): New expander.
17919 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17920 (aarch64_simd_mov_to_<mode>high): Likewise.
17921 (aarch64_simd_mov_from_<mode>low): Likewise.
17922 (aarch64_simd_mov_from_<mode>high): Likewise.
17923 (aarch64_dup_lane<mode>): Update.
17924 (aarch64_dup_lanedi): New instruction pattern.
17925 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17926 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17927
a7a7d10e
BS		 179282013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17929
17930 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17931 (replace_mult_candidate): Remove unnecessary argument; remove
17932 unnecessary parameter from call to introduce_cast_before_cand.
17933 (replace_unconditional_candidate): Remove unnecessary parameter
17934 from call to replace_mult_candidate.
17935 (replace_conditional_candidate): Likewise.
17936 (insert_initializers): Use make_temp_ssa_name.
17937 (introduce_cast_before_cand): Remove unnecessary argument; use
17938 make_temp_ssa_name.
17939 (replace_one_candidate): Remove unnecessary argument; remove
17940 unnecessary parameter from calls to introduce_cast_before_cand.
17941 (replace_profitable_candidates): Remove unnecessary parameters
17942 from calls to replace_one_candidate.
17943
29105868
BS		 179442013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17945
17946 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17947 phi def as possibly hiding a basis for a CAND_ADD whose operands
17948 have been commuted in the analysis.
17949 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17950
4095f9fa
N		 179512013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17952
17953 * config/aarch64/aarch64.md
17954 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17955 shift value between 0-4.
17956
39e843e8
RB		 179572013-05-07 Richard Biener <rguenther@suse.de>
17958
17959 * double-int.h (rshift): New overload.
17960 * double-int.c (rshift): New function.
17961 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17962 (create_reference_ops_from_ref): Remove.
17963 (vn_reference_insert): Use shared ops for constructing the
17964 reference and copy it.
17965
0a1a83cb
RB		 179662013-05-07 Richard Biener <rguenther@suse.de>
17967
17968 PR middle-end/57190
17969 * tree-eh.c (sink_clobbers): Properly propagate
17970 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17971
43303d6f
JJ		 179722013-05-07 Jakub Jelinek <jakub@redhat.com>
17973
ba7e83f8
JJ		 17974 PR tree-optimization/57149
17975 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17976 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17977 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17978 uninit_undefined_value_p instead of ssa_undefined_value_p.
17979
43303d6f
JJ		 17980 PR debug/57184
17981 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17982 for modifier == EXPAND_INITIALIZER.
17983
14523c25
AB		 179842013-05-07 Anton Blanchard <anton@samba.org>
17985
17986 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17987 for powerpc64 little endian.
17988 * configure: Regenerate.
17989
cb7c8be9
GS		 179902013-05-06 Graham Stott <grahams@btinternet.com>
17991
17992 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17993 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17994 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17995 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17996
179972013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS		 17998
17999 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18000 codes which allow non-lvalues.
18001
9a0ee7b0
MG		 180022013-05-06 Marc Glisse <marc.glisse@inria.fr>
18003
18004 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18005 components are all 1s.
18006 (integer_minus_onep): New function.
18007 * tree.h (integer_minus_onep): Declare it.
18008 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18009 integer_minus_onep instead of integer_all_onesp.
18010
f2c17ea9
OE		 180112013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18012
18013 PR target/52933
18014 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18015 variations of these patterns.
18016
f75e6a51
UB		 180172013-05-06 Uros Bizjak <ubizjak@gmail.com>
18018
18019 * config/i386/i386.md (isa): Add x64_sse4 member.
18020 (enabled): Handle x64_sse4.
18021 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18022 instruction for 64bit SSE4_1 targets. Update insn attributes.
18023 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18024 instruction for SSE4_1 targets. Update insn attributes.
18025 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18026 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18027 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18028 const_1 selector.
18029 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18030 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18031 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18032
a986d468
OE		 180332013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18034
18035 PR target/57108
18036 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18037
382522cb
MK		 180382013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18039
18040 * final.c (do_assembler_dialects): Don't handle curly braces and
18041 vertical bar escaped by % as dialect delimiters.
18042 (output_asm_insn): Print curly braces and vertical bar if escaped
18043 by % and ASSEMBLER_DIALECT defined.
18044 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18045 * doc/tm.texi: Regenerated.
18046
fb0d5c60
SB		 180472013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18048
fb0d5c60
SB		 18049 * config/mips/mips.c: Include tree-pass.h.
18050 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18051 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18052 (pass_mips_machine_reorg2): New machine specific pass.
18053 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18054 (mips_option_override): Register the new pass.
18055 * rtl.h (cleanup_barriers): Remove prototype.
18056 (dbr_schedule): Likewise.
18057 * jump.c (cleanup_barriers): Make static.
18058 * reorg.c (dbr_schedule): Likewise.
18059
aa06a978
RB		 180602013-05-06 Richard Biener <rguenther@suse.de>
18061
18062 PR tree-optimization/57185
18063 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18064 SSA names properly.
18065
0b953bec
UB		 180662013-05-06 Uros Bizjak <ubizjak@gmail.com>
18067
18068 PR target/57106
18069 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18070 between "first_arg" and "insn", not "last" and "insn".
18071
28708525
WS		 180722013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18073
18074 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18075 (find_candidates_in_block): Re-enable slsr_process_phi.
18076 (create_phi_basis): Fix double counting of candidate adjustment.
18077
0107dca2
RB		 180782013-05-06 Richard Biener <rguenther@suse.de>
18079
18080 PR middle-end/57147
18081 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18082 the edge is also fallthru, preserve it and just clear the
18083 abnormal flag.
18084 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18085 also complex, preserve that and just clear the fallthru flag.
18086 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18087 update virtual operands.
18088
470d4d13
AM		 180892013-05-06 Alan Modra <amodra@gmail.com>
18090
18091 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18092 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18093 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18094 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18095 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18096 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18097
8f1dbf8d
AM		 180982013-05-06 Alan Modra <amodra@gmail.com>
18099
18100 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18101 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18102 (DEFAULT_ASM_ENDIAN): Define.
18103 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18104 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18105 Update -K PIC clause from sysv4.h.
18106 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18107 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18108
54890767
AM		 181092013-05-06 Alan Modra <amodra@gmail.com>
18110
18111 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18112 twice for little-endian.
18113 (ashrdi3_no_power, ashrdi3): Support little-endian.
18114
2353515d
OE		 181152013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18116
18117 PR target/55303
18118 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18119 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18120 related expanders.
18121 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18122 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18123 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18124 New predicates.
18125
33e67557
SB		 181262013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18127 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18128
18129 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18130 * config/pa/pa.opt: Make mbig-switch a no-op.
18131 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18132 (CASE_VECTOR_MODE): Always return SImode.
18133 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18134 for the !TARGET_BIG_SWITCH case.
18135 * config/pa/pa-linux.h: Likewise.
18136 * config/pa/pa-openbsd.h: Likewise.
18137 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18138 * config/pa/pa.md (short_jump): Remove define_insn.
18139 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18140 (casesi0): Remove define_insn.
18141 (type): Remove btable_branch.
18142 (pa_combine_type): Likewise.
18143 (in_nullified_branch_delay): Likewise.
18144 (in_call_delay): Likewise.
18145 (define_delay): Likewise.
18146 (define_insn_reservation "Z3"): Likewise.
18147 (define_insn_reservation "Z4"): Likewise.
18148 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18149 (pa_adjust_insn_length): Remove adjustment for btable branches.
18150 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18151 and mno-big-switch
18152
3095685e
UB		 181532013-05-05 Uros Bizjak <ubizjak@gmail.com>
18154
18155 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18156 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18157 Add m->r,x alternatives.
18158 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18159 splitters using SWI48x mode iterator.
18160 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18161 TARGET_64BIT. Add m->x alternative.
18162 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18163 Add o->x alternative. Enable for TARGET_SSE.
18164 (sse_storeq): Remove expander.
18165 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18166 with memory input operand.
18167 (*vec_extractv2di_1 splitter): New.
18168 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18169 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18170
4b36ae28
SB		 181712013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18172
18173 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18174 (INT_LOWPART): Delete.
18175 (extract_MB): Adjust.
18176 (extract_ME): Adjust.
18177 (print_operand): Adjust.
18178
da226db2
SB		 181792013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18180
18181 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18182 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18183 (reg_or_logical_cint_operand, easy_fp_constant,
18184 logical_const_operand): Delete "CONST_DOUBLE" case.
18185 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18186 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 18187 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB		 18188 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18189 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18190 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18191 test.
18192 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18193 CONST_DOUBLE DImode/VOIDmode case.
18194 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18195 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18196 CONST_DOUBLE VOIDmode case.
18197 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18198 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18199 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18200 Delete CONST_DOUBLE case.
18201 (splitters for mov FMOVE64 const_double): Delete
18202 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18203 "HOST_BITS_PER_WIDE_INT >= 64" test.
18204 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18205 case.
18206 (mov DI const_double): Delete.
18207
40de22d6
JJ		 182082013-05-04 Jakub Jelinek <jakub@redhat.com>
18209
18210 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18211 on op shows all bits zero in mode of a lowpart subreg, return zero.
18212
5ec6aff2
MM		 182132013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18214
18215 PR target/57150
18216 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18217 to save TFmode registers and DImode to save TImode registers for
18218 caller save operations.
18219 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18220 mark being partially clobbered since they only use the first
18221 double word.
18222
18223 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18224 and TDmode only use the upper 64-bits of each VSX register.
18225
2cefad90
BS		 182262013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18227
18228 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18229 (find_candidates_in_block): Disable slsr_process_phi.
18230
d6d7eee1
GW		 182312013-05-03 Guozhi Wei <carrot@google.com>
18232
18233 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB		 18234 constructor to ...
18235 (build_init_ctor): ... here.
d6d7eee1 18236
9b92d12b
BS		 182372013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18238
18239 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18240 (slsr_cand_d): Redefine def_phi.
18241 (stride_status, phi_adjust_status, count_phis_status): New enums.
18242 (find_phi_def): New.
18243 (find_basis_for_base_expr): New.
18244 (find_basis_for_candidate): Handle hidden bases.
18245 (alloc_cand_and_find_basis): Handle phi candidates.
18246 (slsr_process_phi): New.
18247 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18248 (create_mul_imm_cand): Likewise.
18249 (create_add_ssa_cand): Exclude phi base candidates.
18250 (create_add_imm_cand): Likewise.
18251 (slsr_process_cast): Likewise.
18252 (slsr_process_copy): Likewise.
18253 (find_candidates_in_block): Handle phi candidates.
18254 (dump_candidate): Likewise.
18255 (unconditional_cands): Delete.
18256 (unconditional_cands_with_known_stride_p): Delete.
18257 (phi_dependent_cand_p): New.
18258 (cand_increment): Handle phi-dependent candidates.
18259 (replace_dependent): Delete.
18260 (replace_mult_candidate): New.
18261 (replace_unconditional_candidate): New.
18262 (incr_vec_index): Move to avoid forward reference.
18263 (create_add_on_incoming_edge): New.
18264 (create_phi_basis): New.
18265 (replace_dependents): Delete.
18266 (replace_conditional_candidate): New.
18267 (phi_add_costs): New.
18268 (replace_uncond_cands_and_profitable_phis): New.
18269 (record_increment): Handle phi adjustments.
18270 (record_phi_increments): New.
18271 (record_increments): Handle phi adjustments.
18272 (phi_incr_cost): New.
18273 (lowest_cost_path): Handle phis.
18274 (total_savings): Likewise.
18275 (analyze_increments): Likewise.
18276 (ncd_with_phi): New.
18277 (ncd_of_cand_and_phis): New.
18278 (nearest_common_dominator_for_cands): Handle phi increments.
18279 (all_phi_incrs_profitable): New.
18280 (replace_profitable_candidates): Handle phi-dependent candidates.
18281 (analyze_candidates_and_replace): Likewise.
18282
68f073d4
TJ		 182832013-05-03 Teresa Johnson <tejohnson@google.com>
18284
18285 PR bootstrap/57154
18286 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18287 do not exceed REG_BR_PROB_BASE.
18288
a4ee7cb9
JL		 182892013-05-03 Jeff Law <law@redhat.com>
18290
ade67f70 18291 PR tree-optimization/57144
a4ee7cb9
JL		 18292 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18293 operand of the condition will bit into the new type when eliminating
18294 a cast feeding a condition.
18295
47954c4e
JJ		 182962013-05-03 Jakub Jelinek <jakub@redhat.com>
18297
18298 PR rtl-optimization/57130
3095685e
UB		 18299 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18300 of COMPARE as in_code to the recursive call if needed.
47954c4e 18301
3c21604f
UB		 183022013-05-03 Uros Bizjak <ubizjak@gmail.com>
18303
18304 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18305 (enabled): Handle new members.
18306 * config/i386/sse.md (*vec_concatv2si): Merge from
18307 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18308 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18309
12211b99 183102013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR		 18311
18312 PR tree-optimization/57027
18313 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18314 for fnms opportunity, check we got the prerequisite kind
18315 of tree / gimple before using accessor functions.
18316
07bfc9ec
RB		 183172013-05-03 Richard Biener <rguenther@suse.de>
18318
18319 * double-int.h (lshift): New overload without precision
18320 and arith argument.
18321 (operator *=, operator +=, operator -=): Move ...
18322 * double-int.c (operator *=, operator +=, operator -=): ... here
18323 and implement more efficiently.
18324 (mul_double_with_sign): Remove.
18325 (lshift_double): Adjust to take unsinged shift argument, push
18326 dispatching code to callers.
18327 (mul_double_wide_with_sign): Add early out for callers that
18328 are not interested in high parts or overflow.
18329 (lshift): New function.
18330 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18331 dispatch code here.
18332 (lrotate, rrotate): Use logical shifts.
18333 * expr.c (get_inner_reference): Use lshift.
18334 * fixed-value.c (do_fixed_divide): Likewise.
18335 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18336 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18337 (indirect_refs_may_alias_p): Likewise.
18338 (stmt_kills_ref_p_1): Likewise.
18339
7769bb64
VP		 183402013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18341
18342 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18343
1d0c8e5c
VP		 183442013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18345
18346 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18347 scalar form of FABD instruction.
18348
f15feaf9
VM		 183492013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18350
18351 * lra-constraints.c (process_alt_operands): Add checking alt
18352 number to choose the best alternative.
18353
d90e76d4
RB		 183542013-05-02 Richard Biener <rguenther@suse.de>
18355
18356 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18357 bitmap and its handling.
18358 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18359
9f8e7a96
RB		 183602013-05-02 Richard Biener <rguenther@suse.de>
18361
18362 PR middle-end/57140
18363 * tree-inline.c (copy_loops): Properly handle removed loops.
18364 (copy_cfg_body): Mark destination loops for fixup if source
18365 loops needed fixup.
18366
f3a81b39
GY		 183672013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18368
18369 PR target/56732
18370 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18371 generating simple_return for naked functions.
18372
7b920a9a
MJ		 183732013-05-02 Martin Jambor <mjambor@suse.cz>
18374
18375 PR middle-end/56988
18376 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18377 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18378 flags match.
18379 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18380 ipa_agg_replacement_value structures.
18381 (known_aggs_to_agg_replacement_list): Likewise.
18382 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18383 (read_agg_replacement_chain): Likewise.
18384 (ipcp_transform_function): Also check that by_ref flags match.
18385
2c41c19d
RB		 183862013-05-02 Richard Biener <rguenther@suse.de>
18387
18388 * graphds.h (struct graph): Add obstack member.
18389 * graphds.c (new_graph): Initialize obstack and allocate
18390 vertices from it.
18391 (add_edge): Allocate edge from the obstack.
3c21604f 18392 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 18393
8b47039c
TJ		 183942013-05-02 Teresa Johnson <tejohnson@google.com>
18395
18396 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18397 divides.
18398 * cfg.c (update_bb_profile_for_threading): Ditto.
18399 * tree-inline.c (copy_bb): Ditto.
18400 (copy_edges_for_bb): Ditto.
18401 (initialize_cfun): Ditto.
18402 (copy_cfg_body): Ditto.
18403 (expand_call_inline): Ditto.
18404 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18405 (estimate_node_size_and_time): Ditto.
18406 (inline_merge_summary): Ditto.
18407 * cgraphclones.c (cgraph_clone_edge): Ditto.
18408 (cgraph_clone_node): Ditto.
18409 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18410 (compute_trg_info): Ditto.
18411
da65928c
IB		 184122013-05-02 Ian Bolton <ian.bolton@arm.com>
18413
18414 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18415 S reg when fp attribute set.
18416 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18417
add3c965
IB		 184182013-05-02 Ian Bolton <ian.bolton@arm.com>
18419
18420 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18421 New pattern.
18422 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18423 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18424 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18425
fdd43ac4
RB		 184262013-05-02 Richard Biener <rguenther@suse.de>
18427
18428 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18429 (struct instantiate_cache_entry): New type.
18430 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18431 (struct instantiate_cache_type): New type.
18432 (set_instantiated_value, get_instantiated_value): Remove.
18433 (get_instantiated_value_entry): New function.
18434 (instantiate_scev_name): Use the new cache and adjust.
18435 (instantiate_scev_poly): Adjust.
18436 (instantiate_scev_binary): Likewise.
18437 (instantiate_array_ref): Likewise.
18438 (instantiate_scev_convert): Likewise.
18439 (instantiate_scev_not): Likewise.
18440 (instantiate_scev_3): Likewise.
18441 (instantiate_scev_2): Likewise.
18442 (instantiate_scev_r): Likewise.
18443 (instantiate_scev): Likewise.
18444 (resolve_mixers): Likewise.
18445
36ff9dfb
VM		 184462013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18447
18448 PR target/57091
18449 * lra-constraints.c (best_small_class_operands_num): Remove.
18450 (process_alt_operands): Remove small_class_operands_num. Take
18451 small classes operands into losers and only if the operand is not
18452 matched. Modify debugging output.
18453 (curr_insn_transform): Remove best_small_class_operands_num.
18454 Print insn name.
18455
36054fab
JG		 184562013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18457
18458 * config/aarch64/aarch64-builtins.c
18459 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18460 * config/aarch64/aarch64-simd-builtins.def
18461 (reduc_splus_): Add new modes.
18462 (reduc_uplus_): New.
18463 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18464 (reduc_uplus_v4sf): Likewise.
18465 (reduc_splus_v4sf): Likewise.
18466 (aarch64_addv<mode>): Likewise.
18467 (reduc_uplus_<mode>): Likewise.
18468 (reduc_splus_<mode>): Likewise.
18469 (aarch64_addvv2di): Likewise.
18470 (reduc_uplus_v2di): Likewise.
18471 (reduc_splus_v2di): Likewise.
18472 (aarch64_addvv2si): Likewise.
18473 (reduc_uplus_v2si): Likewise.
18474 (reduc_splus_v2si): Likewise.
18475 (reduc_<sur>plus_<mode>): New.
18476 (reduc_<sur>plus_v2di): Likewise.
18477 (reduc_<sur>plus_v2si): Likewise.
18478 (reduc_<sur>plus_v4sf): Likewise.
18479 (aarch64_addpv4sf): Likewise.
18480 * config/aarch64/arm_neon.h
18481 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18482 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18483 add UNSPEC_SADDV, UNSPEC_UADDV.
18484 (SUADDV): New.
18485 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18486
6dce23a8
JG		 184872013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18488
18489 * config/aarch64/arm_neon.h
18490 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18491
1598945b
JG		 184922013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18493
18494 * config/aarch64/aarch64-builtins
18495 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18496
998eaf97
JG		 184972013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18498
18499 * config/aarch64/aarch64-simd-builtins.def
18500 (reduc_smax_): New.
18501 (reduc_smin_): Likewise.
18502 (reduc_umax_): Likewise.
18503 (reduc_umin_): Likewise.
18504 (reduc_smax_nan_): Likewise.
18505 (reduc_smin_nan_): Likewise.
18506 (fmax): Remove.
18507 (fmin): Likewise.
18508 (smax): Update for V2SF, V4SF and V2DF modes.
18509 (smin): Likewise.
18510 (smax_nan): New.
18511 (smin_nan): Likewise.
18512 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18513 (<su><maxmin><mode>3): ...This, refactor.
18514 (s<maxmin><mode>3): New.
18515 (<maxmin_uns><mode>3): Likewise.
18516 (reduc_<maxmin_uns>_<mode>): Refactor.
18517 (reduc_<maxmin_uns>_v4sf): Likewise.
18518 (reduc_<maxmin_uns>_v2si): Likewise.
18519 (aarch64_<fmaxmin><mode>: Remove.
18520 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18521 new builtin names.
18522 (vmin<q>_f<32,64>): Likewise.
18523 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18524 (FMAXMIN): New.
18525 (su): Add mappings for smax, smin, umax, umin.
18526 (maxmin): New.
18527 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18528 (FMAXMIN): Rename as...
18529 (FMAXMIN_UNS): ...This.
18530 (maxminv): Remove.
18531 (fmaxminv): Likewise.
18532 (fmaxmin): Likewise.
18533 (maxmin_uns): New.
18534 (maxmin_uns_op): Likewise.
18535
bd11644e
JG		 185362013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18537
18538 * config/aarch64/arm_neon.h
18539 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18540 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18541 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18542
75dd5ace
JG		 185432013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18544
18545 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18546 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18547
7c19979f
JG		 185482013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18549
18550 * config/aarch64/aarch64-simd.md
18551 (vcond<mode>_internal): Handle special cases for constant masks.
18552 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18553 (vcondu<mode><mode>): Likewise.
18554 (vcond<v_cmp_result><mode>): New.
18555
bb60efd9
JG		 185562013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18557
18558 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18559 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18560 * config/aarch64/aarch64-simd-builtins.def
18561 (cmeq): Update to BUILTIN_VALLDI.
18562 (cmgt): Likewise.
18563 (cmge): Likewise.
18564 (cmle): Likewise.
18565 (cmlt): Likewise.
18566 * config/aarch64/arm_neon.h
18567 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18568 to builtins or C as appropriate.
18569
889b9412
JG		 185702013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18571
18572 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18573 (cmgeu): ...This.
18574 (cmhi): Rename to...
18575 (cmgtu): ...This.
18576 * config/aarch64/aarch64-simd.md
18577 (simd_mode): Add SF.
18578 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18579 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18580 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18581 (cstore<mode>_neg): ...This.
18582 * config/aarch64/iterators.md
18583 (VALLF): new.
18584 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18585 (COMPARISONS): New.
18586 (UCOMPARISONS): Likewise.
18587 (optab): Add missing comparisons.
18588 (n_optab): New.
18589 (cmp_1): Likewise.
18590 (cmp_2): Likewise.
18591 (CMP): Likewise.
18592 (cmp): Remove.
18593 (VCMP_S): Likewise.
18594 (VCMP_U): Likewise.
18595 (V_cmp_result): Add DF, SF modes.
18596 (v_cmp_result): Likewise.
18597 (v): Likewise.
18598 (vmtype): Likewise.
18599 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18600
0a7dbb76
GY		 186012013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18602
18603 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18604 define_insn to define_insn_and_split.
18605 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18606 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18607 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18608 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18609 (thumb2_negscc): Likewise.
18610
fb614ca6
GY		 186112013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18612
18613 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18614
9e64a0bf
GY		 186152013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18616
18617 * config/arm/thumb2.md: Remove trailing whitespaces.
18618
d6b28156
RS		 186192013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18620
18621 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18622 Use gen_int_mode rather than GEN_INT.
18623
f91674c3
L		 186242013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18625
b0dec607 18626 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L		 18627 debug_gimple_stmt.
18628
3551257c
RB		 186292013-04-30 Richard Biener <rguenther@suse.de>
18630
18631 PR middle-end/57122
3c21604f 18632 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 18633
4c1aff1c
RB		 186342013-04-30 Richard Biener <rguenther@suse.de>
18635
18636 PR middle-end/57107
18637 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18638
54b8379a
AB		 186392013-04-30 Andrey Belevantsev <abel@ispras.ru>
18640
18641 PR rtl-optimization/56957
18642 PR rtl-optimization/57105
54b8379a
AB		 18643 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18644 variable. Use just INSN_UID for determining whether an insn
18645 should be only disconnected from the insn stream.
18646 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18647
abc27962
JJ		 186482013-04-30 Jakub Jelinek <jakub@redhat.com>
18649
18650 PR tree-optimization/57104
18651 * tsan.c (instrument_expr): Don't instrument accesses to
18652 DECL_HARD_REGISTER VAR_DECLs.
18653
0fc822d0
RB		 186542013-04-30 Richard Biener <rguenther@suse.de>
18655
18656 * function.h (loops_for_fn): New inline function.
18657 (set_loops_for_fn): Likewise.
18658 * cfgloop.h (place_new_loop): Add struct function parameter.
18659 (get_loop): Likewise.
18660 (get_loops): Likewise.
18661 (number_of_loops): Likewise.
18662 (fel_next): Adjust.
18663 (fel_init): Likewise.
18664 * cfg.c (get_loop_copy): Adjust.
18665 * cfgloop.c (flow_loops_dump): Likewise.
18666 (record_loop_exits): Likewise.
18667 (verify_loop_structure): Likewise.
18668 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18669 (estimate_reg_pressure_cost): Likewise.
18670 (mark_loop_exit_edges): Likewise.
18671 * cfgloopmanip.c (place_new_loop): Likewise.
18672 (add_loop): Likewise.
18673 (duplicate_loop): Likewise.
18674 * graph.c (draw_cfg_nodes): Likewise.
18675 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18676 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18677 (extract_affine_chrec): Likewise.
18678 (build_scop_iteration_domain): Likewise.
18679 * graphite.c (graphite_initialize): Likewise.
18680 * ira-build.c (create_loop_tree_nodes): Likewise.
18681 (more_one_region_p): Likewise.
18682 (rebuild_regno_allocno_maps): Likewise.
18683 (mark_loops_for_removal): Likewise.
18684 (mark_all_loops_for_removal): Likewise.
18685 (remove_unnecessary_regions): Likewise.
18686 (ira_build): Likewise.
18687 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18688 * loop-init.c (fix_loop_structure): Likewise.
18689 (gate_rtl_move_loop_invariants): Likewise.
18690 (gate_rtl_unswitch): Likewise.
18691 (gate_rtl_unroll_and_peel_loops): Likewise.
18692 (rtl_doloop): Likewise.
18693 * lto-streamer-in.c (input_cfg): Likewise.
18694 * lto-streamer-out.c (output_cfg): Likewise.
18695 * modulo-sched.c (sms_schedule): Likewise.
18696 * predict.c (tree_estimate_probability): Likewise.
18697 (tree_estimate_probability_driver): Likewise.
18698 (estimate_loops): Likewise.
18699 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18700 (move_sese_region_to_fn): Likewise.
18701 (debug_loop_num): Likewise.
18702 * tree-chrec.c (chrec_evaluate): Likewise.
18703 (hide_evolution_in_other_loops_than_loop): Likewise.
18704 (chrec_component_in_loop_num): Likewise.
18705 (reset_evolution_in_loop): Likewise.
18706 (evolution_function_is_invariant_rec_p): Likewise.
18707 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18708 * tree-inline.c (copy_loops): Likewise.
18709 (copy_cfg_body): Likewise.
18710 (tree_function_versioning): Likewise.
18711 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18712 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18713 Likewise.
18714 (add_to_evolution_1): Likewise.
18715 (scev_const_prop): Likewise.
18716 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18717 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18718 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18719 (tree_ssa_lim_initialize): Likewise.
18720 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18721 (verify_loop_closed_ssa): Likewise.
18722 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18723 (tree_ssa_loop_im): Likewise.
18724 (tree_ssa_loop_unswitch): Likewise.
18725 (tree_vectorize): Likewise.
18726 (check_data_deps): Likewise.
18727 (tree_ssa_loop_ivcanon): Likewise.
18728 (tree_ssa_loop_bounds): Likewise.
18729 (tree_complete_unroll): Likewise.
18730 (tree_complete_unroll_inner): Likewise.
18731 (tree_parallelize_loops): Likewise.
18732 (tree_ssa_loop_prefetch): Likewise.
18733 (tree_ssa_loop_ivopts): Likewise.
18734 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18735 * tree-vectorizer.c (vectorize_loops): Likewise.
18736
37953bd3
MF		 187372013-04-29 Mike Frysinger <vapier@gentoo.org>
18738
18739 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18740 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18741 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18742 with EABI_LINK_SPEC.
18743
f9ed28db
UB		 187442013-04-29 Uros Bizjak <ubizjak@gmail.com>
18745
18746 PR target/44578
18747 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18748 alternative.
18749
deca73f5
VM		 187502013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18751
18752 PR target/57097
37953bd3 18753 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM		 18754 using memory for pseudos. Print cost dump for alternatives.
18755 Modify cost values for conflicts with early clobbers.
18756 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18757
fdca7d03
UB		 187582013-04-29 Uros Bizjak <ubizjak@gmail.com>
18759
18760 PR target/57098
18761 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18762
0b064172
IB		 187632013-04-29 Ian Bolton <ian.bolton@arm.com>
18764
18765 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18766 from/to S register.
18767 (movdi_aarch64): Support LDR/STR from/to D register.
18768
473cec55
IB		 187692013-04-29 Ian Bolton <ian.bolton@arm.com>
18770
18771 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18772 or higher by default.
18773
a6f30e66
RB		 187742013-04-29 Richard Biener <rguenther@suse.de>
18775
18776 PR middle-end/57075
18777 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18778 even if not adding abnormal edges for calls that can make
18779 abnormal gotos.
18780
0c2b2040
RB		 187812013-04-29 Richard Biener <rguenther@suse.de>
18782
18783 PR middle-end/57103
18784 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18785 TREE_BLOCK.
18786 (move_stmt_r): Remove redundant checking.
18787
f41f80f9
TJ		 187882013-04-29 Teresa Johnson <tejohnson@google.com>
18789
18790 PR bootstrap/57077
18791 * basic-block.h (apply_scale): New function.
18792 (apply_probability): Use apply_scale.
18793 * gimple-streamer-in.c (input_bb): Ditto.
18794 * lto-streamer-in.c (input_cfg): Ditto.
18795 * lto-cgraph.c (merge_profile_summaries): Ditto.
18796 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 18797 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ		 18798 (copy_edges_for_bb): Ditto.
18799 (copy_cfg_body): Ditto.
18800
315bbd2e
TV		 188012013-04-29 Tom de Vries <tom@codesourcery.com>
18802
18803 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18804 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18805 (tail_merge_optimize): Handle current_loops == NULL.
18806
ebbd90d8
JL		 188072013-04-26 Jeff Law <law@redhat.com>
18808
18809 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18810 (simplify_cond_using_ranges): Generalize code to simplify
18811 COND_EXPRs where one argument is a constant and the other
18812 is an SSA_NAME created by an integral type conversion.
18813
8b9b57eb
KT		 188142013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18815
18816 * config/arm/arm.md (store_minmaxsi): Use only when
18817 optimize_insn_for_size_p.
18818
9498e5dc
CB		 188192013-04-29 Christian Bruel <christian.bruel@st.com>
18820
18821 PR target/57108
18822 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18823
6093bc06
RB		 188242013-04-29 Richard Biener <rguenther@suse.de>
18825
18826 PR middle-end/57089
fdca7d03
UB		 18827 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18828 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB		 18829 (expand_omp_for_generic): Properly add loops.
18830 (expand_omp_for_static_nochunk): Likewise.
18831 (expand_omp_for_static_chunk): Likewise.
18832 (expand_omp_for): For the degenerate case fixup loops.
18833 (expand_omp_sections): Fix default bb placement in loops.
18834 (expand_omp_atomic_pipeline): Properly add loops.
18835
84aacbfd
KT		 188362013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18837
18838 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18839
8dee4479
TV		 188402013-04-29 Tom de Vries <tom@codesourcery.com>
18841
18842 * tree-ssa-tail-merge.c: Update header comment.
18843
47934dc4
JG		 188442013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18845
18846 * config/aarch64/arm_neon.h
18847 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18848 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18849 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18850 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18851 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18852 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18853 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18854 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18855
384be29f
JG		 188562013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18857
18858 * config/aarch64/aarch64-simd.md
18859 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18860 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18861 fix_trunc, fixuns_trunc.
18862 (ftrunc<VDQF:mode>2): New.
18863 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18864 (fix_trunc_optab): New.
18865
0386b123
JG		 188662013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18867
18868 * config/aarch64/aarch64-builtins.c
18869 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18870 iceilf, lround, iroundf.
18871
00fcb892
UB		 188722013-04-29 Uros Bizjak <ubizjak@gmail.com>
18873
18874 PR target/54349
18875 * config/i386/i386.h (enum ix86_tune_indices)
18876 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18877 New, split from X86_TUNE_INTER_UNIT_MOVES.
18878 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18879 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18880 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18881 (TARGET_INTER_UNIT_MOVES): Remove.
18882 * config/i386/i386.c (initial_ix86_tune_features): Update.
18883 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18884 (ix86_expand_convert_uns_didf_sse): Use
18885 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18886 (ix86_expand_vector_init_one_nonzero): Ditto.
18887 (ix86_expand_vector_init_interleave): Ditto.
18888 (inline_secondary_memory_needed): Return true for moves from SSE class
18889 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18890 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18891 * config/i386/constraints.md (Yi, Ym): Depend on
18892 TARGET_INTER_UNIT_MOVES_TO_VEC.
18893 (Yj, Yn): New constraints.
18894 * config/i386/i386.md (*movdi_internal): Change constraints of
18895 operand 1 from Yi to Yj and from Ym to Yn.
18896 (*movsi_internal): Ditto.
18897 (*movdf_internal): Ditto.
18898 (*movsf_internal): Ditto.
18899 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18900 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18901 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18902 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18903 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18904 * config/i386/sse.md (movdi_to_sse): Ditto.
18905 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18906 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18907 TARGET_INTER_UNIT_MOVES.
18908 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18909 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18910 instead of TARGET_INTER_UNIT_MOVES.
18911 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18912 operand 1 from Yi to Yj and from Ym to Yn.
18913
4c871069
JG		 189142013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18915
18916 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18917 (float_truncate_hi_): Likewise.
18918 (float_extend_lo_): Likewise.
18919 (float_truncate_lo_): Likewise.
18920 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18921 (aarch64_float_extend_lo_v2df): Likewise.
18922 (vec_unpacks_hi_v4sf): Likewise.
18923 (aarch64_float_truncate_lo_v2sf): Likewise.
18924 (aarch64_float_truncate_hi_v4sf): Likewise.
18925 (vec_pack_trunc_v2df): Likewise.
18926 (vec_pack_trunc_df): Likewise.
18927
1709ff9b
JG		 189282013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18929
18930 * config/aarch64/aarch64-builtins.c
18931 (aarch64_fold_builtin): Fold float conversions.
18932 * config/aarch64/aarch64-simd-builtins.def
18933 (floatv2si, floatv4si, floatv2di): New.
18934 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18935 * config/aarch64/aarch64-simd.md
18936 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18937 * config/aarch64/iterators.md (FLOATUORS): New.
18938 (optab): Add float, floatuns.
18939 (su_optab): Likewise.
18940
ce966824
JG		 189412013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18942
18943 * config/aarch64/aarch64-builtins.c
18944 (aarch64_builtin_vectorized_function): Use new names for
18945 fcvt builtins.
18946 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18947 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18948 (fcvtzu): Split as...
18949 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18950 (fcvtas): Split as...
18951 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18952 (fcvtau): Split as...
18953 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18954 (fcvtps): Split as...
18955 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18956 (fcvtpu): Split as...
18957 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18958 (fcvtms): Split as...
18959 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18960 (fcvtmu): Split as...
18961 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18962 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18963 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18964 (lfrintnusf, lfrintnudf): Likewise.
18965 * config/aarch64/aarch64-simd.md
18966 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18967 define_insn.
18968 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18969 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18970 (fcvt_pattern): Likewise.
18971
b9de24fe
JG		 189722013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18973
18974 * config/aarch64/aarch64-simd.md
18975 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18976 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18977
77a205be
JG		 189782013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18979
18980 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18981 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18982 (vrnd<a,m,n,p>_f32): Implement using builtins.
18983 (vrnd<i,x><q>_f<32, 64>): New.
18984
0659ce6f
JG		 189852013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18986
18987 * config/aarch64/aarch64-builtins.c
18988 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18989 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18990 (frintz): Rename to...
18991 (btrunc): ...this.
18992 (frintp): Rename to...
18993 (ceil): ...this.
18994 (frintm): Rename to...
18995 (floor): ...this.
18996 (frinti): Rename to...
18997 (nearbyint): ...this.
18998 (frintx): Rename to...
18999 (rint): ...this.
19000 (frinta): Rename to...
19001 (round): ...this.
19002 * config/aarch64/aarch64-simd.md
19003 (aarch64_frint<frint_suffix><mode>): Delete.
19004 (<frint_pattern><mode>2): Convert to insn.
19005 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19006 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19007 (frint_pattern): Likewise.
19008 (frint_suffix): Likewise.
19009
ea78906a
RB		 190102013-04-29 Richard Biener <rguenther@suse.de>
19011
19012 PR tree-optimization/57081
19013 * loop-init.c: Include tree-flow.h.
19014 (loop_optimizer_finalize): Free number of iteration estimates.
19015 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19016
baee1763
JJ		 190172013-04-29 Jakub Jelinek <jakub@redhat.com>
19018
94dc5332
JJ		 19019 PR tree-optimization/57083
19020 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19021 non-singleton shift count range, zero extend low_bound for uns case.
19022
baee1763
JJ		 19023 * config/i386/predicates.md (general_vector_operand): New predicate.
19024 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19025 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19026 if they aren't nonimmediate operands. If their original values
19027 satisfy const_vector_equal_evenodd_p, don't shift them.
19028 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19029 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19030 if not nonimmediate_operand.
19031 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19032 instead of register_operand.
19033 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19034
a9073727 190352013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB		 19036
19037 * stor-layout.c (finalize_size_functions): Allocate a structure and
19038 reset cfun before dumping the functions.
19039
ba8011e6
JJ		 190402013-04-27 Jakub Jelinek <jakub@redhat.com>
19041
d6fde69e
JJ		 19042 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19043
ba8011e6
JJ		 19044 PR target/56866
19045 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19046 use xop_pmacsdqh if uns_p.
19047 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19048 the immediate rotate count.
19049
2c62cbaa
VM		 190502013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19051
19052 * rtl.h (struct rtx_def): Add comment for field jump.
19053 (LRA_SUBREG_P): New macro.
19054 * recog.c (register_operand): Check LRA_SUBREG_P.
19055 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19056 stack frame.
19057 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19058 (lra_final_code_change): Skip subreg change for operators.
19059 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19060 if there are no operand changes.
19061 * lra-constraints.c (curr_insn_set): New.
19062 (match_reload): Set LRA_SUBREG_P.
19063 (emit_spill_move): Ditto.
19064 (check_and_process_move): Use curr_insn_set. Process only single
19065 set insns. Don't initialize sec_mem_p and change_p.
19066 (simplify_operand_subreg): Use LRA_SUBREG_P.
19067 (reg_in_class_p): New function.
19068 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19069 of #ifdef. Add code to remove cycling.
19070 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19071 non-null disp. Reload inner instead of disp when base and index
19072 are null. Try to put lo_sum into register.
19073 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19074 (check_and_process_move): Move code for move cost check to
19075 simple_move_p. Remove equiv_substitution.
19076 (simple_move_p): New function.
19077 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19078 curr_insn_set. Call check_and_process_move only for single set
19079 insns. Use the new function. Move call of check_and_process_move
19080 after operand equiv substitution and address process.
19081
e7d764f3
JJ		 190822013-04-26 Jakub Jelinek <jakub@redhat.com>
19083
19084 PR go/57045
19085 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19086 with nonlocal goto receivers or returns twice calls, ignore
19087 unininitialized values from abnormal edges to nl goto receiver
19088 or returns twice call.
19089
41e10689
JJ		 190902013-04-26 Jakub Jelinek <jakub@redhat.com>
19091
19092 PR tree-optimization/57051
19093 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19094 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19095 bitsize.
19096
d7ed20db
RB		 190972013-04-26 Richard Biener <rguenther@suse.de>
19098
19099 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19100 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19101 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19102 (fixup_loop_arrays_after_move): New function.
19103 (move_sese_region_to_fn): Properly outline the loop tree parts
19104 of the SESE region.
19105
df93505e
UB		 191062013-04-26 Uros Bizjak <ubizjak@gmail.com>
19107
19108 * config/i386/i386.md (type, unit): Fix long lines.
19109
dd366ec3
RB		 191102013-04-26 Richard Biener <rguenther@suse.de>
19111
19112 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19113 (lto-streamer-out.o): Likewise.
19114 * cfgloop.c (init_loops_structure): Export, add struct function
19115 argument and adjust.
19116 (flow_loops_find): Adjust.
19117 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19118 (init_loops_structure): Declare.
19119 * lto-streamer-in.c: Include cfgloop.h.
19120 (input_cfg): Input the loop tree.
19121 * lto-streamer-out.c: Include cfgloop.h.
19122 (output_cfg): Output the loop tree.
19123 (output_struct_function_base): Do not drop PROP_loops.
19124
a9e0d843
RB		 191252013-03-26 Richard Biener <rguenther@suse.de>
19126
19127 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19128 (pass_build_cfg): Provide PROP_loops.
19129 (move_sese_region_to_fn): Remove loops that are outlined into fn
19130 for now.
19131 * tree-inline.c: Include cfgloop.h.
19132 (initialize_cfun): Do not drop PROP_loops.
19133 (copy_loops): New function.
19134 (copy_cfg_body): Copy loop structure.
19135 (tree_function_versioning): Initialize destination loop tree.
19136 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19137 (pass_parallelize_loops): Do IL verification.
19138 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19139 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19140 the CFG make sure we fixup loops as well.
19141 * tree-ssa-tail-merge.c: Include cfgloop.h.
19142 (replace_block_by): When merging loop latches mark loops for fixup.
19143 * lto-streamer-out.c (output_struct_function_base): Drop
19144 PROP_loops for now.
19145 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19146 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19147 * ipa-split.c: Include cfgloop.h.
19148 (split_function): Add the new return block to the loop tree root.
19149 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19150 whether we have removed the forwarder block.
19151 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19152 * cfgloop.h (place_new_loop): Declare.
19153 * cfgloopmanip.c (place_new_loop): Export.
19154 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19155 (tree-switch-conversion.o): Likewise.
19156 (tree-complex.o): Likewise.
19157 (tree-inline.o): Likewise.
19158 (tree-ssa-tailmerge.o): Likewise.
19159 (ipa-split.o): Likewise.
19160 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19161 (tree-ssa-copy.o): Likewise.
19162 * tree-switch-conversion.c: Include cfgloop.h
19163 (process_switch): If we emit a bit-test cascade, schedule loops
19164 for fixup.
19165 * tree-complex.c: Include cfgloop.h.
19166 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19167 * asan.c: Include cfgloop.h.
19168 (create_cond_insert_point): Properly add new basic-blocks to
19169 loops, schedule loop fixup.
19170 * cfgloop.c (verify_loop_structure): Check that looks are not
19171 marked for fixup.
19172 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19173 to loops.
19174 (expand_omp_for_generic): Likewise.
19175 (expand_omp_sections): Likewise.
19176 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19177 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19178 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19179 is initialized, not when loops are present.
19180 * tree-parloops.c (parallelize_loops): Remove checking here.
19181 * passes.c (init_optimization_passes): Schedule a copy-propagation
19182 pass before complete unrolling of inner loops.
19183
e78e8a0b
JJ		 191842013-04-26 Jakub Jelinek <jakub@redhat.com>
19185
a2e836b2
JJ		 19186 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19187 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19188 (colorize_init): Add argument to _WIN32 version.
19189 * toplev.c: Include diagnostic-color.h.
19190 (process_options): Default to -fdiagnostics-color=auto if
19191 GCC_COLORS env var is in the environment.
19192 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB		 19193 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19194 env var is in the environment, the default is auto rather than never.
a2e836b2 19195
e78e8a0b
JJ		 19196 * diagnostic.h (file_name_as_prefix): Add context argument.
19197 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19198 the string as locus.
19199 * langhooks.c (lhd_print_error_function): Adjust caller.
19200
013e5ef9
LC		 192012013-04-25 Lawrence Crowl <crowl@google.com>
19202
19203 * var-tracking.c (shared_hash_def::htab):
19204 Change type to hash_table. Update dependent calls and types.
19205
4a8fb1a1
LC		 192062013-04-25 Lawrence Crowl <crowl@google.com>
19207
19208 * Makefile.in: Update as needed below.
19209
19210 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19211 Move declaration to after the type's method definitons.
19212
19213 * attribs.c (htab_t scoped_attributes::attribute_hash):
19214 Change type to hash_table. Update dependent calls and types.
19215
19216 * bitmap.c (htab_t bitmap_desc_hash):
19217 Change type to hash_table. Update dependent calls and types.
19218
19219 * cselib.c (htab_t cselib_hash_table):
19220 Change type to hash_table. Update dependent calls and types.
19221
19222 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19223 (hash_string_slot_node): Move implementation into lto-streamer.h
19224 struct string_slot_hasher.
19225 (eq_string_slot_node): Likewise.
19226
19227 * data-streamer-out.c: Update output_block::string_hash_table
19228 dependent calls and types.
19229
19230 * dwarf2cfi.c (htab_t trace_index):
19231 Change type to hash_table. Update dependent calls and types.
19232
19233 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19234 Change type to hash_table. Update dependent calls and types.
19235 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19236 (htab_t optimize_external_refs::map): Likewise.
19237 (htab_t output_comp_unit::extern_map): Likewise.
19238 (htab_t output_comdat_type_unit::extern_map): Likewise.
19239 (htab_t output_macinfo::macinfo_htab): Likewise.
19240 (htab_t optimize_location_lists::htab): Likewise.
19241 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19242
19243 * except.c (htab_t ehspec_hash_type):
19244 Change type to hash_table. Update dependent calls and types.
19245 (assign_filter_values::ttypes): Likewise.
19246 (assign_filter_values::ehspec): Likewise.
19247 (sjlj_assign_call_site_values::ar_hash): Likewise.
19248 (convert_to_eh_region_ranges::ar_hash): Likewise.
19249
19250 * gcse.c (htab_t pre_ldst_table):
19251 Change type to hash_table. Update dependent calls and types.
19252
19253 * ggc-common.c (htab_t saving_htab):
19254 Change type to hash_table. Update dependent calls and types.
19255 (htab_t loc_hash): Likewise.
19256 (htab_t ptr_hash): Likewise.
19257 (call_count): Rename ggc_call_count.
19258 (call_alloc): Rename ggc_call_alloc.
19259 (loc_descriptor): Rename make_loc_descriptor.
19260 (add_statistics): Rename ggc_add_statistics.
19261
19262 * ggc-common.c (saving_htab):
19263 Change type to hash_table. Update dependent calls and types.
19264
19265 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19266 (push_gimplify_context): Likewise.
19267 (pop_gimplify_context): Likewise.
19268 (struct gimple_temp_hash_elt): Added.
19269 (struct gimplify_hasher): Likewise.
19270 (struct gimplify_ctx.temp_htab):
19271 Change type to hash_table. Update dependent calls and types.
19272
19273 * gimple-fold.c: Include gimplify-ctx.h.
19274
19275 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19276 Change type to hash_table. Update dependent calls and types.
19277 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19278 avoid potential global name collision.
19279
19280 * gimplify.c: Include gimplify-ctx.h.
19281 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19282 (htab_t gimplify_ctx::temp_htab):
19283 Update dependent calls and types for new type hash_table.
19284 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19285 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19286
19287 * gimplify-ctx.h: New.
19288 (struct gimple_temp_hash_elt): Move from gimplify.c.
19289 (class gimplify_hasher): New.
19290 (struct gimplify_ctx): Move from gimple.h.
19291 (htab_t gimplify_ctx::temp_htab):
19292 Change type to hash_table. Update dependent calls and types.
19293
19294 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19295 (htab_t ivs_params::newivs_index):
19296 Change type to hash_table. Update dependent calls and types.
19297 (htab_t ivs_params::params_index): Likewise.
19298 (htab_t print_generated_program::params_index): Likewise.
19299 (htab_t gloog::newivs_index): Likewise.
19300 (htab_t gloog::params_index): Likewise.
19301
19302 * graphite.c: Include graphite-htab.h.
19303 4htab_t graphite_transform_loops::bb_pbb_mapping):
19304 Change type to hash_table. Update dependent calls and types.
19305
19306 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19307 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19308 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19309
19310 * graphite-dependences.c: Include graphite-htab.h.
19311 (loop_is_parallel_p): Change hash table type of parameter.
19312
19313 * graphite-htab.h: New.
19314 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19315 (extern find_pbb_via_hash): Move from graphite-poly.h.
19316 (extern loop_is_parallel_p): Move from graphite-poly.h.
19317 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19318
19319 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19320 (extern loop_is_parallel_p): Move to graphite-htab.h.
19321 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19322
19323 * haifa-sched.c (htab_t delay_htab):
19324 Change type to hash_table. Update dependent calls and types.
19325 (htab_t delay_htab_i2): Likewise.
19326
19327 * ira-color.c (htab_t allocno_hard_regs_htab):
19328 Change type to hash_table. Update dependent calls and types.
19329
19330 * ira-costs.c (htab_t cost_classes_htab):
19331 Change type to hash_table. Update dependent calls and types.
19332
19333 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19334 Change type to hash_table. Update dependent calls and types.
19335
19336 * loop-iv.c (htab_t bivs):
19337 Change type to hash_table. Update dependent calls and types.
19338
19339 * loop-unroll.c (htab_t opt_info::insns_to_split):
19340 Change type to hash_table. Update dependent calls and types.
19341 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19342
19343 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19344 (struct string_slot_hasher): New.
19345 (htab_t output_block::string_hash_table):
19346 Change type to hash_table. Update dependent calls and types.
19347
19348 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19349 (htab_t file_name_hash_table):
19350 Change type to hash_table. Update dependent calls and types.
19351
19352 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19353 calls and types.
19354
19355 * lto-streamer.c (htab_t tree_htab):
19356 Change type to hash_table. Update dependent calls and types.
19357
19358 * omp-low.c: Include gimplify-ctx.h.
19359
19360 * passes.c (htab_t name_to_pass_map):
19361 Change type to hash_table. Update dependent calls and types.
19362 (pass_traverse): Rename to passes_pass_traverse.
19363
19364 * plugin.c (htab_t event_tab):
19365 Change type to hash_table. Update dependent calls and types.
19366
19367 * postreload-gcse.c (htab_t expr_table):
19368 Change type to hash_table. Update dependent calls and types.
19369 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19370
19371 * sese.c (debug_rename_map_1): Make extern.
19372 (htab_t copy_bb_and_scalar_dependences::rename_map):
19373 Change type to hash_table. Update dependent calls and types.
19374
19375 * sese.h (extern debug_rename_map): Move to .c file.
19376
19377 * store-motion.c (htab_t store_motion_mems_table):
19378 Change type to hash_table. Update dependent calls and types.
19379
19380 * trans-mem.c (htab_t tm_new_mem_hash):
19381 Change type to hash_table. Update dependent calls and types.
19382
19383 * tree-browser.c (htab_t TB_up_ht):
19384 Change type to hash_table. Update dependent calls and types.
19385
19386 * tree-cfg.c (htab_t discriminator_per_locus):
19387 Change type to hash_table. Update dependent calls and types.
19388
19389 * tree-complex.c: Include tree-hasher.h
19390 (htab_t complex_variable_components):
19391 Change type to hash_table. Update dependent calls and types.
19392
19393 * tree-eh.c (htab_t finally_tree):
19394 Change type to hash_table. Update dependent calls and types.
19395
19396 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19397 struct int_tree_hasher.
19398 (extern int_tree_map_eq): Likewise.
19399 (uid_decl_map_hash): Removed.
19400 (extern decl_tree_map_eq): Likewise.
19401
19402 * tree-hasher.h: New.
19403 (struct int_tree_hasher): New.
19404 (typedef int_tree_htab_type): New.
19405
19406 * tree-inline.c: Include gimplify-ctx.h.
19407
19408 * tree-mudflap.c: Include gimplify-ctx.h.
19409
19410 * tree-parloops.c: Include tree-hasher.h.
19411 (htab_t eliminate_local_variables_stmt::decl_address):
19412 Change type to hash_table. Update dependent calls and types.
19413 (htab_t separate_decls_in_region::decl_copies): Likewise.
19414
19415 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19416 Change type to hash_table. Update dependent calls and types.
19417
19418 * tree-sra.c (candidates):
19419 Change type to hash_table. Update dependent calls and types.
19420
19421 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19422 in tree-flow.h.
19423 (int_tree_map_hash): Likewise.
19424
19425 * tree-ssa-dom.c (htab_t avail_exprs):
19426 Change type to hash_table. Update dependent calls and types.
19427
19428 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19429 Change type to hash_table. Update dependent calls and types.
19430
19431 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19432 Change type to hash_table. Update dependent calls and types.
19433
19434 * tree-ssa-phiopt.c (seen_ssa_names):
19435 Change type to hash_table. Update dependent calls and types.
19436
19437 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19438 Change type to hash_table. Update dependent calls and types.
19439
19440 * tree-ssa-uncprop.c (equiv):
19441 Change type to hash_table. Update dependent calls and types.
19442
c5a44004
JJ		 194432013-04-25 Jakub Jelinek <jakub@redhat.com>
19444
19445 PR rtl-optimization/57003
19446 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19447 call note_stores with kill_clobbered_value callback again after
19448 killing regs_invalidated_by_call.
19449
09962a4a
JG		 194502013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19451
19452 * config/aarch64/aarch64-simd.md
19453 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19454 (aarch64_simd_bsl<mode>): Likewise.
19455 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19456
ea28bb0b
MP		 194572013-04-25 Marek Polacek <polacek@redhat.com>
19458
19459 PR tree-optimization/57066
3c21604f 19460 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 19461
96659611
JG		 194622013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19463
19464 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19465
9697e620
JG		 194662013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19467
19468 * config/aarch64/aarch64-builtins.c
19469 (aarch64_fold_builtin): New.
19470 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19471 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19472 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19473 * config/aarch64/arm_neon.h
19474 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19475
0ac198d3
JG		 194762013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19477 Tejas Belagod <tejas.belagod@arm.com>
19478
19479 * config/aarch64/aarch64-builtins.c
19480 (aarch64_gimple_fold_builtin): New.
19481 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19482 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19483 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19484 (addvv4sf): Update.
19485 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19486
58cff58c
N		 194872013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19488
df93505e 19489 * config/aarch64/aarch64.md
58cff58c
N		 19490 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19491
7e0228bf
N		 194922013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19493
19494 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19495 (*ngcsi_uxtw): New pattern.
19496
5819f96f 194972013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 19498 Julian Brown <julian@codesourcery.com>
5819f96f
KT		 19499
19500 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19501 (TB_DREG): Add T_V4HF.
19502 (v4hf_UP): New macro.
19503 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 19504 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19505 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19506 built-ins.
df93505e 19507 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19508 (arm_vector_mode_supported_p): Handle V4HF.
19509 (arm_mangle_map): Handle V4HFmode.
19510 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19511 * config/arm/arm_neon_builtins.def: Add entries for
19512 vcvtv4hfv4sf, vcvtv4sfv4hf.
19513 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19514 (neon_vcvtv4hfv4sf): Likewise.
19515 * config/arm/neon-gen.ml: Handle half-precision floating point
19516 features.
19517 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19518 * config/arm/arm_neon.h: Regenerate.
19519 * config/arm/neon.ml (type elts): Add F16.
19520 (type vectype): Add T_float16x4, T_floatHF.
19521 (type vecmode): Add V4HF.
19522 (type features): Add Requires_FP_bit feature.
19523 (elt_width): Handle F16.
19524 (elt_class): Likewise.
19525 (elt_of_class_width): Likewise.
19526 (mode_of_elt): Refactor.
19527 (type_for_elt): Handle F16, fix error messages.
19528 (vectype_size): Handle T_float16x4.
19529 (vcvt_sh): New function.
19530 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19531 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19532 (string_of_mode): Handle V4HF.
19533 * doc/arm-neon-intrinsics.texi: Regenerate.
19534
1ef395e4
JG		 195352013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19536
19537 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19538 format specifier in 'X' case.
19539
41c34e94
AM		 195402013-04-25 Alan Modra <amodra@gmail.com>
19541
19542 PR target/57052
19543 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19544 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19545 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19546 Repeat for many other rotate/shift and mask patterns using subregs.
19547 Name lshiftrt insns.
19548 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19549 on WORDS_BIG_ENDIAN.
19550
b9a7eb5d
AM		 195512013-04-25 Alan Modra <amodra@gmail.com>
19552
19553 * config.gcc: Support little-endian powerpc-linux targets.
19554 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19555 (LINK_OS_LINUX_SPEC): Define.
19556 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19557 Preserve MASK_LITTLE_ENDIAN.
19558 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19559 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19560 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19561 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19562 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19563 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19564 Correct fp word order for little-endian. Don't shift toc entries
19565 smaller than a word for little-endian.
19566 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19567 (bswapdi2 splits): Correct low-part subreg for little-endian.
19568 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19569 low/high where such is correct only for be.
19570 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19571 little-endian for -mcall-aixdesc.
19572
87f73374
AM		 195732013-04-25 Alan Modra <amodra@gmail.com>
19574
19575 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19576 replace_equiv_address_nv.
19577
cabf91cd
AM		 195782013-04-25 Alan Modra <amodra@gmail.com>
19579
19580 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19581
0ae24cc8
VM		 195822013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19583
19584 Revert:
19585 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19586 * rtl.h (struct rtx_def): ...
cabf91cd 19587
77bce07c
VM		 195882013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19589
19590 PR rtl-optimizations/57046
19591 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19592 for multi-reg splits.
19593
0db63e7f
L		 195942013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19595
19596 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19597
3362b6b6
SA		 195982013-04-24 Sterling Augustine <saugustine@google.com>
19599
19600 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19601 (comp_dir_string, debug_str_dwo_section): New.
19602 (DEBUG_STR_DWO_SECTION): Rename to ...
19603 (DEBUG_DWO_STR_SECTION): ... this.
19604 (DEBUG_NORM_STR_SECTION): Delete.
19605 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19606 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19607 (find_AT_string): Move most logic to ...
19608 (find_AT_string_in_table): ... here. New.
19609 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19610 add_skeleton_AT_string. Delete logic.
19611 (output_skeleton_debug_sections): Remove call to
19612 add_top_level_skeleton_die_attrs.
19613 (add_comp_dir_attribute): Move logic to comp_dir_string.
19614 (dwarf2out_init): Initialize debug_str_dwo_section.
19615 (output_indirect_string): Call find_string_form.
19616 (output_indirect_strings): Rewrite.
19617 (prune_unused_types): Empty skeleton_debug_str_hash.
19618 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19619 (dwarf2out_finish): Call output_indirect_strings.
19620
e93e18e9
PC		 196212013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19622
19623 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19624
f6ce35ac
VM		 196252013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19626
cabf91cd 19627 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM		 19628 (LRA_SUBREG_P): New macro.
19629 * recog.c (register_operand): Check LRA_SUBREG_P.
19630 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19631 stack frame.
19632 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19633 (lra_final_code_change): Skip subreg change for operators.
19634 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19635 if there are no operand changes.
19636 * lra-constraints.c (curr_insn_set): New.
19637 (match_reload): Set LRA_SUBREG_P.
19638 (emit_spill_move): Ditto.
19639 (check_and_process_move): Use curr_insn_set. Process only single
19640 set insns. Don't initialize sec_mem_p and change_p.
19641 (simplify_operand_subreg): Use LRA_SUBREG_P.
19642 (reg_in_class_p): New function.
19643 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19644 of #ifdef. Add code to remove cycling.
19645 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19646 non-null disp. Reload inner instead of disp when base and index
19647 are null. Try to put lo_sum into register.
19648 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 19649 (check_and_process_move): Move code for move cost check to
f6ce35ac 19650 simple_move_p. Remove equiv_substitution.
cabf91cd 19651 (simple_move_p): New function.
f6ce35ac
VM		 19652 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19653 curr_insn_set. Call check_and_process_move only for single set
19654 insns. Use the new function. Move call of check_and_process_move
19655 after operand equiv substitution and address process.
19656
38047d90
JG		 196572013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19658
19659 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19660 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19661 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19662
13f39b2e
PC		 196632013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19664
19665 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19666
97c116dc
MP		 196672013-04-24 Marek Polacek <polacek@redhat.com>
19668
19669 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19670 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19671 (select_loops_exit_conditions): Likewise.
19672 (number_of_iterations_for_all_loops): Likewise.
19673 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19674 (scev_analysis): Likewise.
19675
83082391 196762013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 19677 Chao-ying Fu <fu@mips.com>
83082391 19678
cabf91cd
AM		 19679 * config/mips/micromips.md (jraddiusp): New pattern.
19680 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19681 instruction if possible.
83082391 19682
19e34aa2
AM		 196832013-04-24 Alan Modra <amodra@gmail.com>
19684
19685 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19686
fdb6603c
JB		 196872013-04-24 Julian Brown <julian@codesourcery.com>
19688 Chung-Lin Tang <cltang@codesourcery.com>
19689
19690 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB		 19691 dependency behavior in enumeration type DIE generation. Add TODO note
19692 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 19693
bf190e8d
LC		 196942013-04-23 Lawrence Crowl <crowl@google.com>
19695
19696 * Makefile.in: Update as needed below.
19697
19698 * hash-table.h (class hash_table):
19699 Correct many methods with parameter types compare_type to the correct
19700 value_type. (Correct code was unlikely to notice the change.)
19701 (hash_table::elements_with_deleted) New.
19702 (class hashtable::iterator): New.
19703 (hashtable::begin()): New.
19704 (hashtable::end()): New.
19705 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19706
19707 * statistics.c (statistics_hashes):
19708 Change type to hash_table. Update dependent calls and types.
19709
19710 * tree-into-ssa.c (var_infos):
19711 Change type to hash_table. Update dependent calls and types.
19712
19713 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19714 Change type to hash_table. Update dependent calls and types.
19715
19716 * tree-ssa-loop-im.c (struct mem_ref.refs):
19717 Change type to hash_table. Update dependent calls and types.
19718
19719 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19720 Change type to hash_table. Update dependent calls and types.
19721
19722 * tree-ssa-sccvn.c (vn_tables_s::nary):
19723 Change type to hash_table. Update dependent calls and types.
19724 (vn_tables_s::phis): Likewise.
19725 (vn_tables_s::references): Likewise.
19726
19727 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19728 (vn_reference_eq): Update parameter and return types.
19729
19730 * tree-ssa-structalias.c (pointer_equiv_class_table):
19731 Change type to hash_table. Update dependent calls and types.
19732 (location_equiv_class_table): Likewise.
19733
19734 * tree-vect-data-refs.c: Consequential changes for making
19735 peeling a hash_table.
19736
19737 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19738 (destroy_loop_vec_info): Dependent hash_table update.
19739
19740 * tree-vectorizer.h (peeling_htab):
19741 Change type to hash_table. Update dependent calls and types.
19742
d70a81dd
SC		 197432013-04-23 Shiva Chen <shiva0217@gmail.com>
19744
cabf91cd
AM		 19745 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19746 to check the register content is equal or not.
19747 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19748 to assign register content record.
19749 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 19750 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM		 19751 * lra-int.h (struct lra_reg): Add offset member.
19752 (lra_reg_val_equal_p): New static inline function.
19753 (lra_update_reg_val_offset): New static inline function.
19754 (lra_assign_reg_val): New static inline function.
19755 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19756 to assign register content record.
19757 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 19758
b894a1f3
CM		 197592013-04-23 Catherine Moore <clm@codesourcery.com>
19760
19761 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19762 operands. Record compression.
19763
ecd14de9
XDL		 197642013-04-23 Xinliang David Li <davidxl@google.com>
19765
19766 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19767
92e776e9
RB		 197682013-04-23 Richard Biener <rguenther@suse.de>
19769
19770 PR middle-end/57036
19771 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19772 parameter, only add abnormal goto edges from the copied body
19773 if the call could perform abnormal gotos.
19774 (copy_cfg_body): Adjust.
19775
a15ee567
SN		 197762013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19777
19778 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19779
08c52234
AS		 197802013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19781
19782 * coretypes.h (gimple_stmt_iterator): Add struct to make
19783 compatible with C.
19784
999c1171
RB		 197852013-04-23 Richard Biener <rguenther@suse.de>
19786
19787 PR tree-optimization/57026
19788 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19789 from SSA names occuring in abnormal PHI nodes.
19790
53669259
AK		 197912013-04-22 Andi Kleen <ak@linux.intel.com>
19792
19793 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19794
1a0ad150
AK		 197952013-04-22 Andi Kleen <ak@linux.intel.com>
19796
19797 * lto/lto.c (print_lto_report_1): Declare early.
19798 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19799
057f8f20
AK		 198002013-04-22 Andi Kleen <ak@linux.intel.com>
19801
19802 * common.opt (-flto-report-wpa): Add.
19803 * doc/invoke.texi (-flto-report-wpa): Add.
19804 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19805 (lto_main): dito.
19806
473b1e05
XDL		 198072013-04-22 Xinliang David Li <davidxl@google.com>
19808
19809 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19810 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19811 * Makefile.in: New dependency
19812
cabf91cd 19813 David Daney <ddaney.cavm@gmail.com>
b1485a33 19814
cabf91cd
AM		 19815 * configure.ac (gcc_cv_as_micromips_support): Use the
19816 --fatal-warnings option.
19817 * configure: Regenerate.
b1485a33 19818
829d0168
MP		 198192013-04-22 Marek Polacek <polacek@redhat.com>
19820
19821 PR sanitizer/56990
19822 * tsan.c (instrument_expr): Don't instrument expression
19823 in case its size is zero.
19824
6d9b7208
UB		 198252013-04-22 Uros Bizjak <ubizjak@gmail.com>
19826
19827 PR target/57032
19828 Revert:
19829 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19830
19831 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19832
ea679d55
JG		 198332013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19834
19835 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19836 (gimple_stmt_iterator): New typedef.
19837 * gimple.h (gimple_stmt_iterator): Rename to...
19838 (gimple_stmt_iterator_d): ... This.
19839 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19840 trees be valid for GIMPLE and GENERIC.
19841 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19842 * gimple-fold.c (gimple_fold_call): Call target hook
19843 gimple_fold_builtin.
19844 * hooks.c (hook_bool_gsiptr_false): New.
19845 * hooks.h (hook_bool_gsiptr_false): New.
19846 * target.def (fold_stmt): New.
19847 * doc/tm.texi: Regenerate.
19848
88a581da
VM		 198492013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19850
19851 PR target/57018
19852 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19853 a set sp if no stack realignment.
19854
92be22dc
NC		 198552013-04-22 Nick Clifton <nickc@redhat.com>
19856
19857 * config.gcc (tilegx-linux): Extend extra_objs rather than
19858 overwriting it.
19859 (tilepro-linux): Likewise.
19860
0ddec79f
JG		 198612013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19862
19863 * config/aarch64/aarch64-builtins.c
19864 (CF): Remove.
19865 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19866 (VAR<1-12>): Add MAP parameter.
19867 (BUILTIN_*): Likewise.
19868 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19869 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19870 (aarch64_ushl_n<mode>): Likewise.
19871 (aarch64_sshr_n<mode>): Likewise.
19872 (aarch64_ushr_n<mode>): Likewise.
19873 (aarch64_<maxmin><mode>): Likewise.
19874 (aarch64_sqrt<mode>): Likewise.
19875 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19876 (vshr<q>_n_*): Likewise.
19877
0050faf8
JG		 198782013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19879
19880 * config/aarch64/aarch64-builtins.c
19881 (aarch64_simd_builtin_type_mode): Handle SF types.
19882 (sf_UP): Define.
19883 (BUILTIN_GPF): Define.
19884 (aarch64_init_simd_builtins): Handle SF types.
19885 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19886 (frecps): Likewise.
19887 (frecpx): Likewise.
19888 * config/aarch64/aarch64-simd.md
19889 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19890 (aarch64_frecpe<mode>): New.
19891 (aarch64_frecps<mode>): Likewise.
19892 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19893 (v8type): Add frecp<esx>.
19894 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19895 (aarch64_frecps<mode>): Likewise.
19896 * config/aarch64/iterators.md (FRECP): New.
19897 (frecp_suffix): Likewise.
19898 * config/aarch64/arm_neon.h
19899 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19900
0fad3dbc 199012013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB		 19902
19903 PR target/56995
19904 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19905 (REG_CLASS_NAMES): Idem.
19906 (REG_CLASS_CONTENTS): Idem.
19907 (REGCLASS_HAS_FP_REG): Idem.
19908 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19909 (sh_conditional_register_usage): Idem.
19910
3e8a33f9
JL		 199112013-04-21 Jeff Law <law@redhat.com>
19912
19913 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19914 (ssa_forward_propagate_and_combine): Use it.
19915
f38e1b0a
VM		 199162013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19917
19918 * lra.c: Update the flow chart diagram.
19919
682303da
VM		 199202013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19921
19922 PR rtl-optimization/56847
19923 * lra-constraints.c (process_alt_operands): Discourage alternative
19924 with non-matche doffsettable memory constraint fro memory with
19925 known offset.
19926
f6b64c35
RB		 199272013-04-19 Richard Biener <rguenther@suse.de>
19928
19929 PR tree-optimization/56982
19930 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19931 function.
19932 * gimplify.c (gimplify_call_expr): Notice special calls.
19933 (gimplify_modify_expr): Likewise.
19934 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19935 abnormal control flow receivers.
19936 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19937 in the same way as cfun->has_nonlocal_labels.
19938 (gimple_purge_dead_abnormal_call_edges): Likewise.
19939 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19940 receivers start a basic-block.
19941
01d8bf07
RB		 199422013-04-19 Richard Biener <rguenther@suse.de>
19943
19944 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19945 member ...
19946 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19947 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19948 (SLP_TREE_LOAD_PERMUTATION): Add.
19949 (vect_transform_slp_perm_load): Adjust prototype.
19950 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19951 (vect_free_slp_instance): Likewise.
19952 (vect_create_new_slp_node): Likewise.
19953 (vect_supported_slp_permutation_p): Remove.
19954 (vect_slp_rearrange_stmts): Adjust.
19955 (vect_supported_load_permutation_p): Likewise. Inline
19956 vect_supported_slp_permutation_p here.
19957 (vect_analyze_slp_instance): Compute load permutations per
19958 slp node instead of per instance.
19959 (vect_get_slp_defs): Adjust.
19960 (vect_transform_slp_perm_load): Likewise.
19961 (vect_schedule_slp_instance): Remove redundant code.
19962 (vect_schedule_slp): Remove hack for PR56270, add it ...
19963 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19964 CSE loads for SLP. Adjust.
19965
ede22fc3
GY		 199662013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19967
19968 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19969 spelling in two comments.
19970
67bc84fb
GY		 199712013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19972
19973 PR target/56797
19974 * config/arm/arm.c (load_multiple_sequence): Require SP
19975 as base register for loads if SP is in the register list.
19976
e248d83f
MJ		 199772013-04-19 Martin Jambor <mjambor@suse.cz>
19978
19979 PR tree-optimization/56718
19980 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19981 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19982 and made public. Adjusted all callers.
19983 (ipa_intraprocedural_devirtualization): New function.
19984 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19985 (ipa_intraprocedural_devirtualization): Likewise.
19986 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19987
4891e8f8
RB		 199882013-04-19 Richard Biener <rguenther@suse.de>
19989
19990 PR tree-optimization/57000
19991 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19992
dad89f7c
TG		 199932013-04-19 Terry Guo <terry.guo@arm.com>
19994
19995 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19996 Replace with ...
19997 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19998 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19999 (cortex_m4_fmacs): Use new reservations.
20000 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20001
72ea0d47
VM		 200022013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20003
f1e6512c 20004 PR rtl-optimization/56999
72ea0d47
VM		 20005 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20006 related code.
20007 (lra_coalesce): Remove split_origin_bitmap and related code.
20008 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20009 ranges if necessary.
20010
780a5b71
UB		 200112013-04-18 Uros Bizjak <ubizjak@gmail.com>
20012
20013 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20014 New array.
20015 (ix86_expand_call): Remove clobbered_registers array and use
20016 x86_64_ms_sysv_extra_clobbered_registers instead.
20017 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20018 Declare here.
20019 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20020 predicate.
20021 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20022 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20023 (*call_value_rex64_ms_sysv): Ditto.
20024
6f5a366a
CC		 200252013-04-18 Cary Coutant <ccoutant@google.com>
20026
20027 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20028 parent instead of die_mark.
20029
475b8f37
DN		 200302013-04-18 Diego Novillo <dnovillo@google.com>
20031
20032 * gimple.c (create_gimple_tmp): New.
20033 (get_expr_type): New.
20034 (build_assign): New.
20035 (build_type_cast): New.
20036 * gimple.h (enum ssa_mode): Define.
20037 (gimple_seq_set_location): New.
20038 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 20039 to use build_assign and build_type_cast.
475b8f37 20040
08940f33
RB		 200412013-04-18 Richard Biener <rguenther@suse.de>
20042
20043 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20044 handle negative step. Remove redundant checks.
20045 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20046 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20047 for negative step and grouped loads fail to vectorize.
20048
0e0f87d4
SB		 200492013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20050
20051 * emit-rtl.c (reset_insn_used_flags): New function.
20052 (reset_all_used_flags): Use it.
20053 (verify_insn_sharing): New function.
20054 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20055
4c445590
JJ		 200562013-04-18 Jakub Jelinek <jakub@redhat.com>
20057
20058 PR tree-optimization/56984
20059 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20060 and (x >> M) >= N don't register any assertion if N << M is the
20061 minimum value.
20062
6873ecab
SB		 200632013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20064
20065 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20066 do not delete_insn insns that have not yet been emitted, only
20067 unlink them with remove_insn.
20068 * df-scan.c (df_insn_delete): Revert r197492.
20069
3ccb989e
SB		 200702013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20071
20072 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20073 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20074
200752013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY		 20076
20077 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20078 define_insn_and_split.
20079 (and_scc,ior_scc,negscc): Likewise.
20080 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20081
3ccb989e 200822013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY		 20083
20084 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20085 can benefit from using a sequence of LDRD instructions in epilogue
20086 instead of a single LDM instruction.
20087
6d10a203
MLI		 200882013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20089
20090 PR 45688
20091 * doc/extend.texi: Fix typo.
20092
6983e6b5
RB		 200932013-04-17 Richard Biener <rguenther@suse.de>
20094
20095 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20096 (vect_build_slp_tree): ... here.
20097 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20098 match. Remove special-casing of mismatched complex loads.
20099 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20100 re-try the match with swapped commutative operands.
20101 (vect_supported_load_permutation_p): Remove special-casing of
20102 mismatched complex loads.
20103 (vect_analyze_slp_instance): Adjust.
20104
ef23e6a2
RB		 201052013-04-17 Richard Biener <rguenther@suse.de>
20106
20107 PR rtl-optimization/56921
20108 * cfgloop.h (struct loop): Add simple_loop_desc member.
20109 (struct niter_desc): Mark with GTY(()).
20110 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20111 * loop-iv.c (get_simple_loop_desc): Likewise.
20112 (free_simple_loop_desc): Likewise.
20113
20114 Revert
20115 2013-04-16 Richard Biener <rguenther@suse.de>
20116
20117 PR rtl-optimization/56921
20118 * loop-init.c (pass_rtl_move_loop_invariants): Add
20119 TODO_do_not_ggc_collect to todo_flags_finish.
20120 (pass_rtl_unswitch): Same.
20121 (pass_rtl_unroll_and_peel_loops): Same.
20122 (pass_rtl_doloop): Same.
20123
fc6f94f5
EB		 201242013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20125
20126 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20127 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20128 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20129 references.
20130 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20131 * tree-streamer.c (record_common_node): Adjust reference in comment.
20132
10a88311
TG		 201332013-04-17 Terry Guo <terry.guo@arm.com>
20134
20135 * config/arm/cortex-m4.md: Add a new bypass.
20136
6d9b7208 201372013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N		 20138
20139 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20140 New pattern.
20141 (*subs_<optab><mode>_multp2): New pattern.
20142 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20143 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20144
6d9b7208 201452013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N		 20146
20147 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20148 (*subs_mul_imm_<mode>): New pattern.
20149
18a6701e
DE		 201502013-04-16 David Edelsohn <dje.gcc@gmail.com>
20151
20152 PR target/56948
20153 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20154 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20155 (vsx_movti_32bit): Change j->wa to O->wa.
20156
07c37b2f
RB		 201572013-04-16 Richard Biener <rguenther@suse.de>
20158
20159 PR rtl-optimization/56921
20160 * loop-init.c (pass_rtl_move_loop_invariants): Add
20161 TODO_do_not_ggc_collect to todo_flags_finish.
20162 (pass_rtl_unswitch): Same.
20163 (pass_rtl_unroll_and_peel_loops): Same.
20164 (pass_rtl_doloop): Same.
20165
0e0f87d4 201662013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY		 20167
20168 * config/arm/arm.c (emit_multi_reg_push): New declaration
20169 for an existing function.
20170 (arm_emit_strd_push): New function.
20171 (arm_expand_prologue): Used here.
20172 (arm_emit_ldrd_pop): New function.
20173 (arm_expand_epilogue): Used here.
20174 (arm_get_frame_offsets): Update condition.
20175 (arm_emit_multi_reg_pop): Add a special case for load of a single
20176 register with writeback.
20177
5e8e2af4
UB		 201782013-04-16 Uros Bizjak <ubizjak@gmail.com>
20179
20180 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20181 description.
20182
9fd9ccf7
RB		 201832013-04-16 Richard Biener <rguenther@suse.de>
20184
20185 PR tree-optimization/56756
20186 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20187 (first_mem_ref_loc): New.
20188 (execute_sm): Place the load temporarily before a previous
20189 access instead of in the latch edge to ensure its SSA dependencies
20190 are defined at points dominating the load.
20191
96fba521
SB		 201922013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20193
4c8af858
SB		 20194 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20195 correct fix by moving header and footer insn to the footer of
20196 the merged basic block. Clear BB_END of the merged-away block.
20197
96fba521
SB		 20198 PR middle-end/43631
20199 * emit-rtl.c (make_note_raw): New function.
20200 (link_insn_into_chain): New static inline function.
20201 (add_insn): Use it.
20202 (add_insn_before, add_insn_after): Factor insn chain linking code...
20203 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20204 using link_insn_into_chain.
20205 (note_outside_basic_block_p): New helper function for emit_note_after
20206 and emit_note_before.
20207 (emit_note_after): Use nobb variant of add_insn_after if the note
20208 should not be contained in a basic block.
20209 (emit_note_before): Use nobb variant of add_insn_before if the note
20210 should not be contained in a basic block.
20211 (emit_note_copy): Use make_note_raw.
20212 (emit_note): Likewise.
20213 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20214 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20215 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20216 the moved barrier the tail of the basic block it follows.
20217 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20218
7b8265ba
JJ		 202192013-04-15 Jakub Jelinek <jakub@redhat.com>
20220
20221 PR tree-optimization/56962
20222 * gimple-ssa-strength-reduction.c (record_increment): Only set
20223 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20224 either rhs1 or rhs2 is equal to c->base_expr.
20225
5185d248
RB		 202262013-04-15 Richard Biener <rguenther@suse.de>
20227
20228 PR tree-optimization/56933
20229 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20230 member.
20231 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20232 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20233 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20234 dependence check ...
20235 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20236 ... here.
20237 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20238 GROUP_READ_WRITE_DEPENDENCE.
20239
a24243a0
AK		 202402013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20241
20242 * emit-rtl.c (reset_all_used_flags): New function.
20243 (verify_rtl_sharing): Call reset_all_used_flags before and after
20244 performing the checks.
20245
1c50eada
KT		 202462013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20247
20248 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20249 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20250 * config/arm/constraints.md (De): New constraint.
20251 * config/arm/neon.md (anddi3_neon): Delete.
20252 (neon_vand<mode>): Expand to standard anddi3 pattern.
20253 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20254 Move earlier in the file.
20255 (neon_inv_logic_op2): Likewise.
20256 (arm_anddi_operand_neon): New predicate.
20257
e927b6ad
RO		 202582013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20259
20260 * configure.ac (gcc_cv_ld_as_needed): Set
20261 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20262 Use -z ignore, -z record on *-*-solaris2*.
20263 (HAVE_LD_AS_NEEDED): Update comment.
20264 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20265 * configure: Regenerate.
20266 * config.in: Regenerate.
20267 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20268 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20269 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20270 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20271 equivalents. Fix markup.
20272 * doc/tm.texi: Regenerate.
20273
e0ea8797
AH		 202742013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20275
20276 * config/i386/i386.opt: New option mstack-protector-guard=.
20277 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20278 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20279 TARGET_SSP_TLS_GUARD.
20280 * config/i386/i386.c (ix86_option_override_internal): Set
20281 ix86_stack_protector_guard.
20282 * config/i386/i386.md (stack_protect_set): Enable for
20283 TARGET_SSP_TLS_GUARD only.
20284 (stack_protect_set_<mode>): Ditto.
20285 (stack_protect_test): Ditto.
20286 (stack_protect_test_<mode>): Ditto.
20287 * doc/invoke.texi (i386 Option): Document.
20288
811b72f9
EB		 202892013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20290
20291 PR target/56890
20292 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20293 (S_MODES): Set H_MODE bit.
20294 (SF_MODES): Set only S_MODE and SF_MODE bits.
20295 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20296 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20297 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20298 <MODE_FLOAT>: Likewise.
20299
5529fdd6
JY		 203002013-04-15 Joey Ye <joey.ye@arm.com>
20301
20302 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20303
517b1da2
JY		 203042013-04-15 Joey Ye <joey.ye@arm.com>
20305
20306 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20307 for real far jump.
20308 (thumb_far_jump_used_p): Count instruction size and set
20309 far_jump_used.
20310
01007ae0
EB		 203112013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20312
20313 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20314 * resource.c (mark_target_live_regs): Reformat conditional block.
20315
c46f6580
SB		 203162013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20317
20318 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20319 notes, they are emitted only just before final.
20320 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20321
1f397f45
SB		 203222013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20323
20324 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20325 * cfgrtl.c (delete_insn): Call it here instead.
20326 * lra-spills.c (lra_final_code_change): Use delete_insn.
20327 * haifa-sched.c (sched_remove_insn): Likewise.
20328 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20329 returning to the nop pool.
20330 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20331 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20332
58a51369
SB		 203332013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20334
20335 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20336 * doc/tm.texi: Regenerated.
20337
33159866
UB		 203382013-04-12 Uros Bizjak <ubizjak@gmail.com>
20339
20340 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20341 QImode checks.
20342
226e378f
SB		 203432013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20344
20345 * df-core.c (df_find_def): Compare register numbers.
20346 (df_find_use): Likewise.
20347
fafb9b18
VM		 203482013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20349
20350 PR target/56903
20351 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20352 lra_in_progress for return.
20353
9a946fd6
GY		 203542013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20355
20356 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20357 define_insn into define_insn_and_split and emit movsicc patterns.
20358
41b83758
GY		 203592013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20360
20361 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20362
d3afd9aa
RB		 203632013-04-12 Richard Biener <rguenther@suse.de>
20364
20365 * tree-pass.h (TODO_do_not_ggc_collect): New.
20366 * passes.c (execute_one_ipa_transform_pass): Honor
20367 TODO_do_not_ggc_collect.
20368 (execute_one_pass): Likewise.
20369
20370 Revert
20371 2013-04-10 Richard Biener <rguenther@suse.de>
20372
20373 * passes.c (init_optimization_passes): Remove reload pass.
20374 * ira.c (do_reload): Merge into ...
20375 (ira): ... this.
20376 (rest_of_handle_reload): Remove.
20377 (pass_reload): Likewise.
20378 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20379 of reload for vzeroupper pass placement.
20380
06f9b387
JJ		 203812013-04-12 Jakub Jelinek <jakub@redhat.com>
20382
20383 PR tree-optimization/56918
20384 PR tree-optimization/56920
20385 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20386 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20387 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20388 use wide_mul_with_sign method.
20389
953094d2
RB		 203902013-04-12 Richard Biener <rguenther@suse.de>
20391
20392 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20393 not be considered a gimple constant.
20394
bb506982
MG		 203952013-04-12 Marc Glisse <marc.glisse@inria.fr>
20396
20397 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20398 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20399 operations.
20400
4b84d650
JJ		 204012013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20402 Jakub Jelinek <jakub@redhat.com>
20403
20404 * opts.c: Include diagnostic-color.h.
20405 (common_handle_option): Handle OPT_fdiagnostics_color_.
20406 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20407 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20408 (diagnostic-color.o): New.
20409 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20410 (diagnostic_color_rule): New enum.
20411 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20412 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20413 the location string.
20414 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20415 either NULL, or color kind.
20416 * diagnostic-color.c: New file.
20417 * diagnostic-color.h: New file.
20418 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20419 arguments.
20420 * doc/invoke.texi (-fdiagnostics-color): Document.
20421 * pretty-print.h (pp_show_color): Define.
20422 (struct pretty_print_info): Add show_color field.
20423 * diagnostic.c: Include diagnostic-color.h.
20424 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20425 macros. Colorize error:, warning: etc. strings and also the location
20426 string.
20427 (diagnostic_show_locus): Colorize the caret line.
20428 * pretty-print.c: Include diagnostic-color.h.
20429 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20430 inside of %< %> quotes or quoted through q format modifier.
20431
067a1e71
AK		 204322013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20433
33159866 20434 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 20435
33159866 204362013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N		 20437
20438 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20439 code in CC_NZ mode.
20440 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20441 pattern.
20442
7b55f98f
MP		 204432013-04-11 Marek Polacek <polacek@redhat.com>
20444
20445 PR tree-optimization/48184
33159866 20446 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 20447
966b587e
EB		 204482013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20449
20450 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20451 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20452 (skip_simple_arithmetic): Tidy up.
20453 * tree.h (skip_simple_constant_arithmetic): Declare.
20454
33159866 204552013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N		 20456
20457 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20458
1d42c1ec
RB		 204592013-04-11 Richard Biener <rguenther@suse.de>
20460
20461 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20462 generate vector constants.
20463
4ba5ea11
RB		 204642013-04-11 Richard Biener <rguenther@suse.de>
20465
20466 PR tree-optimization/56878
20467 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20468 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20469 New function.
20470 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20471 Prefer to align the DR with the most invariant base address.
20472
f0defe58
SKS		 204732013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20474
20475 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20476 comment.
20477
d07458be
JG		 204782013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20479
20480 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20481 floating-point vector comparisons against 0.
20482
146b8692
JJ		 204832013-04-11 Jakub Jelinek <jakub@redhat.com>
20484
20485 PR tree-optimization/56899
20486 * fold-const.c (extract_muldiv_1): Apply distributive law
20487 only if TYPE_OVERFLOW_WRAPS (ctype).
20488
b8578ff7
BC		 204892013-04-11 Bin Cheng <bin.cheng@arm.com>
20490
20491 PR target/56124
20492 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20493 loading has side effect.
20494
0ea8a6f9
SB		 204952013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20496
20497 * config/sparc/sparc.c: Include tree-pass.h.
20498 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20499 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20500 head of file. Change return type. Split off gate function.
20501 (sparc_gate_work_around_errata): New function.
20502 (pass_work_around_errata): New pass definition.
20503 (insert_pass_work_around_errata) New pass insert definition to
20504 insert pass_work_around_errata just after delayed-branch scheduling.
20505 (sparc_option_override): Insert the pass.
20506 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20507
42e37616
DM		 205082013-04-10 David S. Miller <davem@davemloft.net>
20509
89deeb3b
DM		 20510 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20511 or -mcpu=hypersparc.
20512
42e37616
DM		 20513 * target.def (cstore_mode): New hook.
20514 * target.h: Include insn-codes.h
20515 * targhooks.c: Likewise.
20516 (default_cstore_mode): New function.
20517 * targhooks.h: Declare it.
20518 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20519 * doc/tm.texi: Rebuild.
20520 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20521 target hook, rather than inspecting the insn_data.
20522 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20523 (TARGET_CSTORE_MODE): Redefine.
20524 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20525 result patterns.
20526 * config/sparc/predicates.md (cstore_result_operand): New special
20527 predicate.
20528 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20529 Use it for operand 0.
20530 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20531 (*snesi_special): Likewise.
20532 (*snesi_zero): Likewise.
20533 (*seqsi_zero): Likewise.
20534 (*sltu_insn): Likewise.
20535 (*sgeu_insn): Likewise.
20536 (*seqdi_special): Make operand 0 and comparison operation be of
20537 DImode.
20538 (*snedi_special): Likewise.
20539 (*snedi_special_vis3): Likewise.
20540 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20541 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20542 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20543 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20544 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20545 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20546 (*sltu_extend_sp64): Likewise.
20547 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20548 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20549 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20550 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20551 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20552
95ca411e
YZ		 205532013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20554
20555 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20556 (aarch64_start_file): Use the new function.
20557
6782438d 205582013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 20559 Jason Merrill <jason@redhat.com>
6782438d
SKS		 20560
20561 * common.opt: Add -gdwarf.
20562 * opts.c (common_handle_option): Handle it.
20563 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20564
bb313b93
RB		 205652013-04-10 Richard Biener <rguenther@suse.de>
20566
20567 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20568 (execute_one_ipa_transform_pass): But unconditionally here.
20569 (execute_one_pass): And here.
20570 (init_optimization_passes): Remove reload pass.
20571 * tree-pass.h (TODO_ggc_collect): Remove.
20572 (pass_reload): Likewise.
20573 * ira.c (do_reload): Merge into ...
20574 (ira): ... this.
20575 (rest_of_handle_reload): Remove.
20576 (pass_reload): Likewise.
20577 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20578 of reload for vzeroupper pass placement.
20579 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20580 and todo_flags_finish of all passes.
20581
793d9a16
RB		 205822013-04-10 Richard Biener <rguenther@suse.de>
20583
20584 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20585 first_const_oprnd field, rename first_def_type to first_op_type.
20586 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20587 (vect_get_and_check_slp_defs): Always use the type of the
20588 operand. Allow mixed vect_external_def, vect_constant_def types.
20589 (vect_get_constant_vectors): Handle mixed vect_external_def,
20590 vect_constant_def types.
20591
12211b99 205922013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR		 20593
20594 PR tree-optimization/55524
20595 * tree-ssa-math-opts.c
20596 (convert_mult_to_fma): Don't use an fms construct
20597 when we don't have an fms operation, but fnma, and it looks
20598 likely that we'll be able to use the latter.
20599
12211b99 206002013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ		 20601
20602 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20603 function.
20604 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20605 inline fail caused by overwritable functions.
20606
34ab4a5b
CJW		 206072013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20608
20609 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20610 unnecessary bits in the constant power of two case.
20611
abf9bfbc
RB		 206122013-04-10 Richard Biener <rguenther@suse.de>
20613
20614 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20615 broken code swapping operands.
20616 (vect_build_slp_tree): Do not compute load permutations here.
20617 (vect_analyze_slp_instance): Compute load permutations here,
20618 after building the SLP tree.
20619
f408477e
CB		 206202013-04-09 Christian Bruel <christian.bruel@st.com>
20621
20622 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20623 of next/prev_real_insn.
20624
5ac42672
JH		 206252013-04-09 Jan Hubicka <jh@suse.cz>
20626
abf9bfbc
RB		 20627 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20628 Drop aliased parameter.
5ac42672
JH		 20629 (function_and_variable_visibility): Do not handle alias pairs.
20630 * cgraph.c (varpool_externally_visible_p): Update prototype.
20631 * varpool.c (varpool_add_new_variable): Update.
20632
5017f1d2
KT		 206332013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20634
20635 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20636
48eecbee
SB		 206372013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20638
a949cf1c
SB		 20639 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20640
48eecbee
SB		 20641 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20642
75ef8e3d
MP		 206432013-04-09 Marek Polacek <polacek@redhat.com>
20644
20645 PR tree-optimization/48762
33159866 20646 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 20647
23847df4
RB		 206482013-04-09 Richard Biener <rguenther@suse.de>
20649
20650 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20651 dealing with cost.
20652 (vect_build_slp_tree): Likewise.
20653 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20654 calculating the cost of a SLP instance.
20655 (vect_analyze_slp_instance): Use it from here, after building
20656 the SLP tree.
20657
ea3a0fde
JJ		 206582013-04-09 Jakub Jelinek <jakub@redhat.com>
20659
20660 PR middle-end/56883
20661 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20662 expand_omp_for_static_chunk): Use simple_p = true in
20663 force_gimple_operand_gsi calls when assigning to addressable decls.
20664
a32dfe9d
JL		 206652013-04-09 Jeff Law <law@redhat.com>
20666
20667 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20668 when the boolean was created by converting a wider object which
20669 had a boolean range.
20670
d755c7ef
RB		 206712013-04-09 Richard Biener <rguenther@suse.de>
20672
20673 * tree-vectorizer.h (slp_void_p): Remove.
20674 (slp_tree): Typedef before _slp_tree declaration.
20675 (struct _slp_tree): Use a vector of slp_tree as children.
20676 (vect_get_place_in_interleaving_chain): Remove.
20677 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20678 Move ...
20679 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20680 and make static.
20681 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20682 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20683 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20684 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20685 Use slp_node instead of slp_void_p and adjust.
20686
3d741091
RB		 206872013-04-09 Richard Biener <rguenther@suse.de>
20688
20689 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20690 work that is not necessary.
20691
39307ba7
JJ		 206922013-04-09 Jakub Jelinek <jakub@redhat.com>
20693
20694 PR tree-optimization/56854
20695 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20696 forward into clobber stmts if it would change MEM_REF lhs into
20697 non-MEM_REF.
20698
343881fd
MK		 206992013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20700
20701 * tree.c (type_hash_lookup, type_hash_add): Make static.
20702 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20703
3922658a
RB		 207042013-04-09 Richard Biener <rguenther@suse.de>
20705
20706 * tree.h (unsave_expr_now): Remove.
20707 * tree-inline.c (mark_local_for_remap_r): Remove.
20708 (unsave_expr_1): Likewise.
20709 (unsave_r): Likewise.
20710 (unsave_expr_now): Likewise.
20711 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20712 (propagate_tree_value): Likewise.
20713
9fb6b620
SB		 207142013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20715
20716 * doc/rtl.texi (sequence): Rewrite documentation to match the
20717 current use of SEQUENCE rtl objects.
20718 * rtl.def (SEQUENCE): Likewise.
20719
20720 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20721 Update documentation.
20722 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20723 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20724
20725 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20726
8ddb5a29
TJ		 207272013-04-08 Teresa Johnson <tejohnson@google.com>
20728
20729 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20730 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 20731 methods.
8ddb5a29
TJ		 20732 (estimate_edge_size_and_time): Add comment to suggest using rounding
20733 methods.
20734 (estimate_node_size_and_time): Ditto.
20735 (remap_edge_change_prob): Use helper rounding divide methods.
20736 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20737 (gimple_mod_pow2_value_transform): Ditto.
20738 (gimple_mod_subtract_transform): Ditto.
20739 (gimple_ic_transform): Ditto.
20740 (gimple_stringops_transform): Ditto.
20741 * stmt.c (conditional_probability): Ditto.
20742 (emit_case_dispatch_table): Ditto.
20743 * lto-cgraph.c (merge_profile_summaries): Ditto.
20744 * tree-optimize.c (execute_fixup_cfg): Ditto.
20745 * cfgcleanup.c (try_forward_edges): Ditto.
20746 * cfgloopmanip.c (scale_loop_profile): Ditto.
20747 (loopify): Ditto.
20748 (duplicate_loop_to_header_edge): Ditto.
20749 (lv_adjust_loop_entry_edge): Ditto.
20750 * tree-vect-loop.c (vect_transform_loop): Ditto.
20751 * profile.c (compute_branch_probabilities): Ditto.
20752 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20753 * lto-streamer-in.c (input_cfg): Ditto.
20754 * gimple-streamer-in.c (input_bb): Ditto.
20755 * ipa-cp.c (update_profiling_info): Ditto.
20756 (update_specialized_profile): Ditto.
20757 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20758 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 20759 rounding methods.
8ddb5a29
TJ		 20760 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20761 (compute_trg_info): Ditto.
20762 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20763 (purge_dead_edges): Ditto.
20764 * loop-unswitch.c (unswitch_loop): Ditto.
20765 * cgraphclones.c (cgraph_clone_edge): Ditto.
20766 (cgraph_clone_node): Ditto.
20767 * tree-inline.c (copy_bb): Ditto.
20768 (copy_edges_for_bb): Ditto.
20769 (initialize_cfun): Ditto.
20770 (copy_cfg_body): Ditto.
20771 (expand_call_inline): Ditto.
20772
661e6bd7
KT		 207732013-04-08 Kai Tietz <ktietz@redhat.com>
20774
20775 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20776 TARGET_CYGWIN64 by TARGET_64BIT.
20777
105766f3
JR		 207782013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20779
20780 * config/epiphany/epiphany.md (GPR_1): New constant.
20781 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20782 * config/epiphany/epiphany.c (gen_compare_reg):
20783 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20784 is already in place.
20785 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20786 Don't require being called during rtl expansion; If y operlaps r0,
20787 return 0.
20788 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20789 (epiphany_expand_epilogue): Likewise.
20790
8afab237
JR		 20791 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20792 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 20793 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 20794
fa7707d6
JR		 20795 * config/epiphany/constraints.md (CnL): New constraint.
20796 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20797 * config/epiphany/predicates.md (add_operand): Allow 1024.
20798
5138e18d
JR		 20799 * config/epiphany/epiphany.md (logical_op): New code iterator.
20800 (op_mnc): New code attribute.
20801 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20802 (mov_f+1, mov_f+2): New peephole2 patterns.
20803
2ccc703d
JR		 20804 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20805 (cstoresi4): Also allow re-use of zero result when doing a NE
20806 comparison to a non-zero operand.
aefb0819 20807 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 20808
093ac0a5
JR		 20809 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20810 Use gen_addsi3_i / gen_subsi3_i.
20811
f223bb13
JJ		 208122013-04-08 Jakub Jelinek <jakub@redhat.com>
20813
20814 PR c++/34949
20815 PR c++/50243
20816 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20817 contain anything but clobbers, at most one __builtin_stack_restore,
20818 optionally debug stmts and final resx, and if it has at least one
20819 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20820 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20821 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20822 which isn't defaut definition, remove them.
20823 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20824 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20825 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20826 with MEM_REF LHS with SSA_NAME address.
20827
4481581f
JL		 208282013-04-08 Jeff Law <law@redhat.com>
20829
20830 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 20831
451dabda
RB		 208322013-04-08 Richard Biener <rguenther@suse.de>
20833
20834 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20835 extra newline.
20836 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20837 determined vector type.
20838 (vect_analyze_data_refs): Likewise.
20839 (vect_get_new_vect_var): Adjust.
20840 (vect_create_destination_var): Preserve SSA name versions.
20841 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20842 not dump anything here.
20843
3b088b47
JR		 208442013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20845
20846 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20847 Add member lr_slot_known.
20848 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20849 if necessary.
20850 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20851 Remove code that sets lr_slot_offset according to what a previous
20852 version of epiphany_emit_save_restore used to do.
20853 (epiphany_emit_save_restore): When doing an lr save or restore,
20854 set/verify lr_slot_known and lr_slot_offset.
20855
d8484d41
XQ		 208562013-04-08 Xinyu Qi <xyqi@marvell.com>
20857
33159866
UB		 20858 PR target/54338
20859 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ		 20860 in ALL_REGS.
20861
481be1c4
RB		 208622013-04-08 Richard Biener <rguenther@suse.de>
20863
20864 * alias.c (find_base_term): Fix thinko in previous change.
20865
401f3a81
JJ		 208662013-04-08 Jakub Jelinek <jakub@redhat.com>
20867
20868 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20869 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20870 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20871 if possible to compute val.
20872 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20873 For QImode integers don't require anything about precision. Use
20874 const_with_all_bytes_same to find out if the constant doesn't have
20875 repeated bytes in it.
20876
03ed99a8
AK		 208772013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20878
20879 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20880 within mode size.
20881
781b2e62
MP		 208822013-04-08 Marek Polacek <polacek@redhat.com>
20883
20884 PR rtl-optimization/48182
20885 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20886 value to 1.
20887
27e430a2
JDA		 208882013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20889
20890 PR target/55487
20891 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20892 nuses, make sure we have a label.
20893
4902aa64
BS		 208942013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20895
20896 PR target/56843
20897 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20898 (rs6000_emit_swdiv_low_precision): Remove.
20899 (rs6000_emit_swdiv): Rewrite to handle between one and four
20900 iterations of Newton-Raphson generally; modify required number of
20901 iterations for some cases.
20902 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20903
7bca81dc
SB		 209042013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20905
20906 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20907 set-but-unused variable.
20908
20909 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20910 basic blocks of released function bodies garbage-collectable.
20911
20912 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20913 (struct rtl_opt_pass): Add TODO_df_finish.
20914
20915 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20916
4542a38a
GY		 209172013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20918
20919 * config/arm/constraints.md (q): New constraint.
20920 * config/arm/ldrdstrd.md: New file.
20921 * config/arm/arm.md (ldrdstrd.md) New include.
20922 (arm_movdi): Use "q" instead of "r" constraint
20923 for double-word memory access.
20924 (movdf_soft_insn): Likewise.
20925 * config/arm/vfp.md (movdi_vfp): Likewise.
20926 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 20927 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY		 20928 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20929 (mem_ok_for_ldrd_strd): Likewise.
20930 (output_move_double): Update assertion.
20931
2385b218
GY		 209322013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20933
20934 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20935
75fe1cb5
GY		 209362013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20937
20938 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20939 define_insn_and_split.
20940 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20941
dd660e8e
GY		 209422013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20943
20944 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20945 define_insn_and_split.
33159866 20946 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY		 20947 (shiftsi3_compare): New pattern.
20948 (rrx): New pattern.
20949 * config/arm/unspecs.md (UNSPEC_RRX): New.
20950
045e472c
GY		 209512013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20952
20953 * config/arm/arm.md (negdi_extendsidi): New pattern.
20954 (negdi_zero_extendsidi): Likewise.
20955
3f3bf1a8
GY		 209562013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20957
20958 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20959 define_insn_and_split.
20960 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20961 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20962
b0b49556
GY		 209632013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20964
20965 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20966 define_insn_and_split.
20967 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20968 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20969
d633dd84
GY		 209702013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20971
20972 * config/arm/arm.md (subsi3_carryin): New pattern.
20973 (subsi3_carryin_const): Likewise.
20974 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20975 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20976
f4499066
GY		 209772013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20978
20979 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20980
ceef6fd9
GY		 209812013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20982
20983 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 20984 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 20985
ddbdd8a7
KT		 209862013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20987
20988 * config/arm/arm.c (arm_expand_builtin): Change fcode
20989 type to unsigned int.
20990
8456d78a
RR		 209912013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20992
20993 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20994
526c230f
ILT		 209952013-04-04 Ian Lance Taylor <iant@google.com>
20996
33159866
UB		 20997 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20998 language standard.
526c230f 20999
3cfbe04d
SB		 210002013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21001
21002 PR middle-end/56729
21003 * df-scan.c (df_insn_delete): Disable failing assert.
21004
dfa3f8d0
KT		 210052013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21006
21007 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21008 New function prototype.
21009 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21010 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21011 (arm_builtin_vectorized_function): New function.
21012
f7837758
KT		 210132013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21014
21015 * config/arm/arm_neon_builtins.def: New file.
21016 * config/arm/arm.c (neon_builtin_data): Move contents to
21017 arm_neon_builtins.def.
21018 (enum arm_builtins): Include neon builtin definitions.
21019 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 21020 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 21021
39c1b6db
MP		 210222013-04-04 Marek Polacek <polacek@redhat.com>
21023
21024 PR tree-optimization/48186
21025 * predict.c (maybe_hot_frequency_p): Return false if
21026 HOT_BB_FREQUENCY_FRACTION is 0.
21027 (cgraph_maybe_hot_edge_p): Likewise.
21028
314f64eb
RB		 210292013-04-04 Richard Biener <rguenther@suse.de>
21030
21031 PR tree-optimization/56826
21032 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21033 more accurately.
21034
51a905b2
RB		 210352013-04-04 Richard Biener <rguenther@suse.de>
21036
21037 PR tree-optimization/56213
21038 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 21039 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 21040
f1bf4f3a
RB		 210412013-04-04 Richard Biener <rguenther@suse.de>
21042
21043 PR tree-optimization/56837
21044 * tree-loop-distribution.c (classify_partition): For non-zero
21045 values require that the value has the same precision as its
21046 mode to be useful as memset value.
21047
0bca7ded
NC		 210482013-04-03 Nick Clifton <nickc@redhat.com>
21049
33159866 21050 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC		 21051 (fmssf4): Use fmsf.s on E3V5 architectures.
21052 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21053 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21054
b4019227
JL		 210552013-04-03 Jeff Law <law@redhat.com>
21056
21057 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21058 (lra-eliminations.o): Likewise.
21059
f57ddb5b
TJ		 210602013-04-03 Teresa Johnson <tejohnson@google.com>
21061
21062 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 21063 compute_working_sets here from profile.c.
f57ddb5b 21064 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB		 21065 (gcov_working_set_t): Moved typedef here from basic-block.h
21066 (compute_working_set): Declare.
f57ddb5b
TJ		 21067 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21068 (get_working_sets): Renamed from compute_working_set,
33159866 21069 replace most of body with call to new compute_working_sets.
f57ddb5b 21070 (get_exec_counts): Replace call to compute_working_sets
33159866
UB		 21071 to get_working_sets.
21072 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 21073 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 21074 to get_working_sets.
f57ddb5b
TJ		 21075 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21076 * gcov-dump.c (dump_working_sets): New function.
21077
12211b99 210782013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ		 21079
21080 * hwint.c (sext_hwi, zext_hwi): New functions.
21081 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21082 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21083 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21084 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21085 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21086 (sext_hwi, zext_hwi): New functions.
0bca7ded 21087
be672e08
JL		 210882013-04-03 Jeff Law <law@redhat.com>
21089
21090 PR tree-optimization/56799
33159866
UB		 21091 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21092 back test for widening conversion erroneously dropped in prior change.
be672e08 21093
9d821fa5
KT		 210942013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21095
21096 PR target/56809
21097 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21098 instead of next_real_insn.
21099
71c581e7
MP		 211002013-04-03 Marek Polacek <polacek@redhat.com>
21101
21102 PR sanitizer/55702
33159866 21103 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 21104
4a32ef80
KT		 211052013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21106
21107 PR target/56809
21108 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21109 next_real_insn.
21110 (thumb1_output_casesi): Likewise.
21111 (thumb2_output_casesi): Likewise.
21112
1b2253d4
RB		 211132013-04-03 Richard Biener <rguenther@suse.de>
21114
21115 PR tree-optimization/56817
21116 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21117 Split out ...
21118 (tree_unroll_loops_completely_1): ... new function to manually
21119 walk the loop tree, properly defering outer loops of unrolled
21120 loops to later iterations.
21121
38000232
MG		 211222013-04-03 Marc Glisse <marc.glisse@inria.fr>
21123
21124 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21125 (vectorizable_load): Likewise.
21126 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21127 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21128
3fa35298
MG		 211292013-04-03 Marc Glisse <marc.glisse@inria.fr>
21130
21131 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21132 BIT_FIELD_REF.
21133
b3d45ff0
UW		 211342013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21135
21136 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21137
ec9202a8
BC		 211382013-04-03 Bin Cheng <bin.cheng@arm.com>
21139
21140 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21141
6805bd36
MG		 211422013-04-03 Marc Glisse <marc.glisse@inria.fr>
21143
21144 PR tree-optimization/56790
33159866
UB		 21145 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21146 folding.
6805bd36 21147
da694a77
MG		 211482013-04-03 Marc Glisse <marc.glisse@inria.fr>
21149
21150 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21151 Handle VEC_MERGE.
21152 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21153 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21154 equal arguments.
21155
4e7d7b3d
JJ		 211562013-04-03 Jakub Jelinek <jakub@redhat.com>
21157
21158 PR c/19449
21159 * tree.h (force_folding_builtin_constant_p): New decl.
21160 * builtins.c (force_folding_builtin_constant_p): New variable.
21161 (fold_builtin_constant_p): Fold immediately also if
21162 force_folding_builtin_constant_p.
21163
e6c9d234
RB		 211642013-04-03 Richard Biener <rguenther@suse.de>
21165
21166 PR tree-optimization/56812
21167 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21168 DRs of the same interleaving chain are independent.
21169
984d07dd
JM		 211702013-04-02 Jason Merrill <jason@redhat.com>
21171
21172 * gdbinit.in (pbb): Use debug fn.
21173
622849c9
LC		 211742013-04-02 Lawrence Crowl <crowl@google.com>
21175
21176 * sese.h (struct ivtype_map_elt_s): Remove unused.
21177 (extern debug_ivtype_map): Remove unused.
21178 (extern eq_ivtype_map_elts): Remove unused.
21179 * sese.c (debug_ivtype_map): Removed unused.
21180 (debug_ivtype_map_1): Removed unused.
21181 (debug_ivtype_elt): Remove unused.
21182 (eq_ivtype_map_elts): Remove unused.
21183
21184
82c0e1a0
KT		 211852013-04-02 Kai Tietz <ktietz@redhat.com>
21186
21187 PR target/52790
21188 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21189 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB		 21190 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21191 function.
82c0e1a0
KT		 21192 (legitimize_pe_coff_symbol): Likewise.
21193 (is_imported_p): New helper-function.
21194 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21195 for Windows x64 targets.
21196 (ix86_expand_prologue): Optimize for pe-coff targets.
21197 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21198 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21199 medium/large code-model.
21200 (legitimize_pic_address): Likewise.
21201 (legitimize_tls_address): Likewise.
21202 (ix86_expand_call): Likewise.
21203 (x86_output_mi_thunk): Likewise.
21204 (get_dllimport_decl): Add new beimport argument.
21205 (construct_plt_address): Don't assert for x64 pe-coff targets.
21206 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21207 targets.
21208 (SYMBOL_FLAG_STUBVAR): New macro.
21209 (SYMBOL_REF_STUBVAR_P): Likewise.
21210 * config/i386/winnt.c (stub_list): New structure.
21211 (stub_head): New local variable.
21212 (i386_pe_record_stub): New function.
21213 (i386_pe_file_end): Emit refptr-stubs.
21214
5d751b0c
JJ		 212152013-04-02 Jakub Jelinek <jakub@redhat.com>
21216
09bb4c99
JJ		 21217 PR rtl-optimization/56745
21218 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21219 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21220
a1d8947a
JJ		 21221 PR c++/34949
21222 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21223 and both of them are MEM_REFs, just compare first argument for
21224 equality and attempt to deal even with differing offsets.
21225
5d751b0c
JJ		 21226 PR c++/34949
21227 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21228 of gimple_clobber_p to be MEM_REF.
21229 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21230 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21231 after gimplification.
21232 * asan.c (get_mem_ref_of_assignment): Don't instrument
21233 gimple_clobber_p stmts.
21234 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21235 gimple_clobber_p stmt if they have MEM_REF lhs and
21236 are dead because of another gimple_clobber_p stmt.
21237 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21238 gimple_clobber_p stmts like debug stmts.
21239 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21240 that refer to unused VAR_DECLs or uninitialized values.
21241 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21242 gimple_clobber_p stmts if they refer to removed parameters.
21243 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21244 formatting.
21245
e594716a
UB		 212462013-04-02 Uros Bizjak <ubizjak@gmail.com>
21247
21248 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21249 using SWI48 mode attribute.
21250
7a80735b
WM		 212512013-04-02 Wei Mi <wmi@google.com>
21252
21253 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21254 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21255 *<rotate_insn><mode>3_mask in i386.md.
21256
f423a9e4
AI		 212572013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21258
21259 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21260
90eb75f2
RB		 212612013-04-02 Richard Biener <rguenther@suse.de>
21262
21263 PR tree-optimization/56778
21264 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21265 Runtime alias tests are not supported for gather loads.
21266 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21267 stmts referenced from SSA operands before updating SSA form.
21268
d8c69a92
IC		 212692013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21270 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21271
21272 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB		 21273 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21274 * config/arm/cortex-a53.md: New file.
21275 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21276 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21277 * config/arm/arm.c (arm_issue_rate): Likewise.
21278 * config/arm/arm-tune.md: Regenerate
21279 * config/arm/arm-tables.opt: Regenerate.
21280 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 21281
239eb04c
ZC		 212822013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21283
21284 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21285 non-static link.
21286
c902d3c8
SN		 212872013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21288
d8c69a92
IC		 21289 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21290 scalar load/store operations using B/H registers.
c902d3c8
SN		 21291 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21292
051d0e2f
SN		 212932013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21294
21295 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21296 scalar move.
21297 * config/aarch64/aarch64.c
21298 (aarch64_simd_scalar_immediate_valid_for_move): New.
21299 * config/aarch64/aarch64-protos.h
21300 (aarch64_simd_scalar_immediate_valid_for_move): New.
21301 * config/aarch64/constraints.md (Dh, Dq): New.
21302 * config/aarch64/iterators.md (hq): New.
21303
0ee1e3d9
EB		 213042013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21305
21306 * reorg.c (get_branch_condition): Deal with conditional returns.
21307 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21308
136fb3f7
WM		 213092013-04-01 Wei Mi <wmi@google.com>
21310
21311 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21312 Truncate operand 2 using %b asm operand modifier.
21313 (*<shift_insn><mode>3_mask): Ditto.
21314 (*<rotate_insn><mode>3_mask): Ditto.
21315
6388c738
SB		 213162013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21317
21318 PR middle-end/56798
21319 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21320
f7a4d826
KK		 213212013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21322
21323 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21324 of next_real_insn.
21325 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21326
4b943a49
LC		 213272013-03-30 Lawrence Crowl <crowl@google.com>
21328
21329 * dse.c (clear_alias_sets): Remove never set.
21330 (disqualified_clear_alias_sets): Remove never set.
21331 (clear_alias_mode_pool): Remove never set.
21332 (dse_step0): Remove condition that is never true.
21333 (canon_address): Remove condition that is never true.
21334 (dse_step7): Remove condition that is never true.
21335 (rest_of_handle_dse): Remove condition that is never true.
21336 (rest_of_handle_dse::did_global): Remove never read from above.
21337 (dse_step2_spill): Remove never called from above.
21338 (dse_step5_spill): Remove never called from above.
21339
39718607
SB		 213402013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21341
da5c6bde
SB		 21342 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21343 JUMP_TABLE_DATA changes.
21344 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21345 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21346 (Insns) <jump_table_data>: New entry.
21347 * doc/tm.texi: Regenerate.
21348
39718607
SB		 21349 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21350
21351 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21352 for table jump at the end of a basic block using tablejump_p.
21353 * targhooks.c (default_invalid_within_doloop): Likewise.
21354 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21355 target hook implementation that is identical to the default hook.
21356 (rs6000_invalid_within_doloop): Remove.
21357
21358 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21359 unused variable from tablejump_p call.
21360
21361 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21362 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21363 (INSN_DELETED_P): Likewise.
21364 (emit_jump_table_data): New prototype.
21365 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21366 after 4th as unused.
21367 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21368 * sched-vis.c (print_insn): Likewise.
21369 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21370 insn for compatibility with back ends that use next_active_insn to
21371 identify jump table data.
21372 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21373 (remove_insn): Likewise.
21374 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21375 to be emitted.
21376 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21377 (emit_jump_table_data): New function.
21378
21379 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21380 basic block, a JUMP_TABLE_DATA never is.
21381 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21382 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21383 off from code handling real insns.
21384 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21385 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21386 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21387 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21388 is not a NONDEBUG_INSN_P.
21389 * ira-costs.c (scan_one_insn): Likewise.
21390 * jump.c (mark_all_labels): Likewise.
21391 (mark_jump_label_1): Likewise.
21392 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21393 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21394 a basic block.
21395 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21396 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21397 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21398 JUMP_TABLE_DATA_P insns.
21399 (calculate_elim_costs_all_insns): Likewise.
21400 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21401 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21402 (delete_output_reload): Code style fixups.
21403 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21404 insn flags on this non-insn.
21405 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21406 as scheduling barriers, for pre-change compatibility.
21407 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21408 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21409
21410 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21411 redundant JUMP_TABLE_DATA_P test.
21412 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21413 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21414 (frv_for_each_packet): Likewise.
21415 * config/i386/i386.c (min_insn_size): Likewise.
21416 (ix86_avoid_jump_mispredicts): Likewise.
21417 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21418 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21419 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21420 (mips16_insn_length): Robustify.
21421 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21422 (mips16_split_long_branches): Likewise.
21423 * config/pa/pa.c (pa_combine_instructions): Likewise.
21424 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21425 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21426 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21427 as contributing to pool range lengths.
21428 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21429 Remove redundant JUMP_TABLE_DATA_P test.
21430 (sh_loop_align): Likewise.
21431 (split_branches): Likewise.
21432 (sh_insn_length_adjustment): Likewise.
21433 * config/spu/spu.c (get_branch_target): Likewise.
21434
0208f7da
JH		 214352013-03-29 Jan Hubicka <jh@suse.cz>
21436
21437 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21438 gcov streaming; stream hot bb threshold to ltrans.
21439 * predict.c (get_hot_bb_threshold): Break out from ....
21440 (maybe_hot_count_p): ... here.
21441 (set_hot_bb_threshold): New function.
21442 * lto-section-in.c (lto_section_name): Add profile.
21443 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21444 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21445 and data-streamer.h
21446 (histogram_entry): New structure.
21447 (histogram, histogram_pool): New global vars.
21448 (histogram_hash): New structure.
21449 (histogram_hash::hash): New method.
21450 (histogram_hash::equal): Likewise.
21451 (account_time_size): New function.
21452 (cmp_counts): New function.
21453 (dump_histogram): New function.
21454 (ipa_profile_generate_summary): New function.
21455 (ipa_profile_write_summary): New function.
21456 (ipa_profile_read_summary): New function.
21457 (ipa_profile): Decide on threshold.
e594716a
UB		 21458 (pass_ipa_profile): Add ipa_profile_write_summary and
21459 ipa_profile_read_summary.
0208f7da
JH		 21460 * Makefile.in (ipa.o): Update dependencies.
21461 * lto-streamer.h (LTO_section_ipa_profile): New section.
21462
5a6ccc94
GDR		 214632013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21464
21465 * tree.h (VAR_P): New.
21466
39385fa6
PC		 214672013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21468
21469 PR lto/56777
21470 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21471
34f0d87a
SB		 214722013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21473
21474 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21475 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21476 (control_flow_insn_p): Likewise.
21477 * cfgrtl.c (duplicate_insn_chain): Likewise.
21478 * final.c (get_attr_length_1): Likewise.
21479 (shorten_branches): Likewise.
21480 (final_scan_insn): Likewise.
21481 * function.c (instantiate_virtual_regs): Likewise.
21482 * gcse.c (insert_insn_end_basic_block): Likewise.
21483 * ira-costs.c (scan_one_insn): Likewise.
21484 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21485 * lra.c (check_rtl): Likewise.
21486 * reload1.c (elimination_costs_in_insn): Likewise.
21487 * reorg.c (follow_jumps): Likewise.
21488
21489 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21490 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21491 (thumb_far_jump_used_p): Likewise.
21492 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21493 (workaround_speculation): Likewise.
21494 (add_sched_insns_for_speculation): Likewise.
21495 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21496 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21497 (frv_for_each_packet): Likewise.
21498 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21499 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21500 (final_emit_insn_group_barriers): Likewise.
21501 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21502 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21503 (mips16_insn_length): Likewise.
21504 * config/pa/pa.c (pa_reorg): Likewise.
21505 (pa_combine_instructions): Likewise.
21506 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21507 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21508 (sh_reorg): Likewise.
21509 (split_branches): Likewise.
21510 * config/spu/spu.c (get_branch_target): Likewise.
21511
21512 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21513 JUMP_TABLE_DATA_P.
21514
4ac761b0
KY		 215152013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21516
39385fa6 21517 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY		 21518 Fix declaration name.
21519
58a49189
LC		 215202013-03-28 Lawrence Crowl <crowl@google.com>
21521
21522 * graphds.h (struct graph.indicies): Remove unused.
21523 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21524 (SCOP_ORIGINAL_PDDRS): Remove unused.
21525 * sese.h (extern insert_loop_close_phis): Removed unused.
21526 (extern insert_guard_phis): Removed unused.
21527 (extern ivtype_map_elt_info): Removed unused.
21528 (new_ivtype_map_elt): Removed unused.
21529 * sese.c (ivtype_map_elt_info): Removed unused.
21530
7b3b6ae4
LC		 215312013-03-28 Lawrence Crowl <crowl@google.com>
21532
21533 * Makefile.in: Add several missing include dependences.
21534 (DUMPFILE_H): New.
21535 (test-dump.o): New. This object is not added to any executable,
21536 but is present for ad-hoc testing.
21537 * bitmap.c
21538 (debug (const bitmap_head_def &)): New.
21539 (debug (const bitmap_head_def *)): New.
21540 * bitmap.h
21541 (extern debug (const bitmap_head_def &)): New.
21542 (extern debug (const bitmap_head_def *)): New.
21543 * cfg.c
21544 (debug (edge_def &)): New.
21545 (debug (edge_def *)): New.
21546 * cfghooks.c
21547 (debug (basic_block_def &)): New.
21548 (debug (basic_block_def *)): New.
21549 * dumpfile.h
21550 (dump_node (const_tree, int, FILE *)): Correct source file.
21551 * dwarf2out.c
21552 (debug (die_struct &)): New.
21553 (debug (die_struct *)): New.
21554 * dwarf2out.h
21555 (extern debug (die_struct &)): New.
21556 (extern debug (die_struct *)): New.
21557 * gimple-pretty-print.c
21558 (debug (gimple_statement_d &)): New.
21559 (debug (gimple_statement_d *)): New.
21560 * gimple-pretty-print.h
21561 (extern debug (gimple_statement_d &)): New.
21562 (extern debug (gimple_statement_d *)): New.
21563 * ira-build.c
21564 (debug (ira_allocno_copy &)): New.
21565 (debug (ira_allocno_copy *)): New.
21566 (debug (ira_allocno &)): New.
21567 (debug (ira_allocno *)): New.
21568 * ira-int.h
21569 (extern debug (ira_allocno_copy &)): New.
21570 (extern debug (ira_allocno_copy *)): New.
21571 (extern debug (ira_allocno &)): New.
21572 (extern debug (ira_allocno *)): New.
21573 * ira-lives.c
21574 (debug (live_range &)): New.
21575 (debug (live_range *)): New.
21576 * lra-int.h
21577 (debug (lra_live_range &)): New.
21578 (debug (lra_live_range *)): New.
21579 * lra-lives.c
21580 (debug (lra_live_range &)): New.
21581 (debug (lra_live_range *)): New.
21582 * omega.c
21583 (debug (omega_pb_d &)): New.
21584 (debug (omega_pb_d *)): New.
21585 * omega.h
21586 (extern debug (omega_pb_d &)): New.
21587 (extern debug (omega_pb_d *)): New.
21588 * print-rtl.c
21589 (debug (const rtx_def &)): New.
21590 (debug (const rtx_def *)): New.
21591 * print-tree.c
21592 (debug_tree (tree): Move within file.
21593 (debug_raw (const tree_node &)): New.
21594 (debug_raw (const tree_node *)): New.
21595 (dump_tree_via_hooks (const tree_node *, int)): New.
21596 (debug (const tree_node &)): New.
21597 (debug (const tree_node *)): New.
21598 (debug_verbose (const tree_node &)): New.
21599 (debug_verbose (const tree_node *)): New.
21600 (debug_head (const tree_node &)): New.
21601 (debug_head (const tree_node *)): New.
21602 (debug_body (const tree_node &)): New.
21603 (debug_body (const tree_node *)): New.
21604 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21605 (debug (vec<tree, va_gc> &)): New.
21606 (debug (vec<tree, va_gc> *)): New.
21607 * rtl.h
21608 (extern debug (const rtx_def &)): New.
21609 (extern debug (const rtx_def *)): New.
21610 * sbitmap.c
21611 (debug_raw (simple_bitmap_def &)): New.
21612 (debug_raw (simple_bitmap_def *)): New.
21613 (debug (simple_bitmap_def &)): New.
21614 (debug (simple_bitmap_def *)): New.
21615 * sbitmap.h
21616 (extern debug (simple_bitmap_def &)): New.
21617 (extern debug (simple_bitmap_def *)): New.
21618 (extern debug_raw (simple_bitmap_def &)): New.
21619 (extern debug_raw (simple_bitmap_def *)): New.
21620 * sel-sched-dump.c
21621 (debug (vinsn_def &)): New.
21622 (debug (vinsn_def *)): New.
21623 (debug_verbose (vinsn_def &)): New.
21624 (debug_verbose (vinsn_def *)): New.
21625 (debug (expr_def &)): New.
21626 (debug (expr_def *)): New.
21627 (debug_verbose (expr_def &)): New.
21628 (debug_verbose (expr_def *)): New.
21629 (debug (vec<rtx> &)): New.
21630 (debug (vec<rtx> *)): New.
21631 * sel-sched-dump.h
21632 (extern debug (vinsn_def &)): New.
21633 (extern debug (vinsn_def *)): New.
21634 (extern debug_verbose (vinsn_def &)): New.
21635 (extern debug_verbose (vinsn_def *)): New.
21636 (extern debug (expr_def &)): New.
21637 (extern debug (expr_def *)): New.
21638 (extern debug_verbose (expr_def &)): New.
21639 (extern debug_verbose (expr_def *)): New.
21640 (extern debug (vec<rtx> &)): New.
21641 (extern debug (vec<rtx> *)): New.
21642 * sel-sched-ir.h
21643 (_list_iter_cond_expr): Make inline instead of static.
21644 * sreal.c
21645 (debug (sreal &)): New.
21646 (debug (sreal *)): New.
21647 * sreal.h
21648 (extern debug (sreal &)): New.
21649 (extern debug (sreal *)): New.
21650 * tree.h
21651 (extern debug_raw (const tree_node &)): New.
21652 (extern debug_raw (const tree_node *)): New.
21653 (extern debug (const tree_node &)): New.
21654 (extern debug (const tree_node *)): New.
21655 (extern debug_verbose (const tree_node &)): New.
21656 (extern debug_verbose (const tree_node *)): New.
21657 (extern debug_head (const tree_node &)): New.
21658 (extern debug_head (const tree_node *)): New.
21659 (extern debug_body (const tree_node &)): New.
21660 (extern debug_body (const tree_node *)): New.
21661 (extern debug (vec<tree, va_gc> &)): New.
21662 (extern debug (vec<tree, va_gc> *)): New.
21663 * tree-cfg.c
21664 (debug (struct loop &)): New.
21665 (debug (struct loop *)): New.
21666 (debug_verbose (struct loop &)): New.
21667 (debug_verbose (struct loop *)): New.
21668 * tree-dump.c: Add header dependence.
21669 * tree-flow.h
21670 (extern debug (struct loop &)): New.
21671 (extern debug (struct loop *)): New.
21672 (extern debug_verbose (struct loop &)): New.
21673 (extern debug_verbose (struct loop *)): New.
21674 * tree-data-ref.c
21675 (debug (data_reference &)): New.
21676 (debug (data_reference *)): New.
21677 (debug (vec<data_reference_p> &)): New.
21678 (debug (vec<data_reference_p> *)): New.
21679 (debug (vec<ddr_p> &)): New.
21680 (debug (vec<ddr_p> *)): New.
21681 * tree-data-ref.h
21682 (extern debug (data_reference &)): New.
21683 (extern debug (data_reference *)): New.
21684 (extern debug (vec<data_reference_p> &)): New.
21685 (extern debug (vec<data_reference_p> *)): New.
21686 (extern debug (vec<ddr_p> &)): New.
21687 (extern debug (vec<ddr_p> *)): New.
21688 * tree-ssa-alias.c
21689 (debug (pt_solution &)): New.
21690 (debug (pt_solution *)): New.
21691 * tree-ssa-alias.h
21692 (extern debug (pt_solution &)): New.
21693 (extern debug (pt_solution *)): New.
21694 * tree-ssa-alias.c
21695 (debug (_var_map &)): New.
21696 (debug (_var_map *)): New.
21697 (debug (tree_live_info_d &)): New.
21698 (debug (tree_live_info_d *)): New.
21699 * tree-ssa-alias.h
21700 (extern debug (_var_map &)): New.
21701 (extern debug (_var_map *)): New.
21702 (extern debug (tree_live_info_d &)): New.
21703 (extern debug (tree_live_info_d *)): New.
21704
be77e1e5
JH		 217052013-03-28 Jan Hubicka <jh@suse.cz>
21706
21707 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21708
777e6976
IB		 217092013-03-28 Ian Bolton <ian.bolton@arm.com>
21710
21711 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21712 record only when desired or required.
21713
9cd347ae
UB		 217142013-03-28 Uros Bizjak <ubizjak@gmail.com>
21715
21716 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21717 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21718
33159866 217192013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N		 21720
21721 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21722 (*andsi3_compare0_uxtw): New pattern.
21723 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21724 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21725
89ab31c1
JH		 217262013-03-28 Jan Hubicka <jh@suse.cz>
21727
21728 * data-streamer-in.c (streamer_read_gcov_count): New function.
21729 * gimple-streamer-out.c: Include value-prof.h.
21730 (output_gimple_stmt): Output histogram.
21731 (output_bb): Use streamer_write_gcov_count.
21732 * value-prof.c: Include data-streamer.h
21733 (dump_histogram_value): Add HIST_TYPE_MAX.
21734 (stream_out_histogram_value): New function.
21735 (stream_in_histogram_value): New function.
21736 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21737 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21738 * data-streamer-out.c (streamer_write_gcov_count): New function.
21739 (streamer_write_gcov_count_stream): New function.
21740 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21741 (lto_output_node): Likewise.
21742 (input_node, input_edge): Likewise.
21743 * lto-streamer-out.c (output_cfg): Update streaming.
21744 * lto-streamer-in.c (input_cfg): Likewise.
21745 * data-streamer.h (streamer_write_gcov_count,
21746 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21747 * gimple-streamer-in.c: Include value-prof.h
21748 (input_gimple_stmt): Input histograms.
21749 (input_bb): Update profile streaming.
21750
e594716a 217512013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 21752
33159866 21753 * genmodes.c (emit_max_int): New function.
8697be17 21754 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC		 21755 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21756 Added doc.
8697be17 21757 * machmode.def: Fixed comment.
89ab31c1 21758
e594716a 217592013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ		 21760
21761 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ		 21762 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21763
74946978
MP		 217642013-03-28 Marek Polacek <polacek@redhat.com>
21765 Richard Biener <rguenther@suse.de>
21766
21767 PR tree-optimization/56695
21768 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21769 build signed result of a vector comparison.
21770 * tree-cfg.c (verify_gimple_comparison): Check that a result
21771 of a vector comparison has signed type.
21772
a64b9c26
RB		 217732013-03-28 Richard Biener <rguenther@suse.de>
21774
21775 PR tree-optimization/37021
21776 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21777 do not restrict gaps between groups.
21778 * tree-vect-stmts.c (vectorizable_load): Properly account for
21779 a gap between groups.
21780
a9dc2a2f
EB		 217812013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21782
21783 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21784 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21785 is not enabled.
21786
53cb97f9
GP		 217872013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21788
21789 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21790 * doc/extend.texi (Named Address Spaces): Ditto.
21791 (Variable Attributes): Ditto.
21792
b802ae5c
KT		 217932013-03-27 Kai Tietz <ktietz@redhat.com>
21794
eddae10a
KT		 21795 * config.build: Add support for cygwin x64 target.
21796 * config.gcc: Likewise.
21797 * config.host: Likewise.
21798 * configure.ac: Likewise
21799 * configure: Regenerated.
b802ae5c 21800
371e77e3 218012013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT		 21802
21803 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21804 * config/i386/t-cygwin-w64: New file.
21805 * config/i386/cygwin-w64.h: New file.
21806 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21807 and add support for x64-cygwin target.
21808 (CPP_SPEC): Likewise.
21809 (CXX_WRAP_SPEC_LIST): Undefine before define.
21810 (LIBGCJ_SONAME): Use 15 as version.
21811
f49b33cb
RB		 218122013-03-27 Richard Biener <rguenther@suse.de>
21813
21814 PR tree-optimization/56716
21815 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21816 dumping for ref nodes.
21817
b37a6ce5
MJ		 218182013-03-27 Martin Jambor <mjambor@suse.cz>
21819
21820 PR tree-optimization/55334
21821 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21822 restricted pointers to arrays.
21823
9469b9b2
GDR		 218242013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21825
21826 * Makefile.in (.SUFFIXES): Add .cc.
21827 (.c.o): Apply same recipe for implicit rule .cc.o.
21828
7d24f650
RB		 218292013-03-27 Richard Biener <rguenther@suse.de>
21830
21831 PR tree-optimization/37021
21832 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21833 REALPART/IMAGPART_EXPRs around the supported refs.
21834 * tree-ssa-structalias.c (find_func_aliases): Assume that
21835 floating-point values are not used to transfer pointers.
21836
2f251a05
AI		 218372013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21838
f43245d1
UB		 21839 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21840 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21841 * doc/tm.texi: Regenerate.
21842 * targhooks.h (default_has_ifunc_p): New.
21843 * targhooks.c (default_has_ifunc_p): Ditto.
21844 * config/linux-protos.h: New file.
21845 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21846 hook for linux which disables support of indirect functions in android.
21847 * config/linux-android.c: New file.
21848 * config/t-linux-android.c: Ditto.
21849 * config.gcc: Added new object file linux-android.o.
21850 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 21851 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 21852 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI		 21853 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21854 doesn't support indirect functions.
21855 * configure: Regenerate.
21856
78b4e425
BC		 218572013-03-27 Bin Cheng <bin.cheng@arm.com>
21858
21859 PR target/56102
21860 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21861 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21862 mult-word mode.
21863
a5ba7b92
AK		 218642013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21865
21866 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21867
be624986
TG		 218682013-03-27 Terry Guo <terry.guo@arm.com>
21869
21870 * config/arm/arm-cores.def: Added core cortex-r7.
21871 * config/arm/arm-tune.md: Regenerated.
21872 * config/arm/arm-tables.opt: Regenerated.
21873 * doc/invoke.texi: Added entry for core cortex-r7.
21874
0a514f47
WL		 218752013-03-27 Walter Lee <walt@tilera.com>
21876
21877 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21878 double-decrement of next_scratch_regno.
21879
9332b0d2
WL		 218802013-03-27 Walter Lee <walt@tilera.com>
21881
801d9b2a 21882 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL		 21883 input operands.
21884 (insn_v1mulus): Ditto.
21885 (insn_v2muls): Ditto.
21886
f54ea5dd
WL		 218872013-03-27 Walter Lee <walt@tilera.com>
21888
e594716a 21889 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL		 21890 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21891
bf60f4ca
WL		 218922013-03-27 Walter Lee <walt@tilera.com>
21893
e594716a 21894 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL		 21895 (*sibcall_value): Ditto.
21896
e3b51eeb
WL		 218972013-03-27 Walter Lee <walt@tilera.com>
21898
21899 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21900 (insn_mnz_v8qi): ... this ...
21901 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21902 vector equivalent.
21903 (insn_v<n>mnz): Replaced by ...
21904 (insn_v1mnz): ... this ...
21905 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21906 equivalent.
21907 (insn_mz_<mode>): Replaced by ...
21908 (insn_mz_v8qi): ... this ...
21909 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21910 vector equivalent.
21911 (insn_v<n>mz): Replaced by ...
21912 (insn_v1mz): ... this ...
21913 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21914 equivalent.
21915
4fc7b145
EB		 219162013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21917
21918 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21919
6f33817e
RM		 219202013-03-26 Roland McGrath <mcgrathr@google.com>
21921
21922 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21923 than fprintf with a non-constant, non-format string.
21924
93a4145b
UB		 219252013-03-26 Uros Bizjak <ubizjak@gmail.com>
21926
21927 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21928 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21929 operand 0 predicate.
21930 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21931 attribute. Use general_x64nomem_operand as operand 1 predicate.
21932 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21933 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21934 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21935 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21936 (mov<mode>_insv_1): Remove expander. Merge insn with
21937 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21938 Use general_x64nomem_operand as operand 1 predicate.
21939 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21940 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21941 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21942 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21943 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21944 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21945 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21946 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21947 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21948 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21949 (general_x64nomem_operand): Ditto.
21950
c6a9ed5a 219512013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH		 21952
21953 * config/rtems.opt: Add -pthread option.
21954
31b0a960
RB		 219552013-03-26 Richard Biener <rguenther@suse.de>
21956
93a4145b
UB		 21957 * alias.c (find_base_term): Avoid redundant and not used recursion.
21958 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB		 21959 (true_dependence_1): Compute and pass base terms to base_alias_check.
21960 (write_dependence_p): Likewise.
21961 (may_alias_p): Likewise.
21962
79517551
SN		 219632013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21964
21965 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21966 PC-relative load in SI modes and above only.
21967
a76213b9
XQ		 219682013-03-26 Xinyu Qi <xyqi@marvell.com>
21969
21970 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB		 21971 * config/arm/iwmmxt.md (WCGR0): Update.
21972 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 21973
37ff9355
UB		 219742013-03-26 Uros Bizjak <ubizjak@gmail.com>
21975
21976 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21977 Use x64 and nox64 isa attributes.
21978
30b0317c
RB		 219792013-03-26 Richard Biener <rguenther@suse.de>
21980
21981 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21982 alignment computations and rely on get_object_alignment_1
21983 for the !TYPE_P case.
21984 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21985
cbcdb140
WL		 219862013-03-26 Walter Lee <walt@tilera.com>
21987
21988 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21989 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21990
dd884525 219912013-03-25 Jeff Law <law@redhat.com>
be672e08 21992
33159866
UB		 21993 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21994 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL		 21995 wrong version of prior patch.
21996
0051d3ec
WL		 219972013-03-25 Walter Lee <walt@tilera.com>
21998
21999 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22000 TILEGX_INSN_SHUFFLEBYTES1.
22001 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22002 shufflebytes1.
22003 (tilegx_builtins): Ditto.
22004 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22005
065a3605
WL		 220062013-03-25 Walter Lee <walt@tilera.com>
22007
22008 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL		 22009 (floatunssisf2): New pattern.
22010 (floatsidf2): New pattern.
22011 (floatunssidf2): New pattern.
065a3605 22012
5b2a3c82
WL		 220132013-03-25 Walter Lee <walt@tilera.com>
22014
22015 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22016 tests for constraint J, K, N, P.
22017
192ea533
WL		 220182013-03-25 Walter Lee <walt@tilera.com>
22019
22020 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22021 Use indirect/pcrel encoding.
22022 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22023 Ditto.
22024
855e0d0b
SE		 220252013-03-25 Steve Ellcey <sellcey@mips.com>
22026
22027 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22028 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22029 * config/mips/mips.c (mips_option_override): Set IMADD default.
22030 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22031 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22032 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22033 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22034 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22035
39e45653
JL		 220362013-03-25 Jeff Law <law@redhat.com>
22037
22038 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22039 slightly to avoid creating and folding useless trees. Simplify
22040 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22041
7482c470
UB		 220422013-03-25 Uros Bizjak <ubizjak@gmail.com>
22043
22044 * config/i386/i386.md (*zero_extendsidi2): Merge with
22045 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22046 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22047 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22048 targets. Clarify comment.
22049
4a53743e
MJ		 220502013-03-25 Martin Jambor <mjambor@suse.cz>
22051
22052 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22053 pass-through jump functions differently.
22054 (ipa_read_jump_function): Likewise. Also use setter functions to set
22055 up jump functions.
22056
162712de
MJ		 220572013-03-25 Martin Jambor <mjambor@suse.cz>
22058
22059 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22060 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22061 process it.
22062 (ipa_get_indirect_edge_target): New function.
22063 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22064 ipa_get_indirect_edge_target. Update all callers.
22065 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22066 ipa_get_indirect_edge_target_1 instead of calling
22067 ipa_get_indirect_edge_target.
22068 (create_specialized_node): Pass aggvlas to
22069 ipcp_discover_new_direct_edges.
22070
a5a4c20a
KT		 220712013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22072
22073 * config/arm/arm.md (f_sels, f_seld): New types.
22074 (*cmov<mode>): New pattern.
22075 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22076 predicate.
22077
f992b9fc
KT		 220782013-03-25 Kai Tietz <ktietz@redhat.com>
22079
22080 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22081 POSIX-printf for mingw-hosted builds.
22082
c16fd676
RB		 220832013-03-25 Richard Biener <rguenther@suse.de>
22084
22085 PR middle-end/56694
22086 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22087 must-not-throw stmt location.
22088
5ad29f12
KT		 220892013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22090
22091 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22092 Emit load-acquire versions when acq is true.
22093 (arm_emit_store_exclusive): Add rel parameter.
22094 Emit store-release versions when rel is true.
22095 (arm_split_compare_and_swap): Use acquire-release instructions
22096 instead.
22097 of barriers when appropriate.
22098 (arm_split_atomic_op): Likewise.
22099 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22100 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22101 (VUNSPEC_SLX): Likewise.
22102 (VUNSPEC_LDA): Likewise.
22103 (VUNSPEC_STL): Likewise.
22104 * config/arm/sync.md (atomic_load<mode>): New pattern.
22105 (atomic_store<mode>): Likewise.
22106 (arm_load_acquire_exclusive<mode>): Likewise.
22107 (arm_load_acquire_exclusivesi): Likewise.
22108 (arm_load_acquire_exclusivedi): Likewise.
22109 (arm_store_release_exclusive<mode>): Likewise.
22110
03a7dddb
CM		 221112013-03-25 Catherine Moore <clm@codesourcery.com>
22112
22113 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22114 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22115 * config/mip/predicates.md (lwsp_swsp_operand,
22116 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22117 sb16_operand, db4_operand, db7_operand, ib3_operand,
22118 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22119 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22120 andi16_operand): New predicates.
22121 * config/mips/mips.md (compression): New attribute.
22122 (enabled): New attribute.
22123 (length): Consider compression in computing length.
22124 (shift_compression): New code attribute.
22125 (*add<mode>3): New operands. Record compression.
22126 (sub<mode>3): Likewise.
22127 (one_cmpl<mode>2): Likewise.
22128 (*and<mode>3): Likewise.
22129 (*ior<mode>3): Likewise.
22130 (unnamed pattern for xor): Likewise.
22131 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22132 (*<optab><mode>3): Likewise.
22133 (*mov<mode>_internal: Likewise.
22134 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22135 (mips_unsigned_immediate_p): New.
22136 (umips_lwsp_swsp_address_p): New.
22137 (m16_based_address_p): New.
22138 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22139 (mips_unsigned_immediate_p): New prototype.
22140 (lwsp_swsp_address_p): New prototype.
22141 (m16_based_address_p): New prototype.
22142 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22143 (mips_signed_immediate_p): New function.
22144 (m16_based_address_p): New function.
22145 (lwsp_swsp_address_p): New function.
22146 (mips_print_operand_punctuation): Recognize short delay slot insns
22147 for microMIPS.add<mode>3"
22148
f35c297f 221492013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 22150
f35c297f
KT		 22151 PR target/56720
22152 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22153 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22154
051b9446
RB		 221552013-03-25 Richard Biener <rguenther@suse.de>
22156
22157 PR tree-optimization/56689
22158 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22159 any edge.
22160
374001cb
RB		 221612013-03-25 Richard Biener <rguenther@suse.de>
22162
22163 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22164 of bitmap.
22165 (memory_references): Likewise.
22166 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22167 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22168 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22169 (gather_mem_refs_in_loops): Fold into ...
22170 (analyze_memory_references): ... this. Move initialization
22171 to tree_ssa_lim_initialize.
22172 (fill_always_executed_in): Rename to ...
22173 (fill_always_executed_in_1): ... this.
22174 (fill_always_executed_in): Move contains_call computation to
22175 this new function from ...
22176 (tree_ssa_lim_initialize): ... here.
22177 (tree_ssa_lim): Call fill_always_executed_in.
22178
57895947
EB		 221792013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22180
22181 * postreload.c (reload_combine): Fix code detecting returns.
22182
1e1b18c1
EB		 221832013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22184
22185 * function.c (emit_use_return_register_into_block): On cc0 targets,
22186 do not emit the sequence between cc0 setter and user.
22187
9216baf1
KT		 221882013-03-25 Kai Tietz <ktietz@redhat.com>
22189
1e1b18c1
EB		 22190 * config/i386/predicates.md (local_symbolic_operand): Interpret
22191 dll-imported symbols as none-local.
9216baf1 22192
76421b44
RB		 221932013-03-25 Richard Biener <rguenther@suse.de>
22194
22195 * tree-ssa-loop-im.c (struct depend): Remove.
22196 (struct lim_aux_data): Make depends a vec of gimples.
22197 (free_lim_aux_data): Adjust.
22198 (add_dependency): Likewise.
22199 (set_level): Likewise.
22200
d154bfa2
RB		 222012013-03-25 Richard Biener <rguenther@suse.de>
22202
22203 PR middle-end/56434
22204 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22205 the pointer returned by calls with ECF_MALLOC set.
22206
2cd9804e
UB		 222072013-03-24 Uros Bizjak <ubizjak@gmail.com>
22208
acef5fe0
UB		 22209 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22210
222112013-03-24 Uros Bizjak <ubizjak@gmail.com>
22212
22213 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB		 22214 using MMXMODE mode iterator.
22215 (*move<mode>_internal): Merge with *movv2sf_internal and
22216 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22217
225ccc68
SB		 222182013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22219
85c0f02d
SB		 22220 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22221 (record_last_mem_set_info): Likewise.
22222
225ccc68
SB		 22223 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22224 of XNEWVEC followed by memset.
22225 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22226
b64925dc
SB		 222272013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22228
22229 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22230 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22231 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22232 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22233 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22234 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22235 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22236 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22237 BARRIER_P instead of GET_CODE.
22238
b9aaf52e
EB		 222392013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22240
22241 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22242 inaccuracy in the probing code.
22243
22244 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22245 (ctrapdi4): Likewise.
22246
66de4d7c
EB		 222472013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22248
22249 * calls.c (expand_call): Add missing guard to code handling return
22250 of non-BLKmode structures in MSB.
22251 * function.c (expand_function_end): Likewise.
22252
4f25c8fb
EB		 222532013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22254
22255 * combine.c (try_combine): Adjust comment. Do not add the set of
22256 insn #0 if the destination indirectly is set or dies in insn #2.
22257 Tidy up code to distribute a new note.
22258
23b7850d
UB		 222592013-03-22 Uros Bizjak <ubizjak@gmail.com>
22260
22261 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22262 also for alternatives 16 and 17.
22263
bed852cf
UB		 222642013-03-22 Uros Bizjak <ubizjak@gmail.com>
22265
22266 * config/i386/sse.md (*mov<mode>_internal): Merge with
22267 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22268 Emit insn template depending on type attribute. Use
22269 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22270 movd instead of movq mnemonic for interunit moves. Rewrite mode
22271 attribute calculation. Remove unit attribute calculation.
22272 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22273 Set prefix_data16 attribute for DImode ssemov types.
22274 Use Ym instead of y for SSE-MMX conversion alternatives.
22275 Reorder operand constraints.
22276
1e8a7937
SB		 222772013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22278
22279 * df.h (df_insn_delete): Adjust prototype.
22280 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22281 and let it decide whether mark the basic block dirty.
22282 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22283 * df-scan.c (df_insn_info_delete): New helper function, split
22284 off from df_insn_delete.
22285 (df_scan_free_bb_info): Use it.
22286 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22287 Likewise.
22288 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22289 that the insn is actually an insn and it has a non-NULL basic block.
22290 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22291
e14d094c
RB		 222922013-03-22 Richard Biener <rguenther@suse.de>
22293
22294 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22295 dep_ref members.
22296 (mem_ref_alloc): Do not allocate them.
22297 (refs_independent_p): Do not query or maintain a cache.
22298
e9cf7316
RB		 222992013-03-22 Richard Biener <rguenther@suse.de>
22300
22301 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22302 (gather_mem_refs_in_loops): Do not compute it.
22303 (analyze_memory_references): Do not allocate it.
22304 (tree_ssa_lim_finalize): Do not free it.
22305 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22306
f046e81b
RB		 223072013-03-22 Richard Biener <rguenther@suse.de>
22308
22309 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 22310
a05c0ddf
IB		 223112013-03-22 Ian Bolton <ian.bolton@arm.com>
22312
22313 * config/aarch64/aarch64.c (aarch64_print_operand): New
22314 format specifier for printing a constant in hex.
22315 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22316 format specifier for printing second operand.
22317
c00217fc
RB		 223182013-03-22 Richard Biener <rguenther@suse.de>
22319
22320 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22321 bitmaps.
22322 (gather_mem_refs_in_loops): Perform store accumulation here.
22323 (create_vop_ref_mapping_loop): Remove.
22324 (create_vop_ref_mapping): Likewise.
22325 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 22326 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB		 22327 (record_indep_loop): Remove.
22328 (record_dep_loop): New function.
22329 (ref_indep_loop_p_1): Adjust to only walk over references
22330 in the loop, not its subloops.
22331 (ref_indep_loop_p): Rename to ...
22332 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22333 maintaining a more fine-grained cache.
22334 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22335 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22336
15d19bf8
RB		 223372013-03-22 Richard Biener <rguenther@suse.de>
22338
22339 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22340 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22341 aggregate mem_ref_loc.
22342 (free_mem_ref_locs): Inline into ...
22343 (memref_free): ... this and adjust.
22344 (mem_ref_alloc): Adjust.
22345 (mem_ref_locs_alloc): Remove.
22346 (record_mem_ref_loc): Adjust.
22347 (get_all_locs_in_loop): Rewrite into ...
22348 (for_all_locs_in_loop): ... this iterator.
22349 (rewrite_mem_ref_loc): New functor.
22350 (rewrite_mem_refs): Use for_all_locs_in_loop.
22351 (sm_set_flag_if_changed): New functor.
22352 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22353 (ref_always_accessed): New functor.
22354 (ref_always_accessed_p): Use for_all_locs_in_loop.
22355
6f37411d
MG		 223562013-03-21 Marc Glisse <marc.glisse@inria.fr>
22357
22358 * tree-pass.h (PROP_gimple_lvec): New.
22359 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22360 (init_optimization_passes): Move pass_lower_vector.
22361 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22362 PROP_gimple_lvec.
22363 (pass_lower_vector): Provide PROP_gimple_lvec.
22364 (pass_lower_vector_ssa): Likewise.
22365 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22366
511ef689
MW		 223672013-03-21 Mark Wielaard <mjw@redhat.com>
22368
22369 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22370
cc1df30b
UB		 223712013-03-21 Uros Bizjak <ubizjak@gmail.com>
22372
22373 * config/i386/i386.md (*movdi_internal): Disparage slightly
22374 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22375 conversion alternatives.
22376
c845cfe1
JJ		 223772013-03-21 Jakub Jelinek <jakub@redhat.com>
22378
37e99116
JJ		 22379 PR middle-end/48087
22380 * diagnostic.def (DK_WERROR): New kind.
22381 * diagnostic.h (werrorcount): Define.
22382 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22383 promoted to DK_ERROR, increment DK_WERROR counter instead of
22384 DK_ERROR counter.
22385 * toplev.c (toplev_main): Call print_ignored_options even if
22386 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22387 even if just werrorcount is non-zero.
22388
c845cfe1
JJ		 22389 PR debug/55608
22390 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22391 on failure.
22392 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22393 (string_cst_pool_decl): New function.
22394 (optimize_one_addr_into_implicit_ptr): New function.
22395 (resolve_addr_in_expr): Optimize DWARF location expression
22396 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22397 which doesn't live in memory, but has DW_AT_location or
22398 DW_AT_const_value, or refers to a string literal, into
22399 DW_OP_GNU_implicit_pointer.
22400 (optimize_location_into_implicit_ptr): New function.
22401 (resolve_addr): If removing DW_AT_location of a variable because
22402 it was DW_OP_addr of address of the variable, but the variable doesn't
22403 live in memory, try to emit const value attribute for the initializer.
22404
08e0cda6
MG		 224052013-03-21 Marc Glisse <marc.glisse@inria.fr>
22406
22407 * tree.h (VECTOR_TYPE_P): New macro.
22408 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22409 TYPE_MODE): Use it.
22410 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22411 VEC_COND_EXPR cannot be lvalues.
22412 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22413
d08633b4
MG		 224142013-03-21 Marc Glisse <marc.glisse@inria.fr>
22415
22416 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22417 Restrict the transformation to equal modes.
22418
e6647190
RB		 224192013-03-21 Richard Biener <rguenther@suse.de>
22420
22421 PR tree-optimization/39326
22422 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22423 (MEM_ANALYZABLE): Adjust.
22424 (record_mem_ref_loc): Move bitmap ops ...
22425 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22426 unanalyzable refs, do not record locations for it.
22427 (analyze_memory_references): Allocate ref zero as shared
22428 unanalyzable ref.
22429 (refs_independent_p): Do not test for unanalyzed mems here.
22430 (ref_indep_loop_p_1): Special-case disambiguation against
22431 the unanalyzed ref.
cc1df30b 22432 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 22433
65074f54
CL		 224342013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22435
22436 * config/arm/arm-protos.h (tune_params): Add
22437 prefer_neon_for_64bits field.
22438 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22439 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22440 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22441 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22442 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22443 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22444 (arm_option_override): Handle -mneon-for-64bits new option.
22445 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22446 (prefer_neon_for_64bits): Declare new variable.
22447 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 22448 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL		 22449 (arch_enabled): Handle new arch types. Remove support for onlya8
22450 and nota8.
22451 (one_cmpldi2): Use new arch names.
bc5faa5b 22452 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL		 22453 * config/arm/arm.opt (mneon-for-64bits): Add option.
22454 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22455 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22456 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22457 of onlya8.
22458 * doc/invoke.texi (-mneon-for-64bits): Document.
22459
5a2d2a79
RB		 224602013-03-21 Richard Biener <rguenther@suse.de>
22461
22462 PR tree-optimization/39326
22463 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22464 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 22465 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 22466
5abe1e05
RB		 224672013-03-21 Richard Biener <rguenther@suse.de>
22468
22469 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22470 (vect_insert_into_interleaving_chain): Likewise.
22471 (vect_drs_dependent_in_basic_block): Inline ...
22472 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22473 split out from ...
22474 (vect_analyze_data_ref_dependence): ... here. Simplify.
22475 (vect_check_interleaving): Simplify.
22476 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22477 (vect_slp_analyze_data_ref_dependences): ... this new function.
22478 (dr_group_sort_cmp): New function.
22479 (vect_analyze_data_ref_accesses): Compute data-reference groups
22480 here instead of in vect_analyze_data_ref_dependence. Use
22481 a more efficient algorithm.
22482 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22483 vect_slp_analyze_data_ref_dependences. Call
22484 vect_analyze_data_ref_accesses earlier.
22485 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22486 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22487 (vect_slp_analyze_data_ref_dependences): New prototype.
22488
cad1735b
RB		 224892013-03-21 Richard Biener <rguenther@suse.de>
22490
22491 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22492 ref is stored in the loop.
22493 (find_refs_for_sm): Walk only over all stores.
22494 (store_motion_loop): Allocate from lim_bitmap_obstack.
22495 (store_motion): Likewise.
22496
141310ef
RB		 224972013-03-21 Richard Biener <rguenther@suse.de>
22498
22499 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22500 Update virtual SSA form.
22501
5022315a
RO		 225022013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22503
22504 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22505 * configure: Regenerate.
22506 * config.in: Regenerate.
22507 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22508 if !HAVE_LD_EH_FRAME_CIEV3.
22509
50d4421c
RB		 225102013-03-21 Richard Biener <rguenther@suse.de>
22511
22512 * tree-cfg.c (verify_expr_no_block): New function.
22513 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22514 nor DECL_VALUE_EXPR have locations with associated blocks.
22515 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22516 (clear_unused_block_pointer): Remove code dealing with
22517 blocks in DECL_DEBUG_EXPR locations.
22518
839b422f
RB		 225192013-03-21 Richard Biener <rguenther@suse.de>
22520
22521 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22522 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22523 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22524 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22525 instead of DECL_DEBUG_EXPR_IS_FROM.
22526 * gimplify.c (gimplify_modify_expr): Likewise.
22527 * tree-cfg.c (verify_expr_location_1): Likewise.
22528 * tree-complex.c (create_one_component_var): Likewise.
22529 * tree-sra.c (create_access_replacement): Likewise.
22530 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22531 (clear_unused_block_pointer): Likewise.
22532 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22533 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22534 * var-tracking.c (var_debug_decl): Likewise.
22535 (track_expr_p): Likewise.
22536 * tree-inline.c (add_local_variables): Likewise. Set
22537 DECL_HAS_DEBUG_EXPR_P after copying it.
22538 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22539 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22540
fe04878d
UB		 225412013-03-21 Uros Bizjak <ubizjak@gmail.com>
22542
22543 PR bootstrap/56656
22544 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22545 * configure: Regenerate.
22546 * config.in: Regenerate.
22547 * config/i386/i386.md (*movdf_internal): Use
22548 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22549 movd instead of movq mnemonic for interunit moves.
22550 (*movdi_internal): Ditto.
22551
bd059b26 225522013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N		 22553
22554 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22555 (abd<mode>_3): New pattern.
22556 (aba<mode>_3): New pattern.
22557 (fabd<mode>_3): New pattern.
22558
bd059b26 225592013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N		 22560
22561 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22562 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22563 occurrence of REGISTER_PREFIX as its empty string.
22564
4f1f78b9
JL		 225652013-03-20 Jeff Law <law@redhat.com>
22566
22567 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22568 addititional equivalences for equality comparisons between an SSA_NAME
22569 and a constant where the SSA_NAME was set from a widening conversion.
22570
327a1118
WL		 225712013-03-20 Walter Lee <walt@tilera.com>
22572
22573 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22574
813e0036
UB		 225752013-03-20 Uros Bizjak <ubizjak@gmail.com>
22576
22577 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22578 depending on type attribute.
22579 (*movti_internal): Ditto.
22580 (*movtf_internal): Ditto.
22581 (*movxf_internal): Ditto.
22582 (*movdf_internal): Ditto.
22583 (*movsf_internal): Ditto.
22584
7cf34aae
UB		 225852013-03-20 Uros Bizjak <ubizjak@gmail.com>
22586
22587 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22588 maybe_vex for sselog1 and ssemov types.
22589 (*movdi_internal): Reorder operand constraints.
22590 (*movsi_internal): Ditto. Set prefix attribute to
22591 maybe_vex for sselog1 and ssemov types.
22592 (*movtf_internal): Set prefix attribute to maybe_vex
22593 for sselog1 and ssemov types.
22594 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22595 DImode ssemov types. Reorder operand constraints.
22596 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22597 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22598 attribute for SImode ssemov types. Reorder operand constraints.
22599
19321415
MJ		 226002013-03-20 Martin Jambor <mjambor@suse.cz>
22601
22602 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22603 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22604
12211b99 226052013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM		 22606
22607 * config/rs6000/predicates.md (indexed_address, update_address_mem
22608 update_indexed_address_mem): New predicates.
22609 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22610 attribute for load/store instructions.
22611 * config/rs6000/dfp.md (movsd_store): Likewise.
22612 (movsd_load): Likewise.
22613 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22614 (unnamed HI->DI extend define_insn): Likewise.
22615 (unnamed SI->DI extend define_insn): Likewise.
22616 (unnamed QI->SI extend define_insn): Likewise.
22617 (unnamed QI->HI extend define_insn): Likewise.
22618 (unnamed HI->SI extend define_insn): Likewise.
22619 (unnamed HI->SI extend define_insn): Likewise.
22620 (extendsfdf2_fpr): Likewise.
22621 (movsi_internal1): Likewise.
22622 (movsi_internal1_single): Likewise.
22623 (movhi_internal): Likewise.
22624 (movqi_internal): Likewise.
22625 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22626 attribute for load/store instructions.
22627 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22628 instructions.
22629 (mov<mode>_softfloat): Likewise.
22630 (mov<mode>_hardfloat32): Likewise.
22631 (mov<mode>_hardfloat64): Likewise.
22632 (mov<mode>_softfloat64): Likewise.
22633 (movdi_internal32): Likewise.
22634 (movdi_internal64): Likewise.
22635 (probe_stack_<mode>): Likewise.
22636
226372013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22638
22639 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22640 floating point, and decimal floating point to reload iterator.
22641
22642 * config/rs6000/constraints.md (wl constraint): New constraints to
22643 return FLOAT_REGS if certain options are used to reduce the number
22644 of separate patterns that exist in the file.
22645 (wx constraint): Likewise.
22646 (wz constraint): Likewise.
22647
22648 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22649 -mdebug=reg, print wg, wl, wx, and wz constraints.
22650 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22651 Initialize the reload functions for 64-bit binary/decimal floating
22652 point types.
22653 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22654 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22655 create the buffer on the stack to overcome not having a 32-bit
22656 load and store.
22657 (rs6000_emit_move): Likewise.
22658 (rs6000_secondary_memory_needed_rtx): Likewise.
22659 (rs6000_alloc_sdmode_stack_slot): Likewise.
22660 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22661 via xxlxor, just like DFmode 0.0.
22662
22663 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22664 define as 1 if we are running on a power7 or newer.
22665 (enum r6000_reg_class_enum): Add new constraints.
22666
22667 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22668 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22669 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 22670 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM		 22671 (movsd splitter): Likewise.
22672 (movsd_hardfloat): Likewise.
22673 (movsd_softfloat): Likewise.
22674
22675 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22676 binary and decimal floating point moves.
22677 (fmove_ok): New attributes to combine binary and decimal floating
22678 point moves, and to combine power6x (mfpgpr) moves along normal
22679 floating moves.
22680 (real_value_to_target): Likewise.
22681 (f32_lr): Likewise.
22682 (f32_lm): Likewise.
22683 (f32_li): Likewise.
22684 (f32_sr): Likewise.
22685 (f32_sm): Likewise.
22686 (f32_si): Likewise.
22687 (movsf): Combine binary and decimal floating point moves. Combine
22688 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 22689 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM		 22690 (mov<mode> for SFmode/SDmode); Likewise.
22691 (SFmode/SDmode splitters): Likewise.
22692 (movsf_hardfloat): Likewise.
22693 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22694 (movsf_softfloat): Likewise.
22695 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22696
22697 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22698 wx and wz constraints.
22699
22700 * config/rs6000/constraints.md (wg constraint): New constraint to
22701 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22702
22703 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22704 constraint.
22705
22706 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22707 -mdebug=reg, print wg, wl, wx, and wz constraints.
22708 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22709 Initialize the reload functions for 64-bit binary/decimal floating
22710 point types.
22711 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22712 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22713 create the buffer on the stack to overcome not having a 32-bit
22714 load and store.
22715 (rs6000_emit_move): Likewise.
22716 (rs6000_secondary_memory_needed_rtx): Likewise.
22717 (rs6000_alloc_sdmode_stack_slot): Likewise.
22718 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22719 via xxlxor, just like DFmode 0.0.
22720
c6d5ff83
MM		 22721 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22722 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22723 with other moves by using conditional constraits (wg). Use LFIWZX
22724 and STFIWX for loading SDmode on power7.
22725 (movdd splitters): Likewise.
22726 (movdd_hardfloat32): Likewise.
22727 (movdd_softfloat32): Likewise.
22728 (movdd_hardfloat64_mfpgpr): Likewise.
22729 (movdd_hardfloat64): Likewise.
22730 (movdd_softfloat64): Likewise.
22731
22732 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22733 64-bit binary and decimal floating point moves.
22734 (FMOVE64X): Likewise.
22735 (movdf): Combine 64-bit binary and decimal floating point moves.
22736 Combine power6x (mfpgpr) moves with other moves by using
22737 conditional constraits (wg).
22738 (mov<mode> for DFmode/DDmode): Likewise.
22739 (DFmode/DDmode splitters): Likewise.
22740 (movdf_hardfloat32): Likewise.
22741 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22742 (movdf_softfloat32): Likewise.
22743 (movdf_hardfloat64_mfpgpr): Likewise.
22744 (movdf_hardfloat64): Likewise.
22745 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22746 (movdf_softfloat64): Likewise.
22747 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22748 (reload_<mode>_load): Move to later in the file so they aren't in
22749 the middle of the floating point move insns.
22750 (reload_<mode>_store): Likewise.
22751
22752 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22753 constraint.
22754
22755 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22756 constraint if -mdebug=reg.
bd059b26
UB		 22757 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22758 Enable using dd reload support if needed.
c6d5ff83
MM		 22759
22760 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22761 binary and decimal floating point moves in rs6000.md.
22762 (movtd_internal): Likewise.
22763
22764 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22765 decimal floating point moves.
22766 (movtf): Likewise.
22767 (movtf_internal): Likewise.
22768 (mov<mode>_internal, TDmode/TFmode): Likewise.
22769 (movtf_softfloat): Likewise.
22770 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22771
22772 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22773 movdi_internal64, using wg constraint for move direct operations.
22774 (movdi_internal64): Likewise.
22775
22776 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22777 MODES_TIEABLE_P for selected modes. Print the numerical value of
22778 the various virtual registers. Use GPR/FPR first/last values,
22779 instead of hard coding the register numbers. Print which modes
22780 have reload functions registered.
bd059b26
UB		 22781 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22782 settings before/after setting cpu, target and subtarget settings.
22783 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22784 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM		 22785 (rs6000_secondary_reload_fail): Likewise.
22786 (rs6000_secondary_reload_inner): Likewise.
22787
22788 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22789 macros for first/last GPR and FPR registers.
22790 (LAST_GPR_REGNO): Likewise.
22791 (FIRST_FPR_REGNO): Likewise.
22792 (LAST_FPR_REGNO): Likewise.
22793
22794 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22795 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22796 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22797 (vcond<mode><mode>): Likewise.
22798 (vcondu<mode><mode>): Likewise.
22799 (vector_gtu<mode>): Likewise.
22800 (vector_gte<mode>): Likewise.
22801 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 22802 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM		 22803 (ior<mode>3): Likewise.
22804 (and<mode>3): Likewise.
22805 (one_cmpl<mode>2): Likewise.
22806 (nor<mode>3): Likewise.
22807 (andc<mode>3): Likewise.
22808
22809 * config/rs6000/constraints.md (wt constraint): New constraint
22810 that returns VSX_REGS if TImode is allowed in VSX registers.
22811
22812 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22813 constant under VSX.
22814
22815 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22816 similar to TImode, but it is restricted to being in the GPRs.
22817
22818 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22819 TImode to occupy a single VSX register.
22820
22821 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22822 -mvsx-timode for power7/power8.
22823 (power7 cpu): Likewise.
22824 (power8 cpu): Likewise.
22825
22826 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22827 sure that TFmode/TDmode take up two registers if they are ever
22828 allowed in the upper VSX registers.
22829 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22830 registers.
22831 (rs6000_init_hard_regno_mode_ok): Likewise.
22832 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22833 constraint. Print if LRA is turned on.
22834 (rs6000_option_override_internal): Give an error if -mvsx-timode
22835 and VSX is not enabled.
22836 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22837 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22838 to reg+offset addressing. Use PTImode when checking offset
22839 addresses for validity.
22840 (reg_offset_addressing_ok_p): Likewise.
22841 (rs6000_legitimate_offset_address_p): Likewise.
22842 (rs6000_legitimize_address): Likewise.
22843 (rs6000_legitimize_reload_address): Likewise.
22844 (rs6000_legitimate_address_p): Likewise.
22845 (rs6000_eliminate_indexed_memrefs): Likewise.
22846 (rs6000_emit_move): Likewise.
22847 (rs6000_secondary_reload): Likewise.
22848 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22849 reloads to fpr registers to continue to use reg+offset addressing,
22850 but 64-bit reloads to altivec registers need reg+reg addressing.
22851 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22852 it. Treat LO_SUM like a PLUS operation.
22853 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 22854 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM		 22855 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22856 registers to share a register with a smaller sized type, since VSX
22857 puts scalars in the upper 64-bits.
22858 (print_operand): Add support for PTImode.
22859 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22860 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22861 registers, but don't have arithmetic support.
22862 (rs6000_memory_move_cost): Add test for VSX.
22863 (rs6000_opt_masks): Add -mvsx-timode.
22864
22865 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22866 for TImode.
22867 (VSs): Likewise.
22868 (VSr): Use wt constraint for TImode.
22869 (VSv): Drop TImode support.
22870 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22871 (vsx_movti_64bit): Likewise.
22872 (vsx_movti_32bit): Likewise.
22873 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22874 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22875 one '?' on the appropriate output constraint. Do not allow TImode
22876 logical operations on 32-bit systems.
22877 (vsx_ior<mode>3): Likewise.
22878 (vsx_xor<mode>3): Likewise.
22879 (vsx_one_cmpl<mode>2): Likewise.
22880 (vsx_nor<mode>3): Likewise.
22881 (vsx_andc<mode>3): Likewise.
22882 (vsx_concat_<mode>): Likewise.
22883 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22884
22885 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22886 OPTION_MASK_VSX_TIMODE.
22887 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22888 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22889
22890 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22891 (TI2 iterator): New iterator for TImode, PTImode.
22892 (wd mode attribute): Add values for vector types.
bd059b26
UB		 22893 (movti_string): Replace TI move operations with operations for TImode
22894 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM		 22895 (mov<mode>_string, TImode/PTImode): Likewise.
22896 (movti_ppc64): Likewise.
22897 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22898 (TI mode splitters): Likewise.
22899
22900 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22901 constraint.
22902
1fc5eced
MG		 229032013-03-20 Marc Glisse <marc.glisse@inria.fr>
22904
22905 PR tree-optimization/56355
22906 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22907 Also handle integers with undefined overflow.
22908
22c4c869
CM		 229092013-03-20 Catherine Moore <clm@codesourcery.com>
22910 Maciej W. Rozycki <macro@codesourcery.com>
22911 Tom de Vries <tom@codesourcery.com>
12211b99 22912 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM		 22913 Iain Sandoe <iain@codesourcery.com>
22914 Nathan Froyd <froydnj@codesourcery.com>
12211b99 22915 Chao-ying Fu <fu@mips.com>
22c4c869
CM		 22916
22917 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 22918 Document new function attributes.
22c4c869
CM		 22919 * doc/invoke.texi (minterlink-compressed, mmicromips,
22920 m14k, m14ke, m14kec): Document new options.
22921 (minterlink-mips16): Update documentation.
22922 * doc/md.texi (ZC, ZD): Document new constraints.
22923 * configure.ac (gcc_cv_as_micromips): Check if linker
22924 supports the .set micromips directive.
22925 * configure: Regenerate.
22926 * config.in: Regenerate.
22927 * config/mips/mips-tables.opt: Regenerate.
22928 * config/mips/micromips.md: New file.
22929 * constraints.md (ZC, ZD): New constraints.
22930 * config/mips/predicates.md (movep_src_register): New predicate.
22931 (movep_src_operand): New predicate.
22932 (non_volatile_mem_operand): New predicate.
22933 * config/mips/mips.md (multimem): New type.
22934 (length): Differentiate between 17-bit and 18-bit branch offsets.
22935 (MOVEP1, MOVEP2): New mode iterator.
33159866 22936 (mov_<load>l): Use ZC constraint.
22c4c869
CM		 22937 (mov_<load>r): Likewise.
22938 (mov_<store>l): Likewise.
22939 (mov_<store>r): Likewise.
22940 (*branch_equality<mode>_inverted): Add microMIPS support.
22941 (*branch_equality<mode>): Likewise.
22942 (*jump_absolute): Likewise.
22943 (indirect_jump_<mode>): Likewise.
22944 (tablejump_<mode>): Likewise.
22945 (<optab>_internal): Likewise.
22946 (sibcall_internal): Likewise.
22947 (sibcall_value_internal): Likewise.
22948 (prefetch): Use constraint ZD.
22949 * config/mips/mips.opt (minterlink-compressed): New option.
22950 (minterlink-mips16): Now an alias for minterlink-compressed.
22951 (mmicromips): New option.
22952 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22953 (compare_and_swap_12): Likewise.
22954 (sync_add<mode>): Likewise.
22955 (sync_<optab>_12): Likewise.
22956 (sync_old_<optab>_12): Likewise.
22957 (sync_new_<optab>_12): Likewise.
22958 (sync_nand_12): Likewise.
22959 (sync_old_nand_12): Likewise.
22960 (sync_new_nand_12): Likewise.
22961 (sync_sub<mode>): Likewise.
22962 (sync_old_add<mode>): Likewise.
22963 (sync_old_sub<mode>): Likewise.
22964 (sync_new_add<mode>): Likewise.
22965 (sync_new_sub<mode>): Likewise.
22966 (sync_<optab><mode>): Likewise.
22967 (sync_old_<optab><mode>): Likewise.
22968 (sync_new_<optab><mode>): Likewise.
22969 (sync_nand<mode>): Likewise.
22970 (sync_old_nand<mode>): Likewise.
22971 (sync_new_nand<mode>): Likewise.
22972 (sync_lock_test_and_set<mode>): Likewise.
22973 (test_and_set_12): Likewise.
22974 (atomic_compare_and_swap<mode>): Likewise.
22975 (atomic_exchange<mode>_llsc): Likewise.
22976 (atomic_fetch_add<mode>_llsc): Likewise.
22977 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22978 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22979 (umips_save_restore_pattern_p): Likewise.
22980 (umips_load_store_pair_p): Likewise.
22981 (umips_output_load_store_pair): Likewise.
22982 (umips_movep_target_p): Likewise.
22983 (umips_12bit_offset_address_p): Likewise.
22984 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22985 (mips_base_mips16): Rename this...
22986 (mips_base_compression_flags): ...to this. Update all uses.
22987 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22988 (mips_mips16_decl_p): Delete.
22989 (mips_nomips16_decl_p): Delete.
33159866 22990 (mips_get_compress_on_flags): New function.
22c4c869
CM		 22991 (mips_get_compress_off_flags): New function.
22992 (mips_get_compress_mode): New function.
22993 (mips_get_compress_on_name): New function.
22994 (mips_get_compress_off_name): New function.
22995 (mips_insert_attributes): Support multiple compression types.
22996 (mips_merge_decl_attributes): Likewise.
22997 (umips_12bit_offset_address_p): New function.
22998 (mips_start_function_definition): Emit .set micromips directive.
22999 (mips_call_may_need_jalx_p): New function.
23000 (mips_function_ok_for_sibcall): Add microMIPS support.
23001 (mips_print_operand_punctuation): Support short delay slots and
23002 compact jumps.
23003 (umips_swm_mask, umips_swm_encoding): New.
23004 (umips_build_save_restore): New function.
23005 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23006 (was_mips16_p): Remove.
23007 (old_compression_mode): New.
23008 (mips_set_compression_mode): New function.
23009 (mips_set_current_function): Add microMIPS support.
23010 (mips_option_override): Likewise.
23011 (umips_save_restore_pattern_p): New function.
23012 (umips_output_save_restore): New function.
23013 (umips_load_store_pair_p_1): New function.
23014 (umips_load_store_pair_p): New function.
23015 (umips_output_load_store_pair_1): New function.
23016 (umips_output_load_store_pair): New function.
23017 (umips_movep_target_p) New function.
23018 (mips_prepare_pch_save): Add microMIPS support.
23019 * config/mips/mips.h (TARGET_COMPRESSION): New.
23020 (TARGET_CPU_CPP_BUILTINS): Update macro
23021 to use new compression flags and to support microMIPS.
23022 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23023 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23024 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23025 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23026 (ASM_SPEC): Support mmicromips and mno-micromips.
23027 (M16STORE_REG_P): New macro.
23028 (MIPS_CALL): Support TARGET_MICROMIPS.
23029 (MICROMIPS_J): New macro.
23030 (mips_base_mips16): Rename this...
23031 (mips_base_compression_flags): ...to this.
23032 (UMIPS_12BIT_OFFSET_P): New macro.
23033 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23034 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB		 230352013-03-20 Richard Biener <rguenther@suse.de>
23036
23037 PR tree-optimization/56661
23038 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23039 the result does not have to be distinct.
23040
54714c68
RB		 230412013-03-20 Richard Biener <rguenther@suse.de>
23042
23043 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23044 remap_gimple_op_r.
23045
cca1130d 230462013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 23047 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS		 23048
23049 PR rtl-optimization/56605
23050 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23051
2a930377
UB		 230522013-03-20 Uros Bizjak <ubizjak@gmail.com>
23053
23054 PR bootstrap/56656
23055 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23056 that require movd instead of movq.
23057
d6d305fe
RB		 230582013-03-20 Richard Biener <rguenther@suse.de>
23059
23060 * tree-ssa-structalias.c (struct variable_info): Add pointer
23061 to the first field of an aggregate with sub-vars. Make
23062 this and the pointer to the next subfield its ID.
23063 (vi_next): New function.
23064 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23065 storedanything_id, integer_id): Increment by one.
23066 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23067 get_call_clobber_vi): Adjust.
23068 (solution_set_expand): Simplify and speedup.
23069 (solution_set_add): Inline into ...
23070 (set_union_with_increment): ... this. Adjust accordingly.
23071 (do_sd_constraint): Likewise.
23072 (do_ds_constraint): Likewise.
23073 (do_complex_constraint): Simplify.
23074 (build_pred_graph): Adjust.
23075 (solve_graph): Likewise. Simplify and speedup.
23076 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23077 get_constraint_for_component_ref, get_constraint_for_1,
23078 first_vi_for_offset, first_or_preceding_vi_for_offset,
23079 create_function_info_for, create_variable_info_for_1,
23080 create_variable_info_for, intra_create_variable_infos): Adjust.
23081 (init_base_vars): Push NULL for ID zero.
23082 (compute_points_to_sets): Adjust.
23083
a271b42d
RB		 230842013-03-20 Richard Biener <rguenther@suse.de>
23085
23086 * cfgloop.c (verify_loop_structure): Streamline and avoid
23087 ICEing on corrupt loop tree.
23088 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23089 loop tree.
23090
ebd65954
RB		 230912013-03-20 Richard Biener <rguenther@suse.de>
23092
23093 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23094 check whether an SSA update is needed.
23095
4547b7ee
RS		 230962013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23097
23098 * config/mips/constraints.md (T): Rename to...
23099 (Yf): ...this.
23100 (U): Rename to...
23101 (Yd): ...this.
23102 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23103 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23104
0b8cdc58
IB		 231052013-03-19 Ian Bolton <ian.bolton@arm.com>
23106
23107 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23108 (*subsi3_carryin_uxtw): Likewise.
23109
5977a10d
IB		 231102013-03-19 Ian Bolton <ian.bolton@arm.com>
23111
23112 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23113 (*rorsi3_insn_uxtw): Likewise.
23114
bd83ff2c
IB		 231152013-03-19 Ian Bolton <ian.bolton@arm.com>
23116
23117 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23118 (*extrsi5_insn_uxtw): Likewise.
23119
ba65123c
RB		 231202013-03-19 Richard Biener <rguenther@suse.de>
23121
23122 PR tree-optimization/56273
23123 * passes.c (init_optimization_passes): Move second VRP after DOM.
23124
2eac0476
UB		 231252013-03-19 Uros Bizjak <ubizjak@gmail.com>
23126
23127 * config/i386/i386.md (*movti_internal): Merge from
23128 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23129 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23130 nox64 isa attributes.
23131
6e55eda7
RB		 231322013-03-18 Richard Biener <rguenther@suse.de>
23133
23134 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23135 (unite): Likewise.
23136 (merge_node_constraints): Likewise.
23137 (build_succ_graph): Likewise.
23138 (valid_graph_edge): Inline into single caller.
23139 (unify_nodes): Likewise. Use bitmap_set_bit return value
23140 and cache varinfo.
23141 (scc_visit): Fix formatting and variable use.
23142 (do_sd_constraint): Use gcc_checking_assert.
23143 (do_ds_constraint): Likewise.
23144 (do_complex_constraint): Likewise.
23145 (condense_visit): Likewise. Cleanup.
23146 (dump_pred_graph): New function.
23147 (perform_var_substitution): Dump the pred-graph before
23148 variable substitution.
23149 (find_equivalent_node): Use gcc_checking_assert.
23150 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23151
4bdd44c4
RB		 231522013-03-18 Richard Biener <rguenther@suse.de>
23153
23154 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23155 Remove cond_expr_stmt_list argument and do not gimplify the
23156 built expression.
23157 (vect_loop_versioning): Adjust.
23158 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23159 Cleanup to use less temporaries.
23160 (vect_create_data_ref_ptr): Cleanup.
23161
38c56a5b
JJ		 231622013-03-18 Jakub Jelinek <jakub@redhat.com>
23163
23164 PR tree-optimization/56635
23165 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23166 require types_compatible_p types.
23167
20b2e6a0
NC		 231682013-03-18 Nick Clifton <nickc@redhat.com>
23169
a6178a25
NC		 23170 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23171 spurious backslash.
23172
20b2e6a0
NC		 23173 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23174 Add missing line to comment describing function.
23175
92608d0e
RB		 231762013-03-18 Richard Biener <rguenther@suse.de>
23177
23178 PR tree-optimization/56210
23179 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23180 Handle string / character search functions.
23181 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23182
31348d52
RB		 231832013-03-18 Richard Biener <rguenther@suse.de>
23184
23185 PR middle-end/56483
23186 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23187 and implement properly.
23188 * gimple.h (gimple_cond_single_var_p): Remove.
23189
fcac74a1
RB		 231902013-03-18 Richard Biener <rguenther@suse.de>
23191
23192 * tree-data-ref.h (find_data_references_in_loop): Declare.
23193 * tree-data-ref.c (get_references_in_stmt): Use a stack
23194 vector pre-allocated in the callers.
23195 (find_data_references_in_stmt): Adjust.
23196 (graphite_find_data_references_in_stmt): Likewise.
23197 (create_rdg_vertices): Likewise.
23198 (find_data_references_in_loop): Export.
23199 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23200 Compute dependences here...
23201 (vect_analyze_data_refs): ...not here. When we encounter
23202 a non-vectorizable data reference in basic-block vectorization
23203 truncate the data reference vector. Do not bother to
23204 fixup data-dependence information for gather loads.
23205 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23206 of data references, as reported.
23207
0d5a1b56
RB		 232082013-03-18 Richard Biener <rguenther@suse.de>
23209
23210 PR tree-optimization/3713
23211 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23212 has_constants and expr.
23213 (stmt_has_constants): Properly valueize SSA names when deciding
23214 whether the stmt has constants.
23215
789c34e3
RB		 232162013-03-18 Richard Biener <rguenther@suse.de>
23217
23218 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23219 whole function when there is nothing to do.
23220 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23221 * tree-vectorizer.c (vectorize_loops): Update virtual and
23222 loop-closed SSA once.
23223 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23224
076b4605
RB		 232252013-03-18 Richard Biener <rguenther@suse.de>
23226
23227 PR middle-end/56113
23228 * domwalk.c (bb_postorder): New global static.
23229 (cmp_bb_postorder): New function.
23230 (walk_dominator_tree): Replace scheme imposing an order for
23231 visiting dominator sons by one sorting them at the time they
23232 are pushed on the stack.
23233
bdb01696
RB		 232342013-03-18 Richard Biener <rguenther@suse.de>
23235
23236 PR tree-optimization/39326
23237 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23238 (struct mem_ref): Replace mem member with ao_ref typed member.
23239 (MEM_ANALYZABLE): Adjust.
23240 (memref_eq): Likewise.
23241 (mem_ref_alloc): Likewise.
23242 (gather_mem_refs_stmt): Likewise.
23243 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23244 (execute_sm_if_changed_flag_set): Adjust.
23245 (execute_sm): Likewise.
23246 (ref_always_accessed_p): Likewise.
23247 (refs_independent_p): Likewise.
23248 (can_sm_ref_p): Likewise.
23249
12d2dc5e
JJ		 232502013-03-18 Jakub Jelinek <jakub@redhat.com>
23251
23252 PR c/56566
23253 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23254 return 1 even for !unsignedp.
23255
286fb677
UB		 232562013-03-17 Uros Bizjak <ubizjak@gmail.com>
23257
23258 * config/i386/i386.md (isa): Add x64 and nox64.
23259 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23260 (*pushtf): Enable *roF alternative for x64 isa only.
23261 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23262 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23263 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23264 (*movtf_internal): Merge from *movtf_internal_rex64 and
23265 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23266 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23267 nox64 isa attributes.
23268 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23269 nox64 isa attributes.
23270 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23271
88b97037
UB		 232722013-03-17 Uros Bizjak <ubizjak@gmail.com>
23273
23274 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23275
9f4f1735
JJ		 232762013-03-17 Jakub Jelinek <jakub@redhat.com>
23277
23278 PR target/56640
23279 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23280 class names. Remove trailing comma after "ALL_REGS".
23281
48f4a6fa
JH		 232822013-03-16 Jan Hubicka <jh@suse.cz>
23283
23284 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23285 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23286 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23287 of cgraph_get_create_node.
23288 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23289
98e81123
JM		 232902013-03-16 Jason Merrill <jason@redhat.com>
23291
23292 PR debug/49090
23293 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23294 with DW_AT_default_value.
23295
8a298c04
JJ		 232962013-03-16 Jakub Jelinek <jakub@redhat.com>
23297
23298 * BASE-VER: Set to 4.9.0.
23299
4323afa0
AK		 233002013-03-14 Andi Kleen <ak@linux.intel.com>
23301
23302 PR target/56619
23303 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23304 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23305 Document _x* TSX intrinsics.
23306
b3c0d469
JJ		 233072013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23308 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI		 23309
23310 * configure.ac: Add MicroBlaze TLS support detection.
23311 * configure: Regenerate.
88b97037
UB		 23312 * config/microblaze/microblaze-protos.h
23313 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23314 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23315 * config/microblaze/microblaze.c (microblaze_address_type): Add
23316 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI		 23317 (microblaze_address_info): Add tls_reloc.
23318 (TARGET_HAVE_TLS): Define.
23319 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23320 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB		 23321 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23322 load_tls_operand, microblaze_call_tls_get_addr,
23323 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI		 23324 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23325 (get_base_reg): Use microblaze_tls_symbol_p.
23326 (microblaze_classify_address): Handle TLS.
88b97037
UB		 23327 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23328 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI		 23329 (microblaze_legitimize_address): Handle TLS.
23330 (microblaze_address_insns): Handle ADDRESS_TLS.
23331 (pic_address_needs_scratch): Handle TLS.
23332 (print_operand_address): Handle TLS.
23333 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23334 (microblaze_expand_move): Handle TLS.
88b97037
UB		 23335 (microblaze_legitimate_constant_p): Check
23336 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI		 23337 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23338 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23339 (PIC_OFFSET_TABLE_REGNUM): Set.
23340 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23341 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23342 (addsi3, movsi_internal2, movdf_internal): Update constraints
23343 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB		 23344 (move_operand): Redefine as move_src_operand,
23345 check microblaze_tls_referenced_p.
8cc9a5a5 23346
d803a491
IB		 233472013-03-14 Ian Bolton <ian.bolton@arm.com>
23348
23349 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23350 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23351
a8504f22
IB		 233522013-03-14 Ian Bolton <ian.bolton@arm.com>
23353
23354 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23355 CC mode for AND.
23356
df2dfaea
JJ		 233572013-03-14 Jakub Jelinek <jakub@redhat.com>
23358
fbd28bc3
JJ		 23359 PR tree-optimization/53265
23360 * common.opt (Waggressive-loop-optimizations): New option.
23361 * tree-ssa-loop-niter.c: Include tree-pass.h.
23362 (do_warn_aggressive_loop_optimizations): New function.
23363 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23364 if number_of_latch_executions returned constant.
23365 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23366 early. If number_of_latch_executions returned constant, set
23367 nb_iterations_upper_bound back to it.
23368 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23369 field.
23370 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23371 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23372
df2dfaea
JJ		 23373 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23374 (MULTILIB_OSDIRNAMES): Set.
23375 * genmultilib: If defaultosdirname doesn't start with :: , set
23376 defaultosdirname2 instead, clear it and emit two . multilib_raw
23377 entries instead of just one.
23378
ee0d2b68
KK		 233792013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23380
23381 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23382 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23383 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23384 (SUBTARGET_OVERRIDE_OPTIONS): New.
23385
decc676e
OE		 233862013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23387
23388 PR target/49880
23389 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23390 (musermode): Convert to Var(TARGET_USERMODE).
23391 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23392 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23393 * config/sh/sh.c (sh_option_override): Use
23394 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23395 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23396 condition.
23397 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23398 TARGET_SH4.
23399 (udivsi3_i4_single, divsi3_i4_single): Use
23400 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23401
f4b719c7
DK		 234022013-03-13 Dave Korn <dave.korn.cygwin@....>
23403
23404 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23405 default setting.
23406
c40eced0
RB		 234072013-03-13 Richard Biener <rguenther@suse.de>
23408
23409 PR tree-optimization/56608
23410 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23411 calls when vectorizing basic-blocks.
23412
1bfa5973
JJ		 234132013-03-13 Jakub Jelinek <jakub@redhat.com>
23414
23415 PR plugins/45078
23416 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23417 tm_file.
23418
6fcf5434
JJ		 234192013-03-12 Jakub Jelinek <jakub@redhat.com>
23420
23421 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23422
4f38fa8c
JH		 234232013-03-11 Jan Hubicka <jh@suse.cz>
23424
23425 PR lto/56557
23426 * lto-streamer-out.c (output_symbol_p): Skip references from
23427 constructors of external variables.
23428
c5c5ba89
JH		 234292013-03-11 Jan Hubicka <jh@suse.cz>
23430
23431 PR middle-end/56571
23432 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23433 from pseudos.
23434 * emit-rtl.c (verify_rtx_sharing): Likewise.
23435 (copy_insn_1): Likewise.
23436 * rtl.c (copy_rtx): Likewise.
23437
c2a939b1
GJL		 234382013-03-11 Georg-Johann Lay <avr@gjlay.de>
23439
23440 PR target/56591
23441 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23442 output_operand_lossage message.
88b97037 23443
3c3279fb
RE		 234442013-03-11 Richard Earnshaw <rearnsha@arm.com>
23445
23446 PR target/56470
23447 * arm.c (shift_op): Validate RTL pattern on the fly.
23448 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23449 the RTL.
23450
aef5ef9d
JDA		 234512013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23452
23453 PR target/56347
23454 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23455 new call patterns that clobber %fr12.
23456 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23457 split and postreload patterns.
23458 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23459 registers %fr12 and %fr12R as call used.
23460
456610d3
SB		 234612013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23462
23463 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23464 (canon_address, record_store, replace_read, check_mem_read_rtx,
23465 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23466 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23467 rest_of_handle_dse): Likewise.
23468
4b1baac8
RS		 234692013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23470
23471 PR middle-end/56524
23472 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23473 Add base_optabs.
23474 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23475 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23476 (save_optabs_if_changed): Replace with...
23477 (init_tree_optimization_optabs): ...this.
23478 * optabs.c (save_optabs_if_changed): Rename to...
23479 (init_tree_optimization_optabs): ...this. Take the optimization node
23480 as argument. Do nothing if the base optabs are already correct.
23481 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23482 to recompute optabs.
23483 * function.h (function): Remove optabs field.
23484 * function.c (invoke_set_current_function_hook): Call
23485 init_tree_optimization_optabs. Use the result to initialize
23486 this_fn_optabs.
23487
b7a78683
AH		 234882013-02-27 Aldy Hernandez <aldyh@redhat.com>
23489
23490 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23491 if GTMA_HAS_NO_INSTRUMENTATION.
23492 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23493 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23494 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23495 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23496 GTMA_HAS_NO_INSTRUMENTATION.
23497
6384c29b
JJ		 234982013-03-08 Jakub Jelinek <jakub@redhat.com>
23499
23500 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23501 libasan_preinit.o.
23502
ecd4f20a
MP		 235032013-03-08 Marek Polacek <polacek@redhat.com>
23504 Jakub Jelinek <jakub@redhat.com>
23505
23506 PR tree-optimization/56478
23507 * predict.c (is_comparison_with_loop_invariant_p): Change the
23508 type of loop_step to tree.
23509 (predict_loops): Adjust.
23510 (predict_iv_comparison): Perform the computations on double_ints.
23511
64366d35
RB		 235122013-03-08 Richard Biener <rguenther@suse.de>
23513
23514 PR tree-optimization/56570
23515 * tree-cfg.c (verify_expr_location_1): Verify locations for
23516 DECL_DEBUG_EXPR.
23517 * tree-sra.c (create_access_replacement): Strip locations
23518 from DECL_DEBUG_EXPRs.
23519
a9d5a059
RB		 235202013-03-08 Richard Biener <rguenther@suse.de>
23521
23522 * tree-inline.c (expand_call_inline): Do not associate
23523 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23524 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23525
b7aa4e9a
RB		 235262013-03-08 Richard Biener <rguenther@suse.de>
23527
23528 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23529 or block changes with -Og. Fix for location / block encoding
23530 changes and PHI arguments with locations.
23531
c4c2f9fa
SB		 235322013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23533
23534 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23535 for all counters.
23536 (struct output_info): Likewise.
23537 (register_overhead): Remove bad gcc_assert.
23538 (bitmap_find_bit): If there is only a single bitmap element, do not
23539 count a miss as a search.
23540 (print_statistics): Update for counter type changes.
23541 (dump_bitmap_statistics): Likewise. Print headers such that they
23542 are properly lined up with the printed counters.
23543
5bf6606a
JJ		 235442013-03-07 Jakub Jelinek <jakub@redhat.com>
23545
23546 PR tree-optimization/56559
23547 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23548 check that it has only a single use.
23549
2c653d46
RB		 235502013-03-07 Richard Biener <rguenther@suse.de>
23551
23552 * doc/invoke.texi (fwhole-program): Discourage use in combination
23553 with -flto.
23554
a72d8780
JJ		 235552013-03-06 Jakub Jelinek <jakub@redhat.com>
23556
01a454df
JJ		 23557 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23558
c1781047
JJ		 23559 PR tree-optimization/56539
23560 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23561 instead of GSI_CONTINUE_LINKING as last argument to
23562 force_gimple_operand_gsi. Adjust function comment.
23563
9772c47a
JJ		 23564 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23565 aarch64-cores.def.
23566
a72d8780
JJ		 23567 PR middle-end/56548
23568 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23569 promoted mode, convert the result back to the original mode.
23570
fa5556de
RB		 235712013-03-06 Richard Biener <rguenther@suse.de>
23572
23573 PR middle-end/56294
23574 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23575 (insert_updated_phi_nodes_compare_uids): New function.
23576 (update_ssa): Sort symbols_to_rename after UID before
23577 traversing it to insert PHI nodes.
23578
010403d1
RB		 235792013-03-06 Richard Biener <rguenther@suse.de>
23580
23581 PR middle-end/50494
23582 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23583 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23584
23585 Revert
23586 2013-02-13 Richard Biener <rguenther@suse.de>
23587
23588 PR lto/50494
23589 * varasm.c (output_constant_def_1): Get the decl representing
23590 the constant as argument.
23591 (output_constant_def): Wrap output_constant_def_1.
23592 (make_decl_rtl): Use output_constant_def_1 with the decl
23593 representing the constant.
23594 (build_constant_desc): Optionally re-use a decl already
23595 representing the constant.
23596 (tree_output_constant_def): Adjust.
23597
3c27ce4c
JY		 235982013-03-06 Joey Ye <joey.ye@arm.com>
23599
23600 PR lto/50293
23601 * gcc.c (convert_white_space): New function.
23602 (main): Handles white space in function name.
23603
8f6d1c86
OE		 236042013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23605
23606 PR target/56529
23607 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23608 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23609 to SH_DIV_CALL_TABLE for TARGET_SH2.
23610 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23611 list.
23612 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23613 call-table options.
23614
798a209f
SA		 236152013-03-05 Sterling Augustine <saugustine@google.com>
23616 Cary Coutant <ccoutant@google.com>
23617
23618 PR debug/55364
23619 * dwarf2out.c (resolve_addr): Don't call
23620 remove_loc_list_addr_table_entries a second time for the same
23621 expression.
23622
6cfa417f 236232013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 23624
6cfa417f
JJ		 23625 PR debug/56510
23626 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23627 (avoid_complex_debug_insns): New function.
23628 (expand_debug_locations): Call it.
23629
23630 PR rtl-optimization/56484
23631 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23632 lifetimes of hard registers on small register class machines.
23633
236342013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23635
23636 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 23637 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 23638 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH		 23639 fast_interrupt.
23640 (microblaze_fast_interrupt_function_p): New function.
23641 (microblaze_is_interrupt_handler): Rename to
23642 microblaze_is_interrupt_variant and add fast_interrupt check.
23643 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23644 (save_restore_insns): Likewise.
23645 (compute_frame_size): Likewise.
23646 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23647 (microblaze_globalize_label): Likewise.
6cfa417f
JJ		 23648 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23649 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH		 23650 microblaze_is_interrupt_variant.
23651
731aefac
KT		 236522013-03-05 Kai Tietz <ktietz@redhat.com>
23653
23654 * sdbout.c (sdbout_one_type): Switch to current function's section
23655 supporting cold/hot.
23656
a72d8780 236572013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH		 23658
23659 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23660 -mxl-reorder.
23661
334e71e8
JJ		 236622013-03-05 Jakub Jelinek <jakub@redhat.com>
23663
0b50e654
JJ		 23664 PR middle-end/56461
23665 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23666 if VALGRIND_GET_VBITS is defined, temporarily make object
23667 memory all defined, and restore previous valgrind addressability
23668 and definability afterwards. Free this_object at the end.
23669
4ccf8f43
JJ		 23670 PR middle-end/56461
23671 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23672 right before calling lra_create_live_ranges, also call it
23673 when clearing live_p. Only call lra_clear_live_ranges
23674 at the end if live_p.
23675
334e71e8
JJ		 23676 PR middle-end/56461
23677 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23678
9ca966ca
RB		 236792013-03-05 Richard Biener <rguenther@suse.de>
23680
23681 PR tree-optimization/56521
23682 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23683 value-id.
23684
d273b176
SB		 236852013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23686
23687 PR c++/55135
23688 * except.h (remove_unreachable_eh_regions): New prototype.
23689 * except.c (remove_eh_handler_splicer): New function, split out
23690 of remove_eh_handler.
23691 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23692 warning about running it on many EH regions one at a time.
23693 (remove_unreachable_eh_regions_worker): New function, walk the
23694 EH tree in depth-first order and remove non-marked regions.
23695 (remove_unreachable_eh_regions): New function.
23696 * tree-eh.c (mark_reachable_handlers): New function, split out
23697 from remove_unreachable_handlers.
23698 (remove_unreachable_handlers): Use mark_reachable_handlers and
23699 remove_unreachable_eh_regions.
23700 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23701 and remove_unreachable_eh_regions.
23702
5e7f1aef
RB		 237032013-03-05 Richard Biener <rguenther@suse.de>
23704
23705 PR middle-end/56525
23706 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23707 not freeing them until the end.
23708
f276b762
AK		 237092013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23710
23711 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23712
12a43ab8
RB		 237132013-03-05 Richard Biener <rguenther@suse.de>
23714
23715 PR tree-optimization/56270
23716 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23717 of loads after scheduling an SLP instance.
23718
93675444
JJ		 237192013-03-05 Jakub Jelinek <jakub@redhat.com>
23720
db4138e3
JJ		 23721 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23722 tic6x.exp.
23723 (check_gcc_parallelize): Run guality.exp as a separate job from
23724 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23725 struct-layout-1.exp with stackalign.exp.
23726
dd3d1ec0
JJ		 23727 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23728
b8d381a3
JJ		 23729 PR middle-end/56461
23730 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23731 load_index sbitmap even if some bit in it isn't set.
23732
b4f9786b
JJ		 23733 PR middle-end/56461
23734 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23735 (discover_iteration_bound_by_body_walk): Change queues to
23736 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23737 spelling in comment. Call safe_push on queues[bound_index] directly.
23738 Release queues[queue_index] in every iteration unconditionally.
23739 Release bounds vector.
23740
93675444
JJ		 23741 PR middle-end/56461
23742 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23743 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23744 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23745 inner_phis vector.
23746
3e492e9c
RB		 237472013-03-05 Richard Biener <rguenther@suse.de>
23748
23749 PR lto/56515
23750 * tree-inline.c (remap_blocks_to_null): New function.
23751 (expand_call_inline): When expanding a call stmt without
23752 an associated block inline remap all callee blocks to NULL.
23753
a5d1569a
JJ		 237542013-03-05 Jakub Jelinek <jakub@redhat.com>
23755
86efb5cd
JJ		 23756 PR rtl-optimization/56494
23757 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23758 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23759 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23760
85f5dbea
JJ		 23761 PR middle-end/56461
23762 * sel-sched-ir.c (free_sched_pools): Release
23763 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23764 if succs_info_pool.max_top isn't -1.
23765
a5d1569a
JJ		 23766 PR bootstrap/56509
23767 * opts.c (opts_obstack, opts_concat): Moved to...
23768 * opts-common.c (opts_obstack, opts_concat): ... here.
23769
4432aa6c
JJ		 237702013-03-04 Jakub Jelinek <jakub@redhat.com>
23771
23772 PR middle-end/56461
23773 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23774
f3a1fb91
MJ		 237752013-03-04 Martin Jambor <mjambor@suse.cz>
23776
23777 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23778 all appropriate places.
23779
22110e6c
EB		 237802013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23781
23782 PR tree-optimization/56424
23783 * ipa-split.c (split_function): Do not set the RSO flag if result is
23784 not by reference and its type is a register type.
23785
a72d8780 237862013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 23787
88b97037 23788 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 23789 (microblaze_legitimate_pic_operand): Likewise
88b97037 23790 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH		 23791 new function microblaze_legitimate_pic_operand
23792 * config/microblaze/microblaze-protos.h
23793 (microblaze_legitimate_pic_operand): Declare.
23794
a72d8780 237952013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 23796
a72d8780 23797 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 23798 New predicate for supported rtx code types.
a72d8780 23799 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI		 23800 call_insn_simple_operand predicate.
23801
541d9ac8
JJ		 238022013-03-04 Jakub Jelinek <jakub@redhat.com>
23803
5eb010bc
JJ		 23804 PR middle-end/56461
23805 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23806 partitions.ordered_remove.
23807
30862efc
JJ		 23808 PR middle-end/56461
23809 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23810 vec_oprnds0.create (1) for modifier == NONE.
23811
8930f723
JJ		 23812 PR middle-end/56461
23813 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23814 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23815 vec_oprnds1 right before pushing anything to it for
23816 scalar_shift_arg.
23817
541d9ac8
JJ		 23818 PR middle-end/56461
23819 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23820 set nbbs to 0 instead of having separate code path.
23821 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23822 instead of false as last argument if returning NULL.
23823
7aa7f2e3
SL		 238242013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23825
23826 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23827 the attribute is now called "target" instead of "option".
23828 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23829 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23830 attribute/pragma name for TARGET_OPTION_VALID_P and
23831 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23832 * doc/tm.texi: Regenerated.
23833
8930f723 238342013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 23835
8930f723 23836 * config/microblaze/microblaze.c:
cb8a1637 23837 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ		 23838 * config/microblaze/microblaze.h: Add -mxl-reorder to
23839 DRIVER_SELF_SPECS.
23840 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23841 instructions emitted if TARGET_REORDER.
88b97037
UB		 23842 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23843 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 23844 separately.
cb8a1637 23845
4dc1d68c
XDL		 238462013-03-01 Xinliang David Li <davidxl@google.com>
23847
23848 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23849 walk length.
23850
689eaba3
JJ		 238512013-03-01 Jakub Jelinek <jakub@redhat.com>
23852
9d676bf6
JJ		 23853 PR middle-end/56461
23854 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23855 vector even when returning true. Fix up function comment formatting.
23856
caff7edf
JJ		 23857 PR middle-end/56461
23858 * ira-build.c (ira_loop_nodes_count): New variable.
23859 (create_loop_tree_nodes): Initialize it.
23860 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23861
b6b9227d
JJ		 23862 PR middle-end/56461
23863 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23864 method on dr_chain and result_chain.
23865 * tree-vect-stmts.c (vectorizable_store): Only call
23866 result_chain.create if j == 0.
23867
689eaba3
JJ		 23868 PR middle-end/56461
23869 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23870 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23871 before overwriting it.
23872
40bf31ed
TB		 238732013-03-01 Tobias Burnus <burnus@net-b.de>
23874
23875 * doc/extended.texi (C Extensions): Change order in @menu
23876 to match @node.
23877 (Other MIPS Built-in Functions): Move last MIPS entry before
23878 "picoChip Built-in Functions".
23879 (SH Built-in Functions): Move after RX Built-in Functions.
23880 * doc/gcc.texi (Introduction): Change order in @menu
23881 to match @node.
23882 * doc/md.texi (Constraints): Ditto.
23883 * gty.texi (Type Information): Ditto.
23884 (User-provided marking routines for template types): Make
23885 subsection.
23886 * doc/invoke.texi (AArch64 Options): Move before
23887 "Adapteva Epiphany Options".
23888
e664c61c
KS		 238892013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23890 Jakub Jelinek <jakub@redhat.com>
23891
23892 PR sanitizer/56454
23893 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23894 no_address_safety_analysis attribute.
23895 * doc/extend.texi (no_address_safety_attribute): Rename to
23896 no_sanitize_address attribute, mention no_address_safety_analysis
23897 attribute as deprecated alias.
23898
37b5ec8f
JJ		 238992013-02-28 Jakub Jelinek <jakub@redhat.com>
23900
23901 PR middle-end/56461
23902 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23903 type to vec<vec<tree> > *.
23904 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23905 to be vec<tree> instead of vec<tree> *, set vec_defs
23906 to vNULL and call vec_defs.create (number_of_vects), adjust other
23907 uses of vec_defs.
23908 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23909 vectorizable_condition): Adjust vect_get_slp_defs callers.
23910
ba96cdfb
JG		 239112013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23912
23913 * config/aarch64/aarch64.c
23914 (aarch64_float_const_representable): Remove unused variable.
23915
6f549691
JG		 239162013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23917
23918 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23919
af55e82d
JG		 239202013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23921
23922 * config/aarch64/aarch64-builtins.c
23923 (aarch64_init_simd_builtins): Make static.
23924
1df3f464
JG		 239252013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23926
23927 * config/aarch64/aarch64.c
23928 (aarch64_simd_make_constant): Make static.
23929
f8f42513
MJ		 239302013-02-28 Martin Jambor <mjambor@suse.cz>
23931
23932 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23933 with no initialization to the RHS of debug statements.
23934
b48b3fc4
MJ		 239352013-02-28 Martin Jambor <mjambor@suse.cz>
23936
23937 PR tree-optimization/56294
23938 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23939 Adjust dumping.
23940 (get_access_replacement): Do not call create_access_replacement.
23941 Assert a replacement exists.
23942 (get_repl_default_def_ssa_name): Create the replacement declaration
23943 itself.
23944
c3ae224c
RR		 239452013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23946
23947 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23948 final_end_function.
23949
45fa0eef
MP		 239502013-02-28 Marek Polacek <polacek@redhat.com>
23951
23952 PR rtl-optimization/56466
23953 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23954 if we're changing a loop.
23955 (peel_loops_completely): Likewise.
23956
502c067d
PC		 239572013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23958
23959 PR c++/55813
23960 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23961
ff24afc8
GJL		 239622013-02-28 Georg-Johann Lay <avr@gjlay.de>
23963
23964 PR target/56445
23965 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23966 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23967 INTX_FTYPE_FX, FX_FTYPE_INTX.
23968 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23969
c7b1fc1b
GJL		 239702013-02-28 Georg-Johann Lay <avr@gjlay.de>
23971
23972 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23973 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23974 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23975 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23976 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23977 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23978 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23979 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23980 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23981 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23982 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23983 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23984 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23985 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23986 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23987 (avrxmega6): Increase max flash segments from 5 to 6.
23988 * config/avr/t-multilib: Regenerate.
23989 * config/avr/avr-tables.opt: Regenerate.
23990 * doc/avr-mmcu.texi: Regenerate.
23991
4a0e3cfe
GJL		 239922013-02-28 Georg-Johann Lay <avr@gjlay.de>
23993
23994 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23995 (avr_device_to_arch): Rename to avr_device_to_ld.
23996 (avr_device_to_as): New prototype.
23997 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23998 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23999 * config/avr/driver-avr.c (avr_device_to_as): New.
24000 (avr_device_to_arch): Rename to avr_device_to_ld.
24001
97785e52
JJ		 240022013-02-27 Jakub Jelinek <jakub@redhat.com>
24003
3f292312
JJ		 24004 PR middle-end/56461
24005 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24006 method on dr_chain and result_chain.
24007
a344216b
JJ		 24008 PR middle-end/56461
24009 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24010 pointer_set_destroy on not_executed_last_iteration.
24011
f121ad02 24012 PR middle-end/56461
88b97037 24013 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 24014
307f83a3
JJ		 24015 PR middle-end/56461
24016 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24017 FOR_EACH_DEFINED_FUNCTION when freeing state.
24018
e19624ee
JJ		 24019 PR middle-end/56461
24020 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24021 pool_free.
24022 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24023 overwriting it.
24024
90e709fd
JJ		 24025 PR middle-end/56461
24026 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24027 known_aggs[i].items and release known_aggs vector.
24028
97785e52
JJ		 24029 PR middle-end/56461
24030 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24031
240322013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 24033
97785e52
JJ		 24034 * config/microblaze/microblaze.c (microblaze_emit_compare):
24035 Use xor for EQ/NE comparisions.
24036 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24037 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 24038
6fa5e0ed
JJ		 240392013-02-27 Jakub Jelinek <jakub@redhat.com>
24040
24041 PR middle-end/56461
24042 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24043 vector.
24044 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24045 vec_safe_push, always update *slot.
24046 (redirect_edge_var_map_clear): Use vec_free.
24047 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24048 (free_var_map_entry): Use vec_free.
24049 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24050 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24051
436a956a
AB		 240522013-02-27 Andrey Belevantsev <abel@ispras.ru>
24053
24054 PR middle-end/45472
436a956a
AB		 24055 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24056 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB		 24057 Reorder tests for speculativeness in the logical and operator.
24058
f5c2caca
JJ		 240592013-02-27 Jakub Jelinek <jakub@redhat.com>
24060
0fcb564b
JJ		 24061 * incpath.c (add_standard_paths): Use reconcat instead of concat
24062 where appropriate and avoid leaking memory.
24063
dc357798
JJ		 24064 * opts.h: Include obstack.h.
24065 (opts_concat): New prototype.
24066 (opts_obstack): New declaration.
24067 * opts.c (opts_concat): New function.
24068 (opts_obstack): New variable.
24069 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24070 (finish_options): Use opts_concat instead of concat
24071 and XOBNEWVEC instead of XNEWVEC.
24072 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24073 generate_option): Likewise.
24074 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24075 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24076
f5c2caca
JJ		 24077 PR target/56455
24078 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24079 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24080
d0163673
JJ		 240812013-02-26 Jakub Jelinek <jakub@redhat.com>
24082
24083 PR middle-end/56461
24084 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24085
b5ad2b8e
JR		 240862013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24087
24088 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24089 (arm_block_move_unaligned_straight): Likewise.
24090 (arm_adjust_block_mem): Likewise.
24091
9b639e2c
JR		 240922013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24093
24094 PR target/48901
24095 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24096 temp, cond and label.
24097 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24098
e13a0ccb
JR		 24099 PR target/52500
24100 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24101 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24102
fbe4f171
JR		 24103 PR target/52501
24104 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24105 of prologue/epilogue functions.
24106
ae006543
JR		 24107 PR target/52550
24108 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24109 Remove unused variable cfa_offset.
24110 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24111
c81369fa
JR		 24112 PR target/54639
24113 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24114 type promotion to unsigned.
24115
f8a8fea7
JR		 24116 PR target/54640
24117 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24118 for HOST_WIDE_INT of 32 bit / same size as int.
24119 (arm_block_move_unaligned_straight): Likewise.
24120 (arm_adjust_block_mem): Likewise.
24121
f8be5169
JR		 24122 PR target/54662
24123 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24124 ALL_CFLAGS.
24125
f1ad3354
MP		 241262013-02-26 Marek Polacek <polacek@redhat.com>
24127
24128 PR tree-optimization/56426
88b97037 24129 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 24130
a6af0f42
RB		 241312013-02-26 Richard Biener <rguenther@suse.de>
24132
24133 PR target/56444
24134 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24135 unused variable loops.
24136
aca43c6c
JJ		 241372013-02-26 Jakub Jelinek <jakub@redhat.com>
24138
cecbe5d9
JJ		 24139 PR tree-optimization/56448
24140 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24141 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24142 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24143 later operands of the references, or even first operand for
24144 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24145
aca43c6c
JJ		 24146 PR tree-optimization/56443
24147 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24148 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24149 to type_for_mode langhook.
24150
53e2e141
MT		 241512013-02-25 Matt Turner <mattst88@gmail.com>
24152
24153 * doc/invoke.texi: Document r4700.
24154
259ee451
RB		 241552013-02-25 Richard Biener <rguenther@suse.de>
24156
24157 PR tree-optimization/56175
24158 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24159 split out from ...
24160 (simplify_bitwise_binary): ... here. Also guard the conversion
24161 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24162
6d65e8f1
CM		 241632013-02-25 Catherine Moore <clm@codesourcery.com>
24164
24165 Revert:
24166 2013-02-24 Catherine Moore <clm@codesourcery.com>
24167 Maciej W. Rozycki <macro@codesourcery.com>
24168 Tom de Vries <tom@codesourcery.com>
a72d8780 24169 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM		 24170 Iain Sandoe <iain@codesourcery.com>
24171 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 24172 Chao-ying Fu <fu@mips.com>
88b97037 24173
6d65e8f1 24174 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 24175 Document new function attributes.
6d65e8f1
CM		 24176 * doc/invoke.texi (minterlink-compressed, mmicromips,
24177 m14k, m14ke, m14kec): Document new options.
24178 (minterlink-mips16): Update documentation.
24179 * doc/md.texi (ZC, ZD): Document new constraints.
24180 * configure.ac (gcc_cv_as_micromips): Check if linker
24181 supports the .set micromips directive.
24182 * configure: Regenerate.
24183 * config.in: Regenerate.
24184 * config/mips/mips-tables.opt: Regenerate.
24185 * config/mips/micromips.md: New file.
24186 * constraints.md (ZC, AD): New constraints.
24187 * config/mips/predicates.md (movep_src_register): New predicate.
24188 (movep_src_operand): New predicate.
24189 (non_volatile_mem_operand): New predicate.
24190 * config/mips/mips.md (multimem): New type.
24191 (length): Differentiate between 17-bit and 18-bit branch offsets.
24192 (MOVEP1, MOVEP2): New mode iterator.
24193 (mov_<load>l): Use ZC constraint.
24194 (mov_<load>r): Likewise.
24195 (mov_<store>l): Likewise.
24196 (mov_<store>r): Likewise.
24197 (*branch_equality<mode>_inverted): Add microMIPS support.
24198 (*branch_equality<mode>): Likewise.
24199 (*jump_absolute): Likewise.
24200 (indirect_jump_<mode>): Likewise.
24201 (tablejump_<mode>): Likewise.
24202 (<optab>_internal): Likewise.
24203 (sibcall_internal): Likewise.
24204 (sibcall_value_internal): Likewise.
24205 (prefetch): Use constraint ZD.
24206 * config/mips/mips.opt (minterlink-compressed): New option.
24207 (minterlink-mips16): Now an alias for minterlink-compressed.
24208 (mmicromips): New option.
24209 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24210 (compare_and_swap_12): Likewise.
24211 (sync_add<mode>): Likewise.
24212 (sync_<optab>_12): Likewise.
24213 (sync_old_<optab>_12): Likewise.
24214 (sync_new_<optab>_12): Likewise.
24215 (sync_nand_12): Likewise.
24216 (sync_old_nand_12): Likewise.
24217 (sync_new_nand_12): Likewise.
24218 (sync_sub<mode>): Likewise.
24219 (sync_old_add<mode>): Likewise.
24220 (sync_old_sub<mode>): Likewise.
24221 (sync_new_add<mode>): Likewise.
24222 (sync_new_sub<mode>): Likewise.
24223 (sync_<optab><mode>): Likewise.
24224 (sync_old_<optab><mode>): Likewise.
24225 (sync_new_<optab><mode>): Likewise.
24226 (sync_nand<mode>): Likewise.
24227 (sync_old_nand<mode>): Likewise.
24228 (sync_new_nand<mode>): Likewise.
24229 (sync_lock_test_and_set<mode>): Likewise.
24230 (test_and_set_12): Likewise.
24231 (atomic_compare_and_swap<mode>): Likewise.
24232 (atomic_exchange<mode>_llsc): Likewise.
24233 (atomic_fetch_add<mode>_llsc): Likewise.
24234 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24235 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24236 (umips_save_restore_pattern_p): Likewise.
24237 (umips_load_store_pair_p): Likewise.
24238 (umips_output_load_store_pair): Likewise.
24239 (umips_movep_target_p): Likewise.
24240 (umips_12bit_offset_address_p): Likewise.
24241 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24242 (mips_base_mips16): Rename this...
24243 (mips_base_compression_flags): ...to this. Update all uses.
24244 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24245 (mips_mips16_decl_p): Delete.
24246 (mips_nomips16_decl_p): Delete.
24247 (mips_get_compress_on_flags): New function.
24248 (mips_get_compress_off_flags): New function.
24249 (mips_get_compress_mode): New function.
24250 (mips_get_compress_on_name): New function.
24251 (mips_get_compress_off_name): New function.
24252 (mips_insert_attributes): Support multiple compression types.
24253 (mips_merge_decl_attributes): Likewise.
24254 (umips_12bit_offset_address_p): New function.
24255 (mips_start_function_definition): Emit .set micromips directive.
24256 (mips_call_may_need_jalx_p): New function.
24257 (mips_function_ok_for_sibcall): Add microMIPS support.
24258 (mips_print_operand_punctuation): Support short delay slots and
24259 compact jumps.
24260 (umips_swm_mask, umips_swm_encoding): New.
24261 (umips_build_save_restore): New function.
24262 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24263 (was_mips16_p): Remove.
24264 (old_compression_mode): New.
24265 (mips_set_compression_mode): New function.
24266 (mips_set_current_function): Add microMIPS support.
24267 (mips_option_override): Likewise.
24268 (umips_save_restore_pattern_p): New function.
24269 (umips_output_save_restore): New function.
24270 (umips_load_store_pair_p_1): New function.
24271 (umips_load_store_pair_p): New function.
24272 (umips_output_load_store_pair_1): New function.
24273 (umips_output_load_store_pair): New function.
24274 (umips_movep_target_p) New function.
24275 (mips_prepare_pch_save): Add microMIPS support.
24276 * config/mips/mips.h (TARGET_COMPRESSION): New.
24277 (TARGET_CPU_CPP_BUILTINS): Update macro
24278 to use new compression flags and to support microMIPS.
24279 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24280 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24281 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24282 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24283 (ASM_SPEC): Support mmicromips and mno-micromips.
24284 (M16STORE_REG_P): New macro.
24285 (MIPS_CALL): Support TARGET_MICROMIPS.
24286 (MICROMIPS_J): New macro.
24287 (mips_base_mips16): Rename this...
24288 (mips_base_compression_flags): ...to this.
24289 (UMIPS_12BIT_OFFSET_P): New macro.
24290 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24291 (MULTILIB_DIRNAMES): Likewise.
24292
5e5df392
TV		 242932013-02-25 Tom de Vries <tom@codesourcery.com>
24294
24295 PR rtl-optimization/56131
24296 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24297 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24298 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24299
3fdb53c1
TB		 243002013-02-25 Tobias Burnus <burnus@net-b.de>
24301
24302 * doc/invoke.texi (-fsanitize=): Move from optimization
24303 to debugging options.
24304
ed358aea
AB		 243052013-02-25 Andrey Belevantsev <abel@ispras.ru>
24306
24307 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24308
f45e9053
AB		 243092013-02-25 Andrey Belevantsev <abel@ispras.ru>
24310 Alexander Monakov <amonakov@ispras.ru>
24311
24312 PR middle-end/56077
24313 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24314 flush pending lists also on non-jumps. Adjust comment.
24315
6941b508
CM		 243162013-02-24 Catherine Moore <clm@codesourcery.com>
24317 Maciej W. Rozycki <macro@codesourcery.com>
24318 Tom de Vries <tom@codesourcery.com>
a72d8780 24319 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM		 24320 Iain Sandoe <iain@codesourcery.com>
24321 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 24322 Chao-ying Fu <fu@mips.com>
6941b508
CM		 24323
24324 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 24325 Document new function attributes.
6941b508
CM		 24326 * doc/invoke.texi (minterlink-compressed, mmicromips,
24327 m14k, m14ke, m14kec): Document new options.
24328 (minterlink-mips16): Update documentation.
24329 * doc/md.texi (ZC, ZD): Document new constraints.
24330 * configure.ac (gcc_cv_as_micromips): Check if linker
24331 supports the .set micromips directive.
24332 * configure: Regenerate.
24333 * config.in: Regenerate.
24334 * config/mips/mips-tables.opt: Regenerate.
24335 * config/mips/micromips.md: New file.
24336 * constraints.md (ZC, AD): New constraints.
24337 * config/mips/predicates.md (movep_src_register): New predicate.
24338 (movep_src_operand): New predicate.
24339 (non_volatile_mem_operand): New predicate.
24340 * config/mips/mips.md (multimem): New type.
24341 (length): Differentiate between 17-bit and 18-bit branch offsets.
24342 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 24343 (mov_<load>l): Use ZC constraint.
6941b508
CM		 24344 (mov_<load>r): Likewise.
24345 (mov_<store>l): Likewise.
24346 (mov_<store>r): Likewise.
24347 (*branch_equality<mode>_inverted): Add microMIPS support.
24348 (*branch_equality<mode>): Likewise.
24349 (*jump_absolute): Likewise.
24350 (indirect_jump_<mode>): Likewise.
24351 (tablejump_<mode>): Likewise.
24352 (<optab>_internal): Likewise.
24353 (sibcall_internal): Likewise.
24354 (sibcall_value_internal): Likewise.
24355 (prefetch): Use constraint ZD.
24356 * config/mips/mips.opt (minterlink-compressed): New option.
24357 (minterlink-mips16): Now an alias for minterlink-compressed.
24358 (mmicromips): New option.
24359 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24360 (compare_and_swap_12): Likewise.
24361 (sync_add<mode>): Likewise.
24362 (sync_<optab>_12): Likewise.
24363 (sync_old_<optab>_12): Likewise.
24364 (sync_new_<optab>_12): Likewise.
24365 (sync_nand_12): Likewise.
24366 (sync_old_nand_12): Likewise.
24367 (sync_new_nand_12): Likewise.
24368 (sync_sub<mode>): Likewise.
24369 (sync_old_add<mode>): Likewise.
24370 (sync_old_sub<mode>): Likewise.
24371 (sync_new_add<mode>): Likewise.
24372 (sync_new_sub<mode>): Likewise.
24373 (sync_<optab><mode>): Likewise.
24374 (sync_old_<optab><mode>): Likewise.
24375 (sync_new_<optab><mode>): Likewise.
24376 (sync_nand<mode>): Likewise.
24377 (sync_old_nand<mode>): Likewise.
24378 (sync_new_nand<mode>): Likewise.
24379 (sync_lock_test_and_set<mode>): Likewise.
24380 (test_and_set_12): Likewise.
24381 (atomic_compare_and_swap<mode>): Likewise.
24382 (atomic_exchange<mode>_llsc): Likewise.
24383 (atomic_fetch_add<mode>_llsc): Likewise.
24384 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24385 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24386 (umips_save_restore_pattern_p): Likewise.
24387 (umips_load_store_pair_p): Likewise.
24388 (umips_output_load_store_pair): Likewise.
24389 (umips_movep_target_p): Likewise.
24390 (umips_12bit_offset_address_p): Likewise.
24391 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24392 (mips_base_mips16): Rename this...
24393 (mips_base_compression_flags): ...to this. Update all uses.
24394 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24395 (mips_mips16_decl_p): Delete.
24396 (mips_nomips16_decl_p): Delete.
3fdb53c1 24397 (mips_get_compress_on_flags): New function.
6941b508
CM		 24398 (mips_get_compress_off_flags): New function.
24399 (mips_get_compress_mode): New function.
24400 (mips_get_compress_on_name): New function.
24401 (mips_get_compress_off_name): New function.
24402 (mips_insert_attributes): Support multiple compression types.
24403 (mips_merge_decl_attributes): Likewise.
24404 (umips_12bit_offset_address_p): New function.
24405 (mips_start_function_definition): Emit .set micromips directive.
24406 (mips_call_may_need_jalx_p): New function.
24407 (mips_function_ok_for_sibcall): Add microMIPS support.
24408 (mips_print_operand_punctuation): Support short delay slots and
24409 compact jumps.
24410 (umips_swm_mask, umips_swm_encoding): New.
24411 (umips_build_save_restore): New function.
24412 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24413 (was_mips16_p): Remove.
24414 (old_compression_mode): New.
24415 (mips_set_compression_mode): New function.
24416 (mips_set_current_function): Add microMIPS support.
24417 (mips_option_override): Likewise.
24418 (umips_save_restore_pattern_p): New function.
24419 (umips_output_save_restore): New function.
24420 (umips_load_store_pair_p_1): New function.
24421 (umips_load_store_pair_p): New function.
24422 (umips_output_load_store_pair_1): New function.
24423 (umips_output_load_store_pair): New function.
24424 (umips_movep_target_p) New function.
24425 (mips_prepare_pch_save): Add microMIPS support.
24426 * config/mips/mips.h (TARGET_COMPRESSION): New.
24427 (TARGET_CPU_CPP_BUILTINS): Update macro
24428 to use new compression flags and to support microMIPS.
24429 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24430 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24431 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24432 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24433 (ASM_SPEC): Support mmicromips and mno-micromips.
24434 (M16STORE_REG_P): New macro.
24435 (MIPS_CALL): Support TARGET_MICROMIPS.
24436 (MICROMIPS_J): New macro.
24437 (mips_base_mips16): Rename this...
24438 (mips_base_compression_flags): ...to this.
24439 (UMIPS_12BIT_OFFSET_P): New macro.
24440 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24441 (MULTILIB_DIRNAMES): Likewise.
24442
cdebbc6d
JJ		 244432013-02-24 Jakub Jelinek <jakub@redhat.com>
24444
24445 PR target/52555
24446 * target-globals.c (save_target_globals): For init_reg_sets and
24447 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24448
18c63565
JG		 244492013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24450
24451 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24452 * config/aarch64/t-aarch64
24453 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24454
2194f7a2
VM		 244552013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24456
24457 PR inline-asm/56148
24458 * lra-constraints.c (process_alt_operands): Reload operand
24459 conflicting with earlier clobber only if no more other conflicting
24460 operands.
24461
7d613735
JJ		 244622013-02-22 Jakub Jelinek <jakub@redhat.com>
24463
24464 PR sanitizer/56393
24465 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24466 if not linking a shared library.
24467
ac8d93a7
SL		 244682013-02-22 Seth LaForge <sethml@google.com>
24469
24470 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24471
e0237780
GY		 244722013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24473
3fdb53c1
TB		 24474 * config/arm/arm.md (split for extendsidi): Update condition.
24475 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24476 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24477 (qhs_zextenddi_cstr): Likewise.
e0237780 24478
d7fde18c
JJ		 244792013-02-21 Jakub Jelinek <jakub@redhat.com>
24480
be63b77d
JJ		 24481 PR middle-end/56420
24482 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24483 avoid signed wrapping.
24484 (expand_mult): Handle properly multiplication by
24485 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24486 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24487 in the compiler if coeff is HOST_WIDE_INT_MIN.
24488 (expand_divmod): Don't make ext_op1 static, change it's type to
24489 uhwi. Avoid undefined behavior in -INTVAL (op1).
24490
d7fde18c
JJ		 24491 PR rtl-optimization/50339
24492 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24493 field.
24494 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24495 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24496 into splitting_ashiftrt field.
24497 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24498 ASHIFTRT.
24499 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24500 choices.
24501
6aad4455
AH		 245022013-02-20 Aldy Hernandez <aldyh@redhat.com>
24503
24504 PR middle-end/56108
24505 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24506 are sure to go irrevocable.
24507
38fe784d
HPN		 245082013-02-21 Hans-Peter Nilsson <hp@axis.com>
24509
24510 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24511 scalars are valid operands.
24512
0fd44da3
MJ		 245132013-02-21 Martin Jambor <mjambor@suse.cz>
24514
24515 PR tree-optimization/56310
24516 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24517 only matching indices and non-negative final offsets.
24518 (intersect_aggregates_with_edge): Pass src_idx to
24519 agg_replacements_to_vector. Pass src_idx insstead of index to
24520 intersect_with_agg_replacements.
24521
7a92038b
MJ		 245222013-02-21 Martin Jambor <mjambor@suse.cz>
24523
24524 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24525 instead of hard-wired defaults.
24526
c0da9c37
MR		 245272013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24528
24529 * doc/invoke.texi (MIPS Options): Update documentation of the
24530 floating-point multiply-accumulate instruction restrictions.
24531
d247ea0c 245322013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS		 24533
24534 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24535 asan_shadow_offset on x86_64 linux.
24536
22deefcb
RB		 245372013-02-21 Richard Biener <rguenther@suse.de>
24538
24539 PR tree-optimization/56415
24540 Revert
24541 2013-02-11 Richard Biener <rguenther@suse.de>
24542
24543 PR tree-optimization/56273
24544 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24545 first VRP run.
24546
7bcc6e75
JJ		 245472013-02-21 Jakub Jelinek <jakub@redhat.com>
24548
a0ad148f
JJ		 24549 PR bootstrap/56258
24550 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24551 instead of @itemx.
24552
7bcc6e75
JJ		 24553 PR inline-asm/56405
24554 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24555 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24556
a0a7b611
JH		 245572013-02-20 Jan Hubicka <jh@suse.cz>
24558
24559 PR tree-optimization/56265
88b97037
UB		 24560 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24561 when target is referenced for first time.
a0a7b611 24562
c0e50f72
RB		 245632013-02-20 Richard Biener <rguenther@suse.de>
24564
24565 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24566 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24567 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24568 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24569 not return anything.
24570 (rename_ssa_copies): Do not remove unused locals.
24571 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 24572 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB		 24573 * passes.c (execute_function_todo): Do not schedule unused locals
24574 removal if cleanup_tree_cfg did something.
24575 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24576 about the number of removed locals.
24577
a52ca739
RB		 245782013-02-20 Richard Biener <rguenther@suse.de>
24579
24580 PR tree-optimization/56398
88b97037 24581 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 24582
ec9c9d1b
MJ		 245832013-02-20 Martin Jambor <mjambor@suse.cz>
24584
24585 PR tree-optimization/55334
24586 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24587 restricted pointers to arrays.
24588
e91c8ed6 245892013-02-20 Richard Biener <rguenther@suse.de>
88b97037 24590 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB		 24591
24592 PR tree-optimization/56396
24593 * tree-ssa-ccp.c (n_const_val): New static variable.
24594 (get_value): Return NULL for SSA names we don't have a lattice
24595 entry for.
24596 (ccp_initialize): Initialize n_const_val.
24597 * tree-ssa-copy.c (n_copy_of): New static variable.
24598 (init_copy_prop): Initialize n_copy_of.
24599 (get_value): Return NULL_TREE for SSA names we don't have a
24600 lattice entry for.
24601
3d916479
MJ		 246022013-02-20 Martin Jambor <mjambor@suse.cz>
24603
24604 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24605
71a86758
RB		 246062013-02-20 Richard Biener <rguenther@suse.de>
24607
24608 * genpreds.c (write_lookup_constraint): Do not compare first
24609 letter of the constraint again.
24610
79836a12
RB		 246112013-02-20 Richard Biener <rguenther@suse.de>
24612
24613 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24614 and ceil_log2.
24615 (get_use_iv_cost): Terminate hashtable walk when coming across
24616 an empty entry.
24617
bbe4fb2c
IZ		 246182013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24619
24620 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24621 reassociation for avx2 targets.
24622
a72d8780 246232012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 24624
c77f83d5
EI		 24625 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24626 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 24627 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI		 24628 version and TARGET_PATTERN_COMPARE check
24629 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 24630
a72d8780 246312012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 24632
a72d8780 24633 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI		 24634 function before branching.
24635
242387fa
AB		 246362012-02-19 Andrey Belevantsev <abel@ispras.ru>
24637
24638 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24639 DUMP_INSN_RTX_UID.
24640 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24641
b4979ab9
AB		 246422012-02-19 Andrey Belevantsev <abel@ispras.ru>
24643
24644 PR middle-end/55889
b4979ab9
AB		 24645 * sel-sched.c: Include ira.h.
24646 (implicit_clobber_conflict_p): New function.
24647 (moveup_expr): Use it.
88b97037 24648 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 24649
24d63016
RB		 246502013-02-19 Richard Biener <rguenther@suse.de>
24651
24652 PR tree-optimization/56384
24653 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24654 (vn_hash_type): Split out from ...
24655 (vn_hash_constant_with_type): ... here.
24656 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24657 (vn_phi_eq): Compare types from vn_phi_s structure.
24658 (vn_phi_lookup): Populate vn_phi_s type.
24659 (vn_phi_insert): Likewise.
24660
a475fd3d
JJ		 246612013-02-19 Jakub Jelinek <jakub@redhat.com>
24662
47cc28f5
JJ		 24663 PR tree-optimization/56350
24664 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24665 if haven't found reduction or nested cycle operand, rather than
24666 asserting we must find it.
24667
a475fd3d
JJ		 24668 PR tree-optimization/56381
24669 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24670 to fold_build3.
24671
246722013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH		 24673 Jakub Jelinek <jakub@redhat.com>
24674
24675 PR target/52555
24676 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24677 (swap_optab_enable): Same.
24678 (init_all_optabs): Use argument instead of global.
88b97037 24679 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH		 24680 * expr.h (init_all_optabs): Add argument to prototype.
24681 (TREE_OPTIMIZATION_OPTABS): New.
24682 (save_optabs_if_changed): Protoize.
24683 * optabs.h: Declare this_fn_optabs.
24684 * optabs.c (save_optabs_if_changed): New.
24685 Declare this_fn_optabs.
24686 (init_optabs): Add argument to init_all_optabs() call.
24687 * function.c (invoke_set_current_function_hook): Handle per
24688 function optabs.
24689 * function.h (struct function): New field optabs.
24690 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24691 optimization_current_node has changed.
24692 * target-globals.h (save_target_globals_default_opts): Protoize.
24693 * target-globals.c (save_target_globals_default_opts): New.
24694
3f587ca3
JDA		 246952013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24696
24697 PR target/56347
24698 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24699 registers %fr12 and %fr12R as call used.
ceaca33e
JDA		 24700
24701 PR target/56214
24702 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB		 24703 and HImode, require all displacements to be an integer multiple of
24704 their mode size.
ceaca33e
JDA		 24705 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24706 only allow QImode and HImode when reload is in progress and strict is
24707 true. Likewise for symbolic addresses. Use base14_operand to check
24708 displacements in REG+BASE addresses.
24709
fe0b4796
RB		 247102013-02-18 Richard Biener <rguenther@suse.de>
24711
24712 PR tree-optimization/56366
24713 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24714 handle sign-conversion of outer-loop initial induction value.
24715
6aaf596b
RB		 247162013-02-18 Richard Biener <rguenther@suse.de>
24717
73db8ff1 24718 PR middle-end/56349
6aaf596b
RB		 24719 * cfghooks.c (merge_blocks): If we merge a latch into another
24720 block adjust references to it.
24721 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24722 (verify_loop_structure): Verify that a recorded latch is in fact
24723 a latch.
24724
5e97dfb6
RB		 247252013-02-18 Richard Biener <rguenther@suse.de>
24726
24727 PR tree-optimization/56321
24728 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24729 order SSA name release and virtual operand unlinking.
24730
825527e8
EI		 247312013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24732
24733 * config/microblaze/microblaze.md (save_stack_block): Define.
24734 (restore_stack_block): Likewise.
24735
debd11d9
EI		 247362013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24737
24738 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24739 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24740 * config/microblaze/microblaze.c (microblaze_option_override):
24741 Bail out early for PIC modes when target does not support PIC.
24742
8ec77be0
EI		 247432013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24744
88b97037 24745 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI		 24746 Replace with a microblaze version.
24747 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 24748 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI		 24749 microblaze.
24750
b41288b3
JJ		 247512013-02-16 Jakub Jelinek <jakub@redhat.com>
24752 Dodji Seketeli <dodji@redhat.com>
24753
24754 PR asan/56330
88b97037 24755 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ		 24756 (instrument_mem_region_access): Do not forget to always put
24757 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24758 0) statement, even for cases where either 'base' or 'base + len'
24759 are not instrumented -- because they have been previously
24760 instrumented. Simplify the logic by putting all the statements
24761 instrument 'base + len' inside a sequence, and then insert that
24762 sequence right before the current insertion point. Then, to
24763 instrument 'base + len', just get an iterator on that statement.
24764 And do not forget to update the pointer to iterator the function
24765 received as argument.
24766
47918951
VM		 247672013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24768
24769 PR rtl-optimization/56348
24770 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24771
3fb7c699
SB		 247722013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24773
24774 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24775 (clean_graph_dump_file): Pass base to start_graph_dump.
24776
576fe41a
RH		 247772013-02-14 Richard Henderson <rth@redhat.com>
24778
24779 PR target/55941
24780 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24781
bb3accfa
SB		 247822013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24783
24784 * collect2-aix.h: Define F_LOADONLY.
24785
03143140
RB		 247862013-02-14 Richard Biener <rguenther@suse.de>
24787
24788 PR lto/50494
24789 * varasm.c (output_constant_def_1): Get the decl representing
24790 the constant as argument.
24791 (output_constant_def): Wrap output_constant_def_1.
24792 (make_decl_rtl): Use output_constant_def_1 with the decl
24793 representing the constant.
24794 (build_constant_desc): Optionally re-use a decl already
24795 representing the constant.
24796 (tree_output_constant_def): Adjust.
24797
8fb06726
DS		 247982013-02-14 Dodji Seketeli <dodji@redhat.com>
24799
24800 Fix an asan crash
24801 * asan.c (instrument_builtin_call): Really put the length of the
24802 second source argument into src1_len.
24803
c4bfe8bf
JJ		 248042013-02-13 Jakub Jelinek <jakub@redhat.com>
24805
24806 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24807 argument. If it is false, don't create edge from then_bb to
24808 fallthru_bb.
24809 (insert_if_then_before_iter): Pass true to it.
24810 (build_check_stmt): Pass false to it.
24811 (transform_statements): Flush hash table only on extended basic
24812 block boundaries, rather than at the beginning of every bb.
24813 Don't flush hash table on nonfreeing_call_p calls.
24814 * tree-flow.h (nonfreeing_call_p): New prototype.
24815 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24816
7afe2801
DM		 248172013-02-13 David S. Miller <davem@davemloft.net>
24818
24819 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24820
70cc3288
VM		 248212013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24822
24823 PR target/56184
24824 * ira.c (max_regno_before_ira): Move from ...
24825 (ira): ... here.
24826 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24827 vec_safe_length.
24828
6422242b
JJ		 248292013-02-13 Jakub Jelinek <jakub@redhat.com>
24830
24831 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24832
976a81ee
RB		 248332013-02-13 Richard Biener <rguenther@suse.de>
24834
24835 PR lto/56295
24836 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24837 globals in MEM_REFs.
24838
c1874a87
RB		 248392013-02-13 Richard Biener <rguenther@suse.de>
24840
24841 * loop-init.c (loop_optimizer_init): Clear loop state when
24842 re-initializing preserved loops.
24843 * loop-unswitch.c (unswitch_single_loop): Return whether
24844 we unswitched the loop. Do not verify loop state here.
88b97037 24845 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 24846
b4ab7d34
KS		 248472013-02-13 Kostya Serebryany <kcc@google.com>
24848
88b97037
UB		 24849 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24850 on x86_64 linux.
b4ab7d34
KS		 24851 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24852
bdcbe80c
DS		 248532013-02-12 Dodji Seketeli <dodji@redhat.com>
24854
24855 Avoid instrumenting duplicated memory access in the same basic block
24856 * Makefile.in (asan.o): Add new dependency on hash-table.h
24857 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24858 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24859 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24860 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24861 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24862 (get_mem_ref_of_assignment): New functions.
24863 (get_mem_refs_of_builtin_call): Extract from
24864 instrument_builtin_call and tweak a little bit to make it fit with
24865 the new signature.
24866 (instrument_builtin_call): Use the new
24867 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24868 of is_gimple_builtin_call.
24869 (instrument_derefs, instrument_mem_region_access): Insert the
24870 instrumented memory reference into the hash table.
24871 (maybe_instrument_assignment): Renamed instrument_assignment into
24872 this, and change it to advance the iterator when instrumentation
24873 actually happened and return true in that case. This makes it
24874 homogeneous with maybe_instrument_assignment, and thus give a
24875 chance to callers to be more 'regular'.
24876 (transform_statements): Clear the memory reference hash table
24877 whenever we enter a new BB, when we cross a function call, or when
24878 we are done transforming statements. Use
24879 maybe_instrument_assignment instead of instrumentation. No more
24880 need to special case maybe_instrument_assignment and advance the
24881 iterator after calling it; it's now handled just like
24882 maybe_instrument_call. Update comment.
24883
4861a1f7
RB		 248842013-02-13 Richard Biener <rguenther@suse.de>
24885
24886 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24887 Fix loop discovery code.
24888
1c86bd80
VM		 248892013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24890
24891 PR inline-asm/56148
24892 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 24893 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM		 24894 the operand is not reloaded. Prefer to reload conflicting operand
24895 if earlyclobber and matching operands are the same.
24896
7cbda518
RB		 248972013-02-12 Richard Biener <rguenther@suse.de>
24898
24899 PR lto/56297
24900 * lto-streamer-out.c (write_symbol): Do not output symbols
24901 for hard register variables.
24902
e68a4ef6
GJL		 249032013-02-12 Georg-Johann Lay <avr@gjlay.de>
24904
24905 PR target/54222
24906 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24907 (umulsidi3_insn, mulsidi3_insn): New insns.
24908
a72d8780 249092013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL		 24910
24911 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24912 (struct tune_params): Add vec_costs field.
24913 * config/arm/arm.c (arm_builtin_vectorization_cost)
24914 (arm_add_stmt_cost): New functions.
24915 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24916 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24917 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24918 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24919 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24920 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24921 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24922
43320568
RB		 249232013-02-12 Richard Biener <rguenther@suse.de>
24924
24925 PR lto/56295
24926 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24927 decls again if possible.
24928
a011aa39
RB		 249292013-02-12 Richard Biener <rguenther@suse.de>
24930
24931 PR middle-end/56288
24932 * tree-ssa.c (verify_ssa_name): Fix check, move
24933 SSA_NAME_IN_FREE_LIST check up.
24934
6da26889
JJ		 249352013-02-12 Jakub Jelinek <jakub@redhat.com>
24936 Steven Bosscher <steven@gcc.gnu.org>
24937
24938 PR rtl-optimization/56151
24939 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24940 equal to op0 or op1, and last_insn pattern is CODE operation
24941 with MEM dest and one of the operands matches that MEM.
24942
f80e0faf
ST		 249432013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24944
24945 * doc/extend.texi: Document Function Multiversioning and "default"
24946 parameter string to target attribute.
24947 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24948 target attribute parameter is "default".
24949 (ix86_compare_version_priority): Remove checks for target attribute.
24950 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24951 Remove check for target attribute equal to NULL. Add assert.
24952 (ix86_generate_version_dispatcher_body): Change error to sorry.
24953
6c59ffd1
IS		 249542013-02-11 Iain Sandoe <iain@codesourcery.com>
24955 Jack Howarth <howarth@bromo.med.uc.edu>
24956 Patrick Marlier <patrick.marlier@gmail.com>
24957
24958 PR libitm/55693
24959 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24960 define ENDFILE_SPEC as TM_DESTRUCTOR.
24961 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24962
a72d8780 249632013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP		 24964 Jack Howarth <howarth@bromo.med.uc.edu>
24965 Jakub Jelinek <jakub@redhat.com>
24966
24967 PR sanitizer/55617
24968 * config/darwin.c (cdtor_record): Rename ctor_record.
24969 (sort_cdtor_records): Rename sort_ctor_records.
24970 (finalize_dtors): New routine to sort destructors by
24971 priority before use in assemble_integer.
24972 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24973
b63fe007
UB		 249742013-02-11 Uros Bizjak <ubizjak@gmail.com>
24975
24976 PR rtl-optimization/56275
24977 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24978 offset is non-negative and less than cmode size before
24979 calling simplify_subreg.
24980
8e89b5b5
RB		 249812013-02-11 Richard Biener <rguenther@suse.de>
24982
24983 PR tree-optimization/56264
24984 * cfgloop.h (fix_loop_structure): Adjust prototype.
24985 * loop-init.c (fix_loop_structure): Return the number of
24986 newly discovered loops.
24987 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24988 are discovered, do a full loop-closed SSA rewrite.
24989
b4a4b56d
RB		 249902013-02-11 Richard Biener <rguenther@suse.de>
24991
24992 PR tree-optimization/56273
24993 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24994 first VRP run.
24995 (check_array_ref): Fix missing newline in dumps.
24996 (search_for_addr_array): Likewise.
24997
0c885229
DE		 249982013-02-09 David Edelsohn <dje.gcc@gmail.com>
24999
25000 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25001
59ac9a55
JJ		 250022013-02-09 Jakub Jelinek <jakub@redhat.com>
25003
25004 PR target/56256
25005 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25006
25bb0bb5
VM		 250072013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25008
25009 PR rtl-optimization/56246
0c885229 25010 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM		 25011 reload pseudo.
25012 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25013 constraints are satisfied.
25014
a698cc03
JL		 250152013-02-08 Jeff Law <law@redhat.com>
25016
25017 PR debug/53948
25018 * emit-rtl.c (reg_is_parm_p): New function.
25019 * regs.h (reg_is_parm_p): New prototype.
25020 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25021 callee-clobbered registers.
25022
e1122ddd
MM		 250232013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25024
25025 PR target/56043
25026 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25027 If there is no implicit builtin declaration, just return NULL.
25028
19c5f6e6
UB		 250292013-02-08 Uros Bizjak <ubizjak@gmail.com>
25030
25031 * config/i386/sse.md (FMAMODEM): New mode iterator.
25032 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25033 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25034
2480f2ca 250352013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 25036
2480f2ca
UB		 25037 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25038 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25039 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25040
250412013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25042
25043 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25044 (microblaze*-*-elf): Likewise.
25045 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25046 LINK_SPEC.
25047 * config/microblaze/microblaze-c.c: Add builtin defines for
25048 _LITTLE_ENDIAN and _BIG_ENDIAN.
25049 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25050 add to TARGET_DEFAULT flags.
76ef61fb 25051 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB		 25052 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25053 * config/microblaze/microblaze.md: Update extendsidi2 and
25054 movdi_internal instructions to use low-order / high-order reg
25055 print_operands.
25056 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25057 options and inversemask / mask of LITTLE_ENDIAN.
25058 * config/microblaze/t-microblaze: Expand multilib options to
25059 include mlittle-endian (le) and update exceptions patterns.
25060
600a5961
JJ		 250612013-02-08 Jakub Jelinek <jakub@redhat.com>
25062
5df81313
JJ		 25063 PR rtl-optimization/56195
25064 * lra-constraints.c (get_reload_reg): Don't reuse regs
25065 if they have smaller mode than requested, if they have
25066 wider mode than requested, try to return a SUBREG.
25067
600a5961
JJ		 25068 PR tree-optimization/56250
25069 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25070 if type is unsigned and code isn't MULT_EXPR.
25071
ff544649
GJL		 250722013-02-08 Georg-Johann Lay <avr@gjlay.de>
25073
25074 PR tree-optimization/56064
25075 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25076 bits according to mode.
25077 * fixed-value.h (fixed_from_double_int)
25078 (const_fixed_from_double_int): Adjust comments.
25079
e45cde98
RB		 250802013-02-08 Richard Biener <rguenther@suse.de>
25081
25082 PR lto/56231
25083 * lto-streamer.h (struct data_in): Remove current_file, current_line
25084 and current_col members.
25085 * lto-streamer-out.c (lto_output_location): Stream changed bits
25086 en-block for efficiency.
25087 * lto-streamer-in.c (clear_line_info): Remove.
25088 (lto_input_location): Cache current file, line and column
25089 globally via local statics. Read changed bits en-block.
25090 (input_function): Do not call clear_line_info.
25091 (lto_read_body): Likewise.
25092 (lto_input_toplevel_asms): Likewise.
25093
c1ca73d8
MM		 250942013-02-08 Michael Matz <matz@suse.de>
25095
25096 PR tree-optimization/52448
25097 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25098 (nt_call_phase): New static.
25099 (add_or_mark_expr): Only mark accesses with newer phase than any
25100 call seen.
25101 (nonfreeing_call_p): New.
25102 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25103 (nt_fini_block): Keep blocks marked as visited.
25104 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25105
57548aa2
RB		 251062013-02-08 Richard Biener <rguenther@suse.de>
25107
25108 * ira.c (ira): Free broken dominator information.
25109
8e10366f
UB		 251102013-02-08 Uros Bizjak <ubizjak@gmail.com>
25111
25112 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25113
f64fb0fa
MP		 251142013-02-08 Marek Polacek <polacek@redhat.com>
25115
8e10366f 25116 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 25117
0375167b
RB		 251182013-02-08 Richard Biener <rguenther@suse.de>
25119
25120 PR middle-end/56181
25121 * cfgloop.h (flow_loops_find): Adjust.
25122 (bb_loop_header_p): Declare.
25123 * cfgloop.c (bb_loop_header_p): New function split out from ...
25124 (flow_loops_find): ... here. Adjust function signature,
25125 support incremental loop structure update.
25126 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25127 * cfgloopmanip.c (fix_loop_structure): Move ...
25128 * loop-init.c (fix_loop_structure): ... here.
25129 (apply_loop_flags): Split out from ...
25130 (loop_optimizer_init): ... here.
25131 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25132 in incremental mode, only remove dead loops here.
25133
85d768f3
GJL		 251342013-02-08 Georg-Johann Lay <avr@gjlay.de>
25135
25136 PR target/54222
25137 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25138 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25139 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25140 (*round<mode>3.libgcc): New insns for fixed-modes.
25141 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25142 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25143 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25144 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25145 implementations. Define to __builtin_avr_absFX,
25146 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25147 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25148 __builtin_avr_countlsFX, respectively.
25149 * config/avr/avr-c.c (target.h): Include it.
25150 (enum avr_builtin_id): New enum.
25151 (avr_resolve_overloaded_builtin): New static function.
25152 (avr_register_target_pragmas): Use it to set
25153 targetm.resolve_overloaded_builtin.
25154 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25155 tree nodes used by DEF_BUILTIN.
25156 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25157 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25158 <AVR_BUILTIN_xxBITS>: Same.
25159
661bc682
RB		 251602013-02-08 Richard Biener <rguenther@suse.de>
25161
25162 * cfgloop.c (verify_loop_structure): Properly handle
25163 a loop exiting to another loop header.
25164 * ira-int.h (ira_loops): Remove.
25165 * ira.c (ira_loops): Remove.
25166 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25167 (do_reload): Use loop_optimizer_finalize.
25168 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25169 number_of_loops to access the loop tree.
25170 (more_one_region_p): Likewise.
25171 (finish_loop_tree_nodes): Likewise.
25172 (rebuild_regno_allocno_maps): Likewise.
25173 (mark_loops_for_removal): Likewise.
25174 (mark_all_loops_for_removal): Likewise.
25175 (remove_unnecessary_regions): Likewise.
25176 (ira_build): Likewise.
25177 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25178
0d5049b2
RB		 251792013-02-08 Richard Biener <rguenther@suse.de>
25180
25181 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25182 * ipa-pure-const.c (analyze_function): Avoid calling
25183 mark_irreducible_loops twice.
8e10366f 25184 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 25185
499f32e8
DM		 251862013-02-07 David S. Miller <davem@davemloft.net>
25187
25188 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25189 on 'reg'.
25190 * var-tracking.c (vt_add_function_parameter): Test the presence of
25191 HAVE_window_save properly and do not remap argument registers when
25192 we have a leaf function.
25193
6edc3e32
UB		 251942013-02-07 Uros Bizjak <ubizjak@gmail.com>
25195
25196 PR bootstrap/56227
25197 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25198 instead of "ll".
25199 * config/i386/i386.c (ix86_print_operand): Ditto.
25200
5306401f
VM		 252012013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25202
6edc3e32 25203 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 25204
027ece11
VM		 252052013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25206
25207 PR rtl-optimization/56225
25208 * lra-constraints.c (process_alt_operands): Check that reload hard
25209 reg can hold value for strict_low_part.
25210
f980dfdb
JJ		 252112013-02-07 Jakub Jelinek <jakub@redhat.com>
25212
25213 PR debug/56154
25214 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25215 dwarf2out_end_function.
25216 (in_first_function_p, maybe_at_text_label_p,
25217 first_loclabel_num_not_at_text_label): New variables.
25218 (dwarf2out_var_location): In the first function find out
6edc3e32 25219 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ		 25220 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25221 functions.
25222
73dd3123
EB		 252232013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25224
25225 PR rtl-optimization/56178
25226 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25227 SUBREG of a register. Tidy up related block of code.
25228 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25229 note if the source is a register or a SUBREG of a register.
25230
e3936f47
JJ		 252312013-02-07 Jakub Jelinek <jakub@redhat.com>
25232
25233 PR target/56228
25234 * config/rs6000/rs6000.md (ptrm): New mode attr.
25235 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25236 call_value_indirect_aix<pttrsize>,
25237 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25238 m in constraints.
25239
d96d674b
MH		 252402013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25241
25242 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25243 if -bnortl. Convert to strcmp and strncmp.
25244
921f2dee
AM		 252452013-02-07 Alan Modra <amodra@gmail.com>
25246
25247 PR target/54009
25248 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25249 addresses won't wrap when offsetting.
25250 (rs6000_secondary_reload): Provide secondary reloads needed for
25251 wrapping LO_SUM addresses.
25252
d09c7dba
TS		 252532013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25254
25255 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25256 MACH, just __MACH__.
25257
a44bbd48
RB		 252582013-02-06 Richard Biener <rguenther@suse.de>
25259
25260 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25261 instead of calling fix_loop_structure.
25262
6d840d99
JJ		 252632013-02-06 Jakub Jelinek <jakub@redhat.com>
25264
25265 PR middle-end/56217
25266 * omp-low.c (use_pointer_for_field): Return false if
25267 lower_send_shared_vars doesn't generate any copy-out code.
25268
0f33baa9
TV		 252692013-02-06 Tom de Vries <tom@codesourcery.com>
25270
25271 PR rtl-optimization/56131
25272 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25273 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25274 of the label is NULL. Add comment.
25275
11f1e3ab
JJ		 252762013-02-05 Jakub Jelinek <jakub@redhat.com>
25277
31502f9f
JJ		 25278 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25279
a74db9bd
JJ		 25280 PR sanitizer/55374
25281 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25282 (STATIC_LIBTSAN_LIBS): Likewise.
25283 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25284 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25285 is defined, don't add anything else beyond that.
25286 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25287 (LINK_COMMAND_SPEC): Use them.
25288
11f1e3ab
JJ		 25289 PR tree-optimization/56205
25290 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25291 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25292 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25293
14ac6aa2
RB		 252942013-02-05 Richard Biener <rguenther@suse.de>
25295
25296 PR tree-optimization/53342
25297 PR tree-optimization/53185
25298 * tree-vectorizer.h (vect_check_strided_load): Remove.
25299 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25300 not disallow peeling for vectorized strided loads.
25301 (vect_check_strided_load): Make static and simplify.
25302 (vect_analyze_data_refs): Adjust.
25303 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25304 correctly when vectorizing strided loads.
25305
6f22445a
RB		 253062013-02-05 Richard Biener <rguenther@suse.de>
25307
25308 * doc/install.texi: Refer to ISL, not PPL.
25309
39f9719e
JH		 253102013-02-05 Jan Hubicka <jh@suse.cz>
25311
ec4224ac
JH		 25312 PR tree-optimization/55789
25313 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25314
253152013-02-05 Jan Hubicka <jh@suse.cz>
25316
25317 PR tree-optimization/55789
39f9719e
JH		 25318 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25319 the dead call anyway.
25320
956d3b33
EB		 253212013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25322
25323 PR sanitizer/55374
25324 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25325
253262013-02-04 Alexander Potapenko <glider@google.com>
25327 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP		 25328 Jakub Jelinek <jakub@redhat.com>
25329
25330 PR sanitizer/55617
25331 * config/darwin.c (sort_ctor_records): Stabilized qsort
25332 on constructor priority by using original position.
25333 (finalize_ctors): New routine to sort constructors by
25334 priority before use in assemble_integer.
25335 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25336
7ac3af38
JJ		 253372013-02-04 Jakub Jelinek <jakub@redhat.com>
25338
25339 PR libstdc++/54314
25340 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25341 about visibility on artificial decls.
25342 * config/sol2.c (solaris_assemble_visibility): Likewise.
25343
152689dc
KT		 253442013-02-04 Kai Tietz <ktietz@redhat.com>
25345
25346 PR target/56186
25347 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25348 argument and improve checking of return-argument types for 16-byte
25349 modes.
25350 (ix86_function_value_1): Add additional valtype argument on call
25351 of function_value_64.
25352 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25353 handling infunction_value_64 function.
25354
2b5987b5
MGD		 253552013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25356
7ac3af38 25357 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 25358
14597080
RB		 253592013-02-04 Richard Biener <rguenther@suse.de>
25360
25361 PR tree-optimization/56188
25362 * tree-ssa-structalias.c (label_visit): Consider case with
25363 initially non-empty points-to set.
25364 (perform_var_substitution): Dump node mapping and clean up.
25365
ed73881e
RG		 253662013-02-04 Richard Guenther <rguenther@suse.de>
25367
25368 PR lto/56168
25369 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25370 node prevail as last resort.
25371 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 25372 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 25373
139a0707
RB		 253742013-02-04 Richard Biener <rguenther@suse.de>
25375
25376 PR tree-optimization/56113
25377 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25378 Merge into ...
25379 (equiv_class_lookup_or_add): ... this.
25380 (label_visit): Adjust and fix error in previous patch.
25381 (perform_var_substitution): Adjust.
25382
50fe8924
OE		 253832013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25384
25385 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25386 * config/sh/sh4-300.md: Likewise.
25387 * config/sh/sh4a.md: Likewise.
25388 * config/sh/constraints.md: Likewise.
25389 * config/sh/sh.md: Likewise.
25390 * config/sh/netbsd-elf.h: Likewise.
25391 * config/sh/predicates.md: Likewise.
25392 * config/sh/sh-protos.h: Likewise.
25393 * config/sh/ushmedia.h: Likewise.
25394 * config/sh/linux.h: Likewise.
25395 * config/sh/sh.c: Likewise.
25396 * config/sh/superh.h: Likewise.
25397 * config/sh/elf.h: Likewise.
25398 * config/sh/sh4.md: Likewise.
25399 * config/sh/sh.h: Likewise.
25400
1a04ac2b
JDA		 254012013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25402
25403 * config/pa/constraints.md: Adjust unused letters. Change "T"
25404 constraint to match_test floating_point_store_memory_operand().
25405 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25406 (base14_operand): New.
25407 (floating_point_store_memory_operand): New.
25408 (integer_store_memory_operand): Revise to use base14_operand and
25409 reg_plus_base_memory_operand.
25410 (move_dest_operand): Allow symbolic_memory_operands.
25411 (symbolic_memory_operand): Check for LO_SOM.
25412 (symbolic_operand): Change default case to break.
25413 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25414 CONST_DOUBLE values to be reloaded by putting them into memory when
25415 the destination is a floating point register.
25416 (movdf): Remove code to handle CONST_DOUBLE.
25417 (movsf): Likewise.
25418 (reload_indf_r1): New.
25419 (reload_insf_r1): New.
25420 Consistently use "Q" and "T" constraints with integer and floating
25421 point move instructions, respectively.
25422 (movdi): Remove FAIL.
25423 Change predicate for source operand unamed DImode move from
25424 general_operand to move_src_operand.
25425 (umulsidi3): Change predicate for destination operand to
25426 register_operand.
25427 Likewise for similar unamed patterns.
25428 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25429 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25430 (hppa_legitimize_address): Simplify mask calculation.
25431 (pa_emit_move_sequence): Revised handling of secondary reloads from
25432 REG+D addresses for floating point loads and stores. Directly handle
25433 loading CONST0_RTX (mode) to a floating point register.
25434 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25435 to floating point registers. Don't restrict secondary reloads to
25436 floating point registers to integer modes. Revise some comments and
25437 cleanup some code.
25438 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25439 (pa_legitimate_address_p): New.
25440 (pa_legitimize_reload_address): New.
25441 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25442 (STRICT_REG_OK_FOR_BASE_P): New.
25443 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25444 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25445
3bf65591
DE		 254462013-02-03 David Edelsohn <dje.gcc@gmail.com>
25447 Andrew Dixie <andrewd@gentrack.com>
25448
25449 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25450 flag set.
25451
c0a8a3e6
RS		 254522013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25453
25454 * expmed.c (extract_bit_field_1): Pass the full width of the
25455 structure to get_best_reg_extraction_insn.
25456
99113dff
DE		 254572013-02-01 David Edelsohn <dje.gcc@gmail.com>
25458
25459 PR target/54601
25460 * configure.ac (use_cxa_atexit): Add AIX.
25461 * configure: Regenerate.
25462
25463 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25464
ddd84654
JJ		 254652013-02-01 Jakub Jelinek <jakub@redhat.com>
25466
25467 PR debug/54793
25468 * final.c (need_profile_function): New variable.
25469 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25470 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25471 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25472 notes, targetm.asm_out.function_prologue doesn't emit anything,
25473 HAVE_prologue and profiler should be emitted before prologue,
25474 set need_profile_function instead of emitting it.
25475 (final_scan_insn): If need_profile_function, emit
25476 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25477 NOTE_INSN_FUNCTION_BEG note.
25478
254792013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH		 25480
25481 * config/rs6000/rs6000.md (smulditi3): New.
25482 (umulditi3): New.
25483
ff2a9d88
RH		 25484 * config/alpha/alpha.md (umulditi3): New.
25485
14d52b90
DE		 254862013-02-01 David Edelsohn <dje.gcc@gmail.com>
25487
25488 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25489 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25490
8c7ca45c
RB		 254912013-02-01 Richard Biener <rguenther@suse.de>
25492
25493 PR tree-optimization/56113
25494 * tree-ssa-structalias.c (label_visit): Reduce work for
25495 single-predecessor nodes.
25496
9f419393
EB		 254972013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25498
25499 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25500 range isn't testing for zero.
25501
c5f4be84
SB		 255022013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25503
25504 PR middle-end/56113
25505 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25506
dbdbd982
NC		 255072013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25508 Nick Clifton <nickc@redhat.com>
25509
25510 * config/v850/constraints.md (Q): Define as a memory constraint.
25511 * config/v850/predicates.md (label_ref_operand): New predicate.
25512 (e3v5_shift_operand): New predicate.
25513 (ior_operator): New predicate.
25514 * config/v850/t-v850: Add e3v5 multilib.
25515 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25516 (v850_gen_movdi): Prototype.
25517 * config/v850/v850.c: Add support for e3v5 architecture.
25518 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25519 TARGET_V850E_UP.
25520 (construct_save_jarl): Add e3v5 long JARL support.
25521 (v850_adjust_insn_length): New function. Adjust length of call
25522 insns when using e3v5 instructions.
25523 (v850_gen_movdi): New function: Generate instructions to move a
25524 DImode value.
25525 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25526 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25527 (TARGET_USE_FPU): Enable for e3v5.
25528 (CONST_OK_FOR_W): New macro.
25529 (ADJUST_INSN_LENGTH): Define.
25530 * config/v850/v850.md (UNSPEC_LOOP): Define.
25531 (attr cpu): Add v850e3v5.
25532 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25533 (movdi): New pattern.
25534 (movdi_internal): New pattern.
25535 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25536 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25537 (cstoresf4): Likewise.
25538 (cstoredf4): Likewise.
25539 (insv): New pattern.
25540 (rotlso3_a): New pattern.
25541 (rotlsi3_b): New pattern
25542 (rotlsi3_v850e3v5): New pattern.
25543 (doloop_begin): New pattern.
25544 (fix_loop_counter): New pattern.
25545 (doloop_end): New pattern.
25546 (branch_normal): Add e3v5 long branch support.
25547 (branch_invert): Likewise.
25548 (branch_z_normal): Likewise.
25549 (branch_z_invert): Likewise.
25550 (branch_nz_normal): Likewise.
25551 (branch_nz_invert): Likewise.
25552 (call_internal_short): Add e3v5 register-indirect JARL support.
25553 (call_internal_long): Likewise.
25554 (call_value_internal_short): Likewise.
25555 (call_value_internal_long): Likewise.
25556 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25557 (mloop): New option.
25558 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 25559 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 25560
73861a41
PK		 255612013-01-31 Paul Koning <ni1d@arrl.net>
25562
25563 PR debug/55059
25564 PR debug/54508
25565 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25566 children if parent is a class.
25567 (prune_unused_types_prune): Don't add DW_AT_declaration.
25568
e44978dc
RB		 255692013-01-31 Richard Biener <rguenther@suse.de>
25570
25571 PR tree-optimization/56157
25572 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25573 match up operand with SLP child.
25574
90b10dec
JM		 255752013-01-31 Jason Merrill <jason@redhat.com>
25576
04d2dadd 25577 PR debug/54410
90b10dec
JM		 25578 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25579 parameters the first time.
25580 (gen_scheduled_generic_parms_dies): Check completeness here.
25581
6e616110
RB		 255822013-01-31 Richard Biener <rguenther@suse.de>
25583
25584 PR middle-end/53073
25585 * common.opt (faggressive-loop-optimizations): New flag,
25586 enabled by default.
25587 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25588 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25589 infer_loop_bounds_from_undefined by it.
25590
636f59cf
RB		 255912013-01-31 Richard Biener <rguenther@suse.de>
25592
25593 PR tree-optimization/56150
25594 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25595 visit virtual operands.
25596 (find_uses_to_rename_bb): Likewise.
25597
255982013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB		 25599
25600 PR tree-optimization/56150
25601 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25602 mixed store non-store stmts.
25603
32887460
JJ		 256042013-01-30 Jakub Jelinek <jakub@redhat.com>
25605
e60e09a0
JJ		 25606 PR sanitizer/55374
25607 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25608 LIBASAN_EARLY_SPEC is defined.
25609 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25610 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25611 before %o.
25612 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25613
32887460
JJ		 25614 PR c++/55742
25615 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25616 invalid args instead of ICEing on it.
25617 (ix86_valid_target_attribute_tree): Return error_mark_node if
25618 ix86_valid_target_attribute_inner_p failed.
25619 (ix86_valid_target_attribute_p): Return false only if
25620 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25621 target("default") attribute.
25622 (sorted_attr_string): Change argument from const char * to tree,
25623 merge in all target attribute arguments rather than just one.
25624 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25625 instead of free. Avoid using strcat.
25626 (ix86_mangle_function_version_assembler_name): Mangle
25627 target("default") as if no target attribute is present. Adjust
25628 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25629 instead of xmalloc and XDELETEVEC instead of free.
25630 (ix86_function_versions): Don't return true if one of the decls
25631 doesn't have target attribute. If they don't and one of the decls
25632 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25633 sorted_attr_string caller. Use XDELETEVEC instead of free.
25634 (ix86_supports_function_versions): Remove.
25635 (make_name): Fix up formatting.
25636 (make_dispatcher_decl): Remove resolver_name and its initialization.
25637 Avoid leaking memory.
25638 (is_function_default_version): Return true if there is
25639 target("default") attribute rather than no target attribute at all.
25640 (make_resolver_func): Avoid leaking memory.
25641 (ix86_generate_version_dispatcher_body): Likewise.
25642 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25643 * target.def (supports_function_versions): Remove.
25644 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25645 * doc/tm.texi: Regenerated.
25646
73cca0cc
VM		 256472013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25648
25649 PR rtl-optimization/56144
25650 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25651 for values with side effects.
25652
11452e7b
RB		 256532013-01-30 Richard Biener <rguenther@suse.de>
25654
25655 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25656 (sparseset_pop): Likewise.
25657 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25658 to be able to use quick_push in the worker loop.
25659
01cb1ef5
MP		 256602013-01-30 Marek Polacek <polacek@redhat.com>
25661
25662 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25663
e892936e
RB		 256642013-01-30 Richard Biener <rguenther@suse.de>
25665
25666 PR lto/56147
6edc3e32 25667 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 25668
cc06c01d
GJL		 256692013-01-30 Georg-Johann Lay <avr@gjlay.de>
25670
25671 PR tree-optimization/56064
25672 * fixed-value.c (fixed_from_double_int): New function.
25673 * fixed-value.h (fixed_from_double_int): New prototype.
25674 (const_fixed_from_double_int): New static inline function.
25675 * fold-const.c (native_interpret_fixed): New static function.
25676 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25677 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25678 (native_encode_fixed): New static function.
25679 (native_encode_expr) <FIXED_CST>: Use it.
25680 (native_interpret_int): Move double_int worker code to...
25681 * double-int.c (double_int::from_buffer): ...this new static method.
25682 * double-int.h (double_int::from_buffer): Prototype it.
25683
d394a308
RB		 256842013-01-30 Richard Biener <rguenther@suse.de>
25685
25686 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25687 New pointer-map and obstack.
25688 (init_alias_vars): Allocate pointer-map and obstack.
25689 (delete_points_to_sets): Free them.
25690 (find_what_var_points_to): Cache result.
25691 (find_what_p_points_to): Adjust for changed interface of
25692 find_what_var_points_to.
25693 (compute_points_to_sets): Likewise.
25694 (ipa_pta_execute): Likewise.
25695
20804d96
RO		 256962013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25697
25698 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25699 * configure: Regenerate.
25700 * config.in: Regenerate.
25701 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25702 #nobits/#progbits if supported.
25703
a7ad88a2
OE		 257042013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25705
25706 PR target/56121
25707 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25708 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25709 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25710
7af79f92
GY		 257112013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25712
91bfca59
OE		 25713 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25714 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 25715
753bcf7b
GY		 257162013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25717
91bfca59
OE		 25718 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25719 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 25720
8cbc2ea8
GY		 257212013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25722
91bfca59
OE		 25723 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25724 declaration.
25725 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25726 * config/arm/cortex-a7.md: New bypasses using
25727 arm_mac_accumulator_is_result.
8cbc2ea8 25728
697a3325
GY		 257292013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25730
91bfca59 25731 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 25732 (cortex_a7_neon_mla): Likewise.
91bfca59
OE		 25733 (cortex_a7_fpfmad): New reservation.
25734 (cortex_a7_fpmacs): Use ffmas and update required units.
25735 (cortex_a7_fpmuld): Update required units and latency.
25736 (cortex_a7_fpmacd): Likewise.
25737 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25738 (cortex_a7_neon). Likewise.
25739 (bypass) Update participating units.
697a3325 25740
29637783
GY		 257412013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25742
91bfca59
OE		 25743 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25744 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25745 from fmac to ffma.
25746 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25747 (vfp_fmul): Use ffmad.
25748 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25749 (cortex_r4_fmacd): Use ffmad.
25750 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25751 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25752 (cortex_a9_fmacd): Use ffmad.
25753 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25754 (cortex_a8_vfp_macd): Use ffmad.
25755 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25756 (cortex_a5_fpmacd): Use ffmad.
25757 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25758 (cortex_a15_vfp_macd): Use ffmad.
25759 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 25760
2ee8a2d5
JM		 257612013-01-29 Jason Merrill <jason@redhat.com>
25762
25763 PR libstdc++/54314
25764 * varasm.c (default_assemble_visibility): Don't warn about
25765 visibility on artificial decls.
25766
78d087bc
RB		 257672013-01-29 Richard Biener <rguenther@suse.de>
25768
25769 PR tree-optimization/56113
25770 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25771 the bitmap leader.
25772 (label_visit): Free duplicate bitmaps and record the leader instead.
25773 (perform_var_substitution): Adjust.
25774
83ba4d6f
RB		 257752013-01-29 Richard Biener <rguenther@suse.de>
25776
25777 PR tree-optimization/55270
25778 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25779 the CFG, schedule loops for fixup.
25780
66dfe13f
NC		 257812013-01-29 Nick Clifton <nickc@redhat.com>
25782
25783 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25784 SP_REG.
25785
5a579c3b
LE		 257862013-01-28 Leif Ekblad <leif@rdos.net>
25787
25788 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25789 * config/i386/i386.h (TARGET_RDOS): New macro.
25790 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25791 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25792 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25793 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25794 DEFAULT_LARGE_SECTION_THRESHOLD.
25795 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25796 * config/i386/rdos.h: New file.
25797 * config/i386/rdos64.h: New file.
25798
51e44392
BS		 257992013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25800
25801 PR other/54814
25802 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25803 TEST_HARD_REG_BIT.
25804
db1fb332
JJ		 258052013-01-28 Jakub Jelinek <jakub@redhat.com>
25806
25807 PR rtl-optimization/56117
25808 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25809 call cselib_lookup_from_insn on the MEM before calling
25810 add_insn_mem_dependence.
25811
16917761
RB		 258122013-01-28 Richard Biener <rguenther@suse.de>
25813
25814 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25815 to a stmt that didn't have one.
25816 (copy_phis_for_bb): Likewise for PHI arguments.
25817 (copy_debug_stmt): Likewise for debug stmts.
25818
b9fc0497
RB		 258192013-01-28 Richard Biener <rguenther@suse.de>
25820
25821 PR tree-optimization/56034
6edc3e32 25822 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB		 25823 (partition_builtin_p): Adjust.
25824 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25825 it is the last partition.
25826 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25827 up the vertex for the definition.
25828 (classify_partition): Classify whether a partition is a
25829 PKIND_REDUCTION, thus has uses outside of the loop.
25830 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25831 Merge all PKIND_REDUCTION partitions into the last partition.
25832 (tree_loop_distribution): Seed partitions from reductions as well.
25833
aa710d25
JJ		 258342013-01-28 Jakub Jelinek <jakub@redhat.com>
25835
0bfbca58
JJ		 25836 PR tree-optimization/56125
25837 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25838 pow(x,c) into sqrt(x) * powi(x, n/2) or
25839 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25840 optimizing for size.
25841 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25842 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25843 integer.
25844
aa710d25
JJ		 25845 PR tree-optimization/56094
25846 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25847 to UNKNOWN_LOCATION while gimplifying expr.
25848
77dc5297
UB		 258492013-01-27 Uros Bizjak <ubizjak@gmail.com>
25850
25851 PR target/56114
25852 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25853 operand 0 in movabs insn template for -masm=intel asm alternative.
25854 (*movabs<mode>_2): Ditto for operand 1.
25855
0bfbca58 258562013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH		 25857
25858 PR target/54663
25859 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25860 of microblaze-c.o
25861
0bfbca58 258622013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI		 25863
25864 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25865 tm_file.
25866
0bfbca58 258672013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N		 25868
25869 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25870 Undef to avoid warning.
25871
478f60f9
MH		 258722013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25873
25874 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25875 * configure: Regenerate.
25876
d7fa6ee2
JJ		 258772013-01-25 Jakub Jelinek <jakub@redhat.com>
25878
25879 PR tree-optimization/56098
25880 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25881 for stmts with volatile ops.
25882 (cond_store_replacement): Don't optimize if assign has volatile ops.
25883 (cond_if_else_store_replacement_1): Don't optimize if either
25884 then_assign or else_assign have volatile ops.
25885 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25886 volatile ops.
25887
f8fe87bd
GJL		 258882013-01-25 Georg-Johann Lay <avr@gjlay.de>
25889
25890 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25891
28398d0d
GJL		 258922013-01-25 Georg-Johann Lay <avr@gjlay.de>
25893
25894 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25895 missing ':' in asm example.
25896
b7d7d917
TB		 258972013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25898
25899 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25900 entries into lane and laneq entries.
77dc5297
UB		 25901 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25902 Remove AdvSIMD scalar modes.
b7d7d917
TB		 25903 (aarch64_sq<r>dmulh_laneq<mode>): New.
25904 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25905 modes.
25906 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25907 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25908 * config/aarch64/iterators.md (VCOND): New.
25909 (VCONQ): New.
25910
556f9906
GJL		 259112013-01-25 Georg-Johann Lay <avr@gjlay.de>
25912
25913 PR target/54222
25914 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25915 Add NULL LIBNAME argument to existing definitions.
25916 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25917 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25918 * config/avr/avr.c (DEF_BUILTIN): Same.
25919 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25920 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25921 implementation is available (DECL_ASSEMBLER_NAME is set).
25922 (avr_fold_absfx): New static function.
25923 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25924 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25925 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25926 AVR_BUILTIN_ABSLLK.
25927 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25928 (abshk, absk, abslk, absllk): Provide as static inline functions.
25929
1f546bbb
MP		 259302013-01-25 Marek Polacek <polacek@redhat.com>
25931
25932 PR tree-optimization/56035
25933 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25934
7e184bd7
UB		 259352012-01-24 Uros Bizjak <ubizjak@gmail.com>
25936
25937 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25938 (*movtf_internal_rex64): Add (!o,C) alternative
25939 (*movxf_internal_rex64): Ditto.
25940 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25941
f5ba49ac
SM		 259422013-01-24 Shenghou Ma <minux.ma@gmail.com>
25943
25944 * doc/invoke.texi: fix typo.
25945 * doc/objc.texi: fix typo.
25946
a6343728
RS		 259472013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25948
25949 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25950 for the first two alternatives.
25951
cd030c07
DN		 259522013-01-24 Diego Novillo <dnovillo@google.com>
25953
77dc5297 25954 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN		 25955 (ggc-zone.o): Remove.
25956 * configure.ac: Remove option --with-gc.
25957 * configure: Re-generate.
25958 * doc/install.texi: Remove documentation for --with-gc.
25959 * gengtype.c (write_enum_defn): Remove. Update all users.
25960 (write_Types_process_field): Remove generation of gt_e_* argument.
25961 (output_type_enum): Remove. Update all users.
25962 (write_enum_defn): Remove. Update all users.
25963 (enum alloc_zone): Remove. Update all users.
77dc5297 25964 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN		 25965 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25966 Update all callers.
25967 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 25968 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN		 25969 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25970 Update all users.
aaf1e810 25971 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN		 25972 (struct alloc_zone): Remove.
25973 (ggc_internal_alloc_zone_stat): Remove.
25974 (ggc_internal_cleared_alloc_zone_stat): Remove.
25975 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB		 25976 (ggc_pch_count_object): Remove last argument. Update all users.
25977 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN		 25978 (struct alloc_zone): Remove.
25979 * ggc-zone.c: Remove.
77dc5297 25980 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN		 25981 (struct alloc_zone): Remove.
25982 (ggc_alloc_typed_stat): Remove.
25983 (ggc_alloc_typed): Remove.
25984 (ggc_splay_alloc): Remove first argument.
25985 (rtl_zone): Remove. Update all users.
25986 (tree_zone): Remove. Update all users.
25987 (tree_id_zone): Remove. Update all users.
25988 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 25989 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 25990 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 25991 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 25992
a861ffa4
GJL		 259932013-01-24 Georg-Johann Lay <avr@gjlay.de>
25994
25995 * config/avr/avr.c (avr_out_fract): Make register numbers that
25996 might be outside of source operand signed.
25997
593c0ddd
UB		 259982013-01-24 Uros Bizjak <ubizjak@gmail.com>
25999
26000 * config/i386/constraints.md (Yf): New constraint.
26001 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26002 of f constraint to conditionaly disable x87 register preferences.
26003 (*movdf_internal): Ditto.
26004 (*movsf_internal): Ditto.
26005
e86c0101
SB		 260062013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26007
26008 PR inline-asm/55934
26009 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26010 that have operands with impossible constraints.
26011 Add a FIXME for a speed-up opportunity.
26012 * lra-constraints.c (process_alt_operands): Verify that a class
26013 selected from constraints on asms is valid for the operand mode.
26014 (curr_insn_transform): Remove incorrect comment.
26015
f6fee35f
DE		 260162013-01-23 David Edelsohn <dje.gcc@gmail.com>
26017
26018 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26019 TOC operand is a valid symbol ref in the constant pool.
26020
260212013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 26022
aaf1e810 26023 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 26024
dc62d7d1
GJL		 260252013-01-23 Georg-Johann Lay <avr@gjlay.de>
26026
26027 PR target/54222
26028 * config/avr/stdfix.h: New file.
26029 * t-avr (stdfix-gcc.h): New rule to build it.
26030 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26031
2660d12d
KS		 260322013-01-23 Kostya Serebryany <kcc@google.com>
26033
77dc5297
UB		 26034 * config/darwin.h: remove dependency on
26035 CoreFoundation (asan on Mac OS).
2660d12d 26036
a70418fc
JJ		 260372013-01-23 Jakub Jelinek <jakub@redhat.com>
26038
26039 PR target/49069
26040 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26041 instead of cmpdi_operand for first comparison operand.
26042 Don't assert that comparison operands aren't both constants.
26043
47876a2a
JW		 260442013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26045
26046 * doc/install.texi (Downloading the Source): Update references to
26047 downloading separate components.
26048
260492013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW		 26050
26051 * doc/extend.texi (__int128): Improve grammar.
26052
47876a2a 260532013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB		 26054
26055 PR target/56028
26056 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26057 alternative to (o,r).
26058 (*movdi_internal_rex64): Remove (!o,n) alternative.
26059 (DImode immediate->memory splitter): Remove.
26060 (DImode immediate->memory peephole2): Remove.
26061 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26062 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26063 alternative to (!o,*r).
26064 (*movtf_internal_sse): New pattern.
26065 (*movxf_internal_rex64): New pattern.
26066 (*movxf_internal): Disable for TARGET_64BIT.
26067 (*movdf_internal_rex64): Remove (!o,F) alternative.
26068
3a984f10
JJ		 260692013-01-22 Jakub Jelinek <jakub@redhat.com>
26070
502498d5
JJ		 26071 PR middle-end/56074
26072 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26073 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26074 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26075 stmt locations where LOCATION_LOCUS of the stmt location is
26076 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26077
3a984f10
JJ		 26078 PR target/55686
26079 * config/i386/i386.md (UNSPEC_STOS): New.
26080 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26081 *strsetqi_1): Add UNSPEC_STOS.
26082
fa817f7f
PC		 260832013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26084
26085 PR c++/56067
26086 * doc/invoke.texi: Remove left over -Wsynth example.
26087
8f498c1b
JJ		 260882013-01-21 Jakub Jelinek <jakub@redhat.com>
26089
26090 PR tree-optimization/56051
26091 * fold-const.c (fold_binary_loc): Don't fold
26092 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26093 a narrowing conversion, or widening conversion from signed
26094 to unsigned.
26095
47876a2a 260962013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB		 26097
26098 PR rtl-optimization/56023
26099 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26100 dependent on debug instruction.
26101
5b9db1bc
MJ		 261022013-01-21 Martin Jambor <mjambor@suse.cz>
26103
26104 PR middle-end/56022
26105 * function.c (allocate_struct_function): Call
26106 invoke_set_current_function_hook earlier.
26107
e8bb7d68
JJ		 261082013-01-21 Jakub Jelinek <jakub@redhat.com>
26109
26110 * reload1.c (init_reload): Only initialize reload_obstack
26111 during the first call.
26112
616a4e32
MP		 261132013-01-21 Marek Polacek <polacek@redhat.com>
26114
26115 * cfgloop.c (verify_loop_structure): Fix up grammar.
26116
4401981b
YHH		 261172013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26118
26119 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26120 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26121
8e87740b
RR		 261222013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26123
26124 PR target/56058
26125 * config/arm/marvell-pj4.md: Update copyright year.
26126 Fix up use of alu to alu_reg and simple_alu_imm.
26127
47876a2a 261282013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB		 26129
26130 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26131
89d56d79
VM		 261322013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26133
26134 PR target/55433
26135 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 26136 insn for secondary memory move when memory mode should be different.
89d56d79 26137
fe603553
JDA		 261382013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26139
26140 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26141 atomic_storedi_1): New patterns.
26142
01284895
VK		 261432013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26144
26145 btver2 pipeline descriptions.
26146 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26147 descriptions.
26148 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 26149 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK		 26150 type attributes.
26151 * config/i386/btver2.md: New file describing btver2 pipelines.
26152
5630e3e1
JL		 261532013-01-19 Andrew Pinski <apinski@cavium.com>
26154
26155 PR tree-optimization/52631
26156 * tree-ssa-sccvn (visit_use): Before looking up the original
26157 statement, try looking up the simplified expression.
26158
650ae806
AG		 261592013-01-19 Anthony Green <green@moxielogic.com>
26160
26161 * config/moxie/moxie.c (moxie_expand_prologue): Set
26162 current_function_static_stack_size.
26163
e300ec2d
JJ		 261642013-01-18 Jakub Jelinek <jakub@redhat.com>
26165
26166 PR tree-optimization/56029
26167 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26168 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26169
a3d7af04
SS		 261702013-01-18 Sharad Singhai <singhai@google.com>
26171
26172 PR tree-optimization/55995
26173 * dumpfile.c (dump_loc): Print location only if available.
26174 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26175
66aa7879
VM		 261762013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26177
26178 PR target/55433
26179 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26180 secondary memory move.
26181 (inherit_reload_reg): Use rclass instead of cl for
26182 check_secondary_memory_needed_p.
26183
3f0fee7b
JJ		 261842013-01-18 Jakub Jelinek <jakub@redhat.com>
26185
26186 PR middle-end/56015
26187 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 26188 the case where writing real complex part of target modifies op1.
3f0fee7b 26189
70c67693
JG		 261902013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26191
26192 * config/aarch64/aarch64-simd.md
26193 (aarch64_vcond_internal<mode>): Handle unordered cases.
26194 * config/aarch64/iterators.md (v_cmp_result): New.
26195
df8de9b3
YHH		 261962013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26197 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26198
26199 * config/arm/marvell-pj4.md: New file.
26200 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26201 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26202 (generic_vfp): Likewise.
26203 * config/arm/arm-cores.def: Add marvell-pj4.
26204 * config/arm/arm-tune.md: Regenerate.
26205 * config/arm/arm-tables.opt: Regenerate.
26206 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26207 * doc/invoke.texi: Document marvell-pj4.
26208
be30c356
TB		 262092013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26210
26211 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26212
0bfbca58 262132013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO		 26214
26215 PR debug/54114
26216 PR debug/54402
26217 PR debug/49888
26218 * var-tracking.c (negative_power_of_two_p): New.
26219 (global_get_addr_cache, local_get_addr_cache): New.
26220 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26221 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26222 heading comment.
26223 (vt_stack_offset_p): Remove.
26224 (vt_canon_true_dep): Always canonicalize loc's address.
26225 (clobber_overlapping_mems): Make sure we have a MEM.
26226 (local_get_addr_clear_given_value): New.
26227 (val_reset): Clear local cached entries.
26228 (compute_bb_dataflow): Create and release the local cache.
26229 Disable duplicate MEMs clobbering.
26230 (emit_notes_in_bb): Clobber MEMs likewise.
26231 (vt_emit_notes): Create and release the local cache.
26232 (vt_initialize, vt_finalize): Create and release the global
26233 cache, respectively.
1f6bc337 26234 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 26235
0bfbca58 262362013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO		 26237
26238 PR libmudflap/53359
26239 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26240 not found in the symtab.
26241
0bfbca58 262422013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 26243
c350ba53 26244 PR debug/56006
3aa03517
AO		 26245 PR rtl-optimization/55547
26246 PR rtl-optimization/53827
26247 PR debug/53671
26248 PR debug/49888
26249 * alias.c (offset_overlap_p): New, factored out of...
26250 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26251 the conservative special case for symbolic constants. Don't
26252 adjust zero sizes on alignment.
26253
c664546f
JL		 262542013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26255
26256 PR rtl-optimization/52573
26257 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26258 REG_UNUSED for the same register.
26259
1bd3f750
MP		 262602013-01-17 Richard Biener <rguenther@suse.de>
26261 Marek Polacek <polacek@redhat.com>
26262
26263 PR rtl-optimization/55833
26264 * loop-unswitch.c (unswitch_loops): Move loop verification...
26265 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26266 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26267 Set it to true when we're removing a loop from hierarchy tree in
26268 an irreducible region.
26269 (fix_bb_placements): Adjust caller.
26270 (fix_loop_placements): Likewise.
26271
e52a8b71
GJL		 262722013-01-17 Georg-Johann Lay <avr@gjlay.de>
26273
26274 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26275 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26276 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26277 Remove ID. Adjust comments.
26278 * config/avr/avr-c.c (avr_builtin_name): Remove.
26279 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26280 * config/avr/avr.c (avr_tolower): New static function.
26281 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26282 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26283 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26284 default expansion.
26285
8386a7ea
JH		 262862013-01-17 Jan Hubicka <jh@suse.cz>
26287
610fb637 26288 PR tree-optimization/55273
8386a7ea
JH		 26289 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26290
47876a2a 262912013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB		 26292
26293 PR target/55981
26294 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26295 store through atomic_store<mode>_1.
26296 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26297
8222c37e
MJ		 262982013-01-17 Martin Jambor <mjambor@suse.cz>
26299
26300 PR tree-optimizations/55264
26301 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26302 for virtual methods.
26303 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26304 virtual methods before inlining is over.
26305 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26306 virtual functions.
26307 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26308 non-virtual.
26309
79f01c76
VM		 263102013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26311
26312 PR rtl-optimization/56005
26313 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26314 pending reads for prefetch.
26315
d0b6bb1b
IB		 263162013-01-16 Ian Bolton <ian.bolton@arm.com>
26317
aaf1e810 26318 * config/aarch64/aarch64.md
d0b6bb1b
IB		 26319 (*cstoresi_neg_uxtw): New pattern.
26320 (*cmovsi_insn_uxtw): New pattern.
26321 (*<optab>si3_uxtw): New pattern.
26322 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26323 (*<optab>si3_insn_uxtw): New pattern.
26324 (*bswapsi2_uxtw): New pattern.
26325
cb9cf03b
RB		 263262013-01-16 Richard Biener <rguenther@suse.de>
26327
26328 * tree-inline.c (tree_function_versioning): Remove set but
26329 never used variable.
26330
2cfc56b9
RB		 263312013-01-16 Richard Biener <rguenther@suse.de>
26332
26333 PR tree-optimization/55964
26334 * tree-flow.h (rename_variables_in_loop): Remove.
26335 (rename_variables_in_bb): Likewise.
26336 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26337 (copy_loop_before): Adjust and delete update-ssa status.
26338 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 26339 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB		 26340 (rename_variables_in_loop): Remove.
26341 (slpeel_update_phis_for_duplicate_loop): Likewise.
26342 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26343 use available cfg machinery instead of duplicating it.
26344 Update PHI nodes and perform poor-mans SSA update here.
26345 (slpeel_tree_peel_loop_to_edge): Adjust.
26346
c25a0c60
RB		 263472013-01-16 Richard Biener <rguenther@suse.de>
26348
26349 PR tree-optimization/54767
26350 PR tree-optimization/53465
26351 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26352 (vrp_visit_phi_node): For PHI arguments coming via backedges
26353 drop all symbolical range information.
26354 (execute_vrp): Compute backedges.
26355
04b535af
RB		 263562013-01-16 Richard Biener <rguenther@suse.de>
26357
26358 * doc/install.texi: Update CLooG and ISL requirements to
26359 0.18.0 and 0.11.1.
26360
8b0a1e0b
CB		 263612013-01-16 Christian Bruel <christian.bruel@st.com>
26362
26363 PR target/55301
26364 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26365 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26366 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26367
263682013-01-16 DJ Delorie <dj@redhat.com>
26369
26370 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26371 (UNSPECV_SP_SWITCH_E): New.
26372 (sp_switch_1): Change to an unspec.
26373 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26374 replace $r15.
26375
263762013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB		 26377
26378 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26379 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26380 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26381 (expand_mem_thread_fence): Ditto.
26382 (expand_mem_signal_fence): Ditto.
26383 (expand_atomic_load): Ditto.
26384 (expand_atomic_store): Ditto.
26385
0bfbca58 263862013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO		 26387
26388 PR rtl-optimization/55547
26389 PR rtl-optimization/53827
26390 PR debug/53671
26391 PR debug/49888
26392 * alias.c (memrefs_conflict_p): Set sizes to negative after
26393 AND adjustments.
26394
305e3ac1
JJ		 263952013-01-15 Jakub Jelinek <jakub@redhat.com>
26396
26397 PR target/55940
26398 * function.c (thread_prologue_and_epilogue_insns): Always
26399 add crtl->drap_reg to set_up_by_prologue.set, even if
26400 stack_realign_drap is false.
26401
f78ac4f2
JBG		 264022013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26403
26404 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26405 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26406 *call): Fix indention.
26407
a78a8cc4
TV		 264082013-01-15 Tom de Vries <tom@codesourcery.com>
26409
26410 PR target/55876
26411 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26412 Update comment.
26413
0e80383f
VM		 264142013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26415
305e3ac1 26416 PR rtl-optimization/55153
0e80383f
VM		 26417 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26418
207b5956
MJ		 264192013-01-15 Martin Jambor <mjambor@suse.cz>
26420
26421 PR tree-optimization/55920
26422 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26423 accesses as grp_to_be_debug_replaced.
26424
a7818b54
JJ		 264252013-01-15 Jakub Jelinek <jakub@redhat.com>
26426
26427 PR tree-optimization/55920
26428 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26429 there is non-useless type conversion needed from debug rhs to lhs,
26430 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26431
b0fe107e
JM		 264322013-01-15 Joseph Myers <joseph@codesourcery.com>
26433 Mikael Pettersson <mikpe@it.uu.se>
26434
26435 PR target/43961
26436 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26437 Thumb.
26438 (ASM_OUTPUT_CASE_LABEL): Remove.
26439 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26440 * final.c (shorten_branches): Update alignment of labels before
26441 jump tables if CASE_VECTOR_SHORTEN_MODE.
26442
34ab62ee
RB		 264432013-01-15 Richard Biener <rguenther@suse.de>
26444
26445 PR bootstrap/55961
26446 * system.h: Do not include gmp.h for building host tools.
26447
783a3a05
RB		 264482013-01-15 Richard Biener <rguenther@suse.de>
26449
26450 PR middle-end/55882
26451 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26452 account for bitpos when computing alignment.
26453
3a579e09
VY		 264542013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26455
26456 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26457 (ix86_target_macros_internal): Likewise.
26458
26459 * config/i386/i386.c (m_CORE2I7): Removed.
26460 (m_CORE_HASWELL): New macro.
26461 (m_CORE_ALL): Likewise.
26462 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26463 (initial_ix86_arch_features): Likewise.
26464 (processor_target_table): Initializations for Core avx2.
26465 (cpu_names): New names "core-avx2".
26466 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26467 PROCESSOR_CORE_HASWELL.
26468 (ix86_issue_rate): New case.
26469 (ia32_multipass_dfa_lookahead): Likewise.
26470 (ix86_sched_init_global): Likewise.
26471
26472 * config/i386/i386.h (TARGET_HASWELL): New macro.
26473 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26474 (processor_type): New PROCESSOR_HASWELL.
26475
955f5a07
JJ		 264762013-01-15 Jakub Jelinek <jakub@redhat.com>
26477
ff784829
JJ		 26478 PR tree-optimization/55955
26479 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26480 *SHIFT_EXPR and *ROTATE_EXPR codes.
26481
955f5a07
JJ		 26482 PR tree-optimization/48766
26483 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26484 -ftrapv disable -fwrapv.
26485
aeb8b4e9
GJL		 264862013-01-14 Georg-Johann Lay <avr@gjlay.de>
26487
26488 PR target/55974
26489 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26490 etc. to 1 and not to __flash.
26491 Use LL suffix for __INT24_MAX__ with -mint8.
26492 Use ULL suffix for __UINT24_MAX__ with -mint8.
26493
1c494c6a
GJL		 264942013-01-14 Georg-Johann Lay <avr@gjlay.de>
26495
26496 * config/avr/avr-arch.h
26497 (struct base_arch_s): Use typedef avr_arch_t instead.
26498 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26499 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26500 * config/avr/avr.c: Same.
26501 * config/avr/avr-devices.c: Same.
26502 * config/avr/driver-avr.c: Same.
26503 * config/avr/gen-avr-mmcu-texi.c: Same.
26504 * config/avr/avr-mcus.def: Adjust comment.
26505
a50344cb
TB		 265062013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26507
88e784e6
UB		 26508 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26509 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 26510
47876a2a 265112013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB		 26512 Andi Kleen <ak@linux.intel.com>
26513
26514 PR target/55948
26515 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26516 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26517 memmodel flag.
26518
00892272
GJL		 265192013-01-14 Georg-Johann Lay <avr@gjlay.de>
26520
26521 * config/avr/avr-stdint.h: Remove trailing blanks.
26522 * config/avr/avr-log.h: Same.
26523 * config/avr/avr-arch.h: Same.
26524 * config/avr/avr-devices.c: Same.
26525 * config/avr/avr-dimode.md: Same.
26526 * config/avr/predicates.md: Same.
26527 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 26528
00892272
GJL		 26529 * config/avr/avr-protos.h: Same. And:
26530 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26531 (init_cumulative_args): Rename to avr_init_cumulative_args.
26532 (expand_prologue): Rename to avr_expand_prologue.
26533 (expand_epilogue): Rename to avr_expand_epilogue.
26534 (adjust_insn_length): Rename to avr_adjust_insn_length.
26535 (notice_update_cc): Rename to avr_notice_update_cc.
26536 (final_prescan_insn): Rename to avr_final_prescan_insn.
26537 * config/avr/avr.c: Same.
26538 * config/avr/avr.h: Same.
26539 * config/avr/avr.md: Remove trailing blanks.
26540 (prologue): Use avr_expand_prologue.
26541 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26542
7bb01996
RB		 265432013-01-14 Richard Biener <rguenther@suse.de>
26544
26545 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26546 verify_location, collect_subblocks): New functions.
26547 (verify_gimple_in_cfg): Verify that locations only reference
26548 BLOCKs in the functions BLOCK tree.
26549
2724573f
RB		 265502013-01-14 Richard Biener <rguenther@suse.de>
26551
26552 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26553 PHI argument.
26554 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26555 unshare reference.
26556 (insert_out_of_ssa_copy_on_edge): Likewise.
26557 (rewrite_close_phi_out_of_ssa): Likewise.
26558 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26559 debug expressions.
26560 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26561 propagated constants.
26562 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26563 can not be shared.
26564
9a0bbab6
GJL		 265652013-01-14 Georg-Johann Lay <avr@gjlay.de>
26566
26567 * config/avr/avr-modes.def: Add GPL copyright notice.
26568
45805f17
UB		 265692013-01-13 Uros Bizjak <ubizjak@gmail.com>
26570
26571 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26572 MEMMODEL_MASK to determine memory model.
26573 (atomic_store<mode>): Ditto from operands[2].
26574 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26575
9d60be38
JJ		 265762013-01-13 Jakub Jelinek <jakub@redhat.com>
26577
26578 PR fortran/55935
45805f17 26579 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ		 26580 (fold_gimple_assign): Don't call unshare_expr here.
26581 (fold_ctor_reference): Call unshare_expr.
26582
e7f49d92
TG		 265832013-01-13 Terry Guo <terry.guo@arm.com>
26584
aaf1e810
EB		 26585 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26586 * doc/fragments.texi: Document MULTILIB_REUSE.
26587 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 26588 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 26589 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG		 26590 (tmpmultilib4): Ditto.
26591 (multilib_reuse): New multilib argument.
26592
fbd03a27
RS		 265932013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26594
26595 * Makefile.in: Update copyright.
26596
b3681f13
TV		 265972013-01-12 Tom de Vries <tom@codesourcery.com>
26598
26599 PR middle-end/55890
26600 * calls.c (expand_call): Check if arg_nr is valid.
26601
3f287e4b
MM		 266022013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26603
26604 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26605 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM		 26606 documentation. Add missing '__' in front of
26607 __builtin_ia32_packssdw256.
3f287e4b 26608
1abcd5eb
AK		 266092013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26610
26611 PR target/55719
26612 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26613 NO_REGS for larl operands.
26614 (s390_reload_larl_operand): Use s390_load_address instead of
26615 emit_move_insn.
26616
980d0812
RB		 266172013-01-11 Richard Biener <rguenther@suse.de>
26618
26619 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26620 (verify_node_sharing): ... here.
26621 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26622
7cb9fd07
EB		 266232013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26624
26625 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26626 Substitute TREECHECKING.
26627 * configure: Regenerate.
26628 * Makefile.in (TREECHECKING): New.
26629
47876a2a 266302013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG		 26631
26632 PR tree-optimization/44061
26633 * tree-vrp.c (extract_range_basic): Compute zero as
26634 value-range for __builtin_constant_p of function parameters.
26635
d1e082c2
RS		 266362013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26637
45805f17 26638 Update copyright years.
d1e082c2 26639
f85021b0
VM		 266402013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26641
9d60be38 26642 PR rtl-optimization/55672
0160442c
VM		 26643 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26644 const to be eliminable.
f85021b0 26645
7a8b1ec4
DE		 266462013-01-10 David Edelsohn <dje.gcc@gmail.com>
26647
26648 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26649 * configure: Regenerate.
26650
ad2e5b71
RB		 266512013-01-10 Richard Biener <rguenther@suse.de>
26652
ddf9322d 26653 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 26654
ffc5b2bb
RB		 266552013-01-10 Richard Biener <rguenther@suse.de>
26656
26657 PR bootstrap/55792
26658 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26659 locations for virtual PHI arguments.
26660 (rewrite_update_phi_arguments): Likewise.
26661
e1f674e4
JS		 266622013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26663
26664 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26665 on to assembler.
26666
a57fc743
JJ		 266672013-01-10 Jakub Jelinek <jakub@redhat.com>
26668
26669 PR tree-optimization/55921
26670 * tree-complex.c (expand_complex_asm): New function.
26671 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26672
0ff4390d
AK		 266732013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26674
26675 PR target/55718
26676 * config/s390/s390.c (s390_symref_operand_p)
26677 (s390_loadrelative_operand_p): Merge the two functions.
26678 (s390_check_qrst_address, print_operand_address): Add parameters
26679 to s390_loadrelative_operand_p invokation.
26680 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26681 (s390_reload_larl_operand, s390_secondary_reload): Use
26682 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26683 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26684
6f557e0e
MS		 266852013-01-09 Mike Stump <mikestump@comcast.net>
26686
26687 * dse.c (record_store): Remove unnecessary assert.
26688
7770c9e9
JH		 266892013-01-09 Jan Hubicka <jh@suse.cz>
26690
26691 PR tree-optimization/55569
26692 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26693 * cfgloop.h (scale_loop_profile): Likewise.
26694
a19b1432
JH		 266952013-01-09 Jan Hubicka <jh@suse.cz>
26696
26697 PR lto/45375
ddf9322d
UB		 26698 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26699 functions.
a19b1432
JH		 26700 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26701
9a002da8
RS		 267022013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26703
26704 PR middle-end/55114
26705 * expr.h (maybe_emit_group_store): Declare.
26706 * expr.c (maybe_emit_group_store): New function.
26707 * builtins.c (expand_builtin_int_roundingfn): Call it.
26708 (expand_builtin_int_roundingfn_2): Likewise.
26709
511dcace
VM		 267102013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26711
e1f2b729 26712 PR rtl-optimization/55829
511dcace
VM		 26713 * lra-constraints.c (match_reload): Add code for absent output.
26714 (curr_insn_transform): Add code for reloads of matched inputs
26715 without output.
26716
7b0fe4f4
UB		 267172013-01-09 Uros Bizjak <ubizjak@gmail.com>
26718
26719 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26720 attribute of movddup insn to DF.
26721 (*vec_interleave_lowv2df): Ditto.
26722 (vec_dupv2df): Ditto.
26723
870ca331
JH		 267242013-01-09 Jan Hubicka <jh@suse.cz>
26725
26726 PR tree-optimiation/55875
26727 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26728 EVERY_ITERATION parameter.
7b0fe4f4 26729 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH		 26730 (idx_infer_loop_bounds): Similarly here.
26731 (n_of_executions_at_most): Simplify
26732 to only test for cases where statement is dominated by the
7b0fe4f4 26733 particular bound; handle correctly the "postdominance" test.
870ca331
JH		 26734 (scev_probably_wraps_p): Use max loop iterations info
26735 as a global bound first.
26736
6f575fe4 267372013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD		 26738 Nick Clifton <nickc@redhat.com>
26739
26740 * config/v850/v850.md (cbranchsf4): New pattern.
26741 (cstoresf4): New pattern.
26742 (cbranchdf4): New pattern.
26743 (cstoredf4): New pattern.
26744 (movsicc): Disallow floating point comparisons.
26745 (cmpsf_le_insn): Fix order of operators.
26746 (cmpsf_lt_insn): Likewise.
26747 (cmpsf_eq_insn): Likewise.
26748 (cmpdf_le_insn): Likewise.
26749 (cmpdf_lt_insn): Likewise.
26750 (cmpdf_eq_insn): Likewise.
26751 (cmpsf_ge_insn): Use LE comparison.
26752 (cmpdf_ge_insn): Likewise.
26753 (cmpsf_gt_insn): Use LT comparison.
26754 (cmpdf_gt_insn): Likewise.
26755 (cmpsf_ne_insn): Delete pattern.
26756 (cmpdf_ne_insn): Delete pattern.
26757 * config/v850/v850.c (v850_gen_float_compare): Use
26758 gen_cmpdf_eq_insn for NE comparison.
26759 (v850_float_z_comparison_operator)
26760 (v850_float_nz_comparison_operator): Move from here ...
26761 * config/v850/predicates.md: ... to here. Move GT and GE
26762 comparisons into v850_float_z_comparison_operator.
26763 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26764 Delete prototype.
26765 (v850_float_nz_comparison_operator): Likewise.
26766
f0d54148
JDA		 267672013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26768
26769 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26770 with calls to gen_insvsi/gen_insvdi.
26771
8f01beca
VK		 267722013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26773
26774 * config/i386/i386.c (initial_ix86_tune_features): Set up
26775 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26776
2aa579ad
JJ		 267772013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26778 Jakub Jelinek <jakub@redhat.com>
26779
26780 PR tree-optimization/48189
26781 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26782 If nitercst is 0, don't predict the exit edge.
26783
6edc3e32 267842013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N		 26785
26786 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26787 in asm_fprintf with reg_names.
26788 (aarch64_print_operand_address): Likewise.
26789 (aarch64_return_addr): Likewise.
26790 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26791
f6f94d94
JDA		 267922013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26793
26794 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26795 (INT_U6_BITS): Likewise.
26796 * config/pa/predicates.md (uint6_operand): New predicate.
26797 (shift5_operand, shift6_operand): Likewise.
26798 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26799 arith32_operand.
26800 (lshrdi3): Use shift6_operand.
26801 (shrpsi4, shrpdi4): New insn patterns.
26802 (extzv): Delete expander.
26803 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26804 predicates in unamed zero extract patterns. Tighten common constraint.
26805 (extv): Delete expander.
26806 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26807 predicates in unamed sign extract patterns. Tighten common constraint.
26808 (insv): Delete expander.
26809 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26810 predicates in unamed insert patterns. Tighten common constraint.
26811 Change uint32_operand predicate to uint6_operand predicate in unamed
26812 DImode pattern to insert constant values of type 1...1xxxx.
26813
36b72910
JH		 268142013-01-04 Jan Hubicka <jh@suse.cz>
26815
26816 PR tree-optimization/55823
7b0fe4f4
UB		 26817 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26818 issue.
36b72910 26819
47876a2a 268202013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ		 26821 Uros Bizjak <ubizjak@gmail.com>
26822
26823 PR rtl-optimization/55845
26824 * df-problems.c (can_move_insns_across): Stop scanning at
26825 volatile_insn_p source instruction or give up if
26826 across_from .. across_to range contains any volatile_insn_p
26827 instructions.
26828
4369c11e
TB		 268292013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26830
7b0fe4f4
UB		 26831 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26832 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26833 Declare.
4369c11e 26834 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 26835 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 26836
aeb7e7c1
JJ		 268372013-01-08 Jakub Jelinek <jakub@redhat.com>
26838
26839 PR fortran/55341
26840 * asan.c (asan_clear_shadow): New function.
26841 (asan_emit_stack_protection): Use it.
26842
a02ad1aa
TB		 268432013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26844
26845 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26846 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26847 with tab instead of space.
26848
f90d87f5
NC		 268492013-01-08 Nick Clifton <nickc@redhat.com>
26850
26851 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26852 register bank 0 at the start of an interrupt handler.
83ffd964
NC		 26853 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26854 MDBH registers.
f90d87f5 26855
385eb93d
JG		 268562013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26857
26858 * config/aarch64/aarch64-simd.md
26859 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26860 (aarch64_simd_bsl): Likewise.
26861 (aarch64_vcond_internal<mode>): Likewise.
26862 (vcond<mode><mode>): Likewise.
26863 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26864 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26865
4dcd1054
JG		 268662013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26867
26868 * config/aarch64/aarch64-builtins.c
26869 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26870
4267a4a6
MJ		 268712013-01-08 Martin Jambor <mjambor@suse.cz>
26872
26873 PR debug/55579
26874 * tree-sra.c (analyze_access_subtree): Return true also after
26875 potentially creating a debug-only replacement.
26876
5f4e6de3
JJ		 268772013-01-08 Jakub Jelinek <jakub@redhat.com>
26878
3138f224
JJ		 26879 PR middle-end/55890
26880 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26881
5f4e6de3
JJ		 26882 PR tree-optimization/54120
26883 * tree-vrp.c (range_fits_type_p): Don't allow
26884 src_precision < precision from signed vr to unsigned_p
26885 if vr->min or vr->max is negative.
26886 (simplify_float_conversion_using_ranges): Test can_float_p
26887 against CODE_FOR_nothing.
26888
f3ef18ff
JJ		 268892013-01-08 Jakub Jelinek <jakub@redhat.com>
26890 Richard Biener <rguenther@suse.de>
26891
26892 PR middle-end/55851
26893 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26894 types instead of just INTEGER_TYPE types.
26895
25c210f9
MK		 268962013-01-07 Mark Kettenis <kettenis@openbsd.org>
26897
26898 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26899 TF_SIZE): Define.
f90d87f5 26900
81826a7b
SE		 269012013-01-07 Steve Ellcey <sellcey@mips.com>
26902
26903 PR target/42661
26904 * config/mips/mips.opt: Change mad to mmad to match documentation.
26905
35678514
GJL		 269062013-01-07 Georg-Johann Lay <avr@gjlay.de>
26907
26908 PR target/55897
26909 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26910 .progmemx.data now.
26911
e5669488
GJL		 269122013-01-07 Georg-Johann Lay <avr@gjlay.de>
26913
26914 PR target/55897
26915 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26916 (avr_addrspace_t): Add .section_name field.
26917 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26918 array size.
26919 (avr_addrspace): Same. Initialize .section_name. Remove last
26920 NULL entry. Put __memx into .progmemx.data.
26921 (progmem_section_prefix): Remove.
26922 (avr_asm_init_sections): No need to initialize progmem_section.
26923 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26924 section name prefix.
26925 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26926 retrieve the progmem section.
26927 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26928 boundary to run over avr_addrspace[].
26929 (avr_register_target_pragmas): Ditto.
26930
656e6f37
JJ		 269312013-01-06 Jakub Jelinek <jakub@redhat.com>
26932
26933 * varasm.c (output_constant_def_contents): For asan_protect_global
26934 protected strings, adjust DECL_ALIGN if needed, before testing for
26935 anchored symbols.
26936 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26937 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26938 normal decls.
26939 (output_object_block): For asan protected decls, emit asan padding
26940 after their contents.
26941 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26942 (asan_finish_file): Test it here instead.
26943
6edc3e32
UB		 269442013-01-07 Nick Clifton <nickc@redhat.com>
26945 Matthias Klose <doko@debian.org>
26946 Doug Kwan <dougkwan@google.com>
26947 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC		 26948
26949 PR driver/55470
26950 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26951
26952 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26953
26954 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26955
7b0fe4f4 26956 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC		 26957
26958 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26959
2a095093
GJL		 269602013-01-07 Georg-Johann Lay <avr@gjlay.de>
26961
f3b503f4 26962 PR target/54461
2a095093
GJL		 26963 * doc/install.texi (Cross-Compiler-Specific Options): Document
26964 --with-avrlibc.
26965
383f9b34
TB		 269662013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26967
26968 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26969 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26970 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26971 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26972 vqmovun_high_s64): Fix source operand number and update copyright.
26973
3626621a
RB		 269742013-01-07 Richard Biener <rguenther@suse.de>
26975
26976 PR middle-end/55890
26977 * gimple.h (gimple_call_builtin_p): New overload.
26978 * gimple.c (validate_call): New function.
26979 (gimple_call_builtin_p): Likewise.
26980 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26981 Use gimple_call_builtin_p.
26982 (find_func_clobbers): Likewise.
26983 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26984 (strlen_optimize_stmt): Likewise.
26985
8b2ea410
JG		 269862013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26987
26988 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26989 (vld1q_dup_*): Likewise.
26990 (vld1_*): Likewise.
26991 (vld1q_*): Likewise.
26992 (vld1_lane_*): Likewise.
26993 (vld1q_lane_*): Likewise.
26994
9713d329
RB		 269952013-01-07 Richard Biener <rguenther@suse.de>
26996
26997 * lto-streamer.h (LTO_minor_version): Bump to 2.
26998
3520f7cc
JG		 269992013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27000
27001 * config/aarch64/aarch64-protos.h
27002 (aarch64_const_double_zero_rtx_p): Rename to...
27003 (aarch64_float_const_zero_rtx_p): ...this.
27004 (aarch64_float_const_representable_p): New.
27005 (aarch64_output_simd_mov_immediate): Likewise.
27006 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27007 move immediate case.
27008 * config/aarch64/aarch64.c
27009 (aarch64_const_double_zero_rtx_p): Rename to...
27010 (aarch64_float_const_zero_rtx_p): ...this.
27011 (aarch64_print_operand): Allow printing of new constants.
27012 (aarch64_valid_floating_const): New.
27013 (aarch64_legitimate_constant_p): Check for valid floating-point
27014 constants.
27015 (aarch64_simd_valid_immediate): Likewise.
27016 (aarch64_vect_float_const_representable_p): New.
27017 (aarch64_float_const_representable_p): Likewise.
27018 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27019 (aarch64_output_simd_mov_immediate): New.
27020 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27021 (*movdf_aarch64): Likewise.
27022 * config/aarch64/constraints.md (Ufc): New.
27023 (Y): call aarch64_float_const_zero_rtx.
27024 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27025
e167c04d
RB		 270262013-01-07 Richard Biener <rguenther@suse.de>
27027
27028 PR tree-optimization/55888
27029 PR tree-optimization/55862
27030 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27031 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27032 not if it is contained therein.
27033
0139aaab
GJL		 270342013-01-07 Georg-Johann Lay <avr@gjlay.de>
27035
27036 * config/avr/t-avr: Typo.
27037
4a176b23
GJL		 270382013-01-07 Georg-Johann Lay <avr@gjlay.de>
27039
27040 PR55243
27041 * config/avr/t-avr: Don't automatically rebuild
27042 $(srcdir)/config/avr/t-multilib
27043 $(srcdir)/config/avr/avr-tables.opt
27044 $(srcdir)/doc/avr-mmcu.texi
27045 (avr-mcus): New phony target to build them on request.
27046 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27047 * avr/avr-mcus.def: Adjust comments.
27048
c7afdc98
UB		 270492013-01-07 Uros Bizjak <ubizjak@gmail.com>
27050
27051 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27052
1ab05c31
RS		 270532013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27054
27055 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27056
488e3acc
RS		 270572013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27058
27059 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27060
a89599d2
DE		 270612013-01-05 David Edelsohn <dje.gcc@gmail.com>
27062
27063 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27064 to generate profiling.
27065 * config/rs6000/aix64.h (LIB_SPEC): Same.
27066
70f09188
AP		 270672013-01-04 Andrew Pinski <apinski@cavium.com>
27068
27069 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27070 New function.
27071 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27072
918d445f
UB		 270732013-01-04 Uros Bizjak <ubizjak@gmail.com>
27074
27075 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27076 unconditionally.
27077 (ix86_expand_move): Ditto.
27078 (ix86_zero_extend_to_Pmode): Ditto.
27079 (ix86_expand_call): Ditto.
27080 (ix86_expand_special_args_builtin): Ditto.
27081 (ix86_expand_builtin): Ditto.
27082
361618ec
RB		 270832013-01-04 Richard Biener <rguenther@suse.de>
27084
27085 PR tree-optimization/55862
27086 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27087 translating them through PHI nodes.
27088
87eab554
MJ		 270892013-01-04 Martin Jambor <mjambor@suse.cz>
27090
27091 PR tree-optimization/55755
27092 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27093 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27094
33879b9f
MP		 270952013-01-04 Marek Polacek <polacek@redhat.com>
27096
27097 PR middle-end/55859
27098 * opts.c (default_options_optimization): Clarify error message.
27099
3068819a
RB		 271002013-01-04 Richard Biener <rguenther@suse.de>
27101
27102 PR middle-end/55863
27103 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27104 reassociation.
27105
0e4ae794
JDA		 271062013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27107
27108 PR target/53789
27109 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA		 27110 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27111 references.
0e4ae794 27112
65c88cef
RH		 271132013-01-03 Richard Henderson <rth@redhat.com>
27114
27115 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27116 after eliminating TLS symbols.
27117
8ac16127
MG		 271182013-01-03 Marc Glisse <marc.glisse@inria.fr>
27119
27120 PR bootstrap/50167
27121 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27122 * graphite-poly.c (debug_gmp_value): Likewise.
27123
bb664f09
UB		 271242013-01-03 Uros Bizjak <ubizjak@gmail.com>
27125
27126 PR target/55712
27127 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27128 selected code model, define __code_mode_small__, __code_model_medium__,
27129 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27130 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27131 xchg temporary register with %k. Declare temporary register as
27132 early clobbered.
27133 [__x86_64__]: For medium and large code models, preserve %rbx register.
27134
bcf1ef00
RB		 271352013-01-03 Richard Biener <rguenther@suse.de>
27136
0506634a 27137 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB		 27138 (dump_subscript): Adjust.
27139 (finalize_ddr_dependent): Do not dump redundant info.
27140 (analyze_siv_subscript): Adjust.
27141 (subscript_dependence_tester): Likewise.
27142 (compute_affine_dependence): Likewise.
27143
59fd17e3
RB		 271442013-01-03 Richard Biener <rguenther@suse.de>
27145
27146 Revert
27147 2013-01-03 Richard Biener <rguenther@suse.de>
27148
27149 PR tree-optimization/55857
27150 * tree-vect-stmts.c (vectorizable_load): Do not setup
27151 re-alignment for invariant loads.
27152
27153 2013-01-02 Richard Biener <rguenther@suse.de>
27154
27155 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 27156 invariant load do not generate a vector load from the scalar location.
59fd17e3 27157
595c2679
RB		 271582013-01-03 Richard Biener <rguenther@suse.de>
27159
27160 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27161 for not vectorizing.
27162 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27163 not build INDIRECT_REFs, call get_name once only.
27164 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27165 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27166
90bb7d7a
RB		 271672013-01-03 Richard Biener <rguenther@suse.de>
27168
27169 PR tree-optimization/55857
27170 * tree-vect-stmts.c (vectorizable_load): Do not setup
27171 re-alignment for invariant loads.
27172
f09b77ca
RB		 271732013-01-03 Richard Biener <rguenther@suse.de>
27174
27175 PR lto/55848
27176 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27177 prefer a built-in decl.
27178
3461a16e
JJ		 271792013-01-03 Jakub Jelinek <jakub@redhat.com>
27180
df8e2b4f
JJ		 27181 * gcc.c (process_command): Update copyright notice dates.
27182 * gcov.c (print_version): Likewise.
27183 * gcov-dump.c (print_version): Likewise.
27184
3461a16e
JJ		 27185 PR rtl-optimization/55838
27186 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27187 iv0.step, iv1.step and step.
27188
8b5546d6
JJ		 271892013-01-03 Jakub Jelinek <jakub@redhat.com>
27190 Marc Glisse <marc.glisse@inria.fr>
27191
27192 PR tree-optimization/55832
27193 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27194 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27195 integer_{one,zero}_node.
27196
8ab1d2e9
JJ		 271972013-01-03 Jakub Jelinek <jakub@redhat.com>
27198
27199 PR debug/54402
27200 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27201 * var-tracking.c (reverse_op): Don't add reverse ops to
27202 VALUEs that have already
0506634a 27203 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 27204
5b9ad1d4
GP		 272052013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27206
27207 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27208
e25a6711
TJ		 272092013-01-02 Teresa Johnson <tejohnson@google.com>
27210
27211 * dumpfile.c (dump_loc): Print filename with location.
27212 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27213 new location_t parameter to emit complete unroll message with
27214 new dump framework.
27215 (canonicalize_loop_induction_variables): Compute loops location
27216 and pass to try_unroll_loop_completely.
27217 * loop-unroll.c (report_unroll_peel): New function.
27218 (peel_loops_completely): Use new dump format with location
27219 for main dumpfile message, and invoke report_unroll_peel on success.
27220 (decide_unrolling_and_peeling): Ditto.
27221 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27222 by report_unroll_peel.
27223 (decide_peel_completely): Ditto.
27224 (decide_unroll_constant_iterations): Ditto.
27225 (decide_unroll_runtime_iterations): Ditto.
27226 (decide_peel_simple): Ditto.
27227 (decide_unroll_stupid): Ditto.
27228 * cfgloop.c (get_loop_location): New function.
27229 * cfgloop.h (get_loop_location): Declare.
27230
77878621
ST		 272312013-01-02 Sriraman Tallam <tmsriram@google.com>
27232
27233 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27234 NULL.
27235
9e65d03e
JDA		 272362013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27237
27238 PR middle-end/55198
27239 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27240 BLKmode objects when EXPAND_MEMORY is specified.
27241
6a7da30f
ST		 272422013-01-02 Sriraman Tallam <tmsriram@google.com>
27243
27244 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27245 in loop predicate.
27246 (fold_builtin_cpu): Do not share cpu model decls across statements.
27247
e78167f2
JM		 272482013-01-02 Jason Merrill <jason@redhat.com>
27249
27250 PR c++/55804
27251 * tree.c (build_array_type_1): Revert earlier change.
27252
8c075fb4
YZ		 272532013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27254
27255 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27256 "cortex-a57".
27257 * config/aarch64/aarch64-tune.md: Re-generate.
27258
0682ed3e
RB		 272592013-01-02 Richard Biener <rguenther@suse.de>
27260
27261 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 27262 invariant load do not generate a vector load from the scalar location.
0682ed3e 27263
b8f6a302
RB		 272642013-01-02 Richard Biener <rguenther@suse.de>
27265
27266 PR bootstrap/55784
27267 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27268 * configure: Regenerate.
27269
04b80dbb
RS		 272702013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27271
27272 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27273 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27274 (expand_builtin_int_roundingfn_2): Keep the original target around
27275 for the fallback case.
27276
635b0b0c
RS		 272772013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27278
27279 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27280 to be clear for sign changes.
27281
42d57399
JH		 272822013-01-01 Jan Hubicka <jh@suse.cz>
27283
27284 * ipa-inline-analysis.c: Fix formatting.
27285
5bb6669d
JJ		 272862013-01-01 Jakub Jelinek <jakub@redhat.com>
27287
27288 PR tree-optimization/55831
27289 * tree-vect-loop.c (get_initial_def_for_induction): Use
27290 gsi_after_labels instead of gsi_start_bb.
ad41bd84 27291\f
86a2db33 27292Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM		 27293
27294Copying and distribution of this file, with or without modification,
27295are permitted in any medium without royalty provided the copyright
27296notice and this notice are preserved.
Mirror of https://gcc.gnu.org/git/gcc.git