]> 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
re PR c++/58816 (ICE have tree_list in private_is_attribute_p)
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
f2f39141
PC		 12013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2
3 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
4 [Ed Smith-Rowland]): New entries.
5 ([Stephen M. Webb]): Update.
6
8223bb78
AM		 72013-10-22 Andrew MacLeod <amacleod@redhat.com>
8
9 * tree-ssa-ter.h: Remove duplicate copy of file contents.
10
1e08df0e
MP		 112013-10-21 Marek Polacek <polacek@redhat.com>
12
13 PR middle-end/58809
14 * fold-const.c (fold_range_test): Return 0 if the type is not
15 an integral type.
16
67b5215c
RS		 172013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
18
19 * system.h: Move hwint.h include further down.
20 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
21 gcc_checking_asserts.
22 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
23
0cad6830
BE		 242013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
25
26 Fix volatile issues in optimize_bit_field_compare.
27 * fold-const.c (optimize_bit_field_compare): Bail out if
28 lvolatilep or rvolatilep.
29
5dfd32cd
BE		 302013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
31
32 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
33 and get_inner_reference returning different pmode for non-volatile
34 bit-field members dependent on flag_strict_volatile_bitfields.
35 * stor-layout.c (layout_decl): Remove special handling of
36 flag_strict_volatile_bitfields.
37 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
38 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
39
0bff3610
PM		 402013-10-21 Paulo Matos <pmatos@broadcom.com>
41
42 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
43 calculations.
44
b5c4ff78
JL		 452013-10-21 Jeff Law <law@redhat.com>
46
f2f39141
PC		 47 * tree-ssa-threadedge.c (thread_through_normal_block): New
48 argument VISITED. Remove VISISTED as a local variable. When we
49 have a threadable jump, verify the destination of the jump has not
50 been visised.
51 (thread_across_edge): Allocate VISITED bitmap once at function
52 scope and use it throughout. Make sure to set appropriate bits in
53 VISITED for E (start of jump thread path).
54 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
55 through a joiner if any edge on the path has a recorded jump thread.
b5c4ff78 56
ce0770fc
ILT		 572013-10-21 Ian Lance Taylor <iant@google.com>
58
59 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
60 don't imply that attributes can solve all problems.
61 (Directory Options): Fix typo.
62
0e7504c8
KT		 632013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
64
65 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
66 extend and extend_add.
67
2445aeac
RB		 682013-10-21 Richard Biener <rguenther@suse.de>
69
70 PR tree-optimization/58794
71 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
72 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
73
d480e6a5
RB		 742013-10-21 Richard Biener <rguenther@suse.de>
75
76 PR middle-end/58742
77 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
78 to (T) X for sign-changing conversions (or no conversion).
79
54f9aba0
UB		 802013-10-20 Uros Bizjak <ubizjak@gmail.com>
81
82 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
83
d6c6ba3c
JH		 842013-10-20 Jan Hubicka <jh@suse.cz>
85
86 * config/i386/i386-tune.def: Add comment; organize into categories
87
d606b917
MZ		 882013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
89
90 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
91 argument. Update function comment.
92 (expand_set_or_movmem_via_rep): New function combining
93 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
94 (expand_movmem_via_rep_mov): Remove.
95 expand_setmem_via_rep_stos): Remove.
96 (expand_movmem_epilogue): Update calls correspondingly.
97 (expand_setmem_epilogue_via_loop): Likewise.
98 (emit_memset): New.
99 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
100 (expand_set_or_movmem_prologue): New function combining
101 expand_movmem_prologue and expand_setmem_prologue.
102 (expand_movmem_prologue): Remove.
103 (expand_setmem_prologue): Remove.
104 (expand_set_or_movmem_constant_prologue): New function combining
105 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
106 (expand_constant_movmem_prologue): Remove.
107 (expand_constant_setmem_prologue): Remove.
108 (promote_duplicated_reg): Allow vector-const0 value.
109 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
110 and ix86_expand_setmem.
111 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
112 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
113
4d648807
DN		 1142013-10-21 Diego Novillo <dnovillo@google.com>
115
116 * asan.c: Include tree.h
117 * bb-reorder.c: Likewise.
118 * cfgcleanup.c: Likewise.
119 * cfgloopmanip.c: Likewise.
120 * data-streamer-in.c: Likewise.
121 * data-streamer-out.c: Likewise.
122 * data-streamer.c: Likewise.
123 * dwarf2cfi.c: Likewise.
124 * graphite-blocking.c: Likewise.
125 * graphite-clast-to-gimple.c: Likewise.
126 * graphite-dependences.c: Likewise.
127 * graphite-interchange.c: Likewise.
128 * graphite-optimize-isl.c: Likewise.
129 * graphite-poly.c: Likewise.
130 * graphite-scop-detection.c: Likewise.
131 * graphite-sese-to-poly.c: Likewise.
132 * graphite.c: Likewise.
133 * ipa-devirt.c: Likewise.
134 * ipa-profile.c: Likewise.
135 * ipa.c: Likewise.
136 * ira.c: Likewise.
137 * loop-init.c: Likewise.
138 * loop-unroll.c: Likewise.
139 * lower-subreg.c: Likewise.
140 * lto/lto-object.c: Likewise.
141 * recog.c: Likewise.
142 * reginfo.c: Likewise.
143 * tree-loop-distribution.c: Likewise.
144 * tree-parloops.c: Likewise.
145 * tree-ssa-strlen.c: Likewise.
146 * tree-streamer.c: Likewise.
147 * value-prof.c: Likewise.
148 * target-globals.c: Likewise.
149 * expr.h: Include tree-core.h instead of tree.h.
150 * gimple.h: Likewise.
151 * ipa-prop.h: Likewise.
152 * ipa-utils.h: Likewise.
153 * lto-streamer.h: Likewise.
154 * streamer-hooks.h: Likewise.
155 * ipa-reference.h: Include cgraph.h instead of tree.h.
156 * cgraph.h: Include basic-block.h instead of tree.h.
157 * tree-streamer.h: Do not include tree.h.
158 * genattrtab.c (write_header): Generate inclusion of
159 tree.h.
160 * genautomata.c (main): Likewise.
161 * genemit.c: Likewise.
162 * genopinit.c: Likewise.
163 * genoutput.c (output_prologue): Likewise.
164 * genpeep.c: Likewise.
165
dd346f1f
BS		 1662013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
167
168 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
169 little endian.
170 (vec_unpacku_hi_v8hi): Likewise.
171 (vec_unpacku_lo_v16qi): Likewise.
172 (vec_unpacku_lo_v8hi): Likewise.
173
915e8e6e
JH		 1742013-10-20 Jan Hubicka <jh@suse.cz>
175
176 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
177 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
178 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
179
0f37323c
RS		 1802013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
181
182 * config/mips/mips.h (ISA_HAS_WSBH): Define.
183 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
184 constants.
185 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
186
84935d25
JDA		 1872013-10-19 John David Anglin <danglin@gcc.gnu.org>
188
189 PR target/58603
190 * system.h: Undef m_slot.
191
d7d1451a
BS		 1922013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
193
194 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
195 all elements for both endian flavors.
196
6eb793dd
UB		 1972013-10-19 Uros Bizjak <ubizjak@gmail.com>
198
199 PR target/58792
200 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
201 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
202 and SI_REG for 64bit SYSV ABI targets.
203
c07757e5
UB		 2042013-10-19 Uros Bizjak <ubizjak@gmail.com>
205
206 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
207 to multi_reg_return. Clarify that we are skipping USEs of multiple
208 return registers. Use bool type where appropriate.
209
41ee845b
JH		 2102013-10-18 Jan Hubicka <jh@suse.cz>
211
212 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
213 for cold functions.
214 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
215 (X86_TUNE_PUSH_MEMORY): Likewise.
216 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
217 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
c07757e5
UB		 218 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
219 New.
41ee845b
JH		 220 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
221 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
222 Remove.
223 (ix86_option_override_internal): Update to use tune features instead
224 of variables.
225
6f978a2a
CH		 2262013-10-18 Cong Hou <congh@google.com>
227
228 PR tree-optimization/58508
229 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
230 statement that contains data refs with zero-step.
231
4484a35a
AM		 2322013-10-18 Andrew MacLeod <amacleod@redhat.com>
233
c07757e5
UB		 234 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
235 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
4484a35a
AM		 236 * gimple-low.c (gimple_check_call_arg,
237 gimple_check_call_matching_types): Move to cgraph.c.
238 * gimple-low.h: Remove prototype.
239 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
240 Relocate from gimple-low.c.
241 * cgraph.h: Add prototype. Don't include basic-block.h.
242 * gimplify.c: Add gimple-low to include list.
243 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
244 * tree-eh.c: Add gimple-low to include list.
245 * tree-nested.c: Likewise.
246 * cfgexpand.c: Add tree-ssa-address.h to include list.
247 * expr.c: Likewise.
248 * gimple-fold.c: Likewise.
249 * gimple-ssa-strength-reduction.c: Likewise.
250 * trans-mem.c: Likewise.
251 * tree-mudflap.c: Likewise.
252 * tree-ssa-loop-ivopts.c: Likewise.
253 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
254 (degenerate_phi_result): Move to tree-phinodes.c.
255 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
256 * tree-ssa-threadedge.c: Likewise.
257 * tree-vrp.c: Likewise.
258 * tree-phinodes.c (degenerate_phi_result): Relocate here.
259 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
260 * tree-phinodes.h (degenerate_phi_result): Add prototype.
261 * tree-ssa-copy.c: Include tree-ssa-dom.h.
262 * tree-ssa-forwprop.c: Likewise.
263 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
264 pass_data_cleanup_cfg_post_optimizing,
265 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
266 * tree-optimize.c: Delete File.
267 * graphite.c: Include tree-cfgcleanup.h.
268 * passes.c: Likewise.
269 * tree-cfg.c: Likewise.
270 * tree-profile.c: Likewise.
271 * tree-ssa-dse.c: Likewise.
272 * tree-ssa-loop-ivcanon.c: Likewise.
273 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
274 * tree-outof-ssa.c: Include sbitmap.h.
275 * tree-ssa-live.c: Likewise.
276 * tree-ssa-propagate.c: Likewise.
277 * tree-ssa-structalias.c: Likewise.
278 * tree-stdarg.c: Likewise.
279 * Makefile.in (OBJS): Delete tree-optimize.o.
280 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
281 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
282 * varasm.c: Include basic-block.h.
283 * cfgloop.h: Include function.h instead of basic-block.h
284 (bb_loop_depth): Move to cfgloop.c.
285 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
286
bfaa17c1
TJ		 2872013-10-18 Teresa Johnson <tejohnson@google.com>
288
289 * predict.c (probably_never_executed): Compare frequency-based
290 count to number of training runs.
291 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
292
f58883bd
KT		 2932013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
294
295 * config/arm/arm.c (cortexa9_extra_costs): New table.
296 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
297
01e127b1
JL		 2982013-10-18 Jeff Law <law@redhat.com>
299
300 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
301
302 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
c07757e5
UB		 303 (register_jump_thread): Add "registered_jump_thread" debug
304 counter support.
01e127b1
JL		 305 * dbgcnt.def (registered_jump_thread): New debug counter.
306
f41b8848
AM		 3072013-10-18 Andrew MacLeod <amacleod@redhat.com>
308
309 * config/rs6000/rs6000.c: Include cgraph.h.
310
adc7a812
TJ		 3112013-10-18 Teresa Johnson <tejohnson@google.com>
312
313 * tree-ssa-tail-merge.c (replace_block_by): Update edge
314 weights during merging.
315
377d9792
AM		 3162013-10-18 Andrew MacLeod <amacleod@redhat.com>
317
318 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
319 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
320 * tree-ssa-operands.h: Remove prototype.
321 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
322 * gimple.c (virtual_operand_p): Relocate from gimple.c.
323 * gimple.h: Add prototype.
324 * gimple-ssa.h: Include tree-ssa-operands.h.
325 * tree-dump.c: Add tree-cfg.h to include list.
326 * tree-ssa-alias.c: Add ipa-reference.h to include list.
327 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
328 * config/i386/i386.c: Don't include tree-flow.h.
329 * config/rs6000/rs6000.c: Likewise.
330
8ad8afaf
JBG		 3312013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
332
333 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
334
c0e081a9
RB		 3352013-10-18 Richard Biener <rguenther@suse.de>
336
337 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
338 or TYPE_UNSIGNED of integral types.
339 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
340 NULL_TREE for zero-precision integral types.
341
054fceb5
JG		 3422013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
343
344 * config/aarch64/arm_neon.h
345 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
346
70e8b89b
ST		 3472013-10-17 Sriraman Tallam <tmsriram@google.com>
348
349 * opth-gen.awk: Define target_flags_explicit.
350
3170766c 3512013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5845f602
MM		 352
353 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
354 fields to the reg_addr array that describes the valid addressing
355 mode for any register, general purpose registers, floating point
356 registers, and Altivec registers.
357 (FIRST_RELOAD_REG_CLASS): Likewise.
358 (LAST_RELOAD_REG_CLASS): Likewise.
359 (struct reload_reg_map_type): Likewise.
360 (reload_reg_map_type): Likewise.
361 (RELOAD_REG_VALID): Likewise.
362 (RELOAD_REG_MULTIPLE): Likewise.
363 (RELOAD_REG_INDEXED): Likewise.
364 (RELOAD_REG_OFFSET): Likewise.
365 (RELOAD_REG_PRE_INCDEC): Likewise.
366 (RELOAD_REG_PRE_MODIFY): Likewise.
367 (reg_addr): Likewise.
368 (mode_supports_pre_incdec_p): New helper functions to say whether
369 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
370 (mode_supports_pre_modify_p): Likewise.
371 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
372 print the valid address mode bits for each mode.
373 (rs6000_debug_print_mode): Likewise.
374 (rs6000_debug_reg_global): Likewise.
375 (rs6000_setup_reg_addr_masks): New function to set up the address
376 mask bits for each type.
377 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
378 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
379 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
380 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
381 PRE_MODIFY are supported.
3170766c
MM		 382 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
383 registers, instead of {src,dest}_av_p.
5845f602
MM		 384 (rs6000_print_options_internal): Tweak the debug output slightly.
385
82e33890
UB		 3862013-10-17 Uros Bizjak <ubizjak@gmail.com>
387
388 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
389 isa attribute.
390
0645c1a2
AM		 3912013-10-17 Andrew MacLeod <amacleod@redhat.com>
392
393 * tree-flow.h (struct omp_region): Move to omp-low.c.
394 Remove omp_ prototypes and variables.
395 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
396 (copy_var_decl): Relocate prototype from tree-flow.h.
397 * gimple.c (copy_var_decl): Relocate from omp-low.c.
398 * tree.h: Move prototype to omp-low.h.
399 * omp-low.h: New File. Relocate prototypes here.
400 * omp-low.c (struct omp_region): Make local here.
401 (root_omp_region): Make static.
402 (copy_var_decl) Move to gimple.c.
403 (new_omp_region): Make static.
404 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
405 * tree-cfg.c: Include omp-low.h.
406 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
407 * gimplify.c: Include omp-low.h.
408 * tree-parloops.c: Likewise.
409
6f1abb55
UB		 4102013-10-17 Uros Bizjak <ubizjak@gmail.com>
411
412 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
413 operands are in memory, prefer to force non-matched operand 1 to
414 the register.
415
1ddbbbc0
MM		 4162013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
417
418 PR target/58673
419 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
420 restrict TImode addresses to single indirect registers if both
421 -mquad-memory and -mvsx-timode are used.
422 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
423 we should emit load/store quad. Remove using %y for quad memory
424 addresses.
425
426 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
427 constraints to allow load/store quad on machines where TImode is
428 not allowed in VSX registers. Use 'n' instead of 'F' constraint
429 for TImode to load integer constants.
430
f541a481
KT		 4312013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
432
433 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
434
51bb310d
MS		 4352013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
436
437 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
438 handling of STACK_REG.
439
f4bed77b
RB		 4402013-10-17 Richard Biener <rguenther@suse.de>
441
442 PR tree-optimization/58143
443 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
444 New function.
445 (rewrite_to_defined_overflow): Likewise.
446 (move_computations_dom_walker::before_dom): Rewrite stmts
447 with undefined signed overflow that are not always executed
448 into unsigned arithmetic.
449
d358fb9d
MM		 4502013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
451
452 PR target/57756
453 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
454 explicit isa flag to be an options variable, instead of using
455 global_options_set. Remove define from rs6000.h.
456 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
457
458 * config/rs6000/rs6000.c (rs6000_option_override_internal):
459 Initialize rs6000_isa_flags_explicit.
460 (rs6000_function_specific_save): Add gcc_options* parameter, so
461 that the powerpc builds after the 2013-10-15 changes.
462 (rs6000_function_specific_restore): Likewise.
463
c71aed47
DD		 4642013-10-16 DJ Delorie <dj@redhat.com>
465
466 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
467 op is a REG before checking REGNO.
468 (rl78_alloc_physical_registers): Verify pattern is a SET before
469 checking SET_SRC.
470
90d3bd51
BS		 4712013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
472
473 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
474 endianness.
475 (vec_unpacks_lo_v4sf): Likewise.
476 (vec_unpacks_float_hi_v4si): Likewise.
477 (vec_unpacks_float_lo_v4si): Likewise.
478 (vec_unpacku_float_hi_v4si): Likewise.
479 (vec_unpacku_float_lo_v4si): Likewise.
480
de75c876
BS		 4812013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
482
483 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
484 (vsx_concat_v2sf): Likewise.
485
21216998
JG		 4862013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
487
488 * config/aarch64/aarch64.md
489 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
490
1ef88893 4912013-10-16 Andrew MacLeod <amacleod@redhat.com>
7ab4168e 492
1ef88893 493 PR tree-optimization/58697
7ab4168e 494 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1ef88893
AM		 495 estimated_loop_iterations_int.
496 (max_stmt_executions_int): Call get_max_loop_iterations_int.
7ab4168e 497 (get_max_loop_iterations_int): New. HWINT version of
1ef88893
AM		 498 get_max_loop_iterations.
499 * cfgloop.h: Add prototypes.
500 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
7ab4168e 501 * loop-unroll.c (decide_peel_once_rolling): Call
1ef88893
AM		 502 get_estimated_loop_iterations_int.
503 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
504 * tree-ssa-loop-niter.h: Tweak prototypes.
505
c22df64f
DM		 5062013-10-16 David Malcolm <dmalcolm@redhat.com>
507
508 * gengtype-parse.c (struct_field_seq): Ignore access-control
509 keywords ("public:" etc).
510
f24bb080
MS		 5112013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
512
513 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
514 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
515
c281bf26
YR		 5162013-10-16 Yvan Roux <yvan.roux@linaro.org>
517
518 * config/arm/arm.opt (mlra): New option.
519 * config/arm/arm.c (arm_lra_p): New function.
520 (TARGET_LRA_P): Define.
521
5806f481
PM		 5222013-10-16 Paulo Matos <pmatos@broadcom.com>
523
524 * tree-core.h (tree_code_name): Remove.
525 * tree.h (get_tree_code_name): New prototype.
526 * tree.c (tree_code_name): Make static.
527 (get_tree_code_name): New function.
528 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
529 tree_class_check_failed, tree_range_check_failed,
530 tree_not_class_check_failed, omp_clause_check_failed,
531 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
532 wrapper get_tree_code_name instead of calling tree_code_name directly.
533 * tree-vrp.c (dump_asserts_for): Likewise.
534 * tree-dump.c (dequeue_and_dump): Likewise.
535 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
536 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
537 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
538 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
539 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
540 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
541 dump_gimple_omp_for): Likewise.
542 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
543 * tree-ssa-pre.c (print_pre_expr): Likewise.
544 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
545 * print-tree.c (print_node_brief, print_node): Likewise.
546 * gimple.c (gimple_check_failed): Likewise.
547 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
548 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
549 * config/mep/mep.c (mep_validate_vliw): Likewise.
550 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
551 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
552
82c9df56
GG		 5532013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
554
555 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
556 for AMD bdver3.
557
b82d0df9
HPN		 5582013-10-16 Hans-Peter Nilsson <hp@axis.com>
559
560 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
561 (MULTILIB_MATCHES): Add multilib for -march=v8.
562
bf7b5747
ST		 5632013-10-15 Sriraman Tallam <tmsriram@google.com>
564
565 PR target/57756
566 * optc-save-gen.awk: Add extra parameter to the save and restore
567 target calls.
568 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
569 * tree.c (build_optimization_node): New parameter. Add extra parameter
570 to call to cl_optimization_save.
571 (build_target_option_node): New parameter. Add extra parameter
572 to call to cl_target_option_save.
573 * tree.h (build_optimization_node): New parameter.
574 (build_target_option_node): New parameter.
575 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
576 build_optimization_node and build_target_option_node.
577 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
578 (handle_pragma_push_options): Ditto.
579 * toplev.c (process_options): Ditto.
580 * opts.c (init_options_struct): Check for opts_set non-null.
581 * target.def (target_option.save): New parameter.
582 (target_option.restore): New parameter.
583 * tm.texi: Generate.
584 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
585 (ix86_pragma_target_parse): Ditto.
586 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
587 parameters.
588 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
589 to build_optimization_node and build_target_option_node.
590 (rs6000_valid_attribute_p): Ditto.
591 (rs6000_pragma_target_parse): Ditto.
592 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
593 data.
594 * config/i386/i386.h:
595 TARGET_64BIT_P: New Macro
596 TARGET_MMX_P: New Macro.
597 TARGET_3DNOW_P: New Macro.
598 TARGET_3DNOW_A_P: New Macro.
599 TARGET_SSE_P: New Macro.
600 TARGET_SSE2_P: New Macro.
601 TARGET_SSE3_P: New Macro.
602 TARGET_SSSE3_P: New Macro.
603 TARGET_SSE4_1_P: New Macro.
604 TARGET_SSE4_2_P: New Macro.
605 TARGET_AVX_P: New Macro.
606 TARGET_AVX2_P: New Macro.
607 TARGET_AVX512F_P: New Macro.
608 TARGET_AVX512PF_P: New Macro.
609 TARGET_AVX512ER_P: New Macro.
610 TARGET_AVX512CD_P: New Macro.
611 TARGET_FMA_P: New Macro.
612 TARGET_SSE4A_P: New Macro.
613 TARGET_FMA4_P: New Macro.
614 TARGET_XOP_P: New Macro.
615 TARGET_LWP_P: New Macro.
616 TARGET_ABM_P: New Macro.
617 TARGET_BMI_P: New Macro.
618 TARGET_BMI2_P: New Macro.
619 TARGET_LZCNT_P: New Macro.
620 TARGET_TBM_P: New Macro.
621 TARGET_POPCNT_P: New Macro.
622 TARGET_SAHF_P: New Macro.
623 TARGET_MOVBE_P: New Macro.
624 TARGET_CRC32_P: New Macro.
625 TARGET_AES_P: New Macro.
626 TARGET_PCLMUL_P: New Macro.
627 TARGET_CMPXCHG16B_P: New Macro.
628 TARGET_FSGSBASE_P: New Macro.
629 TARGET_RDRND_P: New Macro.
630 TARGET_F16C_P: New Macro.
631 TARGET_RTM_P: New Macro.
632 TARGET_HLE_P: New Macro.
633 TARGET_RDSEED_P: New Macro.
634 TARGET_PRFCHW_P: New Macro.
635 TARGET_ADX_P: New Macro.
636 TARGET_FXSR_P: New Macro.
637 TARGET_XSAVE_P: New Macro.
638 TARGET_XSAVEOPT_P: New Macro.
639 TARGET_LP64_P: New Macro.
640 TARGET_X32_P: New Macro.
641 TARGET_FPMATH_DEFAULT_P: New Macro.
642 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
643 * config/i386/i386.c (ix86_option_override_internal): New parameters.
644 opts and opts_set.
645 Change ix86_tune_string to access opts->x_ix86_tune_string.
646 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
647 Change ix86_arch_string to access opts->x_ix86_arch_string.
648 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
649 Change ix86_pmode to access opts->x_ix86_pmode.
650 Change ix86_abi to access opts->x_ix86_abi.
651 Change ix86_cmodel to access opts->x_ix86_cmodel.
652 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
653 Change ix86_isa_flags_explicit to access
654 opts->x_ix86_isa_flags_explicit.
655 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
656 Change ix86_regparm to access opts->x_ix86_regparm.
657 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
658 Change ix86_preferred_stack_boundary_arg to access
659 opts->x_ix86_preferred_stack_boundary_arg.
660 Change ix86_force_align_arg_pointer to access
661 opts->x_ix86_force_align_arg_pointer.
662 Change ix86_incoming_stack_boundar_arg to access
663 opts->x_ix86_incoming_stack_boundar_arg.
664 Change ix86_fpmath to access opts->x_ix86_fpmath.
665 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
666 Change ix86_recip_name to access opts->x_ix86_recip_name.
667 Change ix86_stack_protector_guard to access
668 opts->x_ix86_stack_protector_guard.
669 Change ix86_tune_memcpy_strategy to access
670 opts->x_ix86_tune_memcpy_strategy.
671 Change ix86_tune_memset_strategy to access
672 opts->x_ix86_tune_memset_strategy.
673 Change global_options to access opts.
674 Change global_options_set to access opts_set.
6f1abb55
UB		 675 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
676 Change TARGET_MMX to TARGET_MMX_P (opts->...).
677 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
678 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
679 Change TARGET_SSE to TARGET_SSE_P (opts->...).
680 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
681 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
682 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
683 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
684 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
685 Change TARGET_AVX to TARGET_AVX_P (opts->...).
686 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
687 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
688 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
689 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
690 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
691 Change TARGET_FMA to TARGET_FMA_P (opts->...).
692 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
693 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
694 Change TARGET_XOP to TARGET_XOP_P (opts->...).
695 Change TARGET_LWP to TARGET_LWP_P (opts->...).
696 Change TARGET_ABM to TARGET_ABM_P (opts->...).
697 Change TARGET_BMI to TARGET_BMI_P (opts->...).
698 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
699 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
700 Change TARGET_TBM to TARGET_TBM_P (opts->...).
701 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
702 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
703 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
704 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
705 Change TARGET_AES to TARGET_AES_P (opts->...).
706 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
707 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
708 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
709 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
710 Change TARGET_F16C to TARGET_F16C_P (opts->...).
711 Change TARGET_RTM to TARGET_RTM_P (opts->...).
712 Change TARGET_HLE to TARGET_HLE_P (opts->...).
713 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
714 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
715 Change TARGET_ADX to TARGET_ADX_P (opts->...).
716 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
717 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
718 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
719 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
720 Change TARGET_X32 to TARGET_X32_P (opts->...).
721 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
722 Change TARGET_FLOAT_RETURNS_IN_80387 to
723 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
bf7b5747
ST		 724 (ix86_function_specific_save): New parameter. Use opts-> fields
725 to replace global fields.
726 (ix86_function_specific_restore): Ditto.
727 (ix86_valid_target_attribute_inner_p): New parameters.
728 Fix recursive call.
729 Fix call to ix86_handle_option and set_option.
730 (ix86_valid_target_attribute_tree): New parameters.
731 Change global_options to access opts.
732 Change global_options_set to access opts_set.
733 Fix call to ix86_valid_target_attribute_inner_p.
734 Change ix86_tune_string to access opts->x_ix86_tune_string.
735 Change ix86_arch_string to access opts->x_ix86_arch_string.
736 Change ix86_fpmath to access opts->x_ix86_fpmath
737 Fix call to ix86_option_override_internal.
738 Fix call to ix86_add_new_builtins.
739 Fix calls to build_optimization_node and build_target_option_node.
740 (ix86_valid_target_attribute_p): Remove access to global_options.
741 Use new gcc_options structure func_options.
742 Fix call to ix86_valid_target_attribute_tree.
743 Fix call to build_optimization_node.
744 (get_builtin_code_for_version): Fix call to
745 ix86_valid_target_attribute_tree.
746
459260ec
DM		 7472013-10-15 David Malcolm <dmalcolm@redhat.com>
748
749 * Makefile.in (PICFLAG): New.
750 (enable_host_shared): New.
751 (INTERNAL_CFLAGS): Use PICFLAG.
752 (LIBIBERTY): Use pic build of libiberty.a if configured with
753 --enable-host-shared.
754 * configure.ac: Add --enable-host-shared, setting up new
755 PICFLAG variable.
756 * configure: Regenerate.
6f1abb55 757 * doc/install.texi (--enable-shared): Add note contrasting it with ...
459260ec
DM		 758 (--enable-host-shared): New option.
759
16022041
RS		 7602013-10-15 Richard Biener <rguenther@suse.de>
761
762 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
763 for built-in functions.
764
b114bfb4
ZC		 7652013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
766
767 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
768 (optimize_range_tests_1): New function,
769 extracted from optimize_range_tests.
770 (optimize_range_tests_xor): Similarly.
771 (optimize_range_tests_diff): New function.
772 (optimize_range_tests): Use optimize_range_tests_1.
773
69b8f2f9
CH		 7742013-10-15 Cong Hou <congh@google.com>
775
7ab4168e
MS		 776 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
777 requirement of the reduction pattern so that one operand of the
69b8f2f9
CH		 778 reduction operation can come from outside of the loop.
779
8ff5a4e4
JG		 7802013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
781
782 * config/arm/neon-schedgen.ml: Remove.
783 * config/arm/cortex-a9-neon.md: Remove comment regarding
784 neon-schedgen.ml.
785
52b69ad2
JG		 7862013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
787
788 * config/arm/types: Remove old neon types.
789
e7c6a556
JG		 7902013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
791
792 * config/arm/cortex-a7.md
793 (cortex_a7_neon_type): New.
794 (cortex_a7_neon_mul): Update for new types.
795 (cortex_a7_neon_mla): Likewise.
796 (cortex_a7_neon): Likewise.
797
795dd8d5
JG		 7982013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
799
800 * config/arm/cortex-a15-neon.md
801 (cortex_a15_neon_type): New,
802
803 (cortex_a15_neon_int_1): Remove.
804 (cortex_a15_neon_int_2): Likewise.
805 (cortex_a15_neon_int_3): Likewise.
806 (cortex_a15_neon_int_4): Likewise.
807 (cortex_a15_neon_int_5): Likewise.
808 (cortex_a15_neon_vqneg_vqabs): Likewise.
809 (cortex_a15_neon_vmov): Likewise.
810 (cortex_a15_neon_vaba): Likewise.
811 (cortex_a15_neon_vaba_qqq): Likewise.
812 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
813 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
814 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
815 Likewise.
816 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
817 (cortex_a15_neon_mla_qqq_8_16): Likewise.
818 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
819 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
820 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
821 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
822 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
823 (cortex_a15_neon_shift_1): Likewise.
824 (cortex_a15_neon_shift_2): Likewise.
825 (cortex_a15_neon_shift_3): Likewise.
826 (cortex_a15_neon_vshl_ddd): Likewise.
827 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
828 (cortex_a15_neon_vsra_vrsra): Likewise.
829 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
830 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
831 (cortex_a15_neon_bp_3cycle): Likewise.
832 (cortex_a15_neon_ldm_2): Likewise.
833 (cortex_a15_neon_stm_2): Likewise.
834 (cortex_a15_neon_mcr): Likewise.
835 (cortex_a15_neon_mrc): Likewise.
836 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
837 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
838 (cortex_a15_neon_fp_vmul_ddd): Likewise.
839 (cortex_a15_neon_fp_vmul_qqd): Likewise.
840 (cortex_a15_neon_fp_vmla_ddd): Likewise.
841 (cortex_a15_neon_fp_vmla_qqq): Likewise.
842 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
843 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
844 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
845 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
846 (cortex_a15_neon_bp_simple): Likewise.
847 (cortex_a15_neon_bp_2cycle): Likewise.
848 (cortex_a15_neon_bp_3cycle): Likewise.
849 (cortex_a15_neon_vld1_1_2_regs): Likewise.
850 (cortex_a15_neon_vld1_3_4_regs): Likewise.
851 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
852 (cortex_a15_neon_vld2_4_regs): Likewise.
853 (cortex_a15_neon_vld3_vld4): Likewise.
854 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
855 (cortex_a15_neon_vst1_3_4_regs): Likewise.
856 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
857 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
858 (cortex_a15_neon_vst3_vst4): Rename to...
859 (cortex_a15_neon_vst4): This, update for new attributes.
860 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
861 (cortex_a15_neon_vld3_vld4_lane): Likewise.
862 (cortex_a15_neon_vst1_vst2_lane): Likewise.
863 (cortex_a15_neon_vst3_vst4_lane): Likewise.
864 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
865 (cortex_a15_neon_ldm_2): Likewise.
866 (cortex_a15_neon_stm_2): Likewise.
867 (cortex_a15_neon_mcr): Likewise.
868 (cortex_a15_neon_mcr_2_mcrr): Likewise.
869 (cortex_a15_neon_mrc): Likewise.
870 (cortex_a15_neon_mrrc): Likewise.
871
872 (cortex_a15_neon_abd): New.
873 (cortex_a15_neon_abd_q): Likewise.
874 (cortex_a15_neon_aba): Likewise.
875 (cortex_a15_neon_aba_q): Likewise.
876 (cortex_a15_neon_acc): Likewise.
877 (cortex_a15_neon_acc_q): Likewise.
878 (cortex_a15_neon_arith_basic): Likewise.
879 (cortex_a15_neon_arith_complex): Likewise.
880 (cortex_a15_neon_multiply): Likewise.
881 (cortex_a15_neon_multiply_q): Likewise.
882 (cortex_a15_neon_mla): Likewise.
883 (cortex_a15_neon_mla_q): Likewise.
884 (cortex_a15_neon_sat_mla_long): Likewise.
885 (cortex_a15_neon_shift_acc): Likewise.
886 (cortex_a15_neon_shift_imm_basic): Likewise.
887 (cortex_a15_neon_shift_imm_complex): Likewise.
888 (cortex_a15_neon_shift_reg_basic): Likewise.
889 (cortex_a15_neon_shift_reg_basic_q): Likewise.
890 (cortex_a15_neon_shift_reg_complex): Likewise.
891 (cortex_a15_neon_shift_reg_complex_q): Likewise.
892 (cortex_a15_neon_fp_negabs): Likewise
893 (cortex_a15_neon_fp_arith): Likewise
894 (cortex_a15_neon_fp_arith_q): Likewise
895 (cortex_a15_neon_fp_cvt_int): Likewise
896 (cortex_a15_neon_fp_cvt_int_q): Likewise
897 (cortex_a15_neon_fp_cvt_16): Likewise
898 (cortex_a15_neon_fp_mul): Likewise
899 (cortex_a15_neon_fp_mul_q): Likewise
900 (cortex_a15_neon_fp_mla): Likewise
901 (cortex_a15_neon_fp_mla_q): Likewise
902 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
903 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
904 (cortex_a15_neon_bitops): Likewise.
905 (cortex_a15_neon_bitops_q): Likewise.
906 (cortex_a15_neon_from_gp): Likewise.
907 (cortex_a15_neon_from_gp_q): Likewise.
908 (cortex_a15_neon_tbl3_tbl4): Likewise.
909 (cortex_a15_neon_zip_q): Likewise.
910 (cortex_a15_neon_to_gp): Likewise.
911 (cortex_a15_neon_load_a): Likewise.
912 (cortex_a15_neon_load_b): Likewise.
913 (cortex_a15_neon_load_c): Likewise.
914 (cortex_a15_neon_load_d): Likewise.
915 (cortex_a15_neon_load_e): Likewise.
916 (cortex_a15_neon_load_f): Likewise.
917 (cortex_a15_neon_store_a): Likewise.
918 (cortex_a15_neon_store_b): Likewise.
919 (cortex_a15_neon_store_c): Likewise.
920 (cortex_a15_neon_store_d): Likewise.
921 (cortex_a15_neon_store_e): Likewise.
922 (cortex_a15_neon_store_f): Likewise.
923 (cortex_a15_neon_store_g): Likewise.
924 (cortex_a15_neon_store_h): Likewise.
925 (cortex_a15_vfp_to_from_gp): Likewise.
926
bc5e57e3
JG		 9272013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
928
929 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
930
931 (cortex_a9_neon_vshl_ddd): Remove.
932 (cortex_a9_neon_vst3_vst4): Likewise.
933 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
934
935 (cortex_a9_neon_bit_ops_q): New.
936
937 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
938 (cortex_a9_neon_int_2): Likewise.
939 (cortex_a9_neon_int_3): Likewise.
940 (cortex_a9_neon_int_4): Likewise.
941 (cortex_a9_neon_int_5): Likewise.
942 (cortex_a9_neon_vqneg_vqabs): Likewise.
943 (cortex_a9_neon_vmov): Likewise.
944 (cortex_a9_neon_vaba): Likewise.
945 (cortex_a9_neon_vaba_qqq): Likewise.
946 (cortex_a9_neon_shift_1): Likewise.
947 (cortex_a9_neon_shift_2): Likewise.
948 (cortex_a9_neon_shift_3): Likewise.
949 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
950 (cortex_a9_neon_vsra_vrsra): Likewise.
951 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
952 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
953 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
954 Likewise.
955 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
956 (cortex_a9_neon_mla_qqq_8_16): Likewise.
957 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
958 Likewise.
959 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
960 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
961 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
962 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
963 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
964 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
965 (cortex_a9_neon_fp_vsum): Likewise.
966 (cortex_a9_neon_fp_vmul_ddd): Likewise.
967 (cortex_a9_neon_fp_vmul_qqd): Likewise.
968 (cortex_a9_neon_fp_vmla_ddd): Likewise.
969 (cortex_a9_neon_fp_vmla_qqq): Likewise.
970 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
971 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
972 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
973 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
974 (cortex_a9_neon_bp_simple): Likewise.
975 (cortex_a9_neon_bp_2cycle): Likewise.
976 (cortex_a9_neon_bp_3cycle): Likewise.
977 (cortex_a9_neon_ldr): Likewise.
978 (cortex_a9_neon_str): Likewise.
979 (cortex_a9_neon_vld1_1_2_regs): Likewise.
980 (cortex_a9_neon_vld1_3_4_regs): Likewise.
981 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
982 (cortex_a9_neon_vld2_4_regs): Likewise.
983 (cortex_a9_neon_vld3_vld4): Likewise.
984 (cortex_a9_neon_vld1_vld2_lane): Likewise.
985 (cortex_a9_neon_vld3_vld4_lane): Likewise.
986 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
987 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
988 (cortex_a9_neon_vst1_3_4_regs): Likewise.
989 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
990 (cortex_a9_neon_vst1_vst2_lane): Likewise.
991 (cortex_a9_neon_vst3_vst4_lane): Likewise.
992 (cortex_a9_neon_mcr): Likewise.
993 (cortex_a9_neon_mcr_2_mcrr): Likewise.
994 (cortex_a9_neon_mrc): Likewise.
995 (cortex_a9_neon_mrrc): Likewise.
996
9deda8cb
JG		 9972013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
998
999 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1000
1001 (cortex_a8_neon_vshl_ddd): Remove.
1002 (cortex_a8_neon_vst3_vst4): Likewise.
1003 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1004
1005 (cortex_a8_neon_bit_ops_q): New.
1006
1007 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1008 (cortex_a8_neon_int_2): Likewise..
1009 (cortex_a8_neon_int_3): Likewise.
1010 (cortex_a8_neon_int_5): Likewise.
1011 (cortex_a8_neon_vqneg_vqabs): Likewise.
1012 (cortex_a8_neon_int_4): Likewise.
1013 (cortex_a8_neon_vaba): Likewise.
1014 (cortex_a8_neon_vaba_qqq): Likewise.
1015 (cortex_a8_neon_shift_1): Likewise.
1016 (cortex_a8_neon_shift_2): Likewise.
1017 (cortex_a8_neon_shift_3): Likewise.
1018 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1019 (cortex_a8_neon_vsra_vrsra): Likewise.
1020 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1021 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1022 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1023 Likewise.
1024 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1025 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1026 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1027 Likewise.
1028 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1029 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1030 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1031 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1032 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1033 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1034 (cortex_a8_neon_fp_vsum): Likewise.
1035 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1036 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1037 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1038 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1039 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1040 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1041 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1042 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1043 (cortex_a8_neon_bp_simple): Likewise.
1044 (cortex_a8_neon_bp_2cycle): Likewise.
1045 (cortex_a8_neon_bp_3cycle): Likewise.
1046 (cortex_a8_neon_ldr): Likewise.
1047 (cortex_a8_neon_str): Likewise.
1048 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1049 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1050 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1051 (cortex_a8_neon_vld2_4_regs): Likewise.
1052 (cortex_a8_neon_vld3_vld4): Likewise.
1053 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1054 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1055 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1056 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1057 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1058 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1059 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1060 (cortex_a8_neon_mcr): Likewise.
1061 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1062 (cortex_a8_neon_mrc): Likewise.
1063 (cortex_a8_neon_mrrc): Likewise.
1064
0f686aa9
JG		 10652013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1066
1067 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1068 (fp): New.
1069 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1070 (aarch64_simd_dup<mode>): Add "type" attribute.
1071 (aarch64_dup_lane<mode>): Likewise.
1072 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1073 (*aarch64_simd_mov<mode>): Likewise.
1074 (aarch64_simd_mov_from_<mode>low): Likewise.
1075 (aarch64_simd_mov_from_<mode>high): Likewise.
1076 (orn<mode>3): Likewise.
1077 (bic<mode>3): Likewise.
1078 (add<mode>3): Likewise.
1079 (sub<mode>3): Likewise.
1080 (mul<mode>3): Likewise.
1081 (*aarch64_mul3_elt<mode>): Likewise.
1082 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1083 (*aarch64_mul3_elt_to_128df): Likewise.
1084 (*aarch64_mul3_elt_to_64v2df): Likewise.
1085 (neg<mode>2): Likewise.
1086 (abs<mode>2): Likewise.
1087 (abd<mode>_3): Likewise.
1088 (aba<mode>_3): Likewise.
1089 (fabd<mode>_3): Likewise.
1090 (*fabd_scalar<mode>3): Likewise.
1091 (and<mode>3): Likewise.
1092 (ior<mode>3): Likewise.
1093 (xor<mode>3): Likewise.
1094 (one_cmpl<mode>2): Likewise.
1095 (aarch64_simd_vec_set<mode>): Likewise.
1096 (aarch64_simd_lshr<mode>): Likewise.
1097 (aarch64_simd_ashr<mode>): Likewise.
1098 (aarch64_simd_imm_shl<mode>): Likewise.
1099 (aarch64_simd_reg_sshl<mode): Likewise.
1100 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1101 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1102 (aarch64_simd_vec_setv2di): Likewise.
1103 (aarch64_simd_vec_set<mode>): Likewise.
1104 (aarch64_mla<mode>): Likewise.
1105 (*aarch64_mla_elt<mode>): Likewise.
1106 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1107 (aarch64_mls<mode>): Likewise.
1108 (*aarch64_mls_elt<mode>): Likewise.
1109 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1110 (<su><maxmin><mode>3): Likewise.
1111 (move_lo_quad_<mode>): Likewise.
1112 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1113 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1114 (vec_pack_trunc_<mode>): Likewise.
1115 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1116 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1117 (*aarch64_<su>mlal_lo<mode>): Likewise.
1118 (*aarch64_<su>mlal_hi<mode>): Likewise.
1119 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1120 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1121 (*aarch64_<su>mlal<mode>): Likewise.
1122 (*aarch64_<su>mlsl<mode>): Likewise.
1123 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1124 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1125 (add<mode>3): Likewise.
1126 (sub<mode>3): Likewise.
1127 (mul<mode>3): Likewise.
1128 (div<mode>3): Likewise.
1129 (neg<mode>2): Likewise.
1130 (abs<mode>2): Likewise.
1131 (fma<mode>4): Likewise.
1132 (*aarch64_fma4_elt<mode>): Likewise.
1133 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1134 (*aarch64_fma4_elt_to_128df): Likewise.
1135 (*aarch64_fma4_elt_to_64v2df): Likewise.
1136 (fnma<mode>4): Likewise.
1137 (*aarch64_fnma4_elt<mode>): Likewise.
1138 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1139 (*aarch64_fnma4_elt_to_128df): Likewise.
1140 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1141 (<frint_pattern><mode>2): Likewise.
1142 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1143 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1144 (vec_unpacks_lo_v4sf): Likewise.
1145 (aarch64_float_extend_lo_v2df): Likewise.
1146 (vec_unpacks_hi_v4sf): Likewise.
1147 (aarch64_float_truncate_lo_v2sf): Likewise.
1148 (aarch64_float_truncate_hi_v4sf): Likewise.
1149 (aarch64_vmls<mode>): Likewise.
1150 (<su><maxmin><mode>3): Likewise.
1151 (<maxmin_uns><mode>3): Likewise.
1152 (reduc_<sur>plus_<mode>): Likewise.
1153 (reduc_<sur>plus_v2di): Likewise.
1154 (reduc_<sur>plus_v2si): Likewise.
1155 (reduc_<sur>plus_<mode>): Likewise.
1156 (aarch64_addpv4sf): Likewise.
1157 (clz<mode>2): Likewise.
1158 (reduc_<maxmin_uns>_<mode>): Likewise.
1159 (reduc_<maxmin_uns>_v2di): Likewise.
1160 (reduc_<maxmin_uns>_v2si): Likewise.
1161 (reduc_<maxmin_uns>_<mode>): Likewise.
1162 (reduc_<maxmin_uns>_v4sf): Likewise.
1163 (aarch64_simd_bsl<mode>_internal): Likewise.
1164 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1165 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1166 (aarch64_get_lane<mode>): Likewise.
1167 (*aarch64_combinez<mode>): Likewise.
1168 (aarch64_combine<mode>): Likewise.
1169 (aarch64_simd_combine<mode>): Likewise.
1170 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1171 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1172 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1173 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1174 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1175 (aarch64_<sur>h<addsub><mode>): Likewise.
1176 (aarch64_<sur><addsub>hn<mode>): Likewise.
1177 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1178 (aarch64_pmul<mode>): Likewise.
1179 (aarch64_<su_optab><optab><mode>): Likewise.
1180 (aarch64_<sur>qadd<mode>): Likewise.
1181 (aarch64_sqmovun<mode>): Likewise.
1182 (aarch64_<sur>qmovn<mode>): Likewise.
1183 (aarch64_s<optab><mode>): Likewise.
1184 (aarch64_sq<r>dmulh<mode>): Likewise.
1185 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1186 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1187 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1188 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1189 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1190 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1191 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1192 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1193 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1194 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1195 (aarch64_sqdmull<mode>): Likewise.
1196 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1197 (aarch64_sqdmull_n<mode>): Likewise.
1198 (aarch64_sqdmull2<mode>_internal): Likewise.
1199 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1200 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1201 (aarch64_<sur>shl<mode>): Likewise.
1202 (aarch64_<sur>q<r>shl<mode>
1203 (aarch64_<sur>shll_n<mode>): Likewise.
1204 (aarch64_<sur>shll2_n<mode>): Likewise.
1205 (aarch64_<sur>shr_n<mode>): Likewise.
1206 (aarch64_<sur>sra_n<mode>): Likewise.
1207 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1208 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1209 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1210 (aarch64_cm<optab><mode>): Likewise.
1211 (aarch64_cm<optab>di): Likewise.
1212 (aarch64_cm<optab><mode>): Likewise.
1213 (aarch64_cm<optab>di): Likewise.
1214 (aarch64_cmtst<mode>): Likewise.
1215 (aarch64_cmtstdi): Likewise.
1216 (aarch64_cm<optab><mode>): Likewise.
1217 (*aarch64_fac<optab><mode>): Likewise.
1218 (aarch64_addp<mode>): Likewise.
1219 (aarch64_addpdi): Likewise.
1220 (sqrt<mode>2): Likewise.
1221 (vec_load_lanesoi<mode>): Likewise.
1222 (vec_store_lanesoi<mode>): Likewise.
1223 (vec_load_lanesci<mode>): Likewise.
1224 (vec_store_lanesci<mode>): Likewise.
1225 (vec_load_lanesxi<mode>): Likewise.
1226 (vec_store_lanesxi<mode>): Likewise.
1227 (*aarch64_mov<mode>): Likewise.
1228 (aarch64_ld2<mode>_dreg): Likewise.
1229 (aarch64_ld2<mode>_dreg): Likewise.
1230 (aarch64_ld3<mode>_dreg): Likewise.
1231 (aarch64_ld3<mode>_dreg): Likewise.
1232 (aarch64_ld4<mode>_dreg): Likewise.
1233 (aarch64_ld4<mode>_dreg): Likewise.
1234 (aarch64_tbl1<mode>): Likewise.
1235 (aarch64_tbl2v16qi): Likewise.
1236 (aarch64_combinev16qi): Likewise.
1237 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1238 (aarch64_st2<mode>_dreg): Likewise.
1239 (aarch64_st2<mode>_dreg): Likewise.
1240 (aarch64_st3<mode>_dreg): Likewise.
1241 (aarch64_st3<mode>_dreg): Likewise.
1242 (aarch64_st4<mode>_dreg): Likewise.
1243 (aarch64_st4<mode>_dreg): Likewise.
1244 (*aarch64_simd_ld1r<mode>): Likewise.
1245 (aarch64_frecpe<mode>): Likewise.
1246 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1247 (aarch64_frecps<mode>): Likewise.
1248
f7379e5e
JG		 12492013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1250
1251 * config/arm/iterators.md (V_elem_ch): New.
1252 (q): Likewise.
1253 (VQH_type): Likewise.
1254 * config/arm/arm.md (is_neon_type): New.
1255 (conds): Use is_neon_type.
1256 (anddi3_insn): Update type attribute.
1257 (xordi3_insn): Likewise.
1258 (one_cmpldi2): Likewise.
1259 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1260 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1261 (*movmisalign<mode>_neon_store): Likewise.
1262 (*movmisalign<mode>_neon_load): Likewise.
1263 (vec_set<mode>_internal): Likewise.
1264 (vec_set<mode>_internal): Likewise.
1265 (vec_setv2di_internal): Likewise.
1266 (vec_extract<mode>): Likewise.
1267 (vec_extract<mode>): Likewise.
1268 (vec_extractv2di): Likewise.
1269 (*add<mode>3_neon): Likewise.
1270 (adddi3_neon): Likewise.
1271 (*sub<mode>3_neon): Likewise.
1272 (subdi3_neon): Likewise.
1273 (fma<VCVTF:mode>4): Likewise.
1274 (fma<VCVTF:mode>4_intrinsic): Likewise.
1275 (*fmsub<VCVTF:mode>4): Likewise.
1276 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1277 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1278 (ior<mode>3): Likewise.
1279 (and<mode>3): Likewise.
1280 (orn<mode>3_neon): Likewise.
1281 (orndi3_neon): Likewise.
1282 (bic<mode>3_neon): Likewise.
1283 (bicdi3_neon): Likewise.
1284 (xor<mode>3): Likewise.
1285 (one_cmpl<mode>2): Likewise.
1286 (abs<mode>2): Likewise.
1287 (neg<mode>2): Likewise.
1288 (negdi2_neon): Likewise.
1289 (*umin<mode>3_neon): Likewise.
1290 (*umax<mode>3_neon): Likewise.
1291 (*smin<mode>3_neon): Likewise.
1292 (*smax<mode>3_neon): Likewise.
1293 (vashl<mode>3): Likewise.
1294 (vashr<mode>3_imm): Likewise.
1295 (vlshr<mode>3_imm): Likewise.
1296 (ashl<mode>3_signed): Likewise.
1297 (ashl<mode>3_unsigned): Likewise.
1298 (neon_load_count): Likewise.
1299 (ashldi3_neon_noclobber): Likewise.
1300 (ashldi3_neon): Likewise.
1301 (signed_shift_di3_neon): Likewise.
1302 (unsigned_shift_di3_neon): Likewise.
1303 (ashrdi3_neon_imm_noclobber): Likewise.
1304 (lshrdi3_neon_imm_noclobber): Likewise.
1305 (<shift>di3_neon): Likewise.
1306 (widen_ssum<mode>3): Likewise.
1307 (widen_usum<mode>3): Likewise.
1308 (quad_halves_<code>v4si): Likewise.
1309 (quad_halves_<code>v4sf): Likewise.
1310 (quad_halves_<code>v8hi): Likewise.
1311 (quad_halves_<code>v16qi): Likewise.
1312 (reduc_splus_v2di): Likewise.
1313 (neon_vpadd_internal<mode>): Likewise.
1314 (neon_vpsmin<mode>): Likewise.
1315 (neon_vpsmax<mode>): Likewise.
1316 (neon_vpumin<mode>): Likewise.
1317 (neon_vpumax<mode>): Likewise.
1318 (*ss_add<mode>_neon): Likewise.
1319 (*us_add<mode>_neon): Likewise.
1320 (*ss_sub<mode>_neon): Likewise.
1321 (*us_sub<mode>_neon): Likewise.
1322 (neon_vadd<mode>_unspec): Likewise.
1323 (neon_vaddl<mode>): Likewise.
1324 (neon_vaddw<mode>): Likewise.
1325 (neon_vhadd<mode>): Likewise.
1326 (neon_vqadd<mode>): Likewise.
1327 (neon_vaddhn<mode>): Likewise.
1328 (neon_vmul<mode>): Likewise.
1329 (neon_vfms<VCVTF:mode>): Likewise.
1330 (neon_vmlal<mode>): Likewise.
1331 (neon_vmls<mode>): Likewise.
1332 (neon_vmlsl<mode>): Likewise.
1333 (neon_vqdmulh<mode>): Likewise.
1334 (neon_vqdmlal<mode>): Likewise.
1335 (neon_vqdmlsl<mode>): Likewise.
1336 (neon_vmull<mode>): Likewise.
1337 (neon_vqdmull<mode>): Likewise.
1338 (neon_vsub<mode>_unspec): Likewise.
1339 (neon_vsubl<mode>): Likewise.
1340 (neon_vsubw<mode>): Likewise.
1341 (neon_vqsub<mode>): Likewise.
1342 (neon_vhsub<mode>): Likewise.
1343 (neon_vsubhn<mode>): Likewise.
1344 (neon_vceq<mode>): Likewise.
1345 (neon_vcge<mode>): Likewise.
1346 (neon_vcgeu<mode>): Likewise.
1347 (neon_vcgt<mode>): Likewise.
1348 (neon_vcgtu<mode>): Likewise.
1349 (neon_vcle<mode>): Likewise.
1350 (neon_vclt<mode>): Likewise.
1351 (neon_vcage<mode>): Likewise.
1352 (neon_vcagt<mode>): Likewise.
1353 (neon_vtst<mode>): Likewise.
1354 (neon_vabd<mode>): Likewise.
1355 (neon_vabdl<mode>): Likewise.
1356 (neon_vaba<mode>): Likewise.
1357 (neon_vabal<mode>): Likewise.
1358 (neon_vmax<mode>): Likewise.
1359 (neon_vmin<mode>): Likewise.
1360 (neon_vpaddl<mode>): Likewise.
1361 (neon_vpadal<mode>): Likewise.
1362 (neon_vpmax<mode>): Likewise.
1363 (neon_vpmin<mode>): Likewise.
1364 (neon_vrecps<mode>): Likewise.
1365 (neon_vrsqrts<mode>): Likewise.
1366 (neon_vqabs<mode>): Likewise.
1367 (neon_vqneg<mode>): Likewise.
1368 (neon_vcls<mode>): Likewise.
1369 (clz<mode>2): Likewise.
1370 (popcount<mode>2): Likewise.
1371 (neon_vrecpe<mode>): Likewise.
1372 (neon_vrsqrte<mode>): Likewise.
1373 (neon_vget_lane<mode>_sext_internal): Likewise.
1374 (neon_vget_lane<mode>_zext_internal): Likewise.
1375 (neon_vdup_n<mode>): Likewise.
1376 (neon_vdup_n<mode>): Likewise.
1377 (neon_vdup_nv2di): Likewise.
1378 (neon_vdup_lane<mode>_interal): Likewise.
1379 (*neon_vswp<mode>): Likewise.
1380 (neon_vcombine<mode>): Likewise.
1381 (float<mode><V_cvtto>2): Likewise.
1382 (floatuns<mode><V_cvtto>2): Likewise.
1383 (fix_trunc<mode><V_cvtto>2): Likewise.
1384 (fixuns_trunc<mode><V_cvtto>2
1385 (neon_vcvt<mode>): Likewise.
1386 (neon_vcvt<mode>): Likewise.
1387 (neon_vcvtv4sfv4hf): Likewise.
1388 (neon_vcvtv4hfv4sf): Likewise.
1389 (neon_vcvt_n<mode>): Likewise.
1390 (neon_vcvt_n<mode>): Likewise.
1391 (neon_vmovn<mode>): Likewise.
1392 (neon_vqmovn<mode>): Likewise.
1393 (neon_vqmovun<mode>): Likewise.
1394 (neon_vmovl<mode>): Likewise.
1395 (neon_vmul_lane<mode>): Likewise.
1396 (neon_vmul_lane<mode>): Likewise.
1397 (neon_vmull_lane<mode>): Likewise.
1398 (neon_vqdmull_lane<mode>): Likewise.
1399 (neon_vqdmulh_lane<mode>): Likewise.
1400 (neon_vqdmulh_lane<mode>): Likewise.
1401 (neon_vmla_lane<mode>): Likewise.
1402 (neon_vmla_lane<mode>): Likewise.
1403 (neon_vmlal_lane<mode>): Likewise.
1404 (neon_vqdmlal_lane<mode>): Likewise.
1405 (neon_vmls_lane<mode>): Likewise.
1406 (neon_vmls_lane<mode>): Likewise.
1407 (neon_vmlsl_lane<mode>): Likewise.
1408 (neon_vqdmlsl_lane<mode>): Likewise.
1409 (neon_vext<mode>): Likewise.
1410 (neon_vrev64<mode>): Likewise.
1411 (neon_vrev32<mode>): Likewise.
1412 (neon_vrev16<mode>): Likewise.
1413 (neon_vbsl<mode>_internal): Likewise.
1414 (neon_vshl<mode>): Likewise.
1415 (neon_vqshl<mode>): Likewise.
1416 (neon_vshr_n<mode>): Likewise.
1417 (neon_vshrn_n<mode>): Likewise.
1418 (neon_vqshrn_n<mode>): Likewise.
1419 (neon_vqshrun_n<mode>): Likewise.
1420 (neon_vshl_n<mode>): Likewise.
1421 (neon_vqshl_n<mode>): Likewise.
1422 (neon_vqshlu_n<mode>): Likewise.
1423 (neon_vshll_n<mode>): Likewise.
1424 (neon_vsra_n<mode>): Likewise.
1425 (neon_vsri_n<mode>): Likewise.
1426 (neon_vsli_n<mode>): Likewise.
1427 (neon_vtbl1v8qi): Likewise.
1428 (neon_vtbl2v8qi): Likewise.
1429 (neon_vtbl3v8qi): Likewise.
1430 (neon_vtbl4v8qi): Likewise.
1431 (neon_vtbl1v16qi): Likewise.
1432 (neon_vtbl2v16qi): Likewise.
1433 (neon_vcombinev16qi): Likewise.
1434 (neon_vtbx1v8qi): Likewise.
1435 (neon_vtbx2v8qi): Likewise.
1436 (neon_vtbx3v8qi): Likewise.
1437 (neon_vtbx4v8qi): Likewise.
1438 (*neon_vtrn<mode>_insn): Likewise.
1439 (*neon_vzip<mode>_insn): Likewise.
1440 (*neon_vuzp<mode>_insn): Likewise.
1441 (neon_vld1<mode>): Likewise.
1442 (neon_vld1_lane<mode>): Likewise.
1443 (neon_vld1_lane<mode>): Likewise.
1444 (neon_vld1_dup<mode>): Likewise.
1445 (neon_vld1_dup<mode>): Likewise.
1446 (neon_vld1_dupv2di): Likewise.
1447 (neon_vst1<mode>): Likewise.
1448 (neon_vst1_lane<mode>): Likewise.
1449 (neon_vst1_lane<mode>): Likewise.
1450 (neon_vld2<mode>): Likewise.
1451 (neon_vld2<mode>): Likewise.
1452 (neon_vld2_lane<mode>): Likewise.
1453 (neon_vld2_lane<mode>): Likewise.
1454 (neon_vld2_dup<mode>): Likewise.
1455 (neon_vst2<mode>): Likewise.
1456 (neon_vst2<mode>): Likewise.
1457 (neon_vst2_lane<mode>): Likewise.
1458 (neon_vst2_lane<mode>): Likewise.
1459 (neon_vld3<mode>): Likewise.
1460 (neon_vld3qa<mode>): Likewise.
1461 (neon_vld3qb<mode>): Likewise.
1462 (neon_vld3_lane<mode>): Likewise.
1463 (neon_vld3_lane<mode>): Likewise.
1464 (neon_vld3_dup<mode>): Likewise.
1465 (neon_vst3<mode>): Likewise.
1466 (neon_vst3qa<mode>): Likewise.
1467 (neon_vst3qb<mode>): Likewise.
1468 (neon_vst3_lane<mode>): Likewise.
1469 (neon_vst3_lane<mode>): Likewise.
1470 (neon_vld4<mode>): Likewise.
1471 (neon_vld4qa<mode>): Likewise.
1472 (neon_vld4qb<mode>): Likewise.
1473 (neon_vld4_lane<mode>): Likewise.
1474 (neon_vld4_lane<mode>): Likewise.
1475 (neon_vld4_dup<mode>): Likewise.
1476 (neon_vst4<mode>): Likewise.
1477 (neon_vst4qa<mode>): Likewise.
1478 (neon_vst4qb<mode>): Likewise.
1479 (neon_vst4_lane<mode>): Likewise.
1480 (neon_vst4_lane<mode>): Likewise.
1481 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1482 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1483 (neon_vec_<US>mult_lo_<mode>): Likewise.
1484 (neon_vec_<US>mult_hi_<mode>): Likewise.
1485 (neon_vec_<US>shiftl_<mode>): Likewise.
1486 (neon_unpack<US>_<mode>): Likewise.
1487 (neon_vec_<US>mult_<mode>): Likewise.
1488 (vec_pack_trunc_<mode>): Likewise.
1489 (neon_vec_pack_trunc_<mode>): Likewise.
1490 (neon_vabd<mode>_2): Likewise.
1491 (neon_vabd<mode>_3): Likewise.
1492
a9e66678
JG		 14932013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1494
1495 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1496 (load_pair): Update type attribute.
1497 (store_pair): Update type attribute.
1498 * config/aarch64/iterators.md (q): New.
1499
20445ed6
JG		 15002013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1501
1502 * config/arm/types.md: Add new types for Neon insns.
1503
afb4ac68
AI		 15042013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1505 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1506 Sergey Lega <sergey.s.lega@intel.com>
1507 Anna Tikhonova <anna.tikhonova@intel.com>
1508 Ilya Tocar <ilya.tocar@intel.com>
1509 Andrey Turetskiy <andrey.turetskiy@intel.com>
1510 Ilya Verbin <ilya.verbin@intel.com>
1511 Kirill Yukhin <kirill.yukhin@intel.com>
1512 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1513
1514 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1515 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1516 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1517 (rcp14<mode>): New.
1518 (srcp14<mode>): Ditto.
1519 (rsqrt14<mode>): Ditto.
1520 (rsqrt14<mode>): Ditto.
1521 (avx512f_vmscalef<mode>): Ditto.
1522 (avx512f_scalef<mode>): Ditto.
1523 (avx512f_getexp<mode>): Ditto.
1524 (avx512f_sgetexp<mode>): Ditto.
1525 (avx512f_fixupimm<mode>): Ditto.
1526 (avx512f_sfixupimm<mode>): Ditto.
1527 (avx512f_rndscale<mode>): Ditto.
1528 (*avx512er_exp2<mode>): Ditto.
1529 (*avx512er_rcp28<mode>): Ditto.
1530 (avx512er_rsqrt28<mode>): Ditto.
1531 (avx512f_getmant<mode>): Ditto.
1532 (avx512f_getmant<mode>): Ditto.
1533 (avx512f_rndscale<mode>): Fix formatting.
1534
a1b499a3
JL		 15352013-10-15 Martin Jambor <mjambor@suse.cz>
1536
1537 * ipa-utils.h (ipa_edge_within_scc): Declare.
1538 * ipa-cp.c (edge_within_scc): Moved...
1539 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
afb4ac68 1540
2e2206fa
AI		 15412013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1542 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1543 Sergey Lega <sergey.s.lega@intel.com>
1544 Anna Tikhonova <anna.tikhonova@intel.com>
1545 Ilya Tocar <ilya.tocar@intel.com>
1546 Andrey Turetskiy <andrey.turetskiy@intel.com>
1547 Ilya Verbin <ilya.verbin@intel.com>
1548 Kirill Yukhin <kirill.yukhin@intel.com>
1549 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1550
1551 * config/i386/predicates.md (const_8_to_15_operand): New.
1552 (const_16_to_31_operand): Ditto.
1553 * config/i386/sse.md (V8FI): New.
1554 (V16FI): Ditto.
1555 (reduc_splus_v8df): Ditto.
1556 (reduc_splus_v16sf): Ditto.
1557 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1558 (vec_extract_hi_<mode>): Ditto.
1559 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1560 (vec_set_lo_<mode>): Ditto.
1561 (vec_set_hi_<mode>): Ditto.
1562 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1563 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1564 (avx512f_pshufd_1): Ditto.
1565 (avx512f_broadcast<mode>): Ditto.
1566 (avx512f_broadcast<mode>): Ditto.
1567 (define_split): Split vec_extract_lo into move.
1568 (ssequartermode): Ditto.
1569 (ssedoublemode): Extened with wider modes.
1570 (vec_extract_lo_<mode>): Ditto.
1571
0fe65b75
AI		 15722013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1573 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1574 Sergey Lega <sergey.s.lega@intel.com>
1575 Anna Tikhonova <anna.tikhonova@intel.com>
1576 Ilya Tocar <ilya.tocar@intel.com>
1577 Andrey Turetskiy <andrey.turetskiy@intel.com>
1578 Ilya Verbin <ilya.verbin@intel.com>
1579 Kirill Yukhin <kirill.yukhin@intel.com>
1580 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1581
1582 * config/i386/predicates.md (register_or_constm1_operand): New.
1583 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1584 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1585 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1586 UNSPEC_SCATTER_PREFETCH
1587 (VI48_512): New.
1588 (avx512f_ucmp<mode>3): Ditto.
1589 (avx512f_vternlog<mode>): Ditto.
1590 (avx512f_align<mode>): Ditto.
1591 (<shift_insn><mode>3): Ditto.
1592 (avx512f_<rotate>v<mode>): Ditto.
1593 (avx512f_<rotate><mode>): Ditto.
1594 (avx512f_eq<mode>3): Ditto.
1595 (avx512f_eq<mode>3_1): Ditto.
1596 (avx512f_gt<mode>3): Ditto.
1597 (avx512f_testm<mode>3): Ditto.
1598 (avx512f_testnm<mode>3): Ditto.
1599 (avx512pf_gatherpf<mode>): Ditto.
1600 (*avx512pf_gatherpf<mode>_mask): Ditto.
1601 (*avx512pf_gatherpf<mode>): Ditto.
1602 (avx512pf_scatterpf<mode>): Ditto.
1603 (*avx512pf_scatterpf<mode>_mask): Ditto.
1604 (*avx512pf_scatterpf<mode>): Ditto.
1605 (avx512f_vec_dup_gpr<mode>): Ditto.
1606 (clz<mode>2): Ditto.
1607 (conflict<mode>): Ditto.
1608 (REDUC_SMINMAX_MODE): Extened with wider modes.
1609 (reduc_<code>_<mode>): Ditto.
1610 (vlshr<mode>3): Ditto.
1611 (vashl<mode>3): Ditto.
1612
ab931c71
AI		 16132013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1614 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1615 Sergey Lega <sergey.s.lega@intel.com>
1616 Anna Tikhonova <anna.tikhonova@intel.com>
1617 Ilya Tocar <ilya.tocar@intel.com>
1618 Andrey Turetskiy <andrey.turetskiy@intel.com>
1619 Ilya Verbin <ilya.verbin@intel.com>
1620 Kirill Yukhin <kirill.yukhin@intel.com>
1621 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1622
1623 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1624 UNSPEC_SCATTER.
1625 (VI48F_512): New.
1626 (avx512fmaskmode): Ditto.
1627 (bcstscalarsuff): Ditto.
1628 (avx512f_blendm<mode>): Ditto.
1629 (cmp_imm_predicate): Ditto.
1630 (avx512f_cmp<mode>3): Ditto.
1631 (avx512f_vec_dup<mode>): Ditto.
1632 (avx512f_vec_dup_mem<mode>): Ditto.
1633 (avx512f_vpermi2var<mode>3): Ditto.
1634 (avx512f_vpermt2var<mode>3): Ditto.
1635 (vec_init<mode>): Ditto.
1636 (avx512f_gathersi<mode>): Ditto.
1637 (*avx512f_gathersi<mode>): Ditto.
1638 (*avx512f_gathersi<mode>_2): Ditto.
1639 (avx512f_gatherdi<mode>): Ditto.
1640 (*avx512f_gatherdi<mode>): Ditto.
1641 (*avx512f_gatherdi<mode>_2): Ditto.
1642 (avx512f_scattersi<mode>): Ditto.
1643 (*avx512f_scattersi<mode>): Ditto.
1644 (avx512f_scatterdi<mode>): Ditto.
1645 (*avx512f_scatterdi<mode>): Ditto.
1646 (sseintprefix): Extened with wider modes.
1647 (VEC_GATHER_IDXSI): Ditto.
1648 (VEC_GATHER_IDXDI): Ditto.
1649 (VEC_GATHER_SRCDI): Ditto.
1650
c9acb877
MGD		 16512013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1652 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1653
1654 * config/arm/t-aprofile: New file.
1655 * config.gcc: Handle --with-multilib-list option.
1656
ecbda484
BS		 16572013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1658
6f1abb55
UB		 1659 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1660 the input for r2.
ecbda484 1661
24f161fd
RB		 16622013-10-15 Richard Biener <rguenther@suse.de>
1663
1664 * tree-loop-distribution.c (build_empty_rdg): Inline into
1665 single user.
1666 (rdg_flag_vertex): Inline into single user.
1667 (rdg_flag_vertex_and_dependent): Likewise.
1668 (build_rdg_partition_for_vertex): Remove processed bitmap.
1669 (rdg_build_partitions): Simplify.
1670
5de989ed
RB		 16712013-10-15 Richard Biener <rguenther@suse.de>
1672
1673 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1674 Restructure forwarding through conversions and copies to
1675 avoid performing copy-propagation the wrong way. Adjust
1676 recursion invocations.
1677 (forward_propagate_addr_expr): Add argument stating if we
1678 are recursing from a single-use.
1679 (ssa_forward_propagate_and_combine): Adjust.
1680
47e0da37
DM		 16812013-10-14 David Malcolm <dmalcolm@redhat.com>
1682
1683 * dumpfile.h (gcc::dump_manager): New class, to hold state
1684 relating to dumpfile management.
1685 (get_dump_file_name): Remove in favor of method of dump_manager.
1686 (dump_initialized_p): Likewise.
1687 (dump_start): Likewise.
1688 (dump_finish): Likewise.
1689 (dump_switch_p): Likewise.
1690 (dump_register): Likewise.
1691 (get_dump_file_info): Likewise.
1692 * context.c (gcc::context::context): Construct the dump_manager
1693 instance.
1694 * context.h (gcc::context::get_dumps): New.
1695 (gcc::context::m_dumps): New.
1696 * coverage.c (coverage_init): Port to dump_manager API.
6f1abb55 1697 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
47e0da37
DM		 1698 (extra_dump_files_in_use): Likewise.
1699 (extra_dump_files_alloced): Likewise.
1700 (gcc::dump_manager::dump_manager): New.
1701 (dump_register): Convert to...
1702 (gcc::dump_manager::dump_register): ...method, replacing
1703 function-static next_dump with m_next_dump field.
1704 (get_dump_file_info): Convert to...
1705 (gcc::dump_manager::get_dump_file_info): ...method.
1706 (get_dump_file_name): Convert to...
1707 (gcc::dump_manager::get_dump_file_name): ...method.
1708 (dump_start): Convert to...
1709 (gcc::dump_manager::dump_start): ...method.
1710 (dump_finish): Convert to...
1711 (gcc::dump_manager::dump_finish): ...method.
1712 (dump_begin): Replace body with...
1713 (gcc::dump_manager::dump_begin): ...new method.
1714 (dump_phase_enabled_p): Convert to...
1715 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1716 (dump_phase_enabled_p): Convert to...
1717 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1718 (dump_initialized_p): Convert to...
1719 (gcc::dump_manager::dump_initialized_p): ...method.
1720 (dump_flag_name): Replace body with...
1721 (gcc::dump_manager::dump_flag_name): ...new method.
1722 (dump_enable_all): Convert to...
1723 (gcc::dump_manager::dump_enable_all): ...new method.
1724 (opt_info_enable_passes): Convert to...
1725 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1726 (dump_switch_p_1): Convert to...
1727 (gcc::dump_manager::dump_switch_p_1): ...new method.
1728 (dump_switch_p): Convert to...
1729 (gcc::dump_manager::dump_switch_p): ...new method.
1730 (opt_info_switch_p): Port to dump_manager API.
1731 (enable_rtl_dump_file): Likewise.
1732 * opts-global.c (handle_common_deferred_options): Port to new
1733 dump_manager API.
1734 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1735 (pass_manager::register_one_dump_file): Likewise.
1736 (pass_manager::register_pass): Likewise.
1737 (pass_init_dump_file): Likewise.
1738 (pass_fini_dump_file): Likewise.
1739 * statistics.c (statistics_early_init): Likewise.
1740
83fd5d11
RB		 17412013-10-14 Richard Biener <rguenther@suse.de>
1742
1743 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1744 iterative_hash_canonical_type, gimple_canonical_type_hash,
1745 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1746 gimple_register_canonical_type, print_gimple_types_stats,
1747 free_gimple_type_tables): Move to lto/lto.c
1748 (gt-gimple.h): Do not include.
1749 * gimple.h (gimple_register_canonical_type,
1750 print_gimple_types_stats, free_gimple_type_tables): Remove.
1751 * Makefile.in (GTFILES): Remove gimple.c.
1752
a6dbd0f7
TS		 17532013-10-14 Travis Snoozy <quandary@remstate.com>
1754
1755 PR target/58716
1756 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1757 scanning for msp430x targets.
1758
b5ccb9ed
EB		 17592013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1760
1761 PR bootstrap/58509
1762 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1763 (registers_ok_for_ldd_peep): Move around.
1764 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1765 * config/sparc/sparc.md (widening peepholes): Use it.
1766
fd7e5084
RB		 17672013-10-14 Richard Biener <rguenther@suse.de>
1768
1769 PR middle-end/58712
1770 PR middle-end/55358
1771 * gimple.c (iterative_hash_canonical_type): Make sure to
1772 record the hash into the correct hashtable slot.
1773
3246afaa
EB		 17742013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1775
1776 PR rtl-optimization/58662
1777 * combine.c (try_combine): Take into account death nodes on I2 when
1778 splitting a PARALLEL of two independent SETs. Fix dump message.
1779
5d30dc5b
OE		 17802013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1781
1782 PR target/51244
1783 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1784 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1785 * config/sh/t-sh (sh_treg_combine.o): New entry.
1786 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1787 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1788 (register_sh_passes): New function. Register sh_treg_combine pass.
1789 (sh_option_override): Invoke it.
1790 (sh_canonicalize_comparison): Handle op0_preserve_value.
1791 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1792 opportunities. Canonicalize branch condition.
1793 (nott): Allow only if pseudos can be created for non-SH2A.
1794
37d1c5db
L		 17952013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1796
1797 PR target/58690
1798 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1799 (ix86_expand_movmem): Replace copy_addr_to_reg with
1800 ix86_copy_addr_to_reg.
1801 (ix86_expand_setmem): Likewise.
1802
b86a70ff
AM		 18032013-10-12 Alexander Monakov <amonakov@ispras.ru>
1804
1805 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1806 provided by ix86_fp_compare_mode instead of CCFPUmode.
1807
6e1d5961
JG		 18082013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1809
1810 * config/aarch64/arm_neon.h
1811 (vtbx<1,3>_<psu>8): Fix register constriants.
1812
ef3cfba2
JL		 18132013-10-11 Jeff Law <law@redhat.com>
1814
1815 PR tree-optimization/58640
adcfd489
UB		 1816 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1817 threading paths that cross over two loop entry points.
ef3cfba2 1818
6e8b7d9c
BS		 18192013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1820
1821 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1822 handle vector float as well.
1823 (*vsx_le_perm_load_v4si): Likewise.
1824 (*vsx_le_perm_store_v2di): Likewise.
1825 (*vsx_le_perm_store_v4si): Likewise.
1826
f200869a
BS		 18272013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1828
1829 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1830 directly to circumvent subtract from splat{31} workaround.
1831 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1832 prototype.
1833 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1834 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1835 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1836 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1837 separate big and little endian logic.
1838 (*altivec_vperm_<mode>_internal): New define_insn.
1839 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1840 separate big and little endian logic.
1841 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1842 (vec_permv16qi): Add little endian logic.
1843
8d28e3fc
MG		 18442013-10-11 Marc Glisse <marc.glisse@inria.fr>
1845
1846 * doc/extend.texi (returns_nonnull): Remove arguments.
1847
3c87b77b
AI		 18482013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1849 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1850 Sergey Lega <sergey.s.lega@intel.com>
1851 Anna Tikhonova <anna.tikhonova@intel.com>
1852 Ilya Tocar <ilya.tocar@intel.com>
1853 Andrey Turetskiy <andrey.turetskiy@intel.com>
1854 Ilya Verbin <ilya.verbin@intel.com>
1855 Kirill Yukhin <kirill.yukhin@intel.com>
1856 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1857
1858 * config/i386/sse.md (VI48F_256_512): New.
1859 (avx2_permvar<mode>): Change to ...
1860 (<avx2_avx512f>_permvar<mode>): This.
1861
50e60d7d
AI		 18622013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1863 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1864 Sergey Lega <sergey.s.lega@intel.com>
1865 Anna Tikhonova <anna.tikhonova@intel.com>
1866 Ilya Tocar <ilya.tocar@intel.com>
1867 Andrey Turetskiy <andrey.turetskiy@intel.com>
1868 Ilya Verbin <ilya.verbin@intel.com>
1869 Kirill Yukhin <kirill.yukhin@intel.com>
1870 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1871
6f1abb55 1872 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
50e60d7d
AI		 1873 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1874 * config/i386/sse.md (VI4_AVX): New.
1875 (sf2simodelower): Ditto.
1876 (sse2_cvtps2dq): Change to ...
1877 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1878
f62ce24f
AI		 18792013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1880 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1881 Sergey Lega <sergey.s.lega@intel.com>
1882 Anna Tikhonova <anna.tikhonova@intel.com>
1883 Ilya Tocar <ilya.tocar@intel.com>
1884 Andrey Turetskiy <andrey.turetskiy@intel.com>
1885 Ilya Verbin <ilya.verbin@intel.com>
1886 Kirill Yukhin <kirill.yukhin@intel.com>
1887 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1888
1889 * config/i386/sse.md (V_512): New.
1890 (VI_512): Ditto.
1891 (vcond<V_512:mode><VF_512:mode>): Ditto.
1892 (vcond<V_512:mode><VI_512:mode>): Ditto.
1893 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1894
558d9f79
AI		 18952013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1896 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1897 Sergey Lega <sergey.s.lega@intel.com>
1898 Anna Tikhonova <anna.tikhonova@intel.com>
1899 Ilya Tocar <ilya.tocar@intel.com>
1900 Andrey Turetskiy <andrey.turetskiy@intel.com>
1901 Ilya Verbin <ilya.verbin@intel.com>
1902 Kirill Yukhin <kirill.yukhin@intel.com>
1903 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1904
1905 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1906 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1907 (FMAMODE): Ditto.
1908 (fma<mode>4): Removed condition.
1909 (fms<mode>4): Ditto.
1910 (fnma<mode>4): Ditto.
1911 (fnms<mode>4): Ditto.
1912 (fma4i_fmadd_<mode>): Ditto.
1913 (*fma_fmadd_<mode>): Ditto.
1914 (*fma_fmsub_<mode>): Ditto.
1915 (*fma_fnmadd_<mode>): Ditto.
1916 (*fma_fnmsub_<mode>): Ditto.
1917 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1918 (*fma_fmaddsub_<mode>): Ditto.
1919 (*fma_fmsubadd_<mode>): Ditto.
1920 (*fmai_fmadd_<mode>): Ditto.
1921 (*fmai_fmsub_<mode>): Ditto.
1922 (*fmai_fnmadd_<mode>): Ditto.
1923 (*fmai_fnmsub_<mode>): Ditto.
1924
e8d08206
AI		 19252013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1926 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1927 Sergey Lega <sergey.s.lega@intel.com>
1928 Anna Tikhonova <anna.tikhonova@intel.com>
1929 Ilya Tocar <ilya.tocar@intel.com>
1930 Andrey Turetskiy <andrey.turetskiy@intel.com>
1931 Ilya Verbin <ilya.verbin@intel.com>
1932 Kirill Yukhin <kirill.yukhin@intel.com>
1933 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1934
1935 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1936 (VI124_256): Changed to ...
1937 (VI124_256_48_512): This.
1938 (ssepackmode): Extended with wider modes.
1939 (<code><mode>3): Changed iterator.
1940 (*avx2_<code><mode>3): Ditto.
1941 (vec_pack_trunc_<mode>): Ditto.
1942
16821545
AI		 19432013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1944 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1945 Sergey Lega <sergey.s.lega@intel.com>
1946 Anna Tikhonova <anna.tikhonova@intel.com>
1947 Ilya Tocar <ilya.tocar@intel.com>
1948 Andrey Turetskiy <andrey.turetskiy@intel.com>
1949 Ilya Verbin <ilya.verbin@intel.com>
1950 Kirill Yukhin <kirill.yukhin@intel.com>
1951 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1952
1953 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1954 (VI8F_256_512): Ditto.
1955 (abs<mode>2): Changed iterator.
1956 (avx2_perm<mode>): Changed to ...
1957 (<avx2_avx512f>_perm<mode>): This.
1958 (avx2_perm<mode>_1): Changed to ...
1959 (<avx2_avx512f>_perm<mode>_1): This.
1960
5348cff8
AI		 19612013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1962 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1963 Sergey Lega <sergey.s.lega@intel.com>
1964 Anna Tikhonova <anna.tikhonova@intel.com>
1965 Ilya Tocar <ilya.tocar@intel.com>
1966 Andrey Turetskiy <andrey.turetskiy@intel.com>
1967 Ilya Verbin <ilya.verbin@intel.com>
1968 Kirill Yukhin <kirill.yukhin@intel.com>
1969 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1970
1971 * config/i386/sse.md (VI48_AVX512F): New.
1972 (VI48_AVX2): Changed to ...
1973 (VI48_AVX2_48_AVX512F): This.
1974 (avx2_ashrv<mode>): Changed to ...
1975 (<avx2_avx512f>_ashrv<mode>): This.
1976 (avx2_<shift_insn>v<mode>): Changed to ...
1977 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1978
f5f41d88
AI		 19792013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1980 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1981 Sergey Lega <sergey.s.lega@intel.com>
1982 Anna Tikhonova <anna.tikhonova@intel.com>
1983 Ilya Tocar <ilya.tocar@intel.com>
1984 Andrey Turetskiy <andrey.turetskiy@intel.com>
1985 Ilya Verbin <ilya.verbin@intel.com>
1986 Kirill Yukhin <kirill.yukhin@intel.com>
1987 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1988
1989 * config/i386/sse.md (VI4_AVX512F): New.
1990 (VI8_AVX2_AVX512F): Ditto.
1991 (mul<mode>3): Extended with wider modes.
1992 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1993 (mul<mode>3): Ditto.
1994 (vec_widen_<s>mult_odd_<mode>): Ditto.
1995
3bdf6340
AI		 19962013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1997 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1998 Sergey Lega <sergey.s.lega@intel.com>
1999 Anna Tikhonova <anna.tikhonova@intel.com>
2000 Ilya Tocar <ilya.tocar@intel.com>
2001 Andrey Turetskiy <andrey.turetskiy@intel.com>
2002 Ilya Verbin <ilya.verbin@intel.com>
2003 Kirill Yukhin <kirill.yukhin@intel.com>
2004 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2005
2006 * config/i386/sse.md (VI2_AVX512F): New.
2007 (VI124_AVX512F): Ditto.
2008 (sseunpackmode): Extended with wider modes.
2009 (sseunpackfltmode): Ditto.
2010 (vec_unpacks_float_hi_<mode>): Ditto.
2011 (vec_unpacks_float_lo_<mode>): Ditto.
2012 (vec_unpacku_float_hi_<mode>): Ditto.
2013 (vec_unpacku_float_lo_<mode>): Ditto.
2014 (vec_unpacks_lo_<mode>): Ditto.
2015 (vec_unpacks_hi_<mode>): Ditto.
2016 (vec_unpacku_lo_<mode>): Ditto.
2017 (vec_unpacku_hi_<mode>): Ditto.
2018
67f783cb
AI		 20192013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2020 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2021 Sergey Lega <sergey.s.lega@intel.com>
2022 Anna Tikhonova <anna.tikhonova@intel.com>
2023 Ilya Tocar <ilya.tocar@intel.com>
2024 Andrey Turetskiy <andrey.turetskiy@intel.com>
2025 Ilya Verbin <ilya.verbin@intel.com>
2026 Kirill Yukhin <kirill.yukhin@intel.com>
2027 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2028
2029 * config/i386/i386.md (multdiv): New.
2030 (multdiv_mnemonic): Ditto.
2031 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2032 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2033 (<sse>_vmdiv<mode>3): Removed.
2034
ec5e777c
AI		 20352013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2036 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2037 Sergey Lega <sergey.s.lega@intel.com>
2038 Anna Tikhonova <anna.tikhonova@intel.com>
2039 Ilya Tocar <ilya.tocar@intel.com>
2040 Andrey Turetskiy <andrey.turetskiy@intel.com>
2041 Ilya Verbin <ilya.verbin@intel.com>
2042 Kirill Yukhin <kirill.yukhin@intel.com>
2043 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2044
2045 * config/i386/sse.md (V): Extended with wider modes.
2046 (VF2): Ditto.
2047 (ssehalfvecmode): Ditto.
2048 (i128): Ditto.
2049 (ssepackfltmode): Ditto.
2050 (avx_vec_concat<mode>): Ditto.
2051 (V_256_512): New iterator.
2052 (VF2_512_256): Ditto.
2053 (si2dfmode): New attribute.
2054 (si2dfmodelower): Ditto.
2055 (sf2dfmode): Ditto.
2056 (concat_tg_mode): Ditto.
2057 (floatv4siv4df2): Changed to ...
2058 (float<si2dfmodelower><mode>2): This.
2059 (avx_cvtps2pd256): Changed to ...
2060 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2061 (vec_pack_trunc_v4df): Changed to ...
2062 (vec_pack_trunc_<mode>): This.
2063 (avx_vpermil<mode>): Changed to ...
2064 (<sse2_avx_avx512f>_vpermil<mode>): This.
2065 (<fixsuffix>fix_truncv8dfv8si2): New.
2066 (vec_pack_sfix_trunc_v8df): Ditto.
2067 (avx512f_rndscale<mode>): Ditto.
2068 (avx512f_roundpd512): Ditto.
2069 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2070
a9ccbba2
AI		 20712013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2072 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2073 Sergey Lega <sergey.s.lega@intel.com>
2074 Anna Tikhonova <anna.tikhonova@intel.com>
2075 Ilya Tocar <ilya.tocar@intel.com>
2076 Andrey Turetskiy <andrey.turetskiy@intel.com>
2077 Ilya Verbin <ilya.verbin@intel.com>
2078 Kirill Yukhin <kirill.yukhin@intel.com>
2079 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2080
2081 * config/i386/i386.md (any_fix): New iterator.
2082 (fixsuffix): New attribute.
2083 * config/i386/sse.md (VF1): Extened with wider modes.
2084 (VI): Ditto.
2085 (VI_AVX2): Ditto.
2086 (VI8): Ditto.
2087 (sseintvecmodelower): Ditto.
2088 (ssescalarmode): Ditto.
2089 (ssescalarnum): Ditto.
2090 (VF1_128_256): New.
2091 (ssexmmmode): Ditto.
2092 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2093 (<sse>_rcp<mode>2): Change iterator.
2094 (rsqrt<mode>2): Ditto.
2095 (<sse>_rsqrt<mode>2): Ditto.
2096 (avx2_vec_dup<mode>): Ditto.
2097 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2098 (round<mode>2_sfix): Ditto.
2099 (avx2_pbroadcast<mode>): Ditto.
2100 (*andnot<mode>3): Handle XI mode.
2101 (*<code><mode>3): Ditto.
2102 (AVXTOSSEMODE): Removed.
2103 (avx_vpermil<mode>): Changed to ...
2104 (<sse2_avx_avx512f>_vpermil<mode>): This.
2105
2b1ebb0c
AI		 21062013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2107 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2108 Sergey Lega <sergey.s.lega@intel.com>
2109 Anna Tikhonova <anna.tikhonova@intel.com>
2110 Ilya Tocar <ilya.tocar@intel.com>
2111 Andrey Turetskiy <andrey.turetskiy@intel.com>
2112 Ilya Verbin <ilya.verbin@intel.com>
2113 Kirill Yukhin <kirill.yukhin@intel.com>
2114 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2115
2116 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2117 (<sse>_comi): Ditto.
2118 (<sse>_ucomi): Ditto.
2119 (sse_cvtss2siq_2): Ditto.
2120 (sse2_cvtsd2si): Ditto.
2121 (sse2_cvtsd2siq): Ditto.
2122 (sse2_cvttsd2si): Ditto.
2123 (sse2_cvttsd2siq): Ditto.
2124 (<shift_insn><mode>3): Ditto.
2125 (sse2_cvtsi2sdq): Update constraint and prefix.
2126 (sse_cvtsi2ss): Update prefix.
2127 (sse_cvtsi2ssq): Ditto.
2128
d89124ed
JJ		 21292013-10-11 Jakub Jelinek <jakub@redhat.com>
2130
2131 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2132 ignore internal calls.
2133
e6ad28c3
RB		 21342013-10-11 Richard Biener <rguenther@suse.de>
2135
6f1abb55
UB		 2136 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2137 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2138 TREE_OVERFLOW is set.
e6ad28c3 2139
d1227447
TS		 21402013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2141
1ac4b8e4
TS		 2142 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2143
7d05cebb
TS		 2144 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2145 * gimple.h: Likewise.
2146
adcfd489 2147 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6545f0b3 2148
1a79fb8a
TS		 2149 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2150 macros.
2151
3afd2873
TS		 2152 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2153
adcfd489 2154 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
d1227447
TS		 2155 * configure: Regenerate.
2156
acf0174b
JJ		 21572013-10-11 Jakub Jelinek <jakub@redhat.com>
2158
2159 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2160 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2161 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2162 similarly pp_right_paren.
2163 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2164 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2165 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2166 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2167 clauses.
2168 (omp_declare_simd_clauses_equal,
2169 omp_remove_redundant_declare_simd_attrs): New functions.
2170 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2171 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2172 * tree.h (OMP_LOOP_CHECK): Define.
2173 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2174 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2175 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2176 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2177 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2178 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2179 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2180 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2181 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2182 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2183 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2184 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2185 (omp_remove_redundant_declare_simd_attrs): New prototype.
2186 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2187 GIMPLE_OMP_TEAMS): New codes.
2188 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2189 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2190 fields.
2191 (target_nesting_level): New variable.
2192 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2193 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2194 for collapse > 1 static schedule unless ordered.
2195 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2196 (determine_parallel_type): Adjust get_ws_args_for caller.
2197 (install_var_field): Handle mask & 4 for double indirection.
2198 (scan_sharing_clauses): Ignore shared clause on teams construct.
2199 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2200 (create_omp_child_function): If inside target or declare target
adcfd489 2201 constructs, set "omp declare target" attribute on the child function.
acf0174b
JJ		 2202 (find_combined_for): New function.
2203 (scan_omp_parallel): Handle combined loops.
2204 (scan_omp_target, scan_omp_teams): New functions.
2205 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2206 restrictions and set ctx->cancellable for cancellable constructs.
2207 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2208 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2209 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
adcfd489 2210 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
acf0174b 2211 (omp_clause_aligned_alignment): New function.
6f1abb55 2212 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
acf0174b
JJ		 2213 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2214 on teams constructs. Handle user defined reductions and new
2215 OpenMP 4.0 clauses.
2216 (lower_reduction_clauses): Don't set placeholder to address of ref
2217 if it has already the right type.
2218 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2219 (expand_parallel_call): Use the new non-_start suffixed builtins,
2220 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2221 and GOMP_parallel_end after the call.
2222 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2223 (expand_omp_for_init_counts): Handle combined loops.
2224 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2225 loops.
2226 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2227 end of cancellable loops.
2228 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2229 Likewise. Handle collapse > 1 loops.
2230 (expand_omp_simd): Handle combined loops.
2231 (expand_omp_for): Add inner_stmt argument, adjust callers of
2232 expand_omp_for* functions, use expand_omp_for_static*chunk even
2233 for collapse > 1 unless ordered.
2234 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2235 of cancellable sections.
2236 (expand_omp_single): Remove need_barrier variable, just rely on
2237 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2238 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2239 (expand_omp_atomic_load, expand_omp_atomic_store,
2240 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2241 (expand_omp_target): New function.
2242 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2243 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2244 (build_omp_regions_1): Immediately close region for
2245 GF_OMP_TARGET_KIND_UPDATE.
2246 (maybe_add_implicit_barrier_cancel): New function.
2247 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2248 cancellation.
2249 (lower_omp_single): Likewise. Add clobber after the barrier.
2250 (lower_omp_taskgroup): New function.
2251 (lower_omp_for): Handle combined loops. Adjust
2252 lower_rec_input_clauses caller. Handle cancellation.
2253 (lower_depend_clauses): New function.
2254 (lower_omp_taskreg): Lower depend clauses. Adjust
2255 lower_rec_input_clauses caller. Add clobber after the call. Handle
2256 cancellation.
2257 (lower_omp_target, lower_omp_teams): New functions.
2258 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2259 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2260 and GOMP_cancellation_point calls.
2261 (lower_omp): Fold stmts inside of target region.
2262 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2263 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2264 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2265 (BT_FN_VOID_OMPFN_PTR_UINT,
2266 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2267 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2268 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2269 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2270 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2271 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2272 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2273 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2274 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2275 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2276 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2277 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2278 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2279 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2280 BUILT_IN_GOMP_PARALLEL_END.
2281 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2282 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2283 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2284 GF_OMP_FOR_KIND_DISTRIBUTE.
2285 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2286 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2287 (dump_gimple_omp_return): Print lhs if it has any.
2288 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2289 gimple_omp_atomic_seq_cst_p.
2290 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2291 and GIMPLE_OMP_TEAMS.
2292 * langhooks.c (lhd_omp_mappable_type): New function.
2293 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2294 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2295 hook.
2296 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2297 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6f1abb55 2298 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
acf0174b
JJ		 2299 (struct gimplify_omp_ctx): Add combined_loop field.
2300 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2301 on stmts inside of target region.
2302 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2303 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2304 ORT_TARGET and ORT_TARGET_DATA.
2305 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2306 Handle GOVD_MAP.
2307 (omp_notice_threadprivate_variable): Complain about threadprivate
2308 variables in target region.
2309 (omp_notice_variable): Complain about vars with non-mappable type
2310 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2311 (omp_check_private): Ignore ORT_TARGET* regions.
2312 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2313 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2314 (find_combined_omp_for): New function.
2315 (gimplify_omp_for): Handle gimplification of combined loops.
2316 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2317 OMP_TEAMS.
2318 (gimplify_omp_target_update): New function.
2319 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2320 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2321 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2322 (gimplify_body): If fndecl has "omp declare target" attribute, add
2323 implicit ORT_TARGET context around it.
2324 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2325 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2326 * tree-nested.c (convert_nonlocal_reference_stmt,
2327 convert_local_reference_stmt, convert_gimple_call): Handle
2328 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2329 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2330 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2331 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2332 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2333 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2334 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2335 BUILT_IN_GOMP_LOOP_END_CANCEL,
2336 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2337 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2338 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2339 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2340 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2341 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2342 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2343 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2344 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2345 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2346 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2347 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2348 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2349 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2350 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2351 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2352 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2353 gimple_build_omp_teams): New functions.
2354 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2355 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2356 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2357 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2358 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2359 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2360 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2361 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2362 GF_OMP_ATOMIC_SEQ_CST): New.
2363 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2364 gimple_build_omp_teams): New prototypes.
2365 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2366 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2367 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2368 GIMPLE_OMP_SINGLE as end of range.
2369 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2370 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2371 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2372 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2373 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2374 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2375 gimple_omp_target_kind, gimple_omp_target_set_kind,
2376 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2377 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2378 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2379 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2380 gimple_omp_teams_set_clauses): New inlines.
2381 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2382 and GIMPLE_OMP_TASKGROUP.
adcfd489 2383 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
acf0174b
JJ		 2384 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2385 enum omp_clause_proc_bind_kind): New.
2386 (union omp_clause_subcode): Add depend_kind, map_kind and
2387 proc_bind_kind fields.
2388 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2389 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2390 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2391 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2392 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2393
67fa7880
TJ		 23942013-10-10 Teresa Johnson <tejohnson@google.com>
2395
2396 * predict.c (tree_estimate_probability): Add new parameter
2397 for estimate_bb_frequencies.
2398 (estimate_bb_frequencies): Add new parameter to force estimation.
2399 (rebuild_frequencies): When max frequency in function is small,
2400 recompute counts from frequencies.
2401 * predict.h (estimate_bb_frequencies): New parameter.
2402
b591a8b7
DM		 24032013-10-10 David Malcolm <dmalcolm@redhat.com>
2404
2405 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2406 optimizations are disabled.
2407
782f0db2
DM		 24082013-10-10 David Malcolm <dmalcolm@redhat.com>
2409
2410 * coverage.c (coverage_finish): Fix leak of da_file_name.
2411
13ef00fa
JH		 24122013-10-10 Jan Hubicka <jh@suse.cz>
2413
2414 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2415 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2416 Bobcat and generic.
2417
a1d50386
JJ		 24182013-10-10 Jakub Jelinek <jakub@redhat.com>
2419
2420 PR middle-end/58670
2421 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2422 if any labels are in FALLTHRU_BB, use a special label emitted
2423 immediately after the asm goto insn rather than label_rtx
2424 of the LABEL_DECL.
2425 (expand_asm_stmt): Adjust caller.
2426 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2427 edge if the last insn in predecessor is a jump with single successor,
2428 but it isn't simplejump_p.
2429
e2c2fde2
RB		 24302013-10-10 Richard Biener <rguenther@suse.de>
2431
2432 PR tree-optimization/58656
6f1abb55 2433 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
e2c2fde2 2434
de6a5608
AM		 24352013-10-10 Andrew MacLeod <amacleod@redhat.com>
2436
adcfd489
UB		 2437 * gimplify.c: Include expr.h and tm_p.h for targets with special
2438 va-arg padding requirements.
de6a5608 2439
b184c8f1
AM		 24402013-10-10 Andrew MacLeod <amacleod@redhat.com>
2441
2442 * tree-flow.h: Move some prototypes to gimple.h.
2443 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2444 * gimple.h: Relocate some prototypes from tree-flow.h
2445 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2446 Move to gimplify.c.
2447 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2448 (build_va_arg_indirect_ref): Relocate and make static.
2449 (std_gimplify_va_arg_expr): Relocate here.
2450 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2451 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2452
13b72c22
AK		 24532013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2454
2455 * doc/md.texi: Document the mnemonic attribute.
2456
5db40447
AK		 24572013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2458
2459 PR target/57377
2460 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2461 (set_attr_alternative x ...) when searching for user defined
2462 mnemonic attribute.
2463
ed018d4b
AM		 24642013-10-10 Andrew MacLeod <amacleod@redhat.com>
2465
2466 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2467
4d5b5e9f
ER		 24682013-10-09 Easwaran Raman <eraman@google.com>
2469
2470 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2471 * cfgexpand.c (defer_stack_allocation): ...use here
2472 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2473
c9ef86a1
ZC		 24742013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2475
2476 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2477 (operand_equal_for_value_replacement): New function, extracted from
2478 value_replacement and enhanced to catch more cases.
2479 (value_replacement): Use operand_equal_for_value_replacement.
2480
2431114f
AM		 24812013-10-09 Andrew MacLeod <amacleod@redhat.com>
2482
adcfd489 2483 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2431114f
AM		 2484 get_max_loop_iterations.
2485
1dd2a9a0
KT		 24862013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2487
2488 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2489 Place comment above function.
2490
c1bf2a39
AM		 24912013-10-09 Andrew MacLeod <amacleod@redhat.com>
2492
2493 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2494 are not related to tree-cfg.c.
2495 * tree-ssa-address.h: New file. Relocate prototypes.
2496 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2497 (addr_for_mem_ref): New. Combine call to get_address_description and
2498 return addr_for_mem_ref.
2499 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2500 * tree-ssa-live.h: Adjust prototypes.
2501 * passes.c: Include tree-ssa-live.h.
2502 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2503 * graphite.c (graphite_transform_loops): Make static.
2504 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
adcfd489 2505 make_pass_graphite, pass_data_graphite_transforms,
c1bf2a39
AM		 2506 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2507 * ipa-pure-const.c (warn_function_noreturn): Make static.
2508 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2509 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2510 Relocate from tree-cfg.c
2511 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2512 static.
2513 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2514 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2515 Move to ipa-pure-const.c.
adcfd489
UB		 2516 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2517 Relocate from tree-optimize.c.
c1bf2a39
AM		 2518 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2519 make_pass_fixup_cfg): Move to tree-cfg.c.
2520 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2521 Relocate some prototypes.
2522 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2523 * tree-dump.c (dump_function_to_file): Remove prototype.
2524 Add tree-flow.h to the include file.
2525 * tree-dump.h: Remove prototype.
2526 * tree-parloops.h: New File. Add prototypes.
2527 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2528 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2529 from tree-ssa-loop.c.
2530 * tree-predcom.c (run_tree_predictive_commoning,
2531 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2532 Relocate here from tree-ssa-loop.c.
adcfd489 2533 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
c1bf2a39
AM		 2534 ssa_name_values.release ().
2535 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2536 (ssa_name_values): Relocate from tree-flow.h.
2537 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2538 * tree-ssa-loop.c (run_tree_predictive_commoning,
2539 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2540 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2541 make_pass_graphite, pass_data_graphite_transforms,
2542 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2543 tree_parallelize_loops, pass_data_parallelize_loops,
2544 make_pass_parallelize_loops): Move to other files.
2545 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2546 moved here.
2547 * tree.h: Remove prototypes from tree-address.c.
2548
cc524fc7
AM		 25492013-10-09 Andrew MacLeod <amacleod@redhat.com>
2550
2551 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2552 (struct int_tree_map): Move to tree-hasher.h
2553 (SCALE, LABEL, PERCENT): Move to gimple.h
2554 * tree-flow-inline.h: Delete. Move functions to other files.
2555 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2556 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2557 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2558 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2559 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2560 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2561 * tree-hasher.h: Don't include tree-flow.h.
2562 (struct int_tree_map): Relocate from tree-flow.h.
2563 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2564 tree-flow-inline.h and make static.
adcfd489
UB		 2565 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2566 tree-flow-inline.h.
cc524fc7
AM		 2567 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2568 tree-flow-inline.h and make static.
2569 * tree.h (is_global_var, may_be_aliased): Relocate from
2570 tree-flow-inline.h.
2571 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2572 * value-prof.c: No longer include tree-flow-inline.h.
2573 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2574
71343877
AM		 25752013-10-09 Andrew MacLeod <amacleod@redhat.com>
2576
2577 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2578 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2579 (enum move_pos): Move to tree-ssa-loop-im.h
2580 * cfgloop.h: Move some prototypes.
2581 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2582 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2583 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2584 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2585 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2586 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2587 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2588 (enum move_pos): Relocate here.
2589 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2590 tree-ssa-loop.c.
2591 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
adcfd489
UB		 2592 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2593 make_pass_lim): Relocate here from tree-ssa-loop.c.
2594 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
71343877
AM		 2595 tree-ssa-loop.c.
2596 (loop_edge_to_cancel, unloop_loops): Make static.
2597 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2598 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2599 (tree_complete_unroll, gate_tree_complete_unroll,
2600 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2601 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2602 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2603 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2604 (stmt_invariant_in_loop_p): Remove unused function.
2605 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2606 * tree-ssa-loop-manip.h: New file. Add prototypes.
2607 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2608 (gcov_type_to_double_int): Move to cfgloop.h.
2609 (double_int_cmp, bound_index,
2610 estimate_numbers_of_iterations_loop): Make static.
2611 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2612 (max_loop_iterations): Factor out get_max_loop_iterations.
2613 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2614 cfgloop.c.
2615 * tree-ssa-loop-niter.h: New file. Add prototypes.
2616 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2617 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2618 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2619 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2620 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2621 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2622 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2623 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2624 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2625 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2626 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2627 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2628 pass_data_complete_unroll, make_pass_complete_unroll,
2629 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2630 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2631 tree-ssa-loop-ivcanon.c.
2632 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2633 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2634 tree-ssa-loop-prefetch.c.
2635 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2636 tree-ssa-loop-im.c.
2637 (get_lsm_tmp_name): Relocate and add suffix parameter.
2638 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2639 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2640 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2641 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
adcfd489 2642 (get_estimated_loop_iterations): Factor out accessor from
71343877 2643 estimated_loop_iterations in tree-ssa-loop-niter.c.
adcfd489
UB		 2644 (get_max_loop_iterations): Factor out accessor from
2645 _max_loop_iterations in tree-ssa-niter.c.
71343877
AM		 2646 * loop-unroll.c (decide_unroll_constant_iterations,
2647 decide_unroll_runtime_iterations, decide_peel_simple,
2648 decide_unroll_stupid): Use new get_* accessors.
2649
826cacfe
MG		 26502013-10-09 Marc Glisse <marc.glisse@inria.fr>
2651
2652 PR tree-optimization/20318
2653 * doc/extend.texi (returns_nonnull): New function attribute.
2654 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2655 attribute.
2656 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2657 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2658
378f8976
EB		 26592013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2660
2661 PR middle-end/58570
2662 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2663 false if both components are bitfields.
2664
0fe04f5c
AV		 26652013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2666
2667 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2668 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2669 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2670 * config/aarch64/aarch64.h
2671 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2672 * config/aarch64/aarch64-simd-builtins.def
2673 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2674
bed9bae4
AV		 26752013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2676
2677 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2678
7f3d8b19
AV		 26792013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2680
2681 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2682
7df625a6
AV		 26832013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2684
2685 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2686 (vneg_f64): New intrinsic.
2687 (vneg_s8): Asm replaced with C.
2688 (vneg_s16): Likewise.
2689 (vneg_s32): Likewise.
2690 (vneg_s64): New intrinsic.
2691 (vnegq_f32): Asm replaced with C.
2692 (vnegq_f64): Likewise.
2693 (vnegq_s8): Likewise.
2694 (vnegq_s16): Likewise.
2695 (vnegq_s32): Likewise.
2696 (vnegq_s64): Likewise.
2697
2b86fca7
RL		 26982013-10-09 Renlin Li <Renlin.Li@arm.com>
2699
2700 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2701
e179df83
AK		 27022013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2703
2704 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2705 packed stack special handling.
2706 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2707 back to fixed stack slots for FPRs saved due to stdarg.
2708
74129172
AK		 27092013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2710
2711 * config/s390/s390.c (s390_frame_info): Restructure function.
2712
6455a49e
AK		 27132013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2714
2715 * config/s390/s390.c (struct s390_frame_layout): New field
2716 gpr_save_slots.
2717 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2718 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2719 regs_ever_clobbered to char*.
adcfd489
UB		 2720 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2721 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6455a49e
AK		 2722 (s390_register_info): Enable FPR save slots. Move/Copy some
2723 functionality into ...
2724 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2725 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2726 function.
2727 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2728 not imply a stack frame.
2729 (s390_init_frame_layout): Remove variable clobbered_regs.
2730 (s390_update_register_info): Remove function.
2731 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2732 cannot be used for register renaming.
2733 (s390_hard_regno_scratch_ok): New function.
2734 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2735 (s390_initial_elimination_offset): Change offset calculation of
2736 the return address pointer.
adcfd489 2737 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6455a49e 2738 (restore_gprs): Set frame related flag.
adcfd489 2739 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6455a49e
AK		 2740 (s390_emit_prologue): Call s390_register_info instead of
2741 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2742 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2743 (s390_optimize_prologue): Call s390_optimize_register_info.
2744 Try to remove also FPR slot save/restore INSNs. Remove frame
2745 related flags from restore INSNs.
2746
59d96342
DD		 27472013-10-08 DJ Delorie <dj@redhat.com>
2748
0dc7adad
DD		 2749 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2750 (movhi): Likewise.
2751
59d96342
DD		 2752 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2753 avoid conflict with the MI use of %c.
2754 * config/rl78/rl78-real.md: change %c to %C throughout.
2755 * config/rl78/rl78-virt.md: Likewise.
2756
a81169d8
JH		 27572013-10-08 Jan Hubicka <jh@suse.cz>
2758
2759 * config/i386/i386.c (ix86_option_override_internal): Switch
2760 to SSE math for -ffast-math when target ISA supports SSE2.
2761
481d1b81
AM		 27622013-10-08 Andrew MacLeod <amacleod@redhat.com>
2763
2764 * tree-flow.h: Remove some prototypes.
2765 * tree.h: Remove some protypes, add a couple.
2766 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2767 using_eh_for_cleanups_p): Add interface routines for front ends.
2768 * tree-eh.h: New file. Add protoptyes.
2769 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2770 (add_stmt_to_eh_lp_fn): Make static.
2771 (lower_try_finally): Use new using_eh_for_cleanups_p.
2772 * emit-rtl.c: Include tree-eh.h.
2773 * gimple.h: Include tree-eh.h.
2774
0e6a0e48
MG		 27752013-10-08 Marc Glisse <marc.glisse@inria.fr>
2776
2777 PR tree-optimization/58480
2778 * tree-vrp.c (infer_nonnull_range): New function.
2779 (infer_value_range): Call infer_nonnull_range.
2780
ef6179d1
DC		 27812013-10-08 Dehao Chen <dehao@google.com>
2782
2783 PR tree-optimization/58619
2784 * tree-inline.c (copy_phis_for_bb): Combine location data
2785 only if non-null.
2786
07d964d5
ZC		 27872013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2788
2789 PR target/58423
2790 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2791 RTX_FRAME_RELATED_P on INSN.
2792
19e9b2a3
BS		 27932013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2794
2795 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2796 (altivec_expand_vec_perm_const): Call it.
2797
0cf68694
BS		 27982013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2799
2800 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2801 sequences for LE VSX loads and stores at expand time.
2802 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2803 prototype.
2804 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2805 (rs6000_gen_le_vsx_permute): New.
2806 (rs6000_gen_le_vsx_load): New.
2807 (rs6000_gen_le_vsx_store): New.
2808 (rs6000_gen_le_vsx_move): New.
2809 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2810 (*vsx_le_perm_load_v4si): New.
2811 (*vsx_le_perm_load_v8hi): New.
2812 (*vsx_le_perm_load_v16qi): New.
2813 (*vsx_le_perm_store_v2di): New.
2814 (*vsx_le_perm_store_v4si): New.
2815 (*vsx_le_perm_store_v8hi): New.
2816 (*vsx_le_perm_store_v16qi): New.
2817 (*vsx_xxpermdi2_le_<mode>): New.
2818 (*vsx_xxpermdi4_le_<mode>): New.
2819 (*vsx_xxpermdi8_le_V8HI): New.
2820 (*vsx_xxpermdi16_le_V16QI): New.
2821 (*vsx_lxvd2x2_le_<mode>): New.
2822 (*vsx_lxvd2x4_le_<mode>): New.
2823 (*vsx_lxvd2x8_le_V8HI): New.
2824 (*vsx_lxvd2x16_le_V16QI): New.
2825 (*vsx_stxvd2x2_le_<mode>): New.
2826 (*vsx_stxvd2x4_le_<mode>): New.
2827 (*vsx_stxvd2x8_le_V8HI): New.
2828 (*vsx_stxvd2x16_le_V16QI): New.
2829
9520e1eb
RL		 28302013-10-07 Renlin Li <Renlin.Li@arm.com>
2831
2832 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2833
36c0bd4f
AK		 28342013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2835
2836 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2837 loop to work also for 31bit ABI.
2838 Save the stack pointer for frame_size > 0.
2839
ee163e72
AK		 28402013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2841
2842 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2843 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2844 constraint letters from expanders.
2845 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2846 retry count to general_operand.
2847 ("tabort"): Give operand 0 a mode.
2848 ("tabort_1"): Add mode and constraint letter for operand 0.
2849 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2850
45d99234
JL		 28512013-10-04 Jeff Law <law@redhat.com>
2852
2853 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2854
adcfd489
UB		 2855 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2856 out of ...
45d99234
JL		 2857 (thread_across_edge): Here. Call it.
2858
7c327f7b
CC		 28592013-10-04 Cary Coutant <ccoutant@google.com>
2860
2861 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
adcfd489 2862 discarding a location list expression (or a piece of one).
7c327f7b 2863
6867e128
JH		 28642013-10-03 Jan Hubicka <jh@suse.cz>
2865
65efa7e7
UB		 2866 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2867 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3700bce3 2868 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6867e128 2869
c7f36d55
JH		 28702013-10-03 Jan Hubicka <jh@suse.cz>
2871
65efa7e7 2872 * config/i386/i386.c (ix86_option_override_internal): Do not enable
c7f36d55
JH		 2873 accumulate-outgoing-args when producing unwind info.
2874
4f0bee4c
WM		 28752013-10-03 Wei Mi <wmi@google.com>
2876
2877 * lra-constraints.c (insert_move_for_subreg): New function
2878 extracted from simplify_operand_subreg.
2879 (simplify_operand_subreg): Add reload for paradoxical subreg.
2880
111c3f39
RX		 28812013-10-03 Rong Xu <xur@google.com>
2882
65efa7e7
UB		 2883 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2884 the candidate of builtin_expect such that we should fix the
2885 size/time estimation.
2886 (estimate_function_body_sizes): Do the acutally size/time fix-up
2887 for builtin_expect.
111c3f39 2888
942df739
RX		 28892013-10-03 Rong Xu <xur@google.com>
2890
65efa7e7
UB		 2891 * predict.c (tree_predict_by_opcode): Get the probability
2892 for builtin_expect from param builtin_expect_probability.
2893 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2894 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2895 * doc/invoke.texi: Add documentation for builtin-expect-probability.
942df739 2896
2284b034
MG		 28972013-10-03 Marc Glisse <marc.glisse@inria.fr>
2898
2899 PR c++/19476
2900 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2901 * calls.c (alloca_call_p): Use get_callee_fndecl.
2902 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2903 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2904 Likewise.
2905 (vrp_visit_stmt): Remove duplicated code.
2906
0609bdf2
MM		 29072013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2908
2909 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2910 ceildf2, btruncdf2, instead of vsx_* name.
2911
2912 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2913 iterators to only do V2DF and V4SF here. Move the DF code to
2914 rs6000.md where it is combined with SF mode. Replace <VSv> with
2915 just 'v' since only vector operations are handled with these insns
2916 after moving the DF support to rs6000.md.
2917 (vsx_sub<mode>3): Likewise.
2918 (vsx_mul<mode>3): Likewise.
2919 (vsx_div<mode>3): Likewise.
2920 (vsx_fre<mode>2): Likewise.
2921 (vsx_neg<mode>2): Likewise.
2922 (vsx_abs<mode>2): Likewise.
2923 (vsx_nabs<mode>2): Likewise.
2924 (vsx_smax<mode>3): Likewise.
2925 (vsx_smin<mode>3): Likewise.
2926 (vsx_sqrt<mode>2): Likewise.
2927 (vsx_rsqrte<mode>2): Likewise.
2928 (vsx_fms<mode>4): Likewise.
2929 (vsx_nfma<mode>4): Likewise.
2930 (vsx_copysign<mode>3): Likewise.
2931 (vsx_btrunc<mode>2): Likewise.
2932 (vsx_floor<mode>2): Likewise.
2933 (vsx_ceil<mode>2): Likewise.
2934 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2935 (vsx_sminsf3): Likewise.
2936 (vsx_fmadf4): Likewise.
2937 (vsx_fmsdf4): Likewise.
2938 (vsx_nfmadf4): Likewise.
2939 (vsx_nfmsdf4): Likewise.
2940 (vsx_cmpdf_internal1): Likewise.
2941
2942 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2943 simpler to select whether a target has SPE or traditional floating
2944 point support in iterators.
2945 (TARGET_DF_SPE): Likewise.
2946 (TARGET_SF_FPR): Likewise.
2947 (TARGET_DF_FPR): Likewise.
2948 (TARGET_SF_INSN): Macros to say whether floating point support
2949 exists for a given operation for expanders.
2950 (TARGET_DF_INSN): Likewise.
2951
2952 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2953 combining of SF/DF mode operations, using both traditional and VSX
2954 registers.
2955 (Fvsx): Likewise.
2956 (Ff): Likewise.
2957 (Fv): Likewise.
2958 (Fs): Likewise.
2959 (Ffre): Likewise.
2960 (FFRE): Likewise.
2961 (abs<mode>2): Combine SF/DF modes using traditional floating point
2962 instructions. Add support for using the upper DF registers with
2963 VSX support, and SF registers with power8-vector support. Update
2964 expanders for operations supported by both the SPE and traditional
2965 floating point units.
2966 (abs<mode>2_fpr): Likewise.
2967 (nabs<mode>2): Likewise.
2968 (nabs<mode>2_fpr): Likewise.
2969 (neg<mode>2): Likewise.
2970 (neg<mode>2_fpr): Likewise.
2971 (add<mode>3): Likewise.
2972 (add<mode>3_fpr): Likewise.
2973 (sub<mode>3): Likewise.
2974 (sub<mode>3_fpr): Likewise.
2975 (mul<mode>3): Likewise.
2976 (mul<mode>3_fpr): Likewise.
2977 (div<mode>3): Likewise.
2978 (div<mode>3_fpr): Likewise.
2979 (sqrt<mode>3): Likewise.
2980 (sqrt<mode>3_fpr): Likewise.
2981 (fre<Fs>): Likewise.
2982 (rsqrt<mode>2): Likewise.
2983 (cmp<mode>_fpr): Likewise.
2984 (smax<mode>3): Likewise.
2985 (smin<mode>3): Likewise.
2986 (smax<mode>3_vsx): Likewise.
2987 (smin<mode>3_vsx): Likewise.
2988 (negsf2): Delete SF operations that are merged with DF.
2989 (abssf2): Likewise.
2990 (addsf3): Likewise.
2991 (subsf3): Likewise.
2992 (mulsf3): Likewise.
2993 (divsf3): Likewise.
2994 (fres): Likewise.
2995 (fmasf4_fpr): Likewise.
2996 (fmssf4_fpr): Likewise.
2997 (nfmasf4_fpr): Likewise.
2998 (nfmssf4_fpr): Likewise.
2999 (sqrtsf2): Likewise.
3000 (rsqrtsf_internal1): Likewise.
3001 (smaxsf3): Likewise.
3002 (sminsf3): Likewise.
3003 (cmpsf_internal1): Likewise.
3004 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3005 (negdf2): Delete DF operations that are merged with SF.
3006 (absdf2): Likewise.
3007 (nabsdf2): Likewise.
3008 (adddf3): Likewise.
3009 (subdf3): Likewise.
3010 (muldf3): Likewise.
3011 (divdf3): Likewise.
3012 (fred): Likewise.
3013 (rsqrtdf_internal1): Likewise.
3014 (fmadf4_fpr): Likewise.
3015 (fmsdf4_fpr): Likewise.
3016 (nfmadf4_fpr): Likewise.
3017 (nfmsdf4_fpr): Likewise.
3018 (sqrtdf2): Likewise.
3019 (smaxdf3): Likewise.
3020 (smindf3): Likewise.
3021 (cmpdf_internal1): Likewise.
3022 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3023 (btrunc<mode>2): Delete separate expander, and combine with the
3024 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3025 (btrunc<mode>2_fpr): Likewise.
3026 (ceil<mode>2): Likewise.
3027 (ceil<mode>2_fpr): Likewise.
3028 (floor<mode>2): Likewise.
3029 (floor<mode>2_fpr): Likewise.
3030 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3031 Add support for using the upper registers with VSX and
3032 power8-vector. Move insns to be closer to the define_expands. On
3033 VSX systems, prefer the traditional form of FMA over the VSX
3034 version, since the traditional form allows the target not to
3035 overlap with the inputs.
3036 (fms<mode>4_fpr): Likewise.
3037 (nfma<mode>4_fpr): Likewise.
3038 (nfms<mode>4_fpr): Likewise.
3039
5bea0c6c 30402013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
65efa7e7 3041 Richard Earnshaw <richard.earnshaw@arm.com>
5bea0c6c
KT		 3042
3043 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3044 (struct mult_cost_table): Likewise.
3045 (struct mem_cost_table): Likewise.
3046 (struct fp_cost_table): Likewise.
3047 (struct vector_cost_table): Likewise.
3048 (cpu_cost_table): Likewise.
3049 * config/arm/arm.opt (mold-rts-costs): New option.
3050 (mnew-generic-costs): Likewise.
3051 * config/arm/arm.c (generic_extra_costs): New table.
3052 (cortexa15_extra_costs): Likewise.
3053 (arm_slowmul_tune): Use NULL as new costs.
3054 (arm_fastmul_tune): Likewise.
3055 (arm_strongarm_tune): Likewise.
3056 (arm_xscale_tune): Likewise.
3057 (arm_9e_tune): Likewise.
3058 (arm_v6t2_tune): Likewise.
3059 (arm_cortex_a5_tune): Likewise.
3060 (arm_cortex_a9_tune): Likewise.
3061 (arm_v6m_tune): Likewise.
3062 (arm_fa726te_tune): Likewise.
3063 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3064 (arm_cortex_tune): Use generict_extra_costs.
3065 (shifter_op_p): New function.
3066 (arm_unspec_cost): Likewise.
3067 (LIBCALL_COST): Define.
3068 (arm_new_rtx_costs): New function.
3069 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3070 table is available. Use old costs otherwise unless mnew-generic-costs
3071 is specified.
3072 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3073 (cpu_cost_table): Declare.
3074
24c56925
MS		 30752013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3076
3077 PR target/58460
3078 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3079 (*subs_mul_imm_<mode>)
3080 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3081 (*sub_<shift>_<mode>)
3082 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3083 Remove k constraint.
3084
2ab8f063
IB		 30852013-10-03 Ian Bolton <ian.bolton@arm.com>
3086
3087 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3088 code.
3089 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3090
79221839
TJ		 30912013-10-02 Teresa Johnson <tejohnson@google.com>
3092
3093 * predict.c (probably_never_executed): New function.
3094 (probably_never_executed_bb_p): Invoke probably_never_executed.
3095 (probably_never_executed_edge_p): Ditto.
3096 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3097 Treat profile insanities conservatively.
3098
d441afe0
JDA		 30992013-10-02 John David Anglin <danglin@gcc.gnu.org>
3100
3101 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3102
80f466c4
VM		 31032013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3104
3105 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3106 use it. Use smaller increase for scratch. Don't increase reject
3107 for early clobber scratch.
3108 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3109 setting eliminated regs except setting fp from hfp.
3110 (lra_eliminate): Check lra_insn_recog_data on NULL.
3111
6e228b4b
MM		 31122013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3113
3114 PR target/58587
3115 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
65efa7e7 3116 setting -mvsx-timode by default until the underlying problem is fixed.
6e228b4b
MM		 3117 (RS6000_CPU, power7 defaults): Likewise.
3118
11b54a5a
UB		 31192013-10-02 Uros Bizjak <ubizjak@gmail.com>
3120
3121 * config/x-linux (host-linux.o): Remove header dependencies.
3122 Use $(COMPILE) and $(POSTCOMPILE).
3123 * config/t-linux-android (linux-android.o): Ditto.
3124
bbc02b69
UB		 31252013-10-02 Uros Bizjak <ubizjak@gmail.com>
3126
3127 * Makefile.in (expmed.o-warn): Remove.
3128
440917de
AM		 31292013-10-02 Andrew MacLeod <amacleod@redhat.com>
3130
3131 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3132 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3133
99206ca9
TJ		 31342013-10-02 Teresa Johnson <tejohnson@google.com>
3135
3136 * dojump.c (do_jump_1): Divide probability between
3137 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3138
56e82b14
TT		 31392013-10-02 Tom Tromey <tromey@redhat.com>
3140
3141 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3142
3d9c733e
AM		 31432013-10-02 Andrew MacLeod <amacleod@redhat.com>
3144
3145 * tree-flow.h: Remove some prototypes.
3146 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3147 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3148 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3149 mark_virtual_phi_result_for_renaming): Relocate here.
3150 * tree-into-ssa.h: Add prototypes.
bbc02b69 3151 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3d9c733e
AM		 3152 single_pred_before_succ_order.
3153 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3154 (nonfreeing_call_p) Move to gimple.c.
3155 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3156 tree-ssa-phiopt.c.
3157 * basic-block.h (single_pred_before_succ_order): Add prototype.
3158 * gimple.c (nonfreeing_call_p): Relocate here.
3159 * gimple.h: Add prototype.
3160 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3161 * tree-ssa-dom.h: New file. Relocate prototypes here.
3162 * tree-ssa.h: Include tree-ssa-dom.h.
3163
78cedfb1
UB		 31642013-10-02 Uros Bizjak <ubizjak@gmail.com>
3165
3166 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3167 Use $(COMPILE) and $(POSTCOMPILE).
3168
3169 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3170
744730a4
AM		 31712013-10-02 Andrew MacLeod <amacleod@redhat.com>
3172
3173 * tree-flow.h: Remove some prototypes.
3174 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3175 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3176 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3177 propagate_tree_value*): Move from here to...
3178 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3179 propagate_tree_value*): Relocate here.
3180 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3181 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3182 * gimple-fold.c: Remove gimple-fold.h from include list.
3183 * tree-vrp.c: Remove gimple-fold.h from include list.
3184 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3185 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3186 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3187 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3188 * sese.c: Add tree-ssa-propagate.h to include list.
3189
826a536d
RB		 31902013-10-02 Richard Biener <rguenther@suse.de>
3191
3192 * tree-loop-distribution.c: Include tree-vectorizer.h for
3193 find_loop_location.
3194 (enum partition_kind): Remove PKIND_REDUCTION.
78cedfb1 3195 (struct partition_s): Remove has_writes member, add reduction_p member.
826a536d
RB		 3196 (partition_alloc): Adjust.
3197 (partition_builtin_p): Likewise.
3198 (partition_has_writes): Remove.
3199 (partition_reduction_p): New function.
3200 (partition_merge_into): Likewise.
3201 (generate_code_for_partition): Commonize builtin partition
3202 handling tail.
3203 (rdg_cannot_recompute_vertex_p): Remove.
3204 (already_processed_vertex_p): Likewise.
3205 (rdg_flag_vertex): Do not set has_writes.
3206 (classify_partition): Adjust.
3207 (rdg_build_partitions): Do not set has_writes, treat all
3208 partitions as useful.
78cedfb1 3209 (distribute_loop): Record number of library calls generated. Adjust.
826a536d
RB		 3210 (tree_loop_distribution): Report number of loops and library
3211 calls generated as opt-info.
3212
4b403ece
AM		 32132013-10-02 Andrew MacLeod <amacleod@redhat.com>
3214
3215 * tree-flow.h: Include new .h files. Move prototypes.
3216 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3217 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3218 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3219 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3220 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3221 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3222 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3223 * tree.h (get_ref_base_and_extent): Move prototype out.
78cedfb1 3224 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4b403ece
AM		 3225 tree-dfa.h.
3226 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3227 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3228 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3229 * tree-scalar-evolution.c: Include tree.h.
3230 * sese.c: Include tree.h.
3231 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3232 * dwarf2out.c: Include tree-dfa.h.
3233 * tree-chrec.c: Include tree.h.
3234 * tree-data-ref.c: Include tree.h.
3235
1d2151c6
YZ		 32362013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3237
3238 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3239 Fix whitespace.
3240
441ad147
RO		 32412013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3242
3243 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3244 Use $(COMPILE) and $(POSTCOMPILE).
3245 (sol2-cxx.o): Likewise.
3246 (sol2-stubs.o): Likewise.
3247 (sol2.o): Likewise.
3248 * config/x-solaris (host-solaris.o): Likewise.
3249
3250 * config/sparc/t-sparc (sparc.o): Remove.
3251 (sparc-c.o): Remove header dependencies.
3252 Use $(COMPILE) and $(POSTCOMPILE).
3253 * config/sparc/x-sparc: Likewise.
3254
2deaf8b0
JR		 32552013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3256
3257 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3258 * config/arc/arc700.md: Likewise.
3259 * config/arc/arc-modes.def: Likewise.
3260 * config/arc/arc-simd.h: Likewise.
3261 * config/arc/t-arc-uClibc: Likewise.
3262 * config/arc/t-arc-newlib: Likewise.
3263
f6fe771a
RL		 32642013-10-02 Renlin Li <renlin.li@arm.com>
3265
78cedfb1
UB		 3266 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3267 plus_constant.
f6fe771a
RL		 3268 (aarch64_expand_epilogue): Likewise.
3269
0916f876
YZ		 32702013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3271 Yufeng Zhang <yufeng.zhang@arm.com>
3272
3273 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3274 declaration.
3275 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3276 'base_in' represent a conversion and legal_cast_p_1 holds; set
3277 'base_in' with the returned value from get_unwidened.
3278
ebfcd719
KT		 32792013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3280
3281 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3282 plus_constant is not used.
3283
157ca3e9
WM		 32842013-10-01 Wei Mi <wmi@google.com>
3285
78cedfb1 3286 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
157ca3e9
WM		 3287 * config/i386/i386.md: Add define_peephole2 to
3288 break partial reg stall for cvtss2sd/cvtsd2ss.
3289
cd4dd8f0
JR		 32902013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3291
3292 * config/arc/arc.c (pass_arc_ifcvt::clone):
78cedfb1 3293 Update for ctxt_ -> m_ctxt change.
cd4dd8f0 3294
1465e5cf
JL		 32952013-10-01 Jeff Law <law@redhat.com>
3296
3297 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3298 outgoing_edge and intermediate_edge fields. Instead store the path.
3299 (redirection_data::hash): Hash on the last edge's destination index.
3300 (redirection_data::equal): Check the entire thread path.
3301 (lookup_redirectio_data): Corresponding changes.
3302 (create_edge_and_update_destination_phis): Likewise.
3303 (thread_single_edge): Likewise.
3304
25c606cb 33052013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
0ccbc132 3306 Diego Novillo <dnovillo@google.com>
25c606cb
JR		 3307
3308 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3309 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3310 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3311 versions of these patterns.
3312
0ccbc132
JR		 3313 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3314 (__builtin_arc_aligned): Likewise.
3315
f55d4a20
JR		 3316 * config/arc/arc.md: Expand adc_0 comment stating the intended
3317 purpose and why it isn't ready.
3318 Replace commented out call_value_via_label_mixed with a
3319 plain comment about bl_s.
3320
5719867d 3321 * config/arc/arc.c (stdio.h): Don't include directly.
5719867d
JR		 3322 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3323 Remove if (1) condition.
3324 (arc_encode_section_info): Fix comment.
3325
6462fab0
JR		 33262013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3327
3328 * config/arc/arc.c (arc_conditional_register_usage):
3329 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3330 Also set reg_alloc_order for DMA config regs.
3331
5d5f6720
JR		 33322013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3333 Jeremy Bennett <jeremy.bennett@embecosm.com>
3334
3335 * doc/install.texi (--with-cpu): Mention ARC.
3336 (arc-*-elf32): New paragraph.
3337 (arc-linux-uclibc): Likewise.
3338 * doc/md.texi (Machine Constraints): Add ARC part.
3339 * doc/invoke.texi: (menu): Add ARC Options.
3340 (Machine Dependent Options) <ARC Options>: Add synopsis.
3341 (node ARC Options): Add.
3342 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3343 (ARC Built-in Functions): New section defining
3344 generic ARC built-in functions.
3345 (ARC SIMD Built-in Functions): New section defining SIMD specific
3346 built-in functions.
3347 (Declaring Attributes of Functions): Extended
3348 description of short_call and long_call attributes for ARC and
3349 added index entries.
3350
526b7aee
SV		 33512013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3352 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3353 Joern Rennecke <joern.rennecke@embecosm.com>
3354 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3355 Brendan Kehoe <brendan@zen.org>
3356 Michael Eager <eager@eagercon.com>
3357 Simon Cook <simon.cook@embecosm.com>
3358 Jeremy Bennett <jeremy.bennett@embecosm.com>
3359
3360 * config/arc, common/config/arc: New directories.
3361
53426f6c
JR		 33622013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3363 Brendan Kehoe <brendan@zen.org>
3364 Simon Cook <simon.cook@embecosm.com>
3365
3366 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3367
1df9f5a9
AM		 33682013-10-01 Andrew MacLeod <amacleod@redhat.com>
3369
3370 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3371 * tree-ssa-coalesce.h: New. Move prototype to here.
3372 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3373 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3374 (gimple_can_coalesce_p): Move to...
3375 * gimple.c (gimple_can_coalesce_p): Here.
3376
80560f95
AM		 33772013-10-01 Andrew MacLeod <amacleod@redhat.com>
3378
3379 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3380 (dump_decl_set): Move to gimple.c.
3381 * gimple.h: Don't include tree-ssa-operands.h.
3382 (dump_decl_set): Add prototype.
3383 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3384 Move to gimple-ssa.h.
3385 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3386 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3387 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3388 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3389 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3390 than PHI_ARG_DEF.
3391 (dump_decl_set): Relocate here.
3392 * gimple-ssa.h: New file.
3393 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3394 Relocate from gimple.h.
3395 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3396 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3397 swap_tree_operands and remove non-ssa path.
3398 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3399 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3400 swap_ssa_operands.
3401 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3402 vect_is_simple_reduction_1): Use swap_ssa_operands.
3403 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3404 (enum need_phi_state): Move to tree-into-ssa.c.
3405 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3406 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3407 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3408 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3409 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3410 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3411 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3412 num_imm_uses): Move to ssa-iterators.h.
3413 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
7ab4168e 3414 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
80560f95
AM		 3415 tree-phinodes.h.
3416 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3417 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3418 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3419 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3420 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3421 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3422 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3423 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3424 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3425 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3426 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3427 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3428 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3429 (set_phi_nodes): Move to tree-phinodes.h.
3430 * tree-ssa-operands.h (enum ssa_op_iter_type,
3431 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3432 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3433 (dump_decl_set): Remove prototype.
3434 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3435 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3436 (set_phi_nodes): Relocate from tree-flow-inline.h.
3437 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3438 tree-flow-inline.h
3439 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3440 include list. Temporarily add gimple.h to include list.
3441 * ssa-iterators.h: New file.
3442 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3443 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3444 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3445 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3446 Relocate from tree-ssa-operands.h.
3447 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3448 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3449 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3450 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3451 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3452 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3453 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3454 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3455 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3456 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3457 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3458 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3459 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3460 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3461 Relocate from tree-flow-inline.h.
3462 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3463
8da00d65
VP		 34642013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3465
3466 * aarch64-simd.md
3467 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3468 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3469 Insert '\t' to output template.
3470 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3471 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3472 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3473 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3474
e2ebe1c2
UB		 34752013-10-01 Uros Bizjak <ubizjak@gmail.com>
3476
3477 * doc/install.texi (Host/target specific installation notes for GCC):
3478 Put @anchor before @heading.
3479 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3480 Use @email for email addresses.
3481
aee2d611
JL		 34822013-10-01 Jeff Law <law@redhat.com>
3483
3484 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3485 a vec. Only delete the path if we create one without successfully
3486 registering a jump thread.
3487 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3488 as a pointer.
3489 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3490 (paths): New vector of jump threading paths.
3491 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3492 (THREAD_PATH): New accessor macro for the entire thread path.
3493 (lookup_redirection_data): Get intermediate and final outgoing edge
3494 from the thread path.
3495 (create_edge_and_update_destination_phis): Copy the threading path.
3496 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3497 jump threading path.
3498 (ssa_redirect_edges): Get edges and block types from the jump threading
3499 path. Free the path vector.
3500 (thread_block): Get edges from the jump threading path. Look at the
3501 entire path to see if we thread to a loop exit. If we cancel a jump
3502 thread request, then free the path vector.
3503 (thread_single_edge): Get edges and block types from the jump threading
3504 path. Free the path vector.
3505 (thread_through_loop_header): Get edges and block types from the jump
3506 threading path. Free the path vector.
3507 (mark_threaded_blocks): Iterate over the vector of paths and store
3508 the path on the appropriate edge. Get edges and block types from the
3509 jump threading path.
3510 (mark_threaded_blocks): Get edges and block types from the jump
3511 threading path. Free the path vector.
3512 (thread_through_all_blocks): Use the vector of paths rather than
3513 a vector of 3-edge sets.
3514 (register_jump_thread): Accept pointer to a path vector rather
3515 than the path vector itself. Store the path vector for later use.
3516 Simplify.
3517
966f97ac 35182013-10-01 Jakub Jelinek <jakub@redhat.com>
8da00d65 3519 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
966f97ac
JJ		 3520
3521 PR target/58574
3522 * config/s390/s390.c (s390_split_branches): Modify check for table
3523 jump insns.
3524 (s390_chunkify_start): Rearrange table jump insn check in order to
3525 deal with compare and branch insns correctly.
3526
3a323a38
KV		 35272013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3528
3529 PR target/58578
3530 Revert
3531 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3532 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3533 define_insn_and_split.
3534 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3535 (shiftsi3_compare): New pattern.
3536 (rrx): New pattern.
3537 * config/arm/unspecs.md (UNSPEC_RRX): New.
3538
90444831
AM		 35392013-10-01 Alan Modra <amodra@gmail.com>
3540
3541 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3542 casing inout operands.
3543
818625cf
RB		 35442013-10-01 Richard Biener <rguenther@suse.de>
3545
3546 PR tree-optimization/58553
3547 * tree-loop-distribution.c (struct partition_s): Add niter member.
3548 (classify_partition): Populate niter member for the partition
3549 and properly identify whether the relevant store happens before
3550 or after the loop exit.
3551 (generate_memset_builtin): Use niter member from the partition.
3552 (generate_memcpy_builtin): Likewise.
3553
30f641cd
RS		 35542013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3555
3556 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3557 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3558
65d3284b
RS		 35592013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3560
3561 * basic-block.h (edge_list): Prefix member names with "m_".
3562 * context.h (context): Likewise.
3563 * domwalk.h (dom_walker): Likewise.
3564 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3565 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3566 * hash-table.h (hash_table): Likewise.
3567 * machmode.h (bit_field_mode_iterator): Likewise.
3568 * pass_manager.h (pass_list): Likewise.
3569 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3570 * tree-pass.h (pass_data): Likewise.
3571 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3572 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3573 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3574 * asan.c (pass_data_asan): Update accordingly.
3575 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3576 (control_dependences::control_dependences): Likewise.
3577 (control_dependences::~control_dependences): Likewise.
3578 (control_dependences::~control_dependences): Likewise.
3579 (control_dependences::get_edges_dependent_on): Likewise.
3580 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3581 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3582 * context.c (gcc::context::context): Likewise.
3583 * cprop.c (pass_rtl_cprop::clone): Likewise.
3584 * domwalk.c (dom_walker::walk): Likewise.
3585 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3586 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3587 * mode-switching.c (pass_mode_switching::clone): Likewise.
3588 * passes.c (opt_pass::opt_pass): Likewise.
3589 (pass_manager::pass_manager): Likewise.
3590 * predict.c (pass_strip_predict_hints::clone): Likewise.
3591 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3592 (pass_split_all_insns::clone): Likewise.
3593 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3594 Likewise.
3595 (bit_field_mode_iterator::next_mode): Likewise.
3596 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3597 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3598 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3599 * tree-complex.c (pass_lower_complex::clone): Likewise.
3600 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3601 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3602 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3603 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3604 (pass_fold_builtins::clone): Likewise.
3605 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3606 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3607 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3608 (pass_cd_dce::clone): Likewise.
3609 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3610 (pass_phi_only_cprop::clone): Likewise.
3611 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3612 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3613 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3614 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3615 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3616 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3617 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3618 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3619 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3620 * tree-vrp.c (pass_vrp::clone): Likewise.
3621 * tsan.c (pass_tsan::clone): Likewise.
3622
f66d0891
JJ		 36232013-09-30 Jakub Jelinek <jakub@redhat.com>
3624
d2a365a8
JJ		 3625 PR middle-end/58564
3626 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3627 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3628
f66d0891
JJ		 3629 PR middle-end/58564
3630 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3631 optimization, punt if sign_bit_p looked through any zero extension.
3632
05357ac3
TJ		 36332013-09-30 Teresa Johnson <tejohnson@google.com>
3634
3635 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3636 Update redirected out edge count in joiner case.
3637 (ssa_redirect_edges): Common the joiner and non-joiner cases
3638 so that joiner case gets profile updates.
3639
ca406576
RB		 36402013-09-30 Richard Biener <rguenther@suse.de>
3641
3642 PR tree-optimization/58554
e2ebe1c2
UB		 3643 * tree-loop-distribution.c (classify_partition): Require
3644 unconditionally executed stores for memcpy and memset recognition.
ca406576
RB		 3645 (tree_loop_distribution): Calculate dominance info.
3646
92d649c4
VK		 36472013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3648
3649 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3650 (NO_PROFILE_COUNTERS): Likewise.
3651 (PROFILE_HOOK): Likewise.
3652 (FUNCTION_PROFILER): Likewise.
3653 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3654
bd9534e2
IS		 36552013-09-30 Iain Sandoe <iain@codesourcery.com>
3656
3657 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3658 calls and defines in TARGET_MACHO conditional.
3659 (load_macho_picbase_di): Likewise.
3660 (reload_macho_picbase): Likewise.
3661 (reload_macho_picbase_si): Likewise.
3662 (reload_macho_picbase_di): Likewise.
3663 (nonlocal_goto_receiver): Likewise.
3664
4494fbc9
NC		 36652013-09-30 Nick Clifton <nickc@redhat.com>
3666
3667 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3668 that use the MSP430X ISA.
3669 (msp430_option_override): Scan -mmcu command line option for any
3670 MCU name that supports the MSP430X ISA.
3671 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3672 -mmcu options which enable the MSP430X ISA.
3673
791d9044
RB		 36742013-09-30 Richard Biener <rguenther@suse.de>
3675
3676 PR middle-end/58532
3677 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3678 before looking for setjmp-like calls.
3679
e1c5c877
IS		 36802013-09-29 Iain Sandoe <iain@codesourcery.com>
3681
3682 PR target/10901
3683 * config/darwin-protos.h (machopic_get_function_picbase): New.
3684 * config/darwin.c (machopic_get_function_picbase): New.
3685 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3686 label for a new func. (load_macho_picbase_di): Likewise.
3687 (reload_macho_picbase): New expand.
3688 (reload_macho_picbase_si): New insn.
3689 (reload_macho_picbase_di): New insn.
3690 (nonlocal_goto_receiver): New define and split.
3691 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3692 (unspecv enum): Add UNSPECV_NLGR.
3693
749278c5
IS		 36942013-09-29 Iain Sandoe <iain@codesourcery.com>
3695
3696 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3697 that altivec registers are correctly sized on Darwin.
3698
2c43a51c
IS		 36992013-09-29 Iain Sandoe <iain@codesourcery.com>
3700
3701 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3702 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3703 * config/x-darwin (host-darwin.o): Likewise.
3704 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3705 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3706 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3707
7a5911d3
UB		 37082013-09-29 Uros Bizjak <ubizjak@gmail.com>
3709
3710 * doc/invoke.texi: Fix usage of @tie{} command.
3711
3f67a73d
EB		 37122013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3713
3714 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3715
9ac2f538
JH		 37162013-09-28 Jan Hubicka <jh@suse.cz>
3717
3718 * config/i386/x86-tune.def: Add documentation for each of the options;
3719 add whitespace.
3720
b0ff06da
JH		 37212013-09-28 Jan Hubicka <jh@suse.cz>
3722
7a5911d3
UB		 3723 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3724 generic.
b0ff06da
JH		 3725 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3726 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3727 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3728 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3729 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3730
c3284718
RS		 37312013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3732
3733 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3734 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3735 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3736 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3737 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3738 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3739 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3740 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3741 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3742 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3743 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3744 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3745 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3746 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3747 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3748 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3749 gimple.h, godump.c, graphite-clast-to-gimple.c,
3750 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3751 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3752 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3753 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3754 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3755 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3756 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3757 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3758 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3759 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3760 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3761 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3762 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3763 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3764 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3765 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3766 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3767 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3768 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3769 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3770 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3771 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3772 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3773 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3774 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3775 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3776 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3777 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3778 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3779 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3780 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3781 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3782 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3783 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3784 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3785 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3786 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3787 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3788 whitespace before "(".
3789
c6285bd7
SL		 37902013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3791
3792 * expr.h (extract_bit_field): Remove packedp parameter.
3793 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3794 from forward declaration.
3795 (store_split_bit_field): Remove packedp arg from calls to
3796 extract_fixed_bit_field.
3797 (extract_bit_field_1): Remove packedp parameter and packedp
3798 argument from recursive calls and calls to extract_fixed_bit_field.
3799 (extract_bit_field): Remove packedp parameter and corresponding
3800 arg to extract_bit_field_1.
3801 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3802 to issue warnings.
3803 (extract_split_bit_field): Remove packedp arg from call to
3804 extract_fixed_bit_field.
3805 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3806 (copy_blkmode_from_reg): Likewise.
3807 (copy_blkmode_to_reg): Likewise.
3808 (read_complex_part): Likewise.
3809 (store_field): Likewise.
3810 (expand_expr_real_1): Likewise.
3811 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3812 to extract_bit_field.
3813 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3814 call to extract_bit_field.
3815 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3816 call to extract_bit_field.
3817 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3818 and special packedp behavior from -fstrict-volatile-bitfields
3819 documentation.
3820
522d4efc
JBG		 38212013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3822
3823 * lra-eliminations.c (init_elim_table): Guard value_p.
3824
19cdb489
MM		 38252013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3826
3827 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3828 DFmode, DImode, and SFmode in the upper VSX registers based on the
3829 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3830 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3831 of the VSX register handling.
3832
3833 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3834 (f32_sv): Likewise.
3835 (zero_extendsidi2_lfiwzx): Add support for loading into the
3836 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3837 only do VSX memory options on Altivec registers.
3838 (extendsidi2_lfiwax): Likewise.
3839 (extendsfdf2_fpr): Likewise.
3840 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3841 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3842 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3843 (movdi_internal64): Likewise.
3844
d6d11272
XDL		 38452013-09-27 Xinliang David Li <davidxl@google.com>
3846
3847 * opts.c (finish_options): Adjust parameters
3848 according to vect cost model.
3849 (common_handle_option): Set dynamic vect cost
3850 model for FDO.
3851 targhooks.c (default_add_stmt_cost): Compute stmt cost
3852 unconditionally.
3853 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3854 Use helper function.
3855 * tree-vectorizer.h (unlimited_cost_model): New function.
3856 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3857 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3858 function.
3859 (vect_enhance_data_refs_alignment): Ditto.
3860 * flag-types.h: New enum.
3861 * common/config/i386/i386-common.c (ix86_option_init_struct):
3862 No need to initialize vect_cost_model flag.
3863 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3864 unconditionally.
3865
ac1857a3
DN		 38662013-09-27 Diego Novillo <dnovillo@google.com>
3867
3868 * gimple.h (enum ssa_mode): Remove.
3869
0d0e2af6
PM		 38702013-09-27 Paulo Matos <pmatos@broadcom.com>
3871
3872 * cfgloop.h (number_of_loops): Fix typo in check for null.
3873
09dc585e
JJ		 38742013-09-27 Jakub Jelinek <jakub@redhat.com>
3875
3876 PR middle-end/58551
3877 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3878 are children of outermost saved_cfun's loop, and set it up to
3879 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3880 if loop != loop0 and SESE region contains bbs that belong to loop0.
3881
ec5a3504
RS		 38822013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3883
3884 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3885 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3886 (set_address_segment, set_address_base, set_address_index)
3887 (set_address_disp): Accept the argument unconditionally.
3888 (baseness): Remove must_be_base_p and must_be_index_p checks.
3889 (decompose_normal_address): Classify as much as possible in the
3890 main loop.
3891
f91aec98
RS		 38922013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3893
3894 * cse.c (count_reg_usage): Handle INT_LIST.
3895 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3896 * reginfo.c (reg_scan_mark_refs): Likewise.
3897 * reload1.c (eliminate_regs_1): Likewise.
3898
a9195970
IS		 38992013-09-27 Iain Sandoe <iain@codesourcery.com>
3900
3901 PR middle-end/58547
3902 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3903 signed.
3904
e6f1c509
RB		 39052013-09-27 Richard Biener <rguenther@suse.de>
3906
3907 PR tree-optimization/58459
3908 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3909 restriction not propagating into loops.
3910
84f48495
FW		 39112013-09-26 Florian Weimer <fw@deneb.enyo.de>
3912
3913 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3914 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3915 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3916
74fc8b8a
RB		 39172013-09-26 Richard Biener <rguenther@suse.de>
3918
7a5911d3 3919 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
74fc8b8a 3920
b4ada065
RB		 39212013-09-26 Richard Biener <rguenther@suse.de>
3922
3923 * alias.h (component_uses_parent_alias_set): Rename to ...
3924 (component_uses_parent_alias_set_from): ... this.
3925 * alias.c (component_uses_parent_alias_set): Rename to ...
3926 (component_uses_parent_alias_set_from): ... this and return
3927 the desired parent.
3928 (reference_alias_ptr_type_1): Use the result from
3929 component_uses_parent_alias_set_from instead of stripping
3930 components one at a time.
3931 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3932
78bca40d
AM		 39332013-09-26 Andrew MacLeod <amacleod@redhat.com>
3934
7a5911d3
UB		 3935 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3936 Move prototypes to...
78bca40d
AM		 3937 * tree-ssa-ter.h: New File. Move prototypes here.
3938 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3939 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3940 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7a5911d3 3941 from is_replaceable_p.
78bca40d
AM		 3942 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3943 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3944 (process_replaceable): Use ter_is_replaceable_p.
3945 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3946 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3947 newly refactored ssa_is_replaceable_p.
3948 * cfgexpand.c: Include tree-outof-ssa.h.
3949 * ssaexpand.h: Delete.
3950
ff2a63a7
AM		 39512013-09-26 Andrew MacLeod <amacleod@redhat.com>
3952
3953 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3954 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3955 tree-ssa.c
3956 (create_gimple_tmp): Delete.
3957 (get_expr_type, build_assign, build_type_cast): Move to...
3958 * gimple-builder.c: New File.
3959 (get_expr_type): Relocate from gimple.c.
3960 (build_assign, build_type_cast): Change to only create ssanames.
3961 * gimple.h: Move prototypes to...
3962 * gimple-builder.h: New File. Here.
3963 * tree-ssa.h: And here.
3964 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3965 count_uses_and_derefs): Relocate from gimple.c.
3966 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3967 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3968 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3969 gimple_replace_ssa_lhs.
3970 * asan.c: Include gimple-builder.h.
3971 * Makefile.in: Add gimple-builder.o.
3972
a2544177
RB		 39732013-09-26 Richard Biener <rguenther@suse.de>
3974
3975 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3976 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3977 (loe_visit_block): Use gcc_checking_assert.
3978 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3979 gimple_assign_ssa_name_copy_p.
3980 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3981 change.
3982
6d1ca9a0
DE		 39832013-09-26 David Edelsohn <dje.gcc@gmail.com>
3984
3985 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3986 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3987
46e950db
RB		 39882013-09-26 Richard Biener <rguenther@suse.de>
3989
3990 PR tree-optimization/58539
3991 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3992 the fact that debug statements are not taking part in loop-closed
3993 SSA construction.
3994
40ada30a
NC		 39952013-09-26 Nick Clifton <nickc@redhat.com>
3996
3997 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3998 time warning message.
3999 (msp430_print_operand_raw): Delete unused letter parameter.
4000 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4001 (msp430_print_operand_address): New function.
4002 (msp430_print_operand): Move address printing code from here to
4003 new function.
4004 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4005 assembler.
4006 (zero_extendpsisi2): Likewise.
4007 (extendpsisi2): New pattern.
4008 (andneghi3): New pattern.
4009
38e8f663
YR		 40102013-09-26 Yvan Roux <yvan.roux@linaro.org>
4011
4012 * config/aarch64/aarch64.opt (mlra): New option.
4013 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4014 (TARGET_LRA_P): Define.
4015
4167e8d4
EB		 40162013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4017
4018 * expr.c (expand_assignment): Remove obsolete comment.
4019
5254eac4
JL		 40202013-09-25 Jeff Law <law@redhat.com>
4021
4022 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4023 tree-ssa-threadupdate.h.
4024 (register_jump_thread): Similarly.
4025 * tree-ssa-threadupdate.h: New header file.
4026 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4027 * tree-vrp.c: Likewise.
4028 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4029 (thread_around_empty_blocks): Change type of path vector argument to
4030 an edge,type pair from just an edge. Initialize both elements when
4031 appending to a jump threading path. Tweak references to elements
4032 appropriately.
4033 (thread_across_edge): Similarly. Release memory for the elements
4034 as needed.
4035 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4036 (dump_jump_thread_path): New function broken out from
4037 register_jump_thread.
4038 (register_jump_thread): Use dump_jump_thread_path. Change type of
4039 path vector entries. Search the path for NULL edges and dump
4040 the path if one is found. Tweak the conversion of path to 3-edge
4041 form to use the block copy type information embedded in the path.
4042
20f114a3
YR		 40432013-09-25 Yvan Roux <yvan.roux@linaro.org>
4044
4045 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4046
3936bafc
YR		 40472013-09-25 Yvan Roux <yvan.roux@linaro.org>
4048 Vladimir Makarov <vmakarov@redhat.com>
4049
4050 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4051 from the least significant bit.
4052 (strip_address_mutations): Add bitfield operations handling.
4053 (must_be_index_p): Add shifting and rotate operations handling.
4054 (set_address_base): Use must_be_base_p predicate.
4167e8d4 4055 (set_address_index): Use must_be_index_p predicate.
3936bafc 4056
b86f6e9e
AI		 40572013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4058 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4059 Sergey Lega <sergey.s.lega@intel.com>
4060 Anna Tikhonova <anna.tikhonova@intel.com>
4061 Ilya Tocar <ilya.tocar@intel.com>
4062 Andrey Turetskiy <andrey.turetskiy@intel.com>
4063 Ilya Verbin <ilya.verbin@intel.com>
4064 Kirill Yukhin <kirill.yukhin@intel.com>
4065 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4066
4067 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4068 Use new names.
4069 (ix86_expand_vector_move_misalign): Support new unaligned load and
4070 stores and use new names.
4071 (CODE_FOR_sse2_storedqu): Rename to ...
4072 (CODE_FOR_sse2_storedquv16qi): ... this.
4073 (CODE_FOR_sse2_loaddqu): Rename to ...
4074 (CODE_FOR_sse2_loaddquv16qi): ... this.
4075 (CODE_FOR_avx_loaddqu256): Rename to ...
4076 (CODE_FOR_avx_loaddquv32qi): ... this.
4077 (CODE_FOR_avx_storedqu256): Rename to ...
4078 (CODE_FOR_avx_storedquv32qi): ... this.
4079 * config/i386/i386.md (fpint_logic): New.
4080 * config/i386/sse.md (VMOVE): Extend for AVX512.
4081 (VF): Ditto.
4082 (VF_128_256): New.
4083 (VF_512): Ditto.
4084 (VI_UNALIGNED_LOADSTORE): Ditto.
4085 (sse2_avx_avx512f): Ditto.
4086 (sse2_avx2): Extend for AVX512.
4087 (sse4_1_avx2): Ditto.
4088 (avx2_avx512f): New.
4089 (sse): Extend for AVX512.
4090 (sse2): Ditto.
4091 (sse4_1): Ditto.
4092 (avxsizesuffix): Ditto.
4093 (sseintvecmode): Ditto.
4094 (ssePSmode): Ditto.
4095 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4096 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4097 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4098 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4099 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4100 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4101 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4102 (STORENT_MODE): Extend for AVX512.
4103 (*absneg<mode>2): Replace constraint "x" with "v".
4104 (*mul<mode>3): Ditto.
4105 (*ieee_smin<mode>3): Ditto.
4106 (*ieee_smax<mode>3): Ditto.
4107 (avx_cmp<mode>3): Replace VF with VF_128_256.
4108 (*<sse>_maskcmp<mode>3_comm): Ditto.
4109 (<sse>_maskcmp<mode>3): Ditto.
4110 (<sse>_andnot<mode>3): Extend for AVX512.
4111 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4112 (<code><mode>3, fpint_logic): New.
4113 (*<code><mode>3): Extend for AVX512.
4114 (avx512flogicsuff): New.
4115 (avx512f_<logic><mode>): Ditto.
4116 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4117 VF_128_256.
4118 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4119 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4120 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4121 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4122 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4123 (xop_vpermil2<mode>3): Ditto.
4124 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4125 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4126 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4127 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4128
4d44d03c
TT		 41292013-09-25 Tom Tromey <tromey@redhat.com>
4130
4131 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4132 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4133 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4134 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4135 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4136 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4137 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4138 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4139 (GRAPHITE_HTAB_H): Remove.
4140
86aaf68d
TT		 41412013-09-25 Tom Tromey <tromey@redhat.com>
4142
4143 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4144
c987deb8
TT		 41452013-09-25 Tom Tromey <tromey@redhat.com>
4146
4147 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4148
9439747e
TT		 41492013-09-25 Tom Tromey <tromey@redhat.com>
4150
4151 * config/i386/t-i386 (i386.o): Remove.
4152 (i386-c.o): Use COMPILE and POSTCOMPILE.
4153
aefc31a1
TT		 41542013-09-25 Tom Tromey <tromey@redhat.com>
4155
4156 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4157
b6541edc
TT		 41582013-09-25 Tom Tromey <tromey@redhat.com>
4159
4160 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4161 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4162 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4163 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4164 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4165 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4166 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4167 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4168 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4169 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4170 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4171 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4172 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4173 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4174 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4175 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4176 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4177 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4178 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4179 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4180 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4181 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4182 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4183 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4184 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4185 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4186 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4187 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4188 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4189 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4190 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4191 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4192 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4193 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4194 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4195 (graphite-clast-to-gimple.o, graphite-dependences.o)
4196 (graphite-interchange.o, graphite-poly.o)
4197 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4198 (graphite-optimize-isl.o, tree-vect-loop.o)
4199 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4200 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4201 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4202 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4203 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4204 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4205 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4206 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4207 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4208 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4209 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4210 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4211 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4212 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4213 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4214 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4215 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4216 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4217 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4218 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4219 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4220 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4221 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4222 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4223 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4224 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4225 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4226 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4227 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4228 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4229 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4230 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4231 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4232 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4233 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4234 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4235 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4236 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4237 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4238 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4239 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4240 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4241 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4242 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4243 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4244 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4245 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4246 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4247 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4248 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4249 (gcov-dump.o): Remove.
4250 (default-c.o): Use COMPILE and POSTCOMPILE.
4251 (CFLAGS-gcc.o): New variable.
4252 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4253
8c796d19
TT		 42542013-09-25 Tom Tromey <tromey@redhat.com>
4255
4256 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4257 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4258 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4259 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4260 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4261 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4262 (c-family/array-notation-common.o, c-family/stub-objc.o)
4263 (c-family/c-ubsan.o): Remove.
4264
0a6c2227
TT		 42652013-09-25 Tom Tromey <tromey@redhat.com>
4266
4267 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4268
f3bc55f0
TT		 42692013-09-25 Tom Tromey <tromey@redhat.com>
4270
4271 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4272 to add -DENABLE_SHARED_LIBGCC.
4273 (gcc.o): Don't use subshell.
4274
c11c795e
TT		 42752013-09-25 Tom Tromey <tromey@redhat.com>
4276
4277 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4278 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4279 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4280 * configure, config.in: Rebuild.
4281
043378c3
TT		 42822013-09-25 Tom Tromey <tromey@redhat.com>
4283
4284 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4285 (COMPILE, POSTCOMPILE): New variables.
4286 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4287 (DEPFILES): New variable.
4288 Include ".Po" files.
4289 * configure.ac: Add checks for dependency checking.
4290 * configure, aclocal.m4: Regenerate.
4291
eab34643
TT		 42922013-09-25 Tom Tromey <tromey@redhat.com>
4293
4294 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7a5911d3 4295 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
eab34643 4296
d2db5e91
TT		 42972013-09-25 Tom Tromey <tromey@redhat.com>
4298
4299 * Makefile.in (generated_files): Add options.h,
4300 target-hooks-def.h, insn-opinit.h,
4301 common/common-target-hooks-def.h, pass-instances.def,
4302 c-family/c-target-hooks-def.h.
4303
af33629e
JL		 43042013-09-25 Jeff Law <law@redhat.com>
4305
4306 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4307 than foo[foo.length () - 1] to access last member in a vec.
4308 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4309
ed0d3051
RB		 43102013-09-25 Richard Biener <rguenther@suse.de>
4311
4312 PR middle-end/58521
4313 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4314
123485ca
JH		 43152013-09-25 Jan Hubicka <jh@suse.cz>
4316
4317 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4318 test.
4319
d70b0f1f
MP		 43202013-09-25 Marek Polacek <polacek@redhat.com>
4321
4322 PR sanitizer/58420
4323 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4324 when determining the type name.
4325
e3f02798
OE		 43262013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4327
4328 * config/sh/sh.md: Fix formatting.
4329
7a5911d3 43302013-09-24 Xinliang David Li <davidxl@google.com>
4f17aa0b 4331
7a5911d3
UB		 4332 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4333 max peel iterations parameter.
4f17aa0b
XDL		 4334 * param.def: New parameter.
4335 * doc/invoke.texi: Document New parameter.
4336
0498471b
CL		 43372013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4338
4339 * gimple-pretty-print.c: Various whitespace tweaks.
4340 * tree-core.h: Likewise.
4341 * tree-pretty-print.c: Likewise.
4342 * tree-ssa-alias.c: Likewise.
4343 * tree-ssa-copy.c: Likewise.
4344 * tree-ssanames.c: Likewise.
4345 * tree-ssanames.h: Likewise.
4346 * tree-vrp.c: Likewise.
4347
5a8c2b57
AM		 43482013-09-24 Alan Modra <amodra@gmail.com>
4349
4350 PR middle-end/57134
4351 PR middle-end/57586
7a5911d3
UB		 4352 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4353 for output operands that disallow regs. Don't use EXPAND_WRITE on
4354 inout operands.
5a8c2b57 4355
f40333af
RB		 43562013-09-24 Richard Biener <rguenther@suse.de>
4357
4358 PR middle-end/58513
4359 * tree.c (reference_alias_ptr_type): Move ...
4360 * alias.c (reference_alias_ptr_type): ... here and implement
4361 in terms of the new reference_alias_ptr_type_1.
4362 (ref_all_alias_ptr_type_p): New helper.
4363 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4364 use ref_all_alias_ptr_type_p.
4365 (get_deref_alias_set): Add flag_strict_aliasing check here.
4366 (reference_alias_ptr_type_1): New function, split out from ...
4367 (get_alias_set): ... here.
4368 (alias_ptr_types_compatible_p): New function.
4369 * alias.h (reference_alias_ptr_type): Declare.
4370 (alias_ptr_types_compatible_p): Likewise.
4371 * tree.h (reference_alias_ptr_type): Remove.
4372 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4373 to compare MEM_REF alias types.
4374
583e8bf5
RB		 43752013-09-24 Richard Biener <rguenther@suse.de>
4376
4377 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4378
5e8586d7
MM		 43792013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4380
19c3e797
MM		 4381 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4382 reload helper function arrays into a single array reg_addr.
4383 (reload_fpr_gpr): Likewise.
4384 (reload_gpr_vsx): Likewise.
4385 (reload_vsx_gpr): Likewise.
4386 (struct rs6000_reg_addr): Likewise.
4387 (reg_addr): Likewise.
4388 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4389 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4390 (rs6000_init_hard_regno_mode_ok): Likewise.
4391 (rs6000_secondary_reload_direct_move): Likewise.
4392 (rs6000_secondary_reload): Likewise.
4393
5e8586d7
MM		 4394 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4395 constraints: wu, ww, and wy. Repurpose wv constraint added during
4396 power8 changes. Put wg constraint in alphabetical order.
4397
4398 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4399 for future work to add ISA 2.07 VSX single precision support.
4400 (-mvsx-scalar-double): Change default from -1 to 1, update
4401 documentation comment.
4402 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4403 (-mupper-regs-df): New debug switch to control whether DF values
4404 can go in the traditional Altivec registers.
4405 (-mupper-regs-sf): New debug switch to control whether SF values
4406 can go in the traditional Altivec registers.
4407
4408 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4409 and wy constraints.
4410 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4411 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4412 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4413 (rs6000_debug_legitimate_address_p): Print if we are running
4414 before, during, or after reload.
4415 (rs6000_secondary_reload): Add a comment.
4416 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4417
4418 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4419 constraints. Update documentation string.
4420 (wd constraint): Likewise.
4421 (wf constraint): Likewise.
4422 (wg constraint): Likewise.
4423 (wn constraint): Likewise.
4424 (ws constraint): Likewise.
4425 (wt constraint): Likewise.
4426 (wx constraint): Likewise.
4427 (wz constraint): Likewise.
4428 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4429 instructions.
4430 (ww constraint): Likewise.
4431 (wy constraint): Likewise.
358e1993
MM		 4432 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4433 in the previous submissions.
5e8586d7
MM		 4434 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4435
e5af9ddd
RS		 44362013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4437
4438 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4439 (REG_BR_PROB): Say that the probability is stored in an int_list.
4440 * reg-notes.def: Update commentary to mention INT_LIST.
4441 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4442 (INT_LIST): New rtx.
4443 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4444 * rtlanal.c (int_reg_note_p): New function.
4445 (alloc_reg_note): Assert that the note does not have an int argument.
4446 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4447 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4448 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4449 rather than an INSN_LIST. Handle INT_LIST.
4450 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4451 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4452 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4453 Manipulate them as ints rather than rtxes.
4454 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4455 * regmove.c (copy_src_to_dest): Likewise.
4456 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4457
4458 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4459 into the cases that need it.
4460 * config/arm/arm.c (arm_unwind_emit): Likewise.
4461
4462 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4463 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4464 * loop-doloop.c (add_test, doloop_modify): Likewise.
4465 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4466 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4467 * predict.c (combine_predictions_for_insn): Likewise.
4468 * print-rtl.c (print_rtx): Handle INT_LIST.
4469 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4470 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4471 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4472 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4473 (ix86_split_fp_branch, predict_jump): Likewise.
4474 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4475 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4476 * config/spu/spu.c (ea_load_store_inline): Likewise.
4477
4478 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4479 value of a REG_BR_PROB note.
4480 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4481 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4482 * emit-rtl.c (try_split): Likewise.
4483 * predict.c (br_prob_note_reliable_p): Likewise.
4484 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4485 * reorg.c (mostly_true_jump): Likewise.
4486 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4487 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4488 * config/i386/i386.c (ix86_print_operand): Likewise.
4489 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4490 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4491 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4492 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4493 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4494 * config/sparc/sparc.c (output_cbranch): Likewise.
4495 * config/spu/spu.c (get_branch_target): Likewise.
4496 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4497 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4498
450ad0cd
JH		 44992013-09-23 Jan Hubicka <jh@suse.cz>
4500
4501 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4502 for ipa-devirt.
4503 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4504 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4505 of external calls
4506 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4507 (gimple_fold_call): Dump inconsistent devirtualizations; add
4508 sanity check for type based devirtualizations.
4509 * ipa-prop.c: Include ipa-utils.h
4510 (ipa_intraprocedural_devirtualization): Add sanity check.
4511 (try_make_edge_direct_virtual_call): Likewise.
4512
cefce769
EB		 45132013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4514
4515 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4516 assignment statements.
4517
a895a2b8
KV		 45182013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4519
4520 * gimple-pretty-print.c (dump_ssaname_info): New function.
4521 (dump_gimple_phi): Call it.
4522 (pp_gimple_stmt_1): Likewise.
4523 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4524 (range_info_def): Declare.
4525 * tree-pretty-print.c (pp_double_int): New function.
4526 (dump_generic_node): Call it.
4527 * tree-pretty-print.h (pp_double_int): Declare.
4528 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4529 * tree-ssanames.h (range_info_def): New structure.
4530 (value_range_type): Move definition here.
4531 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4532 Declare.
4533 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4534 initialization.
4535 (set_range_info): New function.
4536 (get_range_info): Likewise.
4537 (duplicate_ssa_name_range_info): Likewise.
4538 (duplicate_ssa_name_fn): Check pointer type and call
4539 duplicate_ssa_name_range_info.
4540 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4541 * tree-vrp.c (value_range_type): Remove definition, now in
4542 tree-ssanames.h.
7a5911d3 4543 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
a895a2b8
KV		 4544 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4545 (SSA_NAME_RANGE_INFO): New macro.
4546
984af6ac
RB		 45472013-09-23 Richard Biener <rguenther@suse.de>
4548
4549 PR tree-optimization/58464
4550 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7a5911d3 4551 (phi_trans_add): Change to add conditionally on being not present.
984af6ac
RB		 4552 (phi_translate_1): Remove recursion detection here.
4553 (phi_translate): Pre-seed the cache with NULL to catch
4554 recursion here in a more generic way.
4555 (bitmap_find_leader): Adjust comment.
4556 (get_representative_for): Dump value-numbers.
4557 (create_expression_by_pieces): Likewise.
4558 (insert_into_preds_of_block): Likewise.
4559
0e26cf79
CB		 45602013-09-23 Christian Bruel <christian.bruel@st.com>
4561
4562 PR target/58475
4563 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4564 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4565
4d0e5b54
JG		 45662013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4567
4568 Revert r202780:
4569 2013-09-20 Renlin Li <renlin.li@arm.com>
4570
7a5911d3
UB		 4571 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4572 plus_constant.
4d0e5b54
JG		 4573 (aarch64_expand_epilogue): Likewise.
4574 (aarch64_legitimize_reload_address): Likewise.
4575
ca081cc8
EB		 45762013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4577
4578 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4579 NULL_TREE before pushing them onto the vector. Likewise for labels.
4580
8fc53a5f
EB		 45812013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4582
4583 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4584 comparison operators when -fno-trapping-math is in effect.
4585 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4586 comparison operators in TFmode and assert that unsupported operators
4587 cannot reach here.
4588 (ia64_print_operand): Likewise.
4589
45902013-09-21 Jan Hubicka <jh@suse.cz>
0ca6c49f
JH		 4591
4592 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4593 (sse_typeless_stores): Enable for core
4594 (sse_load0_by_pxor): Likewise.
4595 (four_jump_limit): Disable for core.
4596 (pad_returns): Likewise.
4597 (avoid_vector_decode): Likewise.
4598 (fuse_cmp_and_branch): Enable for cores.
4599 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4600
578e51e6
JDA		 46012013-09-20 John David Anglin <danglin@gcc.gnu.org>
4602
96ce28a3
JDA		 4603 PR middle-end/56791
4604 * config/pa/pa.c (pa_option_override): Disable auto increment and
4605 decrement instructions until reload is completed.
4606
578e51e6
JDA		 4607 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4608 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4609 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4610
6fcd3a13 46112013-09-20 DJ Delorie <dj@redhat.com>
7a5911d3 4612 Nick Clifton <nickc@redhat.com>
6fcd3a13
DD		 4613
4614 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4615 (need_to_save): Save bank 0 on interrupts.
4616 (characterize_address): Strip far address wrappers.
4617 (rl78_as_legitimate_address): Likewise.
4618 (transcode_memory_rtx): Likewise.
4619 (rl78_peep_movhi_p): Disable this peephole after devirt.
4620 (rl78_propogate_register_origins): Forget all origins when a
4621 CLOBBER is seen.
4622 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4623 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4624 peephole2's.
4625 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4626 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4627 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4628 relocating.
4629 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4630
4a08db33
JDA		 46312013-09-20 John David Anglin <danglin@gcc.gnu.org>
4632
4633 * config/pa/pa.md: In "scc" insn patterns, change output template to
4634 handle const0_rtx in reg_or_0_operand operands.
4635
7691132c
JBG		 46362013-09-20 Martin Husemann <martin@NetBSD.org>
4637
4638 PR target/56875
4639 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4640 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4641
a3cc13cc
RB		 46422013-09-20 Richard Biener <rguenther@suse.de>
4643
4644 PR middle-end/58484
4645 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4646 remembering SSA name version and block index.
4647 (new_scev_info_str): Adjust.
4648 (hash_scev_info): Likewise. Also hash the block index.
4649 (eq_scev_info): Adjust.
4650 (find_var_scev_info): Likewise.
4651 (struct instantiate_cache_entry): Remove.
4652 (struct instantiate_cache_type): Use a htab to map name, block
4653 to chrec.
4654 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4655 (get_instantiated_value_entry): Likewise.
4656 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4657 (instantiate_scev_name): Adjust.
4658
925f3871
JL		 46592013-09-20 Jeff Law <law@redhat.com>
4660
4661 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4662
2888c331
YZ		 46632013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4664
4665 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4666 Call aarch64_simd_expand_args to update op[argc].
4667
0a811e96
BS		 46682013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4669
7a5911d3
UB		 4670 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4671 plugin argument.
0a811e96 4672
0078f462
BS		 46732013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4674
7a5911d3
UB		 4675 * gengtype.c (file_rules): Added rule for *.cc files.
4676 (get_output_file_with_visibility): Give fatal message when no
4677 rules found.
0078f462 4678
d20e48b0
RL		 46792013-09-20 Renlin Li <renlin.li@arm.com>
4680
4681 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4682 (aarch64_expand_epilogue): Likewise.
4683 (aarch64_legitimize_reload_address): Likewise.
4684
59bb1544
BE		 46852013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4686
4687 PR middle-end/57748
4688 * expr.c (expand_assignment): Remove misalignp code path.
4689
a5430019
MP		 46902013-09-20 Marek Polacek <polacek@redhat.com>
4691
4692 PR sanitizer/58413
4693 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4694 TYPE_PRECISION. Add asserts.
4695
9fed7f3a
RB		 46962013-09-20 Richard Biener <rguenther@suse.de>
4697
4698 PR tree-optimization/58453
4699 * tree-loop-distribution.c (distribute_loop): Apply the cost
4700 model for -ftree-loop-distribute-patterns, too.
4701
a820c834
RB		 47022013-09-20 Richard Biener <rguenther@suse.de>
4703
4704 PR middle-end/58473
4705 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4706 make type comparison less strict.
4707
3cbe17f7
AM		 47082013-09-20 Alan Modra <amodra@gmail.com>
4709
4710 * configure: Regenerate.
4711 * aclocal.m4: Regenerate.
4712
87fccdbb
MP		 47132013-09-20 Marek Polacek <polacek@redhat.com>
4714
4715 PR other/58467
4716 * doc/extend.texi: Document that attribute used is meant to be used
4717 on variables with static storage duration.
4718
8928eff3
JJ		 47192013-09-19 Jakub Jelinek <jakub@redhat.com>
4720
4721 PR tree-optimization/58472
4722 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4723 simd_lane_access set inv_p = false.
4724 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4725 the simduid magic VAR_DECL.
4726
b7b28c53
JH		 47272013-09-19 Jan Hubicka <jh@suse.cz>
4728
4729 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4730
83ae86f5
JL		 47312013-09-17 Jeff Law <law@redhat.com>
4732
4733 * tree-ssa-dom.c (record_temporary_equivalences): New function
4734 split out of dom_opt_dom_walker::after_dom_children.
4735 (dom_opt_dom_walker::thread_across_edge): Move common code
4736 in here from dom_opt_dom_walker::after_dom_children.
4737 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4738
9d532162
JH		 47392013-09-19 Jan Hubicka <jh@suse.cz>
4740
4741 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4742 (TARGET_GENERIC): Use PROCESOR_GENERIC
4743 (enum processor_type): Unify generic32 and 64.
4744 * i386.md (cpu): Likewise.
4745 * x86-tune.def (use_leave): Enable for generic32.
4746 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4747 * athlon.md: Change generic64 to generic in all occurences.
4748 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4749 (ix86_target_macros_internal): Likewise.
4750 * driver-i386.c (host_detect_local_cpu): Likewise.
7a5911d3
UB		 4751 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4752 to ..
9d532162
JH		 4753 (generic_memcpy, generic_memset, generic_cost): This one.
4754 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4755 (m_GENERIC32, m_GENERIC64): Remove.
4756 (m_GENERIC): Turn into one flag.
4757 (processor_target): Unify generic tunnings.
4758 (ix86_option_override_internal): Replace generic32/64 by generic.
4759 (ix86_issue_rate): Likewise.
4760 (ix86_adjust_cost): Likewise.
4761
188c7d00
JH		 47622013-09-19 Jan Hubicka <jh@suse.cz>
4763
4764 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4765 of speculative flag.
4766
4befd127
JJ		 47672013-09-19 Jakub Jelinek <jakub@redhat.com>
4768
4769 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4770 GOMP_sections_start, even if !exit_reachable.
4771
8a26ad39
VM		 47722013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4773
4774 * lra-constraints.c (need_for_all_save_p): Use macro
4775 HARD_REGNO_CALL_PART_CLOBBERED.
4776 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4777 set up pseudo conflict hard regs.
4778
09487185
MM		 47792013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4780
4781 PR target/58452
4782 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4783 operands.
4784
080cbf9e
VM		 47852013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4786
4787 PR rtl-optimization/58438
4788 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4789 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4790 from previous subpasses.
4791
a0f70fcb
RE		 47922013-09-18 Richard Earnshaw <rearnsha@arm.com>
4793
4794 * arm.c (arm_get_frame_offsets): Validate architecture supports
4795 LDRD/STRD before accepting the tuning preference.
4796 (arm_expand_prologue): Likewise.
4797 (arm_expand_epilogue): Likewise.
4798
0547c9b6
RB		 47992013-09-18 Richard Biener <rguenther@suse.de>
4800
4801 PR tree-optimization/58417
4802 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4803 have chrecs with symbols defined in the loop as operands.
4804 (chrec_fold_multiply): Likewise.
4805 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4806 parameters before folding binary operations.
4807 (struct instantiate_cache_entry_hasher): Remove.
4808 (struct instantiate_cache_type): Use a pointer-map.
4809 (instantiate_cache_type::instantiate_cache_type): New function.
4810 (instantiate_cache_type::get): Likewise.
4811 (instantiate_cache_type::set): Likewise.
4812 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4813 (get_instantiated_value_entry): Likewise.
4814 (global_cache): New global.
4815 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4816 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4817 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4818 (instantiate_scev_name): Adjust.
4819 (instantiate_scev): Construct global instead of local cache.
4820 (resolve_mixers): Likewise.
4821
309dc1aa
DM		 48222013-09-18 Daniel Morris <danielm@ecoscentric.com>
4823 Paolo Carlini <paolo.carlini@oracle.com>
4824
4825 PR c++/58458
4826 * doc/implement-cxx.texi: Fix references to the C++ standards.
4827
995a1b4a
JJ		 48282013-09-18 Jakub Jelinek <jakub@redhat.com>
4829
4830 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4831 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4832 simd_lane_access drs, update also DR_ALIGNED_TO.
4833
ce6923c5
MP		 48342013-09-18 Marek Polacek <polacek@redhat.com>
4835
4836 PR sanitizer/58411
4837 * doc/extend.texi: Document no_sanitize_undefined attribute.
4838 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4839 no_sanitize_undefined attribute.
4840
d30d00a2
NC		 48412013-09-18 Nick Clifton <nickc@redhat.com>
4842
4843 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4844 (ASM_DECLARE_FUNCTION_NAME): Define.
4845
4d9192b5
TS		 48462013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4847
4848 * compare-elim.c (find_comparison_dom_walker): New class
4849 (find_comparisons_in_bb): Rename to
4850 find_comparison_dom_walker::before_dom_children
4851 (find_comparisons): Adjust
4852 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4853 adjust.
4854 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4855 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4856 (init_walk_dominator_tree): Remove declaration.
4857 (fini_walk_dominator_tree): Remove declaration.
4858 * fwprop.c (single_def_use_dom_walker): New class
4859 (single_def_use_enter_block): Convert to
4860 single_def_use_dom_walker::before_dom_children.
4861 (single_def_use_leave_block): Convert to
4862 single_def_use_dom_walker::after_dom_children.
4863 (build_single_def_use_links): Adjust.
4864 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4865 class.
4866 (find_candidates_in_block): Convert to
4867 find_candidates_dom_walker::before_dom_children.
4868 (execute_strength_reduction): Adjust.
4869 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4870 (sese_dom_walker): New class.
4871 (sese_dom_walker::sese_dom_walker): New constructor.
4872 (sese_dom_walker::~sese_dom_walker): New destructor.
4873 (build_sese_conditions_before): Convert to
4874 sese_dom_walker::before_dom_children.
4875 (build_sese_conditions_after): Convert to
4876 sese_dom_walker::after_dom_children.
4877 (build_poly_scop): Adjust
4878 * tree-into-ssa.c (rewrite_dom_walker): New class
4879 (rewrite_enter_block): Convert to
4880 rewrite_dom_walker::before_dom_children.
4881 (rewrite_leave_block): Convert to
4882 rewrite_dom_walker::after_dom_children.
4883 (rewrite_update_dom_walker): New class.
4884 (rewrite_update_enter_block): Convert to
4885 rewrite_update_dom_walker::before_dom_children.
4886 (rewrite_update_leave_block): Convert to
4887 rewrite_update_dom_walker::after_dom_children.
4888 (rewrite_blocks, rewrite_into_ssa): Adjust.
4889 (mark_def_dom_walker): New class.
4890 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4891 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4892 (mark_def_sites_blocks): Convert to
4893 mark_def_dom_walker::before_dom_children.
4894 (mark_def_site_blocks): Remove.
4895 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4896 (tree_ssa_dominator_optimize): Adjust.
4897 (dom_thread_across_edge): Convert to method
4898 dom_opt_dom_walker::thread_across_edge.
4899 (dom_opt_enter_block): Convert to member function
4900 dom_opt_dom_walker::before_dom_children.
4901 (dom_opt_leave_block): Convert to member function
4902 dom_opt_dom_walker::after_dom_children.
4903 * tree-ssa-dse.c (dse_dom_walker): New class.
4904 (dse_enter_block): Convert to member function
4905 dse_dom_walker::before_dom_children.
4906 (tree_ssa_dse): Adjust.
4907 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4908 (determine_invariantness_stmt): Convert to method
4909 invariantness_dom_walker::before_dom_children.
4910 (determine_invariantness): Remove
4911 (move_computations_dom_walker): New class.
4912 (move_computations_stmt): Convert to method
4913 move_computations_dom_walker::before_dom_children.
4914 (move_computations, tree_ssa_lim): Adjust.
4915 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4916 (nt_init_block): Make method
4917 notrappping_dom_walker::before_dom_children.
4918 (nt_fini_block): Make
4919 method nontrapping_dom_walker::after_dom_children.
4920 (get_non_trapping): Adjust.
4921 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4922 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4923 (eliminate_leave_block): Make method.
4924 eliminate_dom_walker::after_dom_children.
4925 (eliminate): Adjust
4926 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4927 (strlen_enter_block): Make method
4928 strlen_dom_walker::before_dom_children.
4929 (strlen_leave_block): Make
4930 method strlen_dom_walker::after_dom_children.
4931 (tree_ssa_strlen): Adjust.
4932 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4933 (tree_ssa_uncprop): Adjust.
4934 (uncprop_leave_block): Make method
4935 uncprop_dom_walker::after_dom_children.
4936 (uncprop_leave_block): Make method
4937 uncprop_dom_walker::before_dom_children.
4938
3b3cc26b
BC		 49392013-09-18 Bin Cheng <bin.cheng@arm.com>
4940
7a5911d3
UB		 4941 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4942 branch in same basic block. Check both src and dest of the move insn.
3b3cc26b 4943
b5ba3a9b
NC		 49442013-09-17 Nick Clifton <nickc@redhat.com>
4945
4946 * config/rl78/rl78-real.md (bf): New pattern.
4947 (bt): New pattern.
4948 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4949 (rl78_print_operand): Do not put a # before a %B.
4950 * config/rl78/rl78.opt: Tweak doc strings.
4951
09d5094b
DD		 49522013-09-17 DJ Delorie <dj@redhat.com>
4953
4954 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4955 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4956 (rl78_option_override): Likewise, if -mallregs.
4957 (is_virtual_register): Likewise.
4958 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4959 (REGNO_OK_FOR_BASE_P): Likewise.
4960 * config/rl78/rl78.opt (-mallregs): New.
4961
2ba1dca1
NC		 49622013-09-17 Nick Clifton <nickc@redhat.com>
4963
4964 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4965 For interrupt functions: save all call clobbered registers if the
4966 interrupt handler is not a leaf function.
4967 (rl78_expand_prologue): Always recompute the frame information.
4968 For interrupt functions: only select bank 0 if one of the bank 0
4969 registers is going to be psuhed.
4970
2e7c3f21
DD		 49712013-09-17 DJ Delorie <dj@redhat.com>
4972
4973 * config/rl78/constraints.md: For each W* constraint, rename to C*
4974 and create a W* constraint that checks for an optional ES: prefix
4975 pattern also.
4976 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4977 (es_addr): New. Used to wrap far addresses.
4978 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4979 (rl78_es_base): New.
4980 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4981 wrapped far addresses.
4982 (rl78_print_operand_1): Unwrap far addresses before processing.
4983 (rl78_lo16): Wrap far addresses in unspecs.
4984 (rl78_es_addr): New.
4985 (rl78_es_base): New.
4986 (insn_ok_now): Check for not-yet-wrapped far addresses.
4987 (transcode_memory_rtx): Properly re-wrap far addresses.
4988
5c26a69a
SH		 49892013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4990
4991 * config/sparc/t-rtems: Add leon3 multilibs.
4992
181f5f3e
CH		 49932013-09-17 Cong Hou <congh@google.com>
4994
4995 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4996 when checking the dot production pattern. The type of rhs operand
4997 of multiply is now checked correctly.
4998
8d34e421
JL		 49992013-09-17 Jeff Law <law@redhat.com>
5000
5001 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5002 edge implied equivalences into successor phis.
5003 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5004 here from tree-ssa-threadedge.c.
5005 (mark_threaded_blocks): When threading through a joiner, if both
5006 successors of the joiner's clone reach the same block, verify the
5007 PHI arguments are equal. If not, cancel the jump threading request.
5008 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5009 tree-ssa-threadupdate.c
5010 (thread_across_edge): Don't check PHI argument equality when
5011 threading through joiner block here.
7a5911d3 5012
c152901f
AM		 50132013-09-17 Andrew MacLeod <amacleod@redhat.com>
5014
5015 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5016 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7a5911d3
UB		 5017 (warn_uninit, warn_uninitialized_vars,
5018 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5019 Move to tree-ssa-uninit.c.
78cedfb1 5020 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7a5911d3 5021 (has_undefined_value_p): New. Pass dependant parts of
c152901f
AM		 5022 ssa_undefined_value_p.
5023 (uninit_undefined_value_p): Use has_undefined_value_p.
7a5911d3
UB		 5024 (warn_uninit, warn_uninitialized_vars,
5025 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5026 Move from tree-ssa.c.
5027 * tree-ssa.h: Adjust prototypes.
c152901f 5028
b631d45a
JH		 50292013-09-17 Jan Hubicka <jh@suse.cz>
5030
5031 PR middle-end/58332
5032 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5033 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5034 FUNCTION_NOT_OPTIMIZED.
5035 * ipa-inline-analysis.c (compute_inline_parameters): Function
5036 not optimized is not inlinable unless it is alwaysinline.
5037 (inline_analyze_function): Force calls in not optimized
5038 function not inlinable.
5039
5b79657a
JH		 50402013-09-17 Jan Hubicka <jh@suse.cz>
5041
5042 PR middle-end/58329
5043 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5044 to return NULL.
5045 * ipa.c (function_and_variable_visibility): Likewise.
5046 * ipa-profile.c (ipa_profile): Likewise.
5047
27442c24
BE		 50482013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5049
5050 PR ipa/58398
5051 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5052 attribute, and don't inline the resolver in this case.
5053
753b9bc3
TJ		 50542013-09-17 Teresa Johnson <tejohnson@google.com>
5055
5056 * coverage.c (get_coverage_counts): Add missing newline.
5057
e3d3cfb4
KT		 50582013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5059
5060 PR tree-optimization/58088
5061 * fold-const.c (mask_with_trailing_zeros): New function.
5062 (fold_binary_loc): Make sure we don't recurse infinitely
5063 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5064 Use mask_with_trailing_zeros where appropriate.
5065
9a0ac98f
YR		 50662013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5067
5068 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5069 of 'prev' var to get better distance estimation.
5070
78bbd765
EB		 50712013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5072
5073 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5074 * tree-inline.c (is_parameter_of): New predicate.
5075 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5076 a parameter has been remapped.
5077 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5078 (optimize_inline_calls): Initialize transform_parameter.
5079 (copy_gimple_seq_and_replace_locals): Likewise.
5080 (tree_function_versioning): Likewise.
5081 (maybe_inline_call_in_expr): Likewise.
5082
cad055a4
NC		 50832013-09-17 Nick Clifton <nickc@redhat.com>
5084
5085 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5086 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5087 interrupt handlers.
5088 (is_attr_func): New function.
5089 (msp430_is_interrupt_func): New function.
5090 (is_naked_func): New function.
5091 (is_reentrant_func): New function.
5092 (is_critical_func): New function.
5093 (msp430_start_function): Add annotations for function attributes.
5094 (msp430_attr): New function.
5095 (msp430_attribute_table): New.
5096 (msp430_function_section): New function.
5097 (TARGET_ASM_FUNCTION_SECTION): Define.
5098 (msp430_builtin): New enum.
5099 (msp430_init_builtins): New function.
5100 (msp430_builtin_devl): New function.
5101 (msp430_expand_builtin): New function.
5102 (TARGET_INIT_BUILTINS): Define.
5103 (TARGET_EXPAND_BUILTINS): Define.
5104 (TARGET_BUILTIN_DECL): Define.
5105 (msp430_expand_prologue): Add support for naked, interrupt,
5106 critical and reentrant functions.
5107 (msp430_expand_epilogue): Likewise.
5108 (msp430_print_operand): Handle 'O' character.
5109 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5110 NO_TRAMPOLINES.
5111 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5112 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5113 (pushm): Use a 'n' rather than an 'i' constraint.
5114 (msp_return): Add generation of the interrupt return instruction.
5115 (disable_interrupts): New pattern.
5116 (enable_interrupts): New pattern.
5117 (push_intr_state): New pattern.
5118 (pop_intr_state): New pattern.
5119 (bic_SR): New pattern.
5120 (bis_SR): New pattern.
5121 * doc/extend.texi: Document MSP430 function attributes and builtin
5122 functions.
5123
deb6c11a
RB		 51242013-09-17 Richard Biener <rguenther@suse.de>
5125
5126 PR tree-optimization/58432
5127 * tree-loop-distribution.c (tree_loop_distribution): Also
5128 scan PHIs for outside loop uses and seed a partition from them.
5129
c068654b
BC		 51302013-09-17 Bin Cheng <bin.cheng@arm.com>
5131
5132 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5133 (restructure_reference): Call backtrace_base_for_ref.
5134
567969e4
AM		 51352013-09-17 Alan Modra <amodra@gmail.com>
5136
5137 PR target/57589
7a5911d3
UB		 5138 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5139 patch.
567969e4 5140
5c0029de
DD		 51412013-09-16 DJ Delorie <dj@redhat.com>
5142
5143 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5144 vregs location for RL78/G10.
5145 (rl78_expand_prologue): Avoid SEL on G10.
5146 (rl78_expand_epilogue): Likewise.
5147 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5148 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5149 __RL78_G10__ when appropriate.
5150 (ASM_SPEC): Pass -mg10 along to the assembler.
5151 * config/rl78/rl78.md (sel_rb): Disable for G10.
5152 * config/rl78/rl78.opt: Add -mg10 option.
5153 * config/rl78/t-rl78: Add -mg10 multilib.
5154
ea0f3e87
XDL		 51552013-09-16 Xinliang David Li <davidxl@google.com>
5156
5157 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5158 * omp-low.c (omp_max_vf): Ditto.
5159 (expand_omp_simd): Ditto.
5160 * tree-vectorizer.c (vectorize_loops): Ditto.
5161 (gate_vect_slp): Ditto.
5162 (gate_increase_alignment): Ditto.
5163 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5164 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5165 (gate_tree_vectorize): Name change.
5166 (tree_vectorize): Ditto.
5167 (pass_vectorize::gate): Call new function.
5168 (pass_vectorize::execute): Ditto.
7a5911d3 5169 * opts.c: O3 default setting change.
ea0f3e87
XDL		 5170 (finish_options): Check new flag.
5171 * doc/invoke.texi: Document new flags.
5172 * common.opt: New flags.
5173
4f3da2e5
AS		 51742013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5175
5176 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5177 (Condition Code): Don't reference it.
5178 * doc/tm.texi: Regenerate.
5179
aa940f7c
VM		 51802013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5181
5182 PR middle-end/58418
5183 * lra-constraints.c (undo_optional_reloads): Consider all optional
5184 reload even if it did not get a hard reg.
5185
e645e942
TJ		 51862013-09-16 Teresa Johnson <tejohnson@google.com>
5187
5188 * dumpfile.c (dump_loc): Remove newline emission.
5189 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5190 emission to dump_printf_loc calls where missing.
5191 (vect_mark_for_runtime_alias_test): Ditto.
5192 (vect_analyze_data_ref_dependence): Ditto.
5193 (vect_analyze_data_ref_dependences): Ditto.
5194 (vect_slp_analyze_data_ref_dependence): Ditto.
5195 (vect_slp_analyze_data_ref_dependences): Ditto.
5196 (vect_compute_data_ref_alignment): Ditto.
5197 (vect_update_misalignment_for_peel): Ditto.
5198 (vect_verify_datarefs_alignment): Ditto.
5199 (vector_alignment_reachable_p): Ditto.
5200 (vect_get_data_access_cost): Ditto.
5201 (vect_enhance_data_refs_alignment): Ditto.
5202 (vect_find_same_alignment_drs): Ditto.
5203 (vect_analyze_data_refs_alignment): Ditto.
5204 (vect_analyze_group_access): Ditto.
5205 (vect_analyze_data_ref_access): Ditto.
5206 (vect_analyze_data_ref_accesses): Ditto.
5207 (vect_prune_runtime_alias_test_list): Ditto.
5208 (vect_analyze_data_refs): Ditto.
5209 (vect_create_addr_base_for_vector_ref): Ditto.
5210 (vect_create_data_ref_ptr): Ditto.
5211 (vect_grouped_store_supported): Ditto.
5212 (vect_grouped_load_supported): Ditto.
5213 * value-prof.c (check_counter): Ditto.
5214 (check_ic_target): Ditto.
5215 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5216 (vect_recog_widen_mult_pattern): Ditto.
5217 (vect_recog_widen_sum_pattern): Ditto.
5218 (vect_recog_over_widening_pattern): Ditto.
5219 (vect_recog_widen_shift_pattern): Ditto.
5220 (vect_recog_rotate_pattern): Ditto.
5221 (vect_recog_vector_vector_shift_pattern): Ditto.
5222 (vect_recog_divmod_pattern): Ditto.
5223 (vect_recog_mixed_size_cond_pattern): Ditto.
5224 (vect_recog_bool_pattern): Ditto.
5225 (vect_pattern_recog_1): Ditto.
5226 (vect_pattern_recog): Ditto.
5227 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5228 (vect_is_simple_iv_evolution): Ditto.
5229 (vect_analyze_scalar_cycles_1): Ditto.
5230 (vect_get_loop_niters): Ditto.
5231 (vect_analyze_loop_1): Ditto.
5232 (vect_analyze_loop_form): Ditto.
5233 (vect_analyze_loop_operations): Ditto.
5234 (vect_analyze_loop_2): Ditto.
5235 (vect_analyze_loop): Ditto.
5236 (report_vect_op): Ditto.
5237 (vect_is_slp_reduction): Ditto.
5238 (vect_is_simple_reduction_1): Ditto.
5239 (vect_get_known_peeling_cost): Ditto.
5240 (vect_estimate_min_profitable_iters): Ditto.
5241 (vect_model_reduction_cost): Ditto.
5242 (vect_model_induction_cost): Ditto.
5243 (get_initial_def_for_induction): Ditto.
5244 (vect_create_epilog_for_reduction): Ditto.
5245 (vectorizable_reduction): Ditto.
5246 (vectorizable_induction): Ditto.
5247 (vectorizable_live_operation): Ditto.
5248 (vect_loop_kill_debug_uses): Ditto.
5249 (vect_transform_loop): Ditto.
5250 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5251 (vect_stmt_relevant_p): Ditto.
5252 (process_use): Ditto.
5253 (vect_mark_stmts_to_be_vectorized): Ditto.
5254 (vect_model_simple_cost): Ditto.
5255 (vect_model_promotion_demotion_cost): Ditto.
5256 (vect_model_store_cost): Ditto.
5257 (vect_get_store_cost): Ditto.
5258 (vect_model_load_cost): Ditto.
5259 (vect_get_load_cost): Ditto.
5260 (vect_init_vector_1): Ditto.
5261 (vect_get_vec_def_for_operand): Ditto.
5262 (vect_finish_stmt_generation): Ditto.
5263 (vectorizable_call): Ditto.
5264 (vectorizable_conversion): Ditto.
5265 (vectorizable_assignment): Ditto.
5266 (vectorizable_shift): Ditto.
5267 (vectorizable_operation): Ditto.
5268 (vectorizable_store): Ditto.
5269 (vectorizable_load): Ditto.
5270 (vectorizable_condition): Ditto.
5271 (vect_analyze_stmt): Ditto.
5272 (vect_transform_stmt): Ditto.
5273 (vect_is_simple_use): Ditto.
5274 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5275 (vect_can_advance_ivs_p): Ditto.
5276 (vect_update_ivs_after_vectorizer): Ditto.
5277 (vect_do_peeling_for_loop_bound): Ditto.
5278 (vect_gen_niters_for_prolog_loop): Ditto.
5279 (vect_update_inits_of_drs): Ditto.
5280 (vect_create_cond_for_alias_checks): Ditto.
5281 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5282 (vect_build_slp_tree_1): Ditto.
5283 (vect_supported_load_permutation_p): Ditto.
5284 (vect_analyze_slp_instance): Ditto.
5285 (vect_analyze_slp): Ditto.
5286 (vect_make_slp_decision): Ditto.
5287 (vect_detect_hybrid_slp): Ditto.
5288 (vect_bb_vectorization_profitable_p): Ditto.
5289 (vect_slp_analyze_bb_1): Ditto.
5290 (vect_update_slp_costs_according_to_vf): Ditto.
5291 (vect_get_mask_element): Ditto.
5292 (vect_transform_slp_perm_load): Ditto.
5293 (vect_schedule_slp_instance): Ditto.
5294 (vect_schedule_slp): Ditto.
5295 (vect_slp_transform_bb): Ditto.
5296 * profile.c (read_profile_edge_counts): Ditto.
5297 (compute_branch_probabilities): Ditto.
5298 * coverage.c (get_coverage_counts): Ditto.
5299
46f851f3
DN		 53002013-09-16 Diego Novillo <dnovillo@google.com>
5301
7a5911d3 5302 * tree-core.h: Add missing comment lines from refactoring of tree.h.
46f851f3 5303
1632a686
JH		 53042013-09-16 Jan Hubicka <jh@suse.cz>
5305
5306 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7a5911d3 5307 abstract functions; for static functions check the presence of body.
1632a686 5308
828e70c1
JG		 53092013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5310
5311 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5312 * config/aarch64/aarch64-simd.md
5313 (aarch64_mla_elt<mode>): New.
5314 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5315 (aarch64_mls_elt<mode>): Likewise.
5316 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5317 (aarch64_fma4_elt<mode>): Likewise.
5318 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5319 (aarch64_fma4_elt_to_128v2df): Likewise.
5320 (aarch64_fma4_elt_to_64df): Likewise.
5321 (fnma<mode>4): Likewise.
5322 (aarch64_fnma4_elt<mode>): Likewise.
5323 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5324 (aarch64_fnma4_elt_to_128v2df): Likewise.
5325 (aarch64_fnma4_elt_to_64df): Likewise.
5326 * config/aarch64/iterators.md (VDQSF): New.
5327 * config/aarch64/arm_neon.h
5328 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5329 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5330
779aea46
JG		 53312013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5332
5333 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5334 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5335 (aarch64_mul3_elt_to_128df): Likewise.
5336 (aarch64_mul3_elt_to_64v2df): Likewise.
5337 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5338 (VMUL): New.
5339 (VMUL_CHANGE_NLANES) Likewise.
5340 (h_con): Likewise.
5341 (f): Likewise.
5342 * config/aarch64/arm_neon.h
5343 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5344
a407a750
JG		 53452013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5346
5347 * config/aarch64/arm_neon.h
5348 (vcvtx_high_f32_f64): Fix parameters.
5349
f77e2d2d 53502013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
a5bd71fa 5351 Uros Bizjak <ubizjak@gmail.com>
f77e2d2d 5352
a5bd71fa 5353 * config/alpha.c: Include tree-ssanames.h.
f77e2d2d 5354
36875e8f
RB		 53552013-09-16 Richard Biener <rguenther@suse.de>
5356
5357 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5358 (dot_rdg_1): Handle control_dd.
5359 (create_edge_for_control_dependence): New function.
5360 (create_rdg_edges): Add control dependences if asked for.
5361 (build_rdg): Likewise.
5362 (generate_loops_for_partition): If there are not necessary
5363 control stmts remove all their dependencies.
5364 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5365 (distribute_loop): Pass on control dependences.
5366 (tree_loop_distribution): Compute control dependences and remove
5367 restriction on number of loop nodes.
5368
96e24d49
JJ		 53692013-09-16 Jakub Jelinek <jakub@redhat.com>
5370
5371 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5372 for internal calls.
5373
40dbb05c
RS		 53742013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5375
5376 * cse.c (try_const_anchors): Punt on CC modes.
5377
97191ad0
JBG		 53782013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5379
5380 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5381
403a3fb7
JDA		 53822013-09-14 John David Anglin <danglin@gcc.gnu.org>
5383
5384 PR target/58382
5385 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5386 calls to word_mode.
5387
fb5610fb
IS		 53882013-09-14 Iain Sandoe <iain@codesourcery.com>
5389
5390 PR target/48094
5391 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
f77e2d2d
JBG		 5392 seen.
5393 (darwin_objc1_section): Likewise.
fb5610fb
IS		 5394 (darwin_file_end): Emit Image Info section when required.
5395
afdec9bd
JH		 53962013-09-14 Jan Hubicka <jh@suse.cz>
5397
5398 * tree-into-ssa.c (gate_into_ssa): New.
5399 (pass_data_build_ssa): Use it.
5400 * cgraph.h (expand_thunk): Update prototype.
5401 * cgraphunit.c (analyze_function): Expand thunks early.
5402 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5403 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5404 set lowered flag; do not add new function.
5405 (assemble_thunks_and_aliases): Update.
5406 * tree-ssa.c (gate_init_datastructures): New gate.
5407 (pass_data_init_datastructures): Use it.
5408
492d1e0a
IS		 54092013-09-14 Iain Sandoe <iain@codesourcery.com>
5410
5411 PR target/58269
f77e2d2d 5412 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
492d1e0a
IS		 5413 xmm register set described in the psABI.
5414
d13c6b70
EG		 54152013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5416
f77e2d2d
JBG		 5417 * dwarf2out.c (should_emit_struct_debug): Add check
5418 for type_decl variable is not NULL.
d13c6b70 5419
88b0e79e
JC		 54202013-09-13 Jacek Caban <jacek@codeweavers.com>
5421
5422 * config.gcc: Use new winnt-c.c target hooks
5423 * config/t-winnt: New file
5424 * config/winnt-c.c: New file
5425 * doc/tm.texi.in: Document new hook
5426 * doc/tm.texi: Regenerated
5427
19ba6aab
JH		 54282013-09-13 Jan Hubicka <jh@suse.cz>
5429
5430 PR middle-end/58094
5431 * ipa-inline.c (check_callers): New function.
5432 (check_caller_edge): Remove.
5433 (want_inline_function_to_all_callers_p): Also permit alises that are
5434 called dirrectly.
5435 (inline_to_all_callers): Terminate the walk when devirtualization
5436 introduce new calls.
5437
a93c18c8
JH		 54382013-09-13 Jan Hubicka <jh@suse.cz>
5439
5440 * ipa-inline-analysis.c (struct growth_data): Add node.
5441 (do_estimate_growth_1): Fix detection of recursion.
5442
f3b331d1
JJ		 54432013-09-13 Jakub Jelinek <jakub@redhat.com>
5444
5445 PR tree-optimization/58392
5446 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5447 to avoid shadowing of outer loop variable. If
5448 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5449 replace_by_duplicate_decl simduid of loops that have it set and
5450 set dest_cfun->has_simduid_loops and/or
5451 dest_cfun->has_force_vect_loops.
5452 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5453 instead of maybe_lookup_decl.
5454 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5455 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5456 call. Copy over force_vect and copy and remap simduid. Set
5457 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5458 (copy_cfg_body): Remove blocks_to_copy argument. Use
5459 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5460 caller. Don't set cfun->has_simduid_loops and/or
5461 cfun->has_force_vect_loops here.
5462 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5463 caller.
5464 (expand_call_inline, tree_function_versioning): Adjust copy_body
5465 callers.
5466
ae6d0907
MJ		 54672013-09-13 Martin Jambor <mjambor@suse.cz>
5468
5469 PR bootstrap/58388
5470 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5471 the assert if the edge was a speculative one.
5472
2fd5894f
RB		 54732013-09-13 Richard Biener <rguenther@suse.de>
5474
5475 * tree-data-ref.h (known_dependences_p): Move here ...
5476 * tree-loop-distribution.c (known_dependences_p): ... from here.
5477 (dump_rdg_component, debug_rdg_component): Remove.
5478 (dump_rdg): Adjust.
5479 (generate_loops_for_partition): Use gimple_uid instead of
5480 relying on matching stmt visit order.
5481 (rdg_build_partitions): Take starting stmt vector.
5482 (ldist_gen): Merge into ...
5483 (distribute_loop): ... this function. Do not compute starting
5484 vertices vector.
5485 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5486
86eb4bd7
KT		 54872013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5488
5489 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5490 Set type attribute correctly. Set predicable_short_it attribute.
5491 (cmpsi_shiftsi): Remove %? from output template.
5492
83a95546
RB		 54932013-09-13 Richard Biener <rguenther@suse.de>
5494
5495 * tree-loop-distribution.c (struct rdg_component,
5496 rdg_defs_used_in_other_loops_p, free_rdg_components,
5497 rdg_build_components): Remove.
5498 (stmts_from_loop): Do not record virtual PHIs.
5499 (generate_loops_for_partition): Skip virtual PHIs.
5500 (build_rdg_partition_for_component): Rename to ...
5501 (build_rdg_partition_for_vertex): ... this and adjust.
5502 (rdg_build_partitions): Take a vector of starting vertices
5503 instead of components. Remove unnecessary leftover handling.
5504 (ldist_gen): Do not build components or record other stores.
5505 (distribute_loop): Do not distribute loops containing stmts
5506 with side-effects.
5507
c0ecf7f1
CB		 55082013-09-13 Christian Bruel <christian.bruel@st.com>
5509
5510 PR target/58314
5511 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5512
34f56efc
KT		 55132013-09-13 Kai Tietz <ktietz@redhat.com>
5514
740f2bf3
KT		 5515 * config.gcc: Separate cases for mingw and cygwin targets,
5516 and add 64-bit cygwin target case.
5517
34f56efc
KT		 5518 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5519 dll-export inline-functions.
5520 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5521
5522
55232013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL		 5524
5525 PR middle-end/58387
5526 Revert:
5527 2013-09-06 Jeff Law <law@redhat.com>
5528
5529 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5530 edge implied equivalences into successor phis.
5531
aadb52ad
DD		 55322013-09-12 DJ Delorie <dj@redhat.com>
5533
5534 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5535 separators.
5536
90ae7019
BM		 55372013-09-12 Brooks Moses <bmoses@google.com>
5538
5539 PR driver/42955
5540 * Makefile.in: Do not install driver binaries in $(target)/bin.
5541
17b2557c
DD		 55422013-09-12 DJ Delorie <dj@redhat.com>
5543
c5215a9b
DD		 5544 * config/rl78/rl78.opt (mrelax): New.
5545 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5546 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5547
17b2557c
DD		 5548 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5549 between SP and FP.
5550 (rl78_expand_epilogue): Likewise.
5551
ecafcf05
VM		 55522013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5553
5554 PR middle-end/58335
84f6d9fa 5555 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM		 5556 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5557 pointer elimination with using remove_reg_equal_offset_note.
5558
8369f38a
DD		 55592013-09-12 DJ Delorie <dj@redhat.com>
5560
f6a83b4a
DD		 5561 * config/msp430/: New port.
5562 * config.gcc (msp430): Added.
5563 * doc/invoke.texi: Document MSP430 options.
5564 * doc/install.texi: Document msp430-elf
5565 * doc/md.texi: Document msp430-elf
5566 * doc/contrib.texi: Document msp430-elf
5567
8369f38a
DD		 5568 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5569 PSImode.
5570
a854f856
MJ		 55712013-09-12 Martin Jambor <mjambor@suse.cz>
5572
5573 PR ipa/58389
5574 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5575 reference descriptor is NULL.
5576 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5577 edge to NULL.
5578
b12ebd96
AM		 55792013-09-12 Andrew MacLeod <amacleod@redhat.com>
5580
5581 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5582 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5583 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5584 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5585 make_temp_ssa_name): move to tree-ssanames.h
5586 * tree-ssa-alias.h: Move prototype.
5587 * tree-ssa.h: Include tree-ssanames.h.
5588 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5589 * tree-ssanames.h: New. Move items from tree-flow*.h
5590 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5591
e1fb4ad3
RB		 55922013-09-12 Richard Biener <rguenther@suse.de>
5593
5594 PR tree-optimization/58404
5595 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5596 propagate non-invariant addresses into dereferences wrapped
5597 in component references.
5598
154c9188
RB		 55992013-09-12 Richard Biener <rguenther@suse.de>
5600
5601 PR tree-optimization/58402
5602 * passes.def: Move pass_late_warn_uninitialized later.
5603
7a300452
AM		 56042013-09-12 Andrew MacLeod <amacleod@redhat.com>
5605
5606 * tree-ssa.h: New. Move content from tree-flow.h and
5607 tree-flow-inline.h.
5608 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5609 Move prototypes belonging to tree-ssa.c.
f77e2d2d 5610 * tree-flow-inline.h (redirect_edge_var_map_def,
7a300452
AM		 5611 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5612 tree-ssa.h.
5613 * gimple.h: Adjust prototypes.
f77e2d2d
JBG		 5614 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5615 to...
5616 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7a300452
AM		 5617 * tree.h: Move prototype to tree-ssa.h.
5618 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5619 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5620 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5621 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5622 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5623 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5624 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5625 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
f77e2d2d
JBG		 5626 graphite-blocking.c, graphite-clast-to-gimple.c,
5627 graphite-dependences.c, graphite-interchange.c,
5628 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5629 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5630 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5631 ipa-reference.c, ipa-split.c, ipa-utils.c,
7a300452
AM		 5632 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5633 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5634 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5635 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5636 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5637 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5638 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5639 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5640 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5641 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5642 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5643 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5644 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5645 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5646 value-prof.c, var-tracking.c,
5647 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5648
97463b2b
RB		 56492013-09-12 Richard Biener <rguenther@suse.de>
5650
5651 PR tree-optimization/58396
5652 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5653 (build_rdg): Take a loop-nest parameter, fix memleaks.
5654 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5655
80ecd89d
YR		 56562013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5657
5658 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5659 for SLM.
5660
893e85fa
CM		 56612013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5662
5663 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5664 The FMA instruction names should have a 'v' prefix.
5665
174ec470
RB		 56662013-09-12 Richard Biener <rguenther@suse.de>
5667
5668 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5669 (dot_rdg): Use popen instead of system in optional code.
5670 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5671 (already_processed_vertex_p): Adjust.
5672 (has_anti_or_output_dependence, predecessor_has_mem_write,
5673 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5674 rdg_flag_uses): Remove.
5675 (rdg_flag_vertex): Simplify.
5676 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5677 remove recursion.
5678 (build_rdg_partition_for_component): Process the first vertex
5679 of a component only.
5680 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5681
38c89763
AM		 56822013-09-12 Alan Modra <amodra@gmail.com>
5683
5684 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5685
3e3d9d17
DD		 56862013-09-11 DJ Delorie <dj@redhat.com>
5687 Nick Clifton <nickc@redhat.com>
5688
5689 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5690 (rl78_stack_based_mem): New.
5691 * config/rl78/constraints.md (Iv08): New.
5692 (Iv16): New.
5693 (Iv24): New.
5694 (Is09): New.
5695 (Is17): New.
5696 (Is25): New.
5697 (ISsi): New.
5698 (IShi): New.
5699 (ISqi): New.
5700 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5701 (movhi): Likewise.
5702 (movsi): Change from expand to insn-and-split.
5703 (ashrsi3): Clobber AX.
5704 (lshrsi3): New.
5705 (ashlsi3): New.
5706 (cbranchsi4): New.
5707 * config/rl78/rl78.md (CC_REG): Fix.
5708 (addsi3): Allow memory and immediate operands.
5709 (addsi3_internal): Split into...
5710 (addsi3_internal_virt): ...new, and ...
5711 (addsi3_internal_real): ...new.
5712 (subsi): New.
5713 (subsi3_internal_virt): New.
5714 (subsi3_internal_real): New.
5715 (mulsi3): Add memory operand.
5716 (mulsi3_rl78): Likewise.
5717 (mulsi3_g13): Likewise.
5718 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5719 (cbranchqi4_real): Add more constraint options.
5720 (cbranchhi4_real): Expand pattern.
5721 (cbranchhi4_real_signed): New.
5722 (cbranchhi4_real_inverted): New.
5723 (cbranchsi4_real_lt): New.
5724 (cbranchsi4_real_ge): New.
5725 (cbranchsi4_real_signed): New.
5726 (cbranchsi4_real): New.
5727 (peephole2): New.
893e85fa
CM		 5728 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5729 constant shifts.
3e3d9d17
DD		 5730 (lshrsi3_virt): Likewise.
5731 (ashlsi3_virt): Likewise.
5732 (cbranchqi4_virt_signed): New.
5733 (cbranchhi4_virt_signed): New.
5734 (cbranchsi4_virt): New.
5735 * config/rl78/rl78.c: Whitespace fixes throughout.
5736 (move_elim_pass): New.
5737 (pass_data_rl78_move_elim): New.
5738 (pass_rl78_move_elim): New.
5739 (make_pass_rl78_move_elim): New.
5740 (rl78_devirt_info): Run devirt earlier.
5741 (rl78_move_elim_info): New.
5742 (rl78_asm_file_start): Register it.
5743 (rl78_split_movsi): New.
5744 (rl78_as_legitimate_address): Allow virtual base registers when
5745 appropriate.
5746 (rl78_addr_space_convert): Remove spurious debug stuff.
5747 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5748 (rl78_print_operand): More cases for not printing '#'.
5749 (rl78_expand_compare): Remove most of the logic.
5750 (content_memory): New.
5751 (clear_content_memory): New.
5752 (get_content_index): New.
5753 (get_content_name): New.
5754 (display_content_memory): New.
5755 (update_content): New.
5756 (record_content): New.
5757 (already_contains): New.
5758 (insn_ok_now): Re-recog insns with virtual registers.
5759 (add_postponed_content_update): New.
5760 (process_postponed_content_update): New.
5761 (gen_and_emit_move): New.
893e85fa
CM		 5762 (transcode_memory_rtx): Record new location content.
5763 Use gen_and_emit_move.
3e3d9d17
DD		 5764 (force_into_acc): New.
5765 (move_to_acc): Use gen_and_emit_move.
5766 (move_from_acc): Likewise.
5767 (move_acc_to_reg): Likewise.
5768 (move_to_x): Likewise.
5769 (move_to_hl): Likewise.
5770 (move_to_de): Likewise.
5771 (rl78_alloc_physical_registers_op1): Record location content.
5772 (has_constraint): New.
5773 (rl78_alloc_physical_registers_op2): Record location content.
5774 Optimize use of HL.
5775 (rl78_alloc_physical_registers_ro1): Likewise.
5776 (rl78_alloc_physical_registers_cmp): Likewise.
5777 (rl78_alloc_physical_registers_umul): Likewise.
5778 (rl78_alloc_address_registers_macax): New.
5779 (rl78_alloc_physical_registers): Initialize and set location
5780 content memory as needed.
5781 (rl78_reorg): Make sure split2 is called.
5782 (rl78_rtx_costs): New.
5783
088845a5
RS		 57842013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5785
5786 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5787 for (not (neg ...)) and (neg (not ...)) cases.
5788
5c2961cf
RB		 57892013-09-11 Richard Biener <rguenther@suse.de>
5790
5791 PR middle-end/58377
5792 * passes.def: Split critical edges before late uninit warning passes.
5793 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5794
427e6a14
JJ		 57952013-09-11 Jakub Jelinek <jakub@redhat.com>
5796
5797 PR tree-optimization/58385
5798 * fold-const.c (build_range_check): If both low and high are NULL,
5799 use omit_one_operand_loc to preserve exp side-effects.
5800
df45c0cc
KT		 58012013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5802
5803 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5804
80ab0b19
RB		 58052013-09-11 Richard Biener <rguenther@suse.de>
5806
5807 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5808 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5809 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5810 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5811 stmts_from_loop, known_dependences_p, build_empty_rdg,
5812 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5813 * tree-loop-distribution.c: ... here.
5814 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5815 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5816 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5817 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5818 * tree-loop-distribution.c: ... here.
5819 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5820 (struct partition_s): Add loops member.
5821 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5822 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5823 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5824
85a77221
AI		 58252013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5826 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5827 Sergey Lega <sergey.s.lega@intel.com>
5828 Anna Tikhonova <anna.tikhonova@intel.com>
5829 Ilya Tocar <ilya.tocar@intel.com>
5830 Andrey Turetskiy <andrey.turetskiy@intel.com>
5831 Ilya Verbin <ilya.verbin@intel.com>
5832 Kirill Yukhin <kirill.yukhin@intel.com>
5833 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5834
5835 * config/i386/constraints.md (k): New.
5836 (Yk): Ditto.
5837 * config/i386/i386.c (const regclass_map): Add new mask registers.
5838 (dbx_register_map): Ditto.
5839 (dbx64_register_map): Ditto.
5840 (svr4_dbx_register_map): Ditto.
5841 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5842 disabled.
5843 (ix86_preferred_reload_class): Disable constants for mask registers.
5844 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5845 (ix86_hard_regno_mode_ok): Support new mask registers.
5846 (x86_order_regs_for_local_alloc): Ditto.
5847 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5848 (FIXED_REGISTERS): Add new mask registers.
5849 (CALL_USED_REGISTERS): Ditto.
5850 (REG_ALLOC_ORDER): Ditto.
5851 (VALID_MASK_REG_MODE): New.
5852 (FIRST_MASK_REG): Ditto.
5853 (LAST_MASK_REG): Ditto.
5854 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5855 (MAYBE_MASK_CLASS_P): New.
5856 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5857 (REG_CLASS_CONTENTS): Ditto.
5858 (MASK_REGNO_P): New.
5859 (ANY_MASK_REG_P): Ditto.
5860 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM		 5861 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5862 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5863 mask registers.
85a77221
AI		 5864 (attribute "type"): Add mskmov, msklog.
5865 (attribute "length_immediate"): Support them.
5866 (attribute "memory"): Ditto.
5867 (attribute "prefix_0f"): Ditto.
5868 (*movhi_internal): Support new mask registers.
5869 (*movqi_internal): Ditto.
5870 (define_split): Split out clobber pattern is a logic
5871 insn on mask registers.
5872 (*k<logic><mode>): New.
5873 (*andhi_1): Extend to support mask regs.
5874 (*andqi_1): Extend to support mask regs.
5875 (kandn<mode>): New.
5876 (define_split): Split and-not to and and not if operands
5877 are not mask regs.
5878 (*<code><mode>_1): Separate HI mode to new pattern...
5879 (*<code>hi_1): This.
5880 (*<code>qi_1): Extend to support mask regs.
5881 (kxnor<mode>): New.
5882 (kortestzhi): Ditto.
5883 (kortestchi): Ditto.
5884 (kunpckhi): Ditto.
5885 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5886 (*one_cmplhi2_1): ...Here, now with mask registers support.
5887 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM		 5888 (HI/QImode arithmetics splitter): Don't split if mask registers
5889 are used.
85a77221
AI		 5890 (HI/QImode not splitter): Ditto.
5891 * config/i386/predicated.md (mask_reg_operand): New.
5892 (general_reg_operand): Ditto.
5893
7e570821
AI		 58942013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5895
5896 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5897 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5898
9e1376e9
JL		 58992013-09-10 Jeff Law <law@redhat.com>
5900
5901 PR tree-optimization/58380
5902 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5903 of threading through a buried loop header.
5904
5905 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5906 return value for single successor case.
5907
5bbcb888
JH		 59082013-09-10 Jan Hubicka <jh@suse.cz>
5909
5910 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5911
d742ff4b
RE		 59122013-09-10 Richard Earnshaw <rearnsha@arm.com>
5913
5914 PR target/58361
5915 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5916 support conditional execution.
5917 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5918
b0681c9e
VM		 59192013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5920
5921 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5922 constraint pass.
f77e2d2d
JBG		 5923 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5924 Check destination too to check move insn.
b0681c9e 5925 (undo_optional_reloads): Add check that the original peudo did not
f77e2d2d
JBG		 5926 changed its allocation and the optional reload was inherited on last
5927 inheritance pass. Break loop after deciding to keep optional reload.
b0681c9e
VM		 5928 (lra_undo_inheritance): Add check that inherited pseudo still in
5929 memory.
5930
c9b1c957
JG		 59312013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5932
5933 * config/aarch64/aarch64.md (generic_sched): New.
5934 * config/aarch64/aarch64-generic.md (load): Make conditional
5935 on generic_sched attribute.
5936 (nonload): Likewise.
5937
82d618d3
JH		 59382013-09-10 Jan Hubicka <jh@suse.cz>
5939
5940 * lto-cgraph.c: Include ipa-utils.h.
5941 (compute_ltrans_boundary): Also add possible targets into the boundary.
5942
668b6894
JH		 59432013-09-10 Jan Hubicka <jh@suse.cz>
5944
5945 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5946 VAR_DECL of vtable rather than full expression.
5947
4f90d3e0
JH		 59482013-09-10 Jan Hubicka <jh@suse.cz>
5949 Paolo Carlini <paolo.carlini@oracle.com>
5950
5951 * cgraphunit.c (analyze_functions): Save input_location, set it
5952 to UNKNOWN_LOCATION and restore it at the end.
5953
39e87baf
MJ		 59542013-09-10 Martin Jambor <mjambor@suse.cz>
5955
5956 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5957 represented by a thunk.
5958
581aedec
JL		 59592013-09-10 Jeff Law <law@redhat.com>
5960
5961 PR tree-optimization/58343
5962 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5963 jump threading requests through loop headers buried in the middle
5964 of a jump threading path.
5965
5966 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5967 in return value/type.
5968
84cf4ab6
JJ		 59692013-09-10 Jakub Jelinek <jakub@redhat.com>
5970
5971 PR rtl-optimization/58365
5972 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5973 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5974 it differs.
5975
447dd906
RB		 59762013-09-10 Richard Biener <rguenther@suse.de>
5977
5978 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5979 * tree-data-ref.c (create_rdg_vertices): Collect all data
5980 references, signal failure to the caller, use data-ref API.
893e85fa
CM		 5981 (build_rdg): Compute data references only once. Maintain lifetime
5982 of data references and data dependences from within RDG.
447dd906
RB		 5983 (free_rdg): Free dependence relations.
5984 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5985 inventing extra dependences.
5986 (distribute_loop): Update for RDG API changes.
5987
f14a7324
KT		 59882013-09-10 Kai Tietz <ktietz@redhat.com>
5989
5990 * doc/invoke.texi (fms-extensions): Document changed
5991 behavior for ms-abi targets.
5992 * config/i386/i386.c (ix86_option_override_internal):
5993 Set default value of option -fms-extension for ms-abi targets.
5994
ff131f26
MZ		 59952013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5996
5997 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5998
33115a2c
AM		 59992013-09-10 Alan Modra <amodra@gmail.com>
6000
6001 PR target/58330
6002 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6003
b6872a13
AM		 60042013-09-10 Alan Modra <amodra@gmail.com>
6005
6006 * config/rs6000/predicates.md (add_cint_operand): New.
6007 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6008 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6009 using add_cint_operand.
6010 (largetoc_high_plus_aix): Likewise.
6011
28fd0ba2
JJ		 60122013-09-09 Jakub Jelinek <jakub@redhat.com>
6013
6014 PR tree-optimization/58364
6015 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6016 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6017 the current range can't be an unconditional true or false.
6018
552d99fe
JG		 60192013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6020
6021 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6022
f43245d1
UB		 60232013-09-09 Uros Bizjak <ubizjak@gmail.com>
6024
6025 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6026
c9b0866a
PC		 60272013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6028
6029 PR c++/43452
6030 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6031
78d8b9f0
IB		 60322013-09-09 Ian Bolton <ian.bolton@arm.com>
6033
6034 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6035 NO_REGS for immediate that can't be moved directly into FP_REGS.
6036
274b2532
KT		 60372013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6038
6039 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6040 comparison with negated operand.
c9b0866a
PC		 6041 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6042 RTL form.
274b2532 6043
467a3558
RB		 60442013-09-09 Richard Biener <rguenther@suse.de>
6045
6046 PR middle-end/58326
6047 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6048 of a subloop record all its block as affecting loop-closed SSA form.
6049
088c5368
RS		 60502013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6051
6052 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6053 of an rtx/bitpos pair.
6054 (store_fixed_bit_field): Update accordingly.
6055
69db2d57
RS		 60562013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6057
6058 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6059 GEN_INT.
6060 * builtins.c (expand_errno_check): Likewise.
6061 * dwarf2cfi.c (init_return_column_size): Likewise.
6062 * except.c (sjlj_mark_call_sites): Likewise.
6063 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6064 * lra-constraints.c (emit_inc): Likewise.
6065 * ree.c (combine_set_extension): Likewise.
6066 * regmove.c (fixup_match_2): Likewise.
6067 * reload1.c (inc_for_reload): Likewise.
6068
69a59f0f
RS		 60692013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6070
6071 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6072 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6073 Use gen_int_mode with the mode of the associated simplify_* call.
6074 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6075 * expmed.c (expand_shift_1): Likewise.
6076 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6077 * loop-iv.c (iv_number_of_iterations): Likewise.
6078 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6079 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6080
2f1cd2eb
RS		 60812013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6082
6083 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6084 of the associated expand_* call.
6085 (asan_emit_stack_protection): Likewise.
6086 * builtins.c (round_trampoline_addr): Likewise.
6087 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6088 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6089 (emit_store_flag): Likewise.
6090 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6091 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6092 Likewise.
6093 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6094 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6095 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6096 * modulo-sched.c (generate_prolog_epilog): Likewise.
6097 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6098 (expand_ctz, expand_ffs, expand_unop): Likewise.
6099
4789c0ce
RS		 61002013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6101
6102 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6103 of the associated gen_rtx_* call.
6104 * caller-save.c (init_caller_save): Likewise.
6105 * combine.c (find_split_point, make_extraction): Likewise.
6106 (make_compound_operation): Likewise.
6107 * dwarf2out.c (mem_loc_descriptor): Likewise.
6108 * explow.c (plus_constant, probe_stack_range): Likewise.
6109 * expmed.c (expand_mult_const): Likewise.
6110 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6111 * reload1.c (init_reload): Likewise.
6112 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6113 * var-tracking.c (adjust_mems): Likewise.
6114 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6115 rather than gen_rtx_fmt_ee.
6116
a811f7cb
JH		 61172013-09-09 Jan Hubicka <jh@suse.cz>
6118
6119 PR middle-end/58294
6120 * value-prof.c (gimple_ic): Copy also abnormal edges.
6121
dcad1dd3
RS		 61222013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6123
6124 * asan.c (asan_shadow_cst): Use gen_int_mode.
6125
daf5c770
JH		 61262013-09-08 Jan Hubicka <jh@suse.cz>
6127
6128 * ipa-profile.c: Add toplevel comment.
6129 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6130 (contains_hot_call_p): New function.
6131 (ipa_propagate_frequency): Set frequencies based on counts when
6132 profile is read.
6133 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6134 profile; do not tamper with profile after inlining if it is read.
6135
a12cd2db
JH		 61362013-09-08 Jan Hubicka <jh@suse.cz>
6137
6138 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6139 speculative edges.
6140
856a1588
JH		 61412013-09-08 Jan Hubicka <jh@suse.cz>
6142
6143 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6144 summary generation.
6145
4f4b0b73
JL		 61462013-09-08 Jeff Law <law@redhat.com>
6147
6148 PR bootstrap/58340
6149 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6150 of 'found'.
6151
cc4029ee
AK		 61522013-09-08 Andi Kleen <ak@linux.intel.com>
6153
6154 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6155
7b395ddd
JH		 61562013-09-08 Jan Hubicka <jh@suse.cz>
6157
6158 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6159
c34c46dd
RS		 61602013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6161
6162 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6163 for non-debug insns.
6164 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6165 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6166 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6167 accordingly.
6168
3462aa02
JH		 61692013-09-08 Jan Hubicka <jh@suse.cz>
6170
6171 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6172 targets and devirtualize to BUILT_IN_UNREACHABLE.
6173 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6174 * ipa.c (walk_polymorphic_call_targets): New function.
6175 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6176 functions; use the new timevar.
6177 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6178 was removed from the program.
6179 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6180 not consider it in the walk when its vtable is dead.
6181 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6182 record_binfo.
6183 (devirt_variable_node_removal_hook): New function.
6184 (possible_polymorphic_call_targets): Also register
6185 devirt_variable_node_removal_hook.
6186 (ipa_devirt): Do not do non-speculative devirtualization.
6187 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6188
26e5b0fd
JH		 61892013-09-08 Jan Hubicka <jh@suse.cz>
6190
6191 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6192 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6193 varpool_remove_variable_insertion_hook): Declare.
6194 * varpool.c (varpool_node_hook_list): New structure.
6195 (first_varpool_node_removal_hook,
6196 first_varpool_variable_insertion_hook): New variables.
6197 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6198 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6199 varpool_remove_variable_insertion_hook,
6200 varpool_call_variable_insertion_hooks): New functions.
6201 (varpool_remove_node): Use it.
6202
97f1f314 62032013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 6204
97f1f314
PC		 6205 PR c++/54941
6206 * diagnostic.c (diagnostic_build_prefix): When s.file is
6207 "<built-in>" don't output line and column numbers.
6df2ca09 6208
bcb650cb
JH		 62092013-09-06 Jan Hubicka <jh@suse.cz>
6210
6211 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6212 * lto-streamer-out.c: Stream thunks, too.
6213 * lto-streamer-in.c (input_function): Pop cfun here
6214 (lto_read_body): Instead of here.
6215
87f3fea8
CT		 62162013-09-06 Caroline Tice <cmtice@google.com>
6217
6218 * doc/install.texi: Add documentation for the --enable-vtable-verify
6219 and the --disable-libvtv configure options.
6220
5e8d9b45
JL		 62212013-09-06 Jeff Law <law@redhat.com>
6222
6223 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6224 edge implied equivalences into successor phis.
6225
3a001aff
JR		 62262013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6227
6228 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6229
62302013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ		 6231
6232 * resource.c (mark_target_live_regs): Compute resources taking
6233 into account if a call is predicated or not.
6234
2c641f8b
EB		 62352013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6236
6237 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6238 the compiler in the function names.
6239
a81b0a3d
JH		 62402013-09-06 Jan Hubicka <jh@suse.cz>
6241
6242 PR middle-end/58094
6243 * ipa-inline.c (has_caller_p): New function.
6244 (want_inline_function_to_all_callers_p): Use it.
6245 (sum_callers, inline_to_all_callers): Break out from ...
6246 (ipa_inline): ... here.
6247
68e4ca7e
JH		 62482013-09-06 Jan Hubicka <jh@suse.cz>
6249
2c641f8b
EB		 6250 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6251 only when AVX is enabled.
68e4ca7e 6252
3c73e2fe
JG		 62532013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6254
6255 * config/aarch64/aarch64.md
6256 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6257 is fpsimd_<load/store>2.
6258 (load_pair<mode>): Likewise.
6259 (store_pair<mode>): Likewise.
6260
8c48eecd
JG		 62612013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6262
6263 * config/arm/types.md (type): Add "mrs" type.
6264 * config/aarch64/aarch64.md
6265 (aarch64_load_tp_hard): Make type "mrs".
6266 * config/arm/arm.md
6267 (load_tp_hard): Make type "mrs".
6268 * config/arm/cortex-a15.md: Update with new attributes.
6269 * config/arm/cortex-a5.md: Update with new attributes.
6270 * config/arm/cortex-a53.md: Update with new attributes.
6271 * config/arm/cortex-a7.md: Update with new attributes.
6272 * config/arm/cortex-a8.md: Update with new attributes.
6273 * config/arm/cortex-a9.md: Update with new attributes.
6274 * config/arm/cortex-m4.md: Update with new attributes.
6275 * config/arm/cortex-r4.md: Update with new attributes.
6276 * config/arm/fa526.md: Update with new attributes.
6277 * config/arm/fa606te.md: Update with new attributes.
6278 * config/arm/fa626te.md: Update with new attributes.
6279 * config/arm/fa726te.md: Update with new attributes.
6280
9ea187a5
JG		 62812013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6282
6283 * config/aarch64/aarch64.md
6284 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6285 (*movtf_aarch64): Likewise.
6286 * config/arm/arm.md
6287 (thumb1_movdi_insn): Use "multiple" for type where more than one
6288 instruction is used for a move.
6289 (*arm32_movhf): Likewise.
6290 (*thumb_movdf_insn): Likewise.
6291
292b89b3
JG		 62922013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6293
6294 * config/arm/types.md (type): Rename fcpys to fmov.
6295 * config/arm/vfp.md
6296 (*arm_movsi_vfp): Rename type fcpys as fmov.
6297 (*thumb2_movsi_vfp): Likewise
6298 (*movhf_vfp_neon): Likewise
6299 (*movhf_vfp): Likewise
6300 (*movsf_vfp): Likewise
6301 (*thumb2_movsf_vfp): Likewise
6302 (*movsfcc_vfp): Likewise
6303 (*thumb2_movsfcc_vfp): Likewise
6304 * config/aarch64/aarch64-simd.md
6305 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6306 * config/aarch64/aarch64.md
6307 (*movsi_aarch64): Replace type mov_reg with fmovs.
6308 (*movdi_aarch64): Likewise
6309 (*movsf_aarch64): Likewise
6310 (*movdf_aarch64): Likewise
6311 * config/arm/arm.c
6312 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6313 * config/arm/iwmmxt.md
6314 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6315 * config/arm/arm1020e.md: Update with new attributes.
6316 * config/arm/cortex-a15-neon.md: Update with new attributes.
6317 * config/arm/cortex-a5.md: Update with new attributes.
6318 * config/arm/cortex-a53.md: Update with new attributes.
6319 * config/arm/cortex-a7.md: Update with new attributes.
6320 * config/arm/cortex-a8-neon.md: Update with new attributes.
6321 * config/arm/cortex-a9.md: Update with new attributes.
6322 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6323 * config/arm/cortex-r4f.md: Update with new attributes.
6324 * config/arm/marvell-pj4.md: Update with new attributes.
6325 * config/arm/vfp11.md: Update with new attributes.
6326
6a24a83e
JG		 63272013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6328
6329 * config/aarch64/aarch64.md
6330 (*madd<mode>): Fix type attribute.
6331 (*maddsi_uxtw): Likewise.
6332 (*msub<mode>): Likewise.
6333 (*msubsi_uxtw): Likewise.
6334 (<su_optab>maddsidi4): Likewise.
6335 (<su_optab>msubsidi4): Likewise.
6336
b86923f0
JG		 63372013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6338
6339 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6340 * config/arm/arm.md (core_cycles): Remove fdiv.
6341 * config/arm/vfp.md:
6342 (*sqrtsf2_vfp): Update for attribute changes.
6343 (*sqrtdf2_vfp): Likewise.
6344 * config/aarch64/aarch64.md:
6345 (sqrt<mode>2): Update for attribute changes.
6346 * config/arm/arm1020e.md: Update with new attributes.
6347 * config/arm/cortex-a15-neon.md: Update with new attributes.
6348 * config/arm/cortex-a5.md: Update with new attributes.
6349 * config/arm/cortex-a53.md: Update with new attributes.
6350 * config/arm/cortex-a7.md: Update with new attributes.
6351 * config/arm/cortex-a8-neon.md: Update with new attributes.
6352 * config/arm/cortex-a9.md: Update with new attributes.
6353 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6354 * config/arm/cortex-r4f.md: Update with new attributes.
6355 * config/arm/marvell-pj4.md: Update with new attributes.
6356 * config/arm/vfp11.md: Update with new attributes.
6357
7b49c9e1
JG		 63582013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6359
6360 * config/arm/types.md
6361 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6362 * config/aarch64/aarch64.md
6363 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6364 new attributes.
6365 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6366 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6367 (float<GPI:mode><GPF:mode>2): Likewise.
6368 * config/arm/vfp.md
6369 (*truncsisf2_vfp): Update with new attributes.
6370 (*truncsidf2_vfp): Likewise.
6371 (fixuns_truncsfsi2): Likewise.
6372 (fixuns_truncdfsi2): Likewise.
6373 (*floatsisf2_vfp): Likewise.
6374 (*floatsidf2_vfp): Likewise.
6375 (floatunssisf2): Likewise.
6376 (floatunssidf2): Likewise.
6377 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6378 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6379 * config/arm/arm1020e.md: Update with new attributes.
6380 * config/arm/cortex-a15-neon.md: Update with new attributes.
6381 * config/arm/cortex-a5.md: Update with new attributes.
6382 * config/arm/cortex-a53.md: Update with new attributes.
6383 * config/arm/cortex-a7.md: Update with new attributes.
6384 * config/arm/cortex-a8-neon.md: Update with new attributes.
6385 * config/arm/cortex-a9.md: Update with new attributes.
6386 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6387 * config/arm/cortex-r4f.md: Update with new attributes.
6388 * config/arm/marvell-pj4.md: Update with new attributes.
6389 * config/arm/vfp11.md: Update with new attributes.
6390
e7df8af8
JG		 63912013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6392
6393 * config/aarch64/arm_neon.h
6394 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6395 (vqtbx<1,2,3,4><q>_s8): Likewise.
6396
594726e4
JG		 63972013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6398
f43245d1
UB		 6399 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6400 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG		 6401 (add<mode>3): Add type attribute.
6402 (add<mode>3): Likewise.
6403 (usadd<mode>3): Likewise.
6404 (ssadd<mode>3): Likewise.
6405 (sub<mode>3): Likewise.
6406 (sub<mode>3): Likewise.
6407 (ussub<mode>3): Likewise.
6408 (sssub<mode>3): Likewise.
6409 (ssmulsa3): Likewise.
6410 (usmulusa3): Likewise.
6411 (arm_usatsihi): Likewise.
6412 * config/arm/vfp.md
6413 (*movdi_vfp): Add types for all instructions.
6414 (*movdi_vfp_cortexa8): Likewise.
6415 (*movhf_vfp_neon): Likewise.
6416 (*movhf_vfp): Likewise.
6417 (*movdf_vfp): Likewise.
6418 (*thumb2_movdf_vfp): Likewise.
6419 (*thumb2_movdfcc_vfp): Likewise.
6420 * config/arm/arm.md: Add type attribute to all insn patterns.
6421 (*thumb1_adddi3): Add type attribute.
6422 (*arm_adddi3): Likewise.
6423 (*adddi_sesidi_di): Likewise.
6424 (*adddi_zesidi_di): Likewise.
6425 (*thumb1_addsi3): Likewise.
6426 (addsi3_compare0): Likewise.
6427 (*addsi3_compare0_scratch): Likewise.
6428 (*compare_negsi_si): Likewise.
6429 (cmpsi2_addneg): Likewise.
6430 (*addsi3_carryin_<optab>): Likewise.
6431 (*addsi3_carryin_alt2_<optab>): Likewise.
6432 (*addsi3_carryin_clobercc_<optab>): Likewise.
6433 (*subsi3_carryin): Likewise.
6434 (*subsi3_carryin_const): Likewise.
6435 (*subsi3_carryin_compare): Likewise.
6436 (*subsi3_carryin_compare_const): Likewise.
6437 (*arm_subdi3): Likewise.
6438 (*thumb_subdi3): Likewise.
6439 (*subdi_di_zesidi): Likewise.
6440 (*subdi_di_sesidi): Likewise.
6441 (*subdi_zesidi_di): Likewise.
6442 (*subdi_sesidi_di): Likewise.
6443 (*subdi_zesidi_ze): Likewise.
6444 (thumb1_subsi3_insn): Likewise.
6445 (*arm_subsi3_insn): Likewise.
6446 (*anddi3_insn): Likewise.
6447 (*anddi_zesidi_di): Likewise.
6448 (*anddi_sesdi_di): Likewise.
6449 (*ne_zeroextracts): Likewise.
6450 (*ne_zeroextracts): Likewise.
6451 (*ite_ne_zeroextr): Likewise.
6452 (*ite_ne_zeroextr): Likewise.
6453 (*anddi_notdi_di): Likewise.
6454 (*anddi_notzesidi): Likewise.
6455 (*anddi_notsesidi): Likewise.
6456 (andsi_notsi_si): Likewise.
6457 (thumb1_bicsi3): Likewise.
6458 (*iordi3_insn): Likewise.
6459 (*iordi_zesidi_di): Likewise.
6460 (*iordi_sesidi_di): Likewise.
6461 (*thumb1_iorsi3_insn): Likewise.
6462 (*xordi3_insn): Likewise.
6463 (*xordi_zesidi_di): Likewise.
6464 (*xordi_sesidi_di): Likewise.
6465 (*arm_xorsi3): Likewise.
6466 (*andsi_iorsi3_no): Likewise.
6467 (*smax_0): Likewise.
6468 (*smax_m1): Likewise.
6469 (*arm_smax_insn): Likewise.
6470 (*smin_0): Likewise.
6471 (*arm_smin_insn): Likewise.
6472 (*arm_umaxsi3): Likewise.
6473 (*arm_uminsi3): Likewise.
6474 (*minmax_arithsi): Likewise.
6475 (*minmax_arithsi_): Likewise.
6476 (*satsi_<SAT:code>): Likewise.
6477 (arm_ashldi3_1bit): Likewise.
6478 (arm_ashrdi3_1bit): Likewise.
6479 (arm_lshrdi3_1bit): Likewise.
6480 (*arm_negdi2): Likewise.
6481 (*thumb1_negdi2): Likewise.
6482 (*arm_negsi2): Likewise.
6483 (*thumb1_negsi2): Likewise.
6484 (*negdi_extendsid): Likewise.
6485 (*negdi_zero_extend): Likewise.
6486 (*arm_abssi2): Likewise.
6487 (*thumb1_abssi2): Likewise.
6488 (*arm_neg_abssi2): Likewise.
6489 (*thumb1_neg_abss): Likewise.
6490 (one_cmpldi2): Likewise.
6491 (extend<mode>di2): Likewise.
6492 (*compareqi_eq0): Likewise.
6493 (*arm_extendhisi2addsi): Likewise.
6494 (*arm_movdi): Likewise.
6495 (*thumb1_movdi_insn): Likewise.
6496 (*arm_movt): Likewise.
6497 (*thumb1_movsi_insn): Likewise.
6498 (pic_add_dot_plus_four): Likewise.
6499 (pic_add_dot_plus_eight): Likewise.
6500 (tls_load_dot_plus_eight): Likewise.
6501 (*thumb1_movhi_insn): Likewise.
6502 (*thumb1_movsf_insn): Likewise.
6503 (*movdf_soft_insn): Likewise.
6504 (*thumb_movdf_insn): Likewise.
6505 (cbranchsi4_insn): Likewise.
6506 (cbranchsi4_scratch): Likewise.
6507 (*negated_cbranchsi4): Likewise.
6508 (*tbit_cbranch): Likewise.
6509 (*tlobits_cbranch): Likewise.
6510 (*tstsi3_cbranch): Likewise.
6511 (*cbranchne_decr1): Likewise.
6512 (*addsi3_cbranch): Likewise.
6513 (*addsi3_cbranch_scratch): Likewise.
6514 (*arm_cmpdi_insn): Likewise.
6515 (*arm_cmpdi_unsig): Likewise.
6516 (*arm_cmpdi_zero): Likewise.
6517 (*thumb_cmpdi_zero): Likewise.
6518 (*deleted_compare): Likewise.
6519 (*mov_scc): Likewise.
6520 (*mov_negscc): Likewise.
6521 (*mov_notscc): Likewise.
6522 (*cstoresi_eq0_thumb1_insn): Likewise.
6523 (cstoresi_nltu_thumb1): Likewise.
6524 (cstoresi_ltu_thu): Likewise.
6525 (thumb1_addsi3_addgeu): Likewise.
6526 (*arm_jump): Likewise.
6527 (*thumb_jump): Likewise.
6528 (*check_arch2): Likewise.
6529 (arm_casesi_internal): Likewise.
6530 (thumb1_casesi_dispatch): Likewise.
6531 (*arm_indirect_jump): Likewise.
6532 (*thumb1_indirect_jump): Likewise.
6533 (nop): Likewise.
6534 (*and_scc): Likewise.
6535 (*ior_scc): Likewise.
6536 (*compare_scc): Likewise.
6537 (*cond_move): Likewise.
6538 (*cond_arith): Likewise.
6539 (*cond_sub): Likewise.
6540 (*cmp_ite0): Likewise.
6541 (*cmp_ite1): Likewise.
6542 (*cmp_and): Likewise.
6543 (*cmp_ior): Likewise.
6544 (*ior_scc_scc): Likewise.
6545 (*ior_scc_scc_cmp): Likewise.
6546 (*and_scc_scc): Likewise.
6547 (*and_scc_scc_cmp): Likewise.
6548 (*and_scc_scc_nod): Likewise.
6549 (*negscc): Likewise.
6550 (movcond_addsi): Likewise.
6551 (movcond): Likewise.
6552 (*ifcompare_plus_move): Likewise.
6553 (*if_plus_move): Likewise.
6554 (*ifcompare_move_plus): Likewise.
6555 (*if_move_plus): Likewise.
6556 (*ifcompare_arith_arith): Likewise.
6557 (*if_arith_arith): Likewise.
6558 (*ifcompare_arith_move): Likewise.
6559 (*if_arith_move): Likewise.
6560 (*ifcompare_move_arith): Likewise.
6561 (*if_move_arith): Likewise.
6562 (*ifcompare_move_not): Likewise.
6563 (*if_move_not): Likewise.
6564 (*ifcompare_not_move): Likewise.
6565 (*if_not_move): Likewise.
6566 (*ifcompare_shift_move): Likewise.
6567 (*if_shift_move): Likewise.
6568 (*ifcompare_move_shift): Likewise.
6569 (*if_move_shift): Likewise.
6570 (*ifcompare_shift_shift): Likewise.
6571 (*ifcompare_not_arith): Likewise.
6572 (*ifcompare_arith_not): Likewise.
6573 (*if_arith_not): Likewise.
6574 (*ifcompare_neg_move): Likewise.
6575 (*if_neg_move): Likewise.
6576 (*ifcompare_move_neg): Likewise.
6577 (*if_move_neg): Likewise.
6578 (prologue_thumb1_interwork): Likewise.
6579 (*cond_move_not): Likewise.
6580 (*sign_extract_onebit): Likewise.
6581 (*not_signextract_onebit): Likewise.
6582 (stack_tie): Likewise.
6583 (align_4): Likewise.
6584 (align_8): Likewise.
6585 (consttable_end): Likewise.
6586 (consttable_1): Likewise.
6587 (consttable_2): Likewise.
6588 (consttable_4): Likewise.
6589 (consttable_8): Likewise.
6590 (consttable_16): Likewise.
6591 (*thumb1_tablejump): Likewise.
6592 (prefetch): Likewise.
6593 (force_register_use): Likewise.
6594 (thumb_eh_return): Likewise.
6595 (load_tp_hard): Likewise.
6596 (load_tp_soft): Likewise.
6597 (tlscall): Likewise.
6598 (*arm_movtas_ze): Likewise.
6599 (*arm_rev): Likewise.
6600 (*arm_revsh): Likewise.
6601 (*arm_rev16): Likewise.
6602 * config/arm/thumb2.md
6603 (*thumb2_smaxsi3): Likewise.
6604 (*thumb2_sminsi3): Likewise.
6605 (*thumb32_umaxsi3): Likewise.
6606 (*thumb2_uminsi3): Likewise.
6607 (*thumb2_negdi2): Likewise.
6608 (*thumb2_abssi2): Likewise.
6609 (*thumb2_neg_abss): Likewise.
6610 (*thumb2_movsi_insn): Likewise.
6611 (tls_load_dot_plus_four): Likewise.
6612 (*thumb2_movhi_insn): Likewise.
6613 (*thumb2_mov_scc): Likewise.
6614 (*thumb2_mov_negs): Likewise.
6615 (*thumb2_mov_negs): Likewise.
6616 (*thumb2_mov_nots): Likewise.
6617 (*thumb2_mov_nots): Likewise.
6618 (*thumb2_movsicc_): Likewise.
6619 (*thumb2_movsfcc_soft_insn): Likewise.
6620 (*thumb2_indirect_jump): Likewise.
6621 (*thumb2_and_scc): Likewise.
6622 (*thumb2_ior_scc): Likewise.
6623 (*thumb2_ior_scc_strict_it): Likewise.
6624 (*thumb2_cond_move): Likewise.
6625 (*thumb2_cond_arith): Likewise.
6626 (*thumb2_cond_ari): Likewise.
6627 (*thumb2_cond_sub): Likewise.
6628 (*thumb2_negscc): Likewise.
6629 (*thumb2_movcond): Likewise.
6630 (thumb2_casesi_internal): Likewise.
6631 (thumb2_casesi_internal_pic): Likewise.
6632 (*thumb2_alusi3_short): Likewise.
6633 (*thumb2_mov<mode>_shortim): Likewise.
6634 (*thumb2_addsi_short): Likewise.
6635 (*thumb2_subsi_short): Likewise.
6636 (thumb2_addsi3_compare0): Likewise.
6637 (*thumb2_cbz): Likewise.
6638 (*thumb2_cbnz): Likewise.
6639 (*thumb2_one_cmplsi2_short): Likewise.
6640 (*thumb2_negsi2_short): Likewise.
6641 (*orsi_notsi_si): Likewise.
6642 * config/arm/arm1020e.md: Update with new attributes.
6643 * config/arm/arm1026ejs.md: Update with new attributes.
6644 * config/arm/arm1136jfs.md: Update with new attributes.
6645 * config/arm/arm926ejs.md: Update with new attributes.
6646 * config/arm/cortex-a15.md: Update with new attributes.
6647 * config/arm/cortex-a5.md: Update with new attributes.
6648 * config/arm/cortex-a53.md: Update with new attributes.
6649 * config/arm/cortex-a7.md: Update with new attributes.
6650 * config/arm/cortex-a8.md: Update with new attributes.
6651 * config/arm/cortex-a9.md: Update with new attributes.
6652 * config/arm/cortex-m4.md: Update with new attributes.
6653 * config/arm/cortex-r4.md: Update with new attributes.
6654 * config/arm/fa526.md: Update with new attributes.
6655 * config/arm/fa606te.md: Update with new attributes.
6656 * config/arm/fa626te.md: Update with new attributes.
6657 * config/arm/fa726te.md: Update with new attributes.
6658
1c83b673
JG		 66592013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6660
6661 * config/aarch64/aarch64-simd.md
6662 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6663 <vwx> iterator to ensure correct register choice.
6664 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6665 (aarch64_sqdmull_n<mode>): Likewise.
6666 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6667 * config/aarch64/arm_neon.h
6668 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6669 (vml<as><q>_n_<su>16): Likewise.
6670 (vml<as>l_high_lane<q>_<su>16): Likewise.
6671 (vml<as>l_high_n_<su>16): Likewise.
6672 (vml<as>l_lane<q>_<su>16): Likewise.
6673 (vml<as>l_n_<su>16): Likewise.
6674 (vmul<q>_lane<q>_<su>16): Likewise.
6675 (vmul<q>_n_<su>16): Likewise.
6676 (vmull_lane<q>_<su>16): Likewise.
6677 (vmull_n_<su>16): Likewise.
6678 (vmull_high_lane<q>_<su>16): Likewise.
6679 (vmull_high_n_<su>16): Likewise.
6680 (vqrdmulh<q>_n_s16): Likewise.
6681
bb1ae543
TB		 66822013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6683
6684 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6685 have the correct lane parameter.
6686
f23c0742
RB		 66872013-09-06 Richard Biener <rguenther@suse.de>
6688
6689 * cfganal.c (control_dependences::~control_dependences):
6690 Properly free all of the vector.
6691
c1d49770
KY		 66922013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6693
6694 PR target/58269
6695 * config/i386/i386.c (ix86_conditional_register_usage):
6696 Proper initialize extended SSE registers.
6697
6da47f52
JH		 66982013-09-06 Jan Hubicka <jh@suse.cz>
6699
6700 PR tree-optimization/58311
6701 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6702
67032013-09-06 Jan Hubicka <jh@suse.cz>
6704
6705 * Makefile.in (tree-sra.o): Update dependencies.
6706 * tree-sra.c: Include ipa-utils.h
6707 (scan_function): Use recursive_call_p.
6708 (has_caller_p): New function.
6709 (cgraph_for_node_and_aliases): Count also callers of aliases.
6710
fc11f321
JH		 67112013-09-06 Jan Hubicka <jh@suse.cz>
6712
6713 PR middle-end/58094
6714 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6715 * tree-tailcall.c: Include ipa-utils.h.
6716 (find_tail_calls): Use it.
6717 * ipa-pure-const.c (check_call): Likewise.
6718 * ipa-utils.c (recursive_call_p): New function.
6719 * ipa-utils.h (recursive_call_p): Dclare.
6720 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6721 (symtab_semantically_equivalent_p): New function.
6722 * Makefile.in (tree-tailcall.o): Update dependencies.
6723
9a6af450
EB		 67242013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6725
6726 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6727 non-inlinable part.
6728
35f32ad4
RB		 67292013-09-06 Richard Biener <rguenther@suse.de>
6730
6731 * lto-streamer.h (lto_global_var_decls): Remove.
6732 * Makefile.in (OBJS): Remove lto-symtab.o.
6733 (lto-symtab.o): Remove.
6734 (GTFILES): Remove lto-symtab.c
6735 * lto-symtab.c: Move to lto/
6736
d12a76f3
AK		 67372013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6738
6739 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6740 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6741 (UNSPEC_FPINT_RINT): New constant definitions.
6742 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6743 definition with 2 attributes.
6744 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6745 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6746 definitions.
6747
67482013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6749
f43245d1 6750 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK		 6751 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6752 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6753 attribute to "z196_alone".
6754 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6755 "zEC12_simple".
6756
c8e9d8c3
RB		 67572013-09-06 Richard Biener <rguenther@suse.de>
6758
6759 * basic-block.h (class control_dependences): New.
6760 * tree-ssa-dce.c (control_dependence_map): Remove.
6761 (cd): New global.
6762 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6763 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6764 find_pdom, find_control_dependence, find_all_control_dependences):
6765 Move to cfganal.c.
f43245d1
UB		 6766 (mark_control_dependent_edges_necessary,
6767 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6768 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB		 6769 * cfganal.c (set_control_dependence_map_bit,
6770 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6771 find_all_control_dependences): Move from tree-ssa-dce.c and
6772 implement as methods of control_dependences class.
6773 (control_dependences::control_dependences): New.
6774 (control_dependences::~control_dependences): Likewise.
6775 (control_dependences::get_edges_dependent_on): Likewise.
6776 (control_dependences::get_edge): Likewise.
6777
7d2ba471
JH		 67782013-09-04 Jan Hubicka <jh@suse.cz>
6779
6780 * tree.c (types_same_for_odr): Drop overactive check.
6781 * ipa-devirt.c (hash_type_name): Likewise.
6782
e18412fc
JH		 67832013-09-04 Jan Hubicka <jh@suse.cz>
6784
6785 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6786 (analyze_functions): ... here.
6787
d352b245
JH		 67882013-09-04 Jan Hubicka <jh@suse.cz>
6789
6790 PR middle-end/58201
6791 * cgraphunit.c (analyze_functions): Clear AUX fields
6792 after processing; initialize assembler name has.
6793
770da076
JL		 67942013-09-05 Jeff Law <law@redhat.com>
6795
6796 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6797 from thread_around_empty_block. Record threading path into PATH.
6798 Recurse if threading through the initial block is successful.
6799 (thread_across_edge): Corresponding changes to slightly simplify.
6800
fe6f68e2
JG		 68012013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6802
6803 * config/aarch64/aarch64.md
6804 (type): Remove frecpe, frecps, frecpx.
6805 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6806 fix to be a TARGET_SIMD instruction.
6807 (aarch64_frecps): Remove.
6808 * config/aarch64/aarch64-simd.md
6809 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6810 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6811
6e4150e1
JG		 68122013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6813 Sofiane Naci <sofiane.naci@arm.com>
6814
f43245d1 6815 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 6816 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB		 6817 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6818 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6819 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6820 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6821 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6822 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6823 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 6824 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 6825 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG		 6826 shift* types.
6827 * config/arm/arm-fixed.md: Update for attribute changes
6828 all occurrences of arlo_* types.
6829 * config/arm/thumb2.md: Update for attribute changes all occurrences
6830 of arlo_* types.
6831 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6832 (cortexa7_older_only): Likewise.
6833 (cortexa7_younger): Likewise.
6834 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6835 (1020alu_shift_op): Likewise.
6836 (1020alu_shift_reg_op): Likewise.
6837 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6838 (alu_shift_op): Likewise.
6839 (alu_shift_reg_op): Likewise.
f43245d1 6840 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG		 6841 (11_alu_shift_op): Likewise.
6842 (11_alu_shift_reg_op): Likewise.
6843 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6844 (9_alu_shift_reg_op): Likewise.
6845 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6846 attribute changes.
6847 (cortex_a15_alu_shift): Likewise.
6848 (cortex_a15_alu_shift_reg): Likewise.
6849 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6850 attribute changes.
6851 (cortex_a5_alu_shift): Likewise.
f43245d1
UB		 6852 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6853 attribute changes.
6e4150e1 6854 (cortex_a53_alu_shift): Likewise.
f43245d1
UB		 6855 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6856 attribute changes.
6e4150e1
JG		 6857 (cortex_a7_alu_reg): Likewise.
6858 (cortex_a7_alu_shift): Likewise.
f43245d1
UB		 6859 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6860 attribute changes.
6e4150e1
JG		 6861 (cortex_a8_alu_shift): Likewise.
6862 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 6863 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 6864 (cortex_a9_dp_shift): Likewise.
f43245d1
UB		 6865 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6866 attribute changes.
6e4150e1
JG		 6867 * config/arm/cortex-r4.md
6868 (cortex_r4_alu): Update for attribute changes.
6869 (cortex_r4_mov): Likewise.
6870 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 6871 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 6872 (526_alu_shift_op): Likewise.
f43245d1
UB		 6873 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6874 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 6875 (626te_alu_shift_op): Likewise.
f43245d1 6876 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG		 6877 (726te_alu_shift_op): Likewise.
6878 (726te_alu_shift_reg_op): Likewise.
6879 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6880 (mp626_alu_shift_op): Likewise.
6881 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6882 (pj4_alu_conds): Likewise.
6883 (pj4_shift): Likewise.
6884 (pj4_shift_conds): Likewise.
6885 (pj4_alu_shift): Likewise.
6886 (pj4_alu_shift_conds): Likewise.
6887 * config/aarch64/aarch64.md: Update for attribute change
6888 all occurrences of arlo_* and shift* types.
6889
e6523306
MS		 68902013-09-05 Mike Stump <mikestump@comcast.net>
6891
6892 * tree.h: Move documentation for tree_function_decl to tree-core.h
6893 with the declaration.
6894
80a18161
PB		 68952013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6896
6897 PR target/58139
6898 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6899 looking for widest mode.
6900
6994430a
EB		 69012013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6902
6903 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6904
e67f39f7
RB		 69052013-09-05 Richard Biener <rguenther@suse.de>
6906
6907 PR tree-optimization/58137
6908 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6909 Do not create vectors of pointers.
6910 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6911 types for the components of the vector initializer.
6912 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6913 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6914
568cda29
MJ		 69152013-09-05 Martin Jambor <mjambor@suse.cz>
6916
6917 * ipa-prop.c (remove_described_reference): Accept missing references,
6918 return false if that hppens, otherwise return true.
6919 (cgraph_node_for_jfunc): New function.
6920 (try_decrement_rdesc_refcount): Likewise.
6921 (try_make_edge_direct_simple_call): Use them.
6922 (ipa_edge_removal_hook): Remove references from rdescs.
6923 (ipa_edge_duplication_hook): Clone rdescs and their references
6924 when the new edge has the same caller as the old one.
6925 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6926 reference before removing any edges.
6927
5a200acb
RE		 69282013-09-05 Richard Earnshaw <rearnsha@arm.com>
6929
6930 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6931 initial store.
6932 * thumb2.md (thumb2_storewb_parisi): New pattern.
6933
5922847b
YZ		 69342013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6935
6936 * config/aarch64/aarch64-option-extensions.def: Add
6937 AARCH64_OPT_EXTENSION of 'crc'.
6938 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6939 (AARCH64_ISA_CRC): Ditto.
6940 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6941 description of the CRC extension.
6942
996746aa
AI		 69432013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6944
6945 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6946 * config/rs6000/linux.h: Ditto.
6947 * alpha/linux.h: Ditto.
6948 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6949 no_c99_libc_has_function.
6950 * config/c6x/uclinux-elf.h: Ditto.
6951 * config/lm32/uclinux-elf.h: Ditto.
6952 * config/m68k/uclinux.h: Ditto.
6953 * config/moxie/uclinux.h: Ditto.
6954 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6955 (crisv32-*-linux*, cris-*-linux*): Ditto.
6956 * config/bfin/bfin.c: Include "tm_p.h".
6957
c370d99f
RB		 69582013-09-05 Richard Biener <rguenther@suse.de>
6959
6960 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6961 check for a definition without a basic-block.
6962
003bb7f3
JG		 69632013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6964 Sofiane Naci <sofiane.naci@arm.com>
6965
6966 * config/aarch64/aarch64.md
6967 (*movti_aarch64): Rename r_2_f and f_2_r.
6968 (*movsf_aarch64): Likewise.
6969 (*movdf_aarch64): Likewise.
6970 (*movtf_aarch64): Likewise.
6971 (aarch64_movdi_<mode>low): Likewise.
6972 (aarch64_movdi_<mode>high): Likewise.
6973 (aarch64_mov<mode>high_di): Likewise.
6974 (aarch64_mov<mode>low_di): Likewise.
6975 (aarch64_movtilow_tilow): Likewise.
6976 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6977 values to config/arm/types.md
6978 (attribute "conds"): Update for attribute change.
6979 (anddi3_insn): Likewise.
6980 (iordi3_insn): Likewise.
6981 (xordi3_insn): Likewise.
6982 (one_cmpldi2): Likewise.
6983 * config/arm/types.md (type): Add Neon types.
6984 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6985 use "type" attribute.
6986 (movmisalign<mode>_neon_store): Likewise.
6987 (movmisalign<mode>_neon_load): Likewise.
6988 (vec_set<mode>_internal): Likewise.
6989 (vec_setv2di_internal): Likewise.
6990 (vec_extract<mode>): Likewise.
6991 (vec_extractv2di): Likewise.
6992 (add<mode>3_neon): Likewise.
6993 (adddi3_neon): Likewise.
6994 (sub<mode>3_neon): Likewise.
6995 (subdi3_neon): Likewise.
6996 (mul<mode>3_neon): Likewise.
6997 (mul<mode>3add<mode>_neon): Likewise.
6998 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6999 (fma<VCVTF:mode>4)): Likewise.
7000 (fma<VCVTF:mode>4_intrinsic): Likewise.
7001 (fmsub<VCVTF:mode>4)): Likewise.
7002 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7003 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7004 (ior<mode>3): Likewise.
7005 (and<mode>3): Likewise.
7006 (anddi3_neon): Likewise.
7007 (orn<mode>3_neon): Likewise.
7008 (orndi3_neon): Likewise.
7009 (bic<mode>3_neon): Likewise.
7010 (bicdi3_neon): Likewise.
7011 (xor<mode>3): Likewise.
7012 (one_cmpl<mode>2): Likewise.
7013 (abs<mode>2): Likewise.
7014 (neg<mode>2): Likewise.
7015 (umin<mode>3_neon): Likewise.
7016 (umax<mode>3_neon): Likewise.
7017 (smin<mode>3_neon): Likewise.
7018 (smax<mode>3_neon): Likewise.
7019 (vashl<mode>3): Likewise.
7020 (vashr<mode>3_imm): Likewise.
7021 (vlshr<mode>3_imm): Likewise.
7022 (ashl<mode>3_signed): Likewise.
7023 (ashl<mode>3_unsigned): Likewise.
7024 (neon_load_count): Likewise.
7025 (ashldi3_neon_noclobber): Likewise.
7026 (signed_shift_di3_neon): Likewise.
7027 (unsigned_shift_di3_neon): Likewise.
7028 (ashrdi3_neon_imm_noclobber): Likewise.
7029 (lshrdi3_neon_imm_noclobber): Likewise.
7030 (widen_ssum<mode>3): Likewise.
7031 (widen_usum<mode>3): Likewise.
7032 (quad_halves_<code>v4si): Likewise.
7033 (quad_halves_<code>v4sf): Likewise.
7034 (quad_halves_<code>v8hi): Likewise.
7035 (quad_halves_<code>v16qi): Likewise.
7036 (reduc_splus_v2di): Likewise.
7037 (neon_vpadd_internal<mode>): Likewise.
7038 (neon_vpsmin<mode>): Likewise.
7039 (neon_vpsmax<mode>): Likewise.
7040 (neon_vpumin<mode>): Likewise.
7041 (neon_vpumax<mode>): Likewise.
7042 (ss_add<mode>_neon): Likewise.
7043 (us_add<mode>_neon): Likewise.
7044 (ss_sub<mode>_neon): Likewise.
7045 (us_sub<mode>_neon): Likewise.
7046 (neon_vadd<mode>_unspec): Likewise.
7047 (neon_vaddl<mode>): Likewise.
7048 (neon_vaddw<mode>): Likewise.
7049 (neon_vhadd<mode>): Likewise.
7050 (neon_vqadd<mode>): Likewise.
7051 (neon_vaddhn<mode>): Likewise.
7052 (neon_vmul<mode>): Likewise.
7053 (neon_vmla<mode>): Likewise.
7054 (neon_vmlal<mode>): Likewise.
7055 (neon_vmls<mode>): Likewise.
7056 (neon_vmlsl<mode>): Likewise.
7057 (neon_vqdmulh<mode>): Likewise.
7058 (neon_vqdmlal<mode>): Likewise.
7059 (neon_vqdmlsl<mode>): Likewise.
7060 (neon_vmull<mode>): Likewise.
7061 (neon_vqdmull<mode>): Likewise.
7062 (neon_vsub<mode>_unspec): Likewise.
7063 (neon_vsubl<mode>): Likewise.
7064 (neon_vsubw<mode>): Likewise.
7065 (neon_vqsub<mode>): Likewise.
7066 (neon_vhsub<mode>): Likewise.
7067 (neon_vsubhn<mode>): Likewise.
7068 (neon_vceq<mode>): Likewise.
7069 (neon_vcge<mode>): Likewise.
7070 (neon_vcgeu<mode>): Likewise.
7071 (neon_vcgt<mode>): Likewise.
7072 (neon_vcgtu<mode>): Likewise.
7073 (neon_vcle<mode>): Likewise.
7074 (neon_vclt<mode>): Likewise.
7075 (neon_vcage<mode>): Likewise.
7076 (neon_vcagt<mode>): Likewise.
7077 (neon_vtst<mode>): Likewise.
7078 (neon_vabd<mode>): Likewise.
7079 (neon_vabdl<mode>): Likewise.
7080 (neon_vaba<mode>): Likewise.
7081 (neon_vabal<mode>): Likewise.
7082 (neon_vmax<mode>): Likewise.
7083 (neon_vmin<mode>): Likewise.
7084 (neon_vpaddl<mode>): Likewise.
7085 (neon_vpadal<mode>): Likewise.
7086 (neon_vpmax<mode>): Likewise.
7087 (neon_vpmin<mode>): Likewise.
7088 (neon_vrecps<mode>): Likewise.
7089 (neon_vrsqrts<mode>): Likewise.
7090 (neon_vqabs<mode>): Likewise.
7091 (neon_vqneg<mode>): Likewise.
7092 (neon_vcls<mode>): Likewise.
7093 (clz<mode>2): Likewise.
7094 (popcount<mode>2): Likewise.
7095 (neon_vrecpe): Likewise.
7096 (neon_vrsqrte): Likewise.
7097 (neon_vget_lane<mode>_sext_internal): Likewise.
7098 (neon_vget_lane<mode>_zext_internal): Likewise.
7099 (neon_vdup_n<mode>): Likewise.
7100 (neon_vdup_nv2di): Likewise.
7101 (neon_vdpu_lane<mode>_internal): Likewise.
7102 (neon_vswp<mode>): Likewise.
7103 (float<mode><V_cvtto>2): Likewise.
7104 (floatuns<mode><V_cvtto>2): Likewise.
7105 (fix_trunc<mode><V_cvtto>)2): Likewise
7106 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7107 (neon_vcvt<mode>): Likewise.
7108 (neon_vcvtv4sfv4hf): Likewise.
7109 (neon_vcvtv4hfv4sf): Likewise.
7110 (neon_vcvt_n<mode>): Likewise.
7111 (neon_vmovn<mode>): Likewise.
7112 (neon_vqmovn<mode>): Likewise.
7113 (neon_vqmovun<mode>): Likewise.
7114 (neon_vmovl<mode>): Likewise.
7115 (neon_vmul_lane<mode>): Likewise.
7116 (neon_vmull_lane<mode>): Likewise.
7117 (neon_vqdmull_lane<mode>): Likewise.
7118 (neon_vqdmulh_lane<mode>): Likewise.
7119 (neon_vmla_lane<mode>): Likewise.
7120 (neon_vmlal_lane<mode>): Likewise.
7121 (neon_vqdmlal_lane<mode>): Likewise.
7122 (neon_vmls_lane<mode>): Likewise.
7123 (neon_vmlsl_lane<mode>): Likewise.
7124 (neon_vqdmlsl_lane<mode>): Likewise.
7125 (neon_vext<mode>): Likewise.
7126 (neon_vrev64<mode>): Likewise.
7127 (neon_vrev32<mode>): Likewise.
7128 (neon_vrev16<mode>): Likewise.
7129 (neon_vbsl<mode>_internal): Likewise.
7130 (neon_vshl<mode>): Likewise.
7131 (neon_vqshl<mode>): Likewise.
7132 (neon_vshr_n<mode>): Likewise.
7133 (neon_vshrn_n<mode>): Likewise.
7134 (neon_vqshrn_n<mode>): Likewise.
7135 (neon_vqshrun_n<mode>): Likewise.
7136 (neon_vshl_n<mode>): Likewise.
7137 (neon_vqshl_n<mode>): Likewise.
7138 (neon_vqshlu_n<mode>): Likewise.
7139 (neon_vshll_n<mode>): Likewise.
7140 (neon_vsra_n<mode>): Likewise.
7141 (neon_vsri_n<mode>): Likewise.
7142 (neon_vsli_n<mode>): Likewise.
7143 (neon_vtbl1v8qi): Likewise.
7144 (neon_vtbl2v8qi): Likewise.
7145 (neon_vtbl3v8qi): Likewise.
7146 (neon_vtbl4v8qi): Likewise.
7147 (neon_vtbx1v8qi): Likewise.
7148 (neon_vtbx2v8qi): Likewise.
7149 (neon_vtbx3v8qi): Likewise.
7150 (neon_vtbx4v8qi): Likewise.
7151 (neon_vtrn<mode>_internal): Likewise.
7152 (neon_vzip<mode>_internal): Likewise.
7153 (neon_vuzp<mode>_internal): Likewise.
7154 (neon_vld1<mode>): Likewise.
7155 (neon_vld1_lane<mode>): Likewise.
7156 (neon_vld1_dup<mode>): Likewise.
7157 (neon_vld1_dupv2di): Likewise.
7158 (neon_vst1<mode>): Likewise.
7159 (neon_vst1_lane<mode>): Likewise.
7160 (neon_vld2<mode>): Likewise.
7161 (neon_vld2_lane<mode>): Likewise.
7162 (neon_vld2_dup<mode>): Likewise.
7163 (neon_vst2<mode>): Likewise.
7164 (neon_vst2_lane<mode>): Likewise.
7165 (neon_vld3<mode>): Likewise.
7166 (neon_vld3qa<mode>): Likewise.
7167 (neon_vld3qb<mode>): Likewise.
7168 (neon_vld3_lane<mode>): Likewise.
7169 (neon_vld3_dup<mode>): Likewise.
7170 (neon_vst3<mode>): Likewise.
7171 (neon_vst3qa<mode>): Likewise.
7172 (neon_vst3qb<mode>): Likewise.
7173 (neon_vst3_lane<mode>): Likewise.
7174 (neon_vld4<mode>): Likewise.
7175 (neon_vld4qa<mode>): Likewise.
7176 (neon_vld4qb<mode>): Likewise.
7177 (neon_vld4_lane<mode>): Likewise.
7178 (neon_vld4_dup<mode>): Likewise.
7179 (neon_vst4<mode>): Likewise.
7180 (neon_vst4qa<mode>): Likewise.
7181 (neon_vst4qb<mode>): Likewise.
7182 (neon_vst4_lane<mode>): Likewise.
7183 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7184 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7185 (neon_vec_<US>mult_lo_<mode>): Likewise.
7186 (neon_vec_<US>mult_hi_<mode>): Likewise.
7187 (neon_vec_<US>shiftl_<mode>): Likewise.
7188 (neon_unpack<US>_<mode>): Likewise.
7189 (neon_vec_<US>mult_<mode>): Likewise.
7190 (vec_pack_trunc_<mode>): Likewise.
7191 (neon_vec_pack_trunk_<mode>): Likewise.
7192 (neon_vabd<mode>_2): Likewise.
7193 (neon_vabd<mode>_3): Likewise.
7194 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7195 (thumb2_movsi_vfp): Likewise.
7196 (movdi_vfp): Likewise.
7197 (movdi_vfp_cortexa8): Likewise.
7198 (movhf_vfp_neon): Likewise.
7199 (movhf_vfp): Likewiwse.
7200 (movsf_vfp): Likewiwse.
7201 (thumb2_movsf_vfp): Likewiwse.
7202 (movdf_vfp): Likewise.
7203 (thumb2_movdf_vfp): Likewise.
7204 (movsfcc_vfp): Likewise.
7205 (thumb2_movsfcc_vfp): Likewise.
7206 (movdfcc_vfp): Likewise.
7207 (thumb2_movdfcc_vfp): Likewise.
7208 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7209 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7210 (v10_v2c): Likewise.
7211 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7212 attribute change.
7213 (cortex_a15_neon_int_2): Likewise.
7214 (cortex_a15_neon_int_3): Likewise.
7215 (cortex_a15_neon_int_4): Likewise.
7216 (cortex_a15_neon_int_5): Likewise.
7217 (cortex_a15_neon_vqneg_vqabs): Likewise.
7218 (cortex_a15_neon_vmov): Likewise.
7219 (cortex_a15_neon_vaba): Likewise.
7220 (cortex_a15_neon_vaba_qqq): Likewise.
7221 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7222 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7223 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7224 scalar_64_32_long_scalar): Likewise.
7225 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7226 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7227 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7228 lotype_qdd_64_32_long): Likewise.
7229 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7230 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7231 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7232 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7233 (cortex_a15_neon_shift_1): Likewise.
7234 (cortex_a15_neon_shift_2): Likewise.
7235 (cortex_a15_neon_shift_3): Likewise.
7236 (cortex_a15_neon_vshl_ddd): Likewise.
7237 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7238 (cortex_a15_neon_vsra_vrsra): Likewise.
7239 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7240 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7241 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7242 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7243 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7244 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7245 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7246 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7247 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7248 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7249 (cortex_a15_neon_bp_simple): Likewise.
7250 (cortex_a15_neon_bp_2cycle): Likewise.
7251 (cortex_a15_neon_bp_3cycle): Likewise.
7252 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7253 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7254 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7255 (cortex_a15_neon_vld2_4_regs): Likewise.
7256 (cortex_a15_neon_vld3_vld4): Likewise.
7257 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7258 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7259 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7260 (cortex_a15_neon_vst3_vst4): Likewise.
7261 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7262 (cortex_a15_neon_vld3_vld4_lane" 10
7263 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7264 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7265 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7266 (cortex_a15_neon_ldm_2): Likewise.0
7267 (cortex_a15_neon_stm_2): Likewise.
7268 (cortex_a15_neon_mcr): Likewise.
7269 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7270 (cortex_a15_neon_mrc): Likewise.
7271 (cortex_a15_neon_mrrc): Likewise.
7272 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7273 change.
7274 (cortex_a15_alu_shift): Likewise.
7275 (cortex_a15_alu_shift_reg): Likewise.
7276 (cortex_a15_mult32): Likewise.
7277 (cortex_a15_mult64): Likewise.
7278 (cortex_a15_block): Likewise.
7279 (cortex_a15_branch): Likewise.
7280 (cortex_a15_load1): Likewise.
7281 (cortex_a15_load3): Likewise.
7282 (cortex_a15_store1): Likewise.
7283 (cortex_a15_store3): Likewise.
7284 (cortex_a15_call): Likewise.
f43245d1 7285 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG		 7286 (cortex_a5_f2r): Likewise.
7287 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7288 change.
7289 (cortex_a53_f2r): Likewise.
7290 * config/arm/cortex-a7.md
7291 (cortex_a7_branch): Update for attribute change.
7292 (cortex_a7_call): Likewise.
7293 (cortex_a7_alu_imm): Likewise.
7294 (cortex_a7_alu_reg): Likewise.
7295 (cortex_a7_alu_shift): Likewise.
7296 (cortex_a7_mul): Likewise.
7297 (cortex_a7_load1): Likewise.
7298 (cortex_a7_store1): Likewise.
7299 (cortex_a7_load2): Likewise.
7300 (cortex_a7_store2): Likewise.
7301 (cortex_a7_load3): Likewise.
7302 (cortex_a7_store3): Likewise.
7303 (cortex_a7_load4): Likewise.
7304 (cortex_a7_store4): Likewise.
7305 (cortex_a7_fpalu): Likewise.
7306 (cortex_a7_fconst): Likewise.
7307 (cortex_a7_fpmuls): Likewise.
7308 (cortex_a7_neon_mul): Likewise.
7309 (cortex_a7_fpmacs): Likewise.
7310 (cortex_a7_neon_mla: Likewise.
7311 (cortex_a7_fpmuld: Likewise.
7312 (cortex_a7_fpmacd: Likewise.
7313 (cortex_a7_fpfmad: Likewise.
7314 (cortex_a7_fdivs: Likewise.
7315 (cortex_a7_fdivd: Likewise.
7316 (cortex_a7_r2f: Likewise.
7317 (cortex_a7_f2r: Likewise.
7318 (cortex_a7_f_flags: Likewise.
7319 (cortex_a7_f_loads: Likewise.
7320 (cortex_a7_f_loadd: Likewise.
7321 (cortex_a7_f_stores: Likewise.
7322 (cortex_a7_f_stored: Likewise.
7323 (cortex_a7_neon): Likewise.
7324 * config/arm/cortex-a8-neon.md
7325 (cortex_a8_neon_mrc): Update for attribute change.
7326 (cortex_a8_neon_mrrc): Likewise.
7327 (cortex_a8_neon_int_1): Likewise.
7328 (cortex_a8_neon_int_2): Likewise.
7329 (cortex_a8_neon_int_3): Likewise.
7330 (cortex_a8_neon_int_4): Likewise.
7331 (cortex_a8_neon_int_5): Likewise.
7332 (cortex_a8_neon_vqneg_vqabs): Likewise.
7333 (cortex_a8_neon_vmov): Likewise.
7334 (cortex_a8_neon_vaba): Likewise.
7335 (cortex_a8_neon_vaba_qqq): Likewise.
7336 (cortex_a8_neon_vsma): Likewise.
7337 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7338 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 7339 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7340 long_scalar): Likewise.
003bb7f3
JG		 7341 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7342 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7343 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7344 long_scalar_qdd_64_32_long): Likewise.
7345 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7346 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7347 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7348 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7349 (cortex_a8_neon_shift_1): Likewise.
7350 (cortex_a8_neon_shift_2): Likewise.
7351 (cortex_a8_neon_shift_3): Likewise.
7352 (cortex_a8_neon_vshl_ddd): Likewise.
7353 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7354 (cortex_a8_neon_vsra_vrsra): Likewise.
7355 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7356 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7357 (cortex_a8_neon_fp_vsum): Likewise.
7358 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7359 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7360 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7361 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7362 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7363 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7364 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7365 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7366 (cortex_a8_neon_bp_simple): Likewise.
7367 (cortex_a8_neon_bp_2cycle): Likewise.
7368 (cortex_a8_neon_bp_3cycle): Likewise.
7369 (cortex_a8_neon_ldr): Likewise.
7370 (cortex_a8_neon_str): Likewise.
7371 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7372 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7373 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7374 (cortex_a8_neon_vld2_4_regs): Likewise.
7375 (cortex_a8_neon_vld3_vld4): Likewise.
7376 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7377 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7378 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7379 (cortex_a8_neon_vst3_vst4): Likewise.
7380 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7381 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7382 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7383 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7384 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7385 (cortex_a8_neon_mcr): Likewise.
7386 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 7387 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG		 7388 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7389 change.
7390 (ca9_neon_mrrc): Likewise.
7391 (cortex_a9_neon_int_1): Likewise.
7392 (cortex_a9_neon_int_2): Likewise.
7393 (cortex_a9_neon_int_3): Likewise.
7394 (cortex_a9_neon_int_4): Likewise.
7395 (cortex_a9_neon_int_5): Likewise.
7396 (cortex_a9_neon_vqneg_vqabs): Likewise.
7397 (cortex_a9_neon_vmov): Likewise.
7398 (cortex_a9_neon_vaba): Likewise.
7399 (cortex_a9_neon_vaba_qqq): Likewise.
7400 (cortex_a9_neon_vsma): Likewise.
7401 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7402 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB		 7403 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7404 long_scalar): Likewise.
003bb7f3
JG		 7405 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7406 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7407 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7408 long_scalar_qdd_64_32_long): Likewise.
7409 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7410 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7411 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7412 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7413 (cortex_a9_neon_shift_1): Likewise.
7414 (cortex_a9_neon_shift_2): Likewise.
7415 (cortex_a9_neon_shift_3): Likewise.
7416 (cortex_a9_neon_vshl_ddd): Likewise.
7417 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7418 (cortex_a9_neon_vsra_vrsra): Likewise.
7419 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7420 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7421 (cortex_a9_neon_fp_vsum): Likewise.
7422 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7423 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7424 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7425 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7426 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7427 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7428 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7429 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7430 (cortex_a9_neon_bp_simple): Likewise.
7431 (cortex_a9_neon_bp_2cycle): Likewise.
7432 (cortex_a9_neon_bp_3cycle): Likewise.
7433 (cortex_a9_neon_ldr): Likewise.
7434 (cortex_a9_neon_str): Likewise.
7435 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7436 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7437 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7438 (cortex_a9_neon_vld2_4_regs): Likewise.
7439 (cortex_a9_neon_vld3_vld4): Likewise.
7440 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7441 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7442 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7443 (cortex_a9_neon_vst3_vst4): Likewise.
7444 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7445 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7446 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7447 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7448 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7449 (cortex_a9_neon_mcr): Likewise.
7450 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7451 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7452 (cortex_a9_fps): Likewise.
7453 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7454 change.
7455 (cortex_m4_fmuls): Likewise.
7456 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7457 change.
7458 (cortex_r4_mrc): Likewise.
7459 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 7460 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 7461 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB		 7462 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7463 attribute change.
003bb7f3
JG		 7464 (pj4_core_to_vfp): Likewise.
7465 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7466 attribute change.
7467 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7468 (vfp_fstore): Likewise.
7469 * doc/md.texi: Change references to neon_type to refer to type.
7470
fbe0dc78
DS		 74712013-09-04 Dodji Seketeli <dodji@redhat.com>
7472
7473 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7474
c167bc5b
DE		 74752013-09-04 David Edelsohn <dje.gcc@gmail.com>
7476
7477 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7478 lglobl if not weak.
7479
035cb59f
ER		 74802013-09-04 Easwaran Raman <eraman@google.com>
7481
7482 PR middle-end/57370
7483 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7484 (build_and_add_sum): Use it.
7485 (appears_later_in_bb): Simplify code.
7486
71191083
TJ		 74872013-09-04 Teresa Johnson <tejohnson@google.com>
7488
7489 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7490
0c488213
JG		 74912013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7492
7493 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7494
61a74079
JH		 74952013-09-04 Jan Hubicka <jh@suse.cz>
7496
7497 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7498 * ipa-devirt.c: Include diganostic.h
7499 (odr_type_d): Add types and types_set.
7500 (hash_type_name): Work for types with vtables during LTO.
7501 (odr_hasher::remove): Fix comment; destroy types_set.
7502 (add_type_duplicate): New function,
7503 (get_odr_type): Use it.
7504 (dump_type_inheritance_graph): Dump type duplicates.
7505 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7506 graph.
7507 * tree.c (types_same_for_odr): Give exact answers on types with
7508 virtual tables.
7509
24ba8195
DS		 75102013-09-04 Dodji Seketeli <dodji@redhat.com>
7511
7512 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7a5911d3 7513 explaining their differences.
24ba8195 7514
0a04869e
SKS		 75152013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7516
7517 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7518
3b18bc42
JL		 75192013-09-03 Jeff Law <law@redhat.com>
7520
7521 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7522 when not threading through a joiner block. Pass joiner/no joiner
7523 state to register_jump_thread.
7524 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7525 state from argument rather than implying on path length.
7526 Dump the entire jump thread path into debugging dump.
7527 * tree-flow.h (register_jump_thread): Update prototype.
7528
ae84e151
XDL		 75292013-08-29 Xinliang David Li <davidxl@google.com>
7530
7531 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7532 Remove a trivial gcc_assert.
7533
c716e67f
XDL		 75342013-08-29 Xinliang David Li <davidxl@google.com>
7535
7536 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7537 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7538 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7539 Delay base decl alignment adjustment.
7540 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7541 * tree-vectorizer.h: New data structure.
7542 (set_dr_misalignment): New function.
7543 (dr_misalignment): Ditto.
7544 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7545 (vectorizable_load): Ditto.
7546 (ensure_base_align): New function.
7547 (vectorize_loops): Add dbg_cnt support.
7548 (execute_vect_slp): Ditto.
7549 * dbgcnt.def: New debug counter.
7550 * Makefile: New dependency.
7551
db8800bc
MI		 75522013-09-03 Meador Inge <meadori@codesourcery.com>
7553
7554 Revert:
7555
7556 2013-08-30 Meador Inge <meadori@codesourcery.com>
7557
7558 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7559
0d9b4f5c
DE		 75602013-09-03 David Edelsohn <dje.gcc@gmail.com>
7561
7562 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7563 function descriptor.
7564
29e810b9
RB		 75652013-09-03 Richard Biener <rguenther@suse.de>
7566
7567 * tree-affine.c (add_elt_to_tree): Fix association issue,
7568 avoid useless converts and make sure to always return a
7569 properly typed result.
7570
dbfc2894
RB		 75712013-09-03 Richard Biener <rguenther@suse.de>
7572
7573 PR middle-end/57656
7574 * fold-const.c (negate_expr_p): Fix division case.
7575 (negate_expr): Likewise.
7576
fa3bf4ce
RB		 75772013-09-03 Richard Biener <rguenther@suse.de>
7578
7579 PR lto/58285
7580 * tree-streamer-out.c: Include tm.h.
7581 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7582
a64bbb3f
JH		 75832013-09-03 Jan Hubicka <jh@suse.cz>
7584
7585 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7586
93ed07e2
AM		 75872013-09-03 Alan Modra <amodra@gmail.com>
7588
7589 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7590 multilibs to big-endian target and vice versa.
7591 * config/rs6000/t-linux64: Use := assignment on all vars.
7592 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7593 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7594 * config/rs6000/t-linux64le: New file.
7595 * config/rs6000/t-linux64bele: New file.
7596 * config/rs6000/t-linux64lebe: New file.
7597
07d6cd64
JH		 75982013-09-02 Jan Hubicka <jh@suse.cz>
7599
7600 * ipa-inline-transform.c (inline_transform): Do not
7601 optimize_inline_calls when not optimizing.
7602
8e4c9a10
JH		 76032013-09-02 Jan Hubicka <jh@suse.cz>
7604
7605 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7606 duplicated nodes for assembler names.
7607 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7608 hard registers from assembler name hash.
7609
c91061e6
JH		 76102013-09-02 Jan Hubicka <jh@suse.cz>
7611
7612 * ipa-split.c (execute_split_functions): Split externally visible
7613 functions called once.
7614
2fd0985c
MJ		 76152013-09-02 Martin Jambor <mjambor@suse.cz>
7616
7617 PR ipa/58106
7618 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7619 linked list. When finding the correct duplicate, also consider also
7620 the caller in additon to its inlined_to node.
7621
91bd4114
JG		 76222013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7623
7624 * config/aarch64/aarch64-simd-builtins.def
7625 (dup_lane_scalar): Remove.
7626 * config/aarch64/aarch64-simd.md
7627 (aarch64_simd_dup): Add 'w->w' alternative.
7628 (aarch64_dup_lane<mode>): Allow for VALL.
7629 (aarch64_dup_lane_scalar<mode>): Remove.
7630 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7631 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7632 (aarch64_get_lane_unsigned<mode>): Likewise.
7633 (aarch64_get_lane<mode>): Likewise.
7634 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7635 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7636 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7637 (VCON): Change container of V2SF.
7638 (vswap_width_name): Likewise.
7639 * config/aarch64/arm_neon.h
7640 (__aarch64_vdup_lane_any): New.
7641 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7642 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7643 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7644
d617d2d8
EB		 76452013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7646
7647 PR middle-end/56382
7648 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7649 the source or the destination is a single hard register.
7650
63fdb7be
RB		 76512013-09-02 Richard Biener <rguenther@suse.de>
7652
7653 PR middle-end/57511
7654 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7655 non-linear SCEVs.
7656
78de2333
RB		 76572013-09-02 Richard Biener <rguenther@suse.de>
7658
7659 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7660 arithmetic to sizetype.
7661
85ff4ec6
BC		 76622013-09-02 Bin Cheng <bin.cheng@arm.com>
7663
7664 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7665 Find auto-increment use both before and after candidate.
7666
576016fe
MP		 76672013-09-02 Marek Polacek <polacek@redhat.com>
7668
7669 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7670
08f835dc
JH		 76712013-09-01 Jan Hubicka <jh@suse.cz>
7672
7673 * Makefile.in: Add ipa-profile.o
7674 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7675 * cgraph.c (struct cgraph_propagate_frequency_data,
7676 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7677 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7678 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB		 7679 * ipa-inline-analysis.c: Include ipa-utils.h;
7680 drop duplicated cfgloop.h.
08f835dc
JH		 7681 (inline_update_callee_summaries): Update.
7682 * ipa-profile.c: New file.
7683 * ipa-utils.h (ipa_propagate_frequency): Declare.
7684 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 7685 data-streamer.h, value-prof.h.
08f835dc
JH		 7686 (symtab_remove_unreachable_nodes): Update profile.
7687 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7688 account_time_size, cmp_counts, dump_histogram,
7689 ipa_profile_generate_summary, ipa_profile_write_summary,
7690 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7691 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 7692 Move to ipa-profile.c.
08f835dc 7693
cec8583c
JDA		 76942013-09-01 John David Anglin <danglin@gcc.gnu.org>
7695
7696 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7697
bbc9396b
JH		 76982013-09-01 Jan Hubicka <jh@suse.cz>
7699
7700 * common.opt (fdevirtualize-speculatively): New function.
7701 * invoke.texi (fdevirtualize-speculatively): Document.
7702 * ipa-devirt.c: Include ipa-inline.h
7703 (likely_target_p): New function.
7704 (ipa_devirt): New function.
7705 (gate_ipa_devirt): New function.
7706 (pass_data_ipa_devirt): New static var.
7707 (pass_ipa_devirt): Likewise.
7708 (make_pass_ipa_devirt): New function.
7709 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7710 (common_handle_option): Disable devirtualization when
7711 value range profiling is available.
7712 * passes.def (pass_ipa_devirt): Add.
7713 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 7714 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 7715
0cea211e
IS		 77162013-09-01 Iain Sandoe <iain@codesourcery.com>
7717
7718 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7719 include sanitize(undefined).
7720
5e351e96
DN		 77212013-08-31 Diego Novillo <dnovillo@google.com>
7722
7723 * Makefile.in (TREE_CORE_H): Define.
7724 (TREE_H): Use.
7725 (GTFILES): Add tree-core.h.
7726 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7727 size the array.
7728 * tree-core.h: New file.
7729 Move all data structures, enum, typedefs, global
7730 declarations and constants from ...
7731 * tree.h: ... here.
7732
63bf9a90
JH		 77332013-08-31 Jan Hubicka <jh@suse.cz>
7734
7735 * bulitins.c (expand_builtin): Do not early exit for gcov
7736 instrumented functions.
7737
5c0abd6a
MP		 77382013-08-31 Marek Polacek <polacek@redhat.com>
7739
7740 * ubsan.c: Include tm_p.h.
7741
4843f032
JH		 77422013-08-31 Jan Hubicka <jh@suse.cz>
7743
8356c89c
JH		 7744 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7745 warning.
7746
4843f032
JH		 7747 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7748 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 7749 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH		 7750 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7751 (ipa_merge_profiles): New function.
7752 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7753 (lto_input_function_body): Likewise.
5e581212
JH		 7754 * ipa-utils.h (ipa_merge_profiles): Declare.
7755 * lto-streamer.h (lto_input_function_body): Update prototype.
7756 (emit_label_in_global_context_p): Remove.
7757 * lto-symtab.c: Include ipa-utils.h
7758 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 7759
57292ce9
JH		 77602013-08-31 Jan Hubicka <jh@suse.cz>
7761
7762 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7763
b9cbfeeb
JH		 77642013-08-31 Jan Hubicka <jh@suse.cz>
7765
7766 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7767
41dedebd
UB		 77682013-08-31 Uros Bizjak <ubizjak@gmail.com>
7769
7770 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7771 "cmp" RTX before signed_comparison_operator check to account
7772 for "code" changes.
7773
06d65050
JH		 77742013-08-30 Jan Hubicka <jh@suse.cz>
7775
41dedebd 7776 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH		 7777 (detect_type_change_1): Rename to ...
7778 (detect_type_change): ... this one; early return on non-polymorphic
7779 types.
41dedebd 7780 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH		 7781 use of detect_type_change.
7782 (compute_complex_assign_jump_func): Add param_type parameter;
7783 update use of detect_type_change_ssa.
7784 (compute_complex_ancestor_jump_func): Likewise.
7785 (ipa_get_callee_param_type): New function.
7786 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7787 update calls to the jump function computation functions.
7788
600b5b1d
TJ		 77892013-08-30 Teresa Johnson <tejohnson@google.com>
7790 Steven Bosscher <steven@gcc.gnu.org>
7791
7792 * cfgrtl.c (fixup_new_cold_bb): New routine.
7793 (commit_edge_insertions): Invoke fixup_partitions.
7794 (find_partition_fixes): New routine.
7795 (fixup_partitions): Ditto.
7796 (verify_hot_cold_block_grouping): Update comments.
7797 (rtl_verify_edges): Invoke find_partition_fixes.
7798 (rtl_verify_bb_pointers): Update comments.
7799 (rtl_verify_bb_layout): Ditto.
7800 * basic-block.h (probably_never_executed_edge_p): Declare.
7801 (fixup_partitions): Ditto.
7802 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7803 * bb-reorder.c (sanitize_hot_paths): New function.
7804 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7805 sanitize_hot_paths.
7806 * predict.c (probably_never_executed_edge_p): New routine.
7807 * cfg.c (check_bb_profile): Add partition insanity warnings.
7808
0a6ea5c9
MI		 78092013-08-30 Meador Inge <meadori@codesourcery.com>
7810
7811 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7812
de5a5fa1
MP		 78132013-08-30 Marek Polacek <polacek@redhat.com>
7814
7815 * Makefile.in (ubsan.o): Add.
7816 (c-family/c-ubsan.o): Add.
7817 (builtins.o): Add ubsan.h dependency.
7818 * ubsan.h: New file.
7819 * ubsan.c: New file.
7820 * common.opt: Add -fsanitize=undefined option.
7821 (flag_sanitize): Add variable.
7822 (fsanitize=): Add option. Add Driver.
7823 (fsanitize=thread): Remove option.
7824 (fsanitize=address): Likewise.
7825 (static-libubsan): New option.
7826 * doc/invoke.texi: Document the new flag and -static-libubsan.
7827 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7828 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7829 * builtin-attrs.def (ATTR_COLD): Define.
7830 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7831 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7832 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7833 * flag-types.h (sanitize_code): New enum.
7834 * opts.c (common_handle_option): Parse command line arguments
7835 of -fsanitize=. Add -fsanitize=unreachable option.
7836 * varasm.c (get_variable_section): Adjust.
7837 (assemble_noswitch_variable): Likewise.
7838 (assemble_variable): Likewise.
7839 (output_constant_def_contents): Likewise.
7840 (categorize_decl_for_section): Likewise.
7841 (place_block_symbol): Likewise.
7842 (output_object_block): Likewise.
7843 * builtins.def: Likewise.
7844 * toplev.c (compile_file): Likewise.
7845 (process_options): Likewise.
7846 * cppbuiltin.c: Likewise.
7847 * tsan.c (tsan_pass): Likewise.
7848 (tsan_gate): Likewise.
7849 (tsan_gate_O0): Likewise.
7850 * cfgexpand.c (partition_stack_vars): Likewise.
7851 (expand_stack_vars): Likewise.
7852 (defer_stack_allocation): Likewise.
7853 (expand_used_vars): Likewise.
7854 * cfgcleanup.c (old_insns_match_p): Likewise.
7855 * asan.c (asan_finish_file): Likewise.
7856 (asan_instrument): Likewise.
7857 (gate_asan): Likewise.
7858 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7859 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7860 (asan_global_struct): Use pointer_sized_int_node instead
7861 calling build_nonstandard_integer_type.
7862 (initialize_sanitizer_builtins): Likewise.
7863 (asan_finish_file): Likewise.
7864 * gcc.c: Document %{%:function(args):X}.
7865 (static_spec_functions): Add sanitize.
7866 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7867 store funcval != NULL there.
7868 (do_spec_1): Adjust handle_spec_function caller.
7869 (handle_braces): Allow %:function(args) as condition.
7870 (sanitize_spec_function): New function.
7871 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7872 (LIBUBSAN_SPEC): Likewise.
7873 (LIBUBSAN_EARLY_SPEC): Likewise.
7874 (SANITIZER_SPEC): Handle libubsan.
7875 (SANITIZER_EARLY_SPEC): Likewise.
7876 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7877 instead of fsanitize=address.
7878 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7879 instead of fsanitize=address*.
7880 * builtins.c: Include ubsan.h.
7881 (fold_builtin_0): Instrument __builtin_unreachable.
7882 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7883 instead of flag_asan.
7884 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7885 (pointer_sized_int_node): Define.
41dedebd 7886 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 7887
f07f30cf
MS		 78882013-08-30 Mike Stump <mikestump@comcast.net>
7889
7890 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7891 with RE patterns.
7892
8a41354f
JH		 78932013-08-29 Jan Hubicka <jh@suse.cz>
7894
7895 * cgraph.c (cgraph_function_body_availability): Handle weakref
7896 correctly.
7897 * passes.def: Remove pass_fixup_cfg.
7898 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7899 track when we need to remove functions.
7900 (gate_ipa_inline): Execute inlining always; add comment why.
7901 (pass_data_ipa_inline): Remove TODO_remove_functions.
7902 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7903 do not produce summaries.
7904 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7905 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7906 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7907 constant pool and vtable.
7908
dc8d7a0f
TB		 79092013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7910
41dedebd
UB		 7911 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7912 New arm_neon.h's internal macros to specify 64-bit constants.
7913 Avoid using stdint.h's macros.
dc8d7a0f 7914
ac7eacd2
JR		 79152013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7916
7917 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7918 before applying REG_P.
7919
8b29fd4e
JJ		 79202013-08-30 Jakub Jelinek <jakub@redhat.com>
7921
7922 PR tree-optimization/58277
7923 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7924 after seeing too many stmts with vdef in between dombb and current
7925 bb, invalidate everything.
7926
26d75703
RB		 79272013-08-30 Richard Biener <rguenther@suse.de>
7928
7929 * fold-const.c (fold_single_bit_test): Fix overflow test.
7930
f7e088e7
EB		 79312013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7932
7933 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7934 and which can live in a register, always retrieve the value on entry.
7935 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7936 passed by invisible reference specially.
7937 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7938 (vt_add_function_parameter): Correctly deal with a parameter passed by
7939 invisible reference.
7940
79412013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH		 7942
7943 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7944
6e8dad05
RB		 79452013-08-30 Richard Biener <rguenther@suse.de>
7946
7947 PR tree-optimization/58228
7948 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7949 allow invariant loads in nested loop vectorization.
7950
062ef2c8
RB		 79512013-08-30 Richard Biener <rguenther@suse.de>
7952
7953 PR tree-optimization/58223
7954 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7955 (has_anti_or_output_dependence): ... this and adjust to also
7956 look for output dependences.
7957 (mark_nodes_having_upstream_mem_writes): Adjust.
7958 (rdg_flag_uses): Likewise.
7959
7a764c60
RB		 79602013-08-30 Richard Biener <rguenther@suse.de>
7961
7962 PR tree-optimization/58010
7963 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7964 assert that we have a loop-closed PHI.
7965
82e9d642
JH		 79662013-08-29 Jan Hubicka <jh@suse.cz>
7967
7968 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7969 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7970 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7971 (lto_free_function_in_decl_state_for_node): New function.
7972
9cc1fb4b
XDL		 79732013-08-29 Xinliang David Li <davidxl@google.com>
7974
41dedebd 7975 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL		 7976 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7977 Emit alignment peeling message with default -fopt-info.
7978 (vect_loop_versioning): Emit loop version info message.
41dedebd 7979 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL		 7980 (execute_vect_slp): Ditto.
7981
440a5082
EB		 79822013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7983
7984 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7985 of the clone from the DECL_NAME of the original function.
7986
3fa3690d
OE		 79872013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7988
7989 * passes.c (register_pass): Add overload.
7990 * tree-pass.h (register_pass): Forward declare it. Add comment.
7991
0170f33c
JH		 79922013-08-29 Jan Hubicka <jh@suse.cz>
7993
41dedebd
UB		 7994 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7995 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7996 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7997 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 7998 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7999 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 8000 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH		 8001 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8002
103ff0d6
TJ		 80032013-08-29 Teresa Johnson <tejohnson@google.com>
8004
8005 * dumpfile.c (dump_loc): Output column number.
8006 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8007 * doc/invoke.texi: Document optall -fopt-info flag.
8008 * profile.c (read_profile_edge_counts): Use new dump framework.
8009 (compute_branch_probabilities): Ditto.
8010 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8011 when pass not in any opt group.
8012 * pass_manager.h (pass_manager::get_pass_profile): New method.
8013 * value-prof.c (check_counter): Use new dump framework.
8014 (check_ic_target): Ditto.
8015 * coverage.c (get_coverage_counts): Ditto.
8016 (coverage_init): Setup new dump framework.
8017
301bbc16
RB		 80182013-08-29 Richard Biener <rguenther@suse.de>
8019
8020 PR tree-optimization/58246
8021 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8022 handle the dominance check inside a basic-block.
8023
1b275000
RB		 80242013-08-29 Richard Biener <rguenther@suse.de>
8025
8026 PR middle-end/57287
8027 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8028 of default defs that appear in abnormal PHI nodes.
8029
6b1184ba
RB		 80302013-08-29 Richard Biener <rguenther@suse.de>
8031
8032 PR tree-optimization/57685
8033 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8034 single-use operands to avoid exponential complexity.
8035
1b08b734
DC		 80362013-08-28 Dehao Chen <dehao@google.com>
8037
8038 * ipa-inline.c (edge_badness): Fix integer underflow.
8039
48a3fa69
UB		 80402013-08-28 Uros Bizjak <ubizjak@gmail.com>
8041
8042 * gtm-builtins.def (_ITM_free): Declare leaf.
8043
215f73e6
JJ		 80442013-08-28 Jakub Jelinek <jakub@redhat.com>
8045
8046 PR target/58067
8047 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8048 (*tls_local_dynamic_base_64_largepic): Likewise.
8049 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8050 Remove predicate from call operand.
8051 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8052 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8053
34554d1a
JL		 80542013-08-28 Jeff Law <law@redhat.com>
8055
8056 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8057 checks for the number of predecessors and successors allowed.
8058 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8059 which require copying a joiner block if there is a request which
8060 is a subpath that requires no joiner block copying.
8061
c01c111b
JH		 80622013-08-28 Jan Hubicka <jh@suse.cz>
8063
8064 * lto-streamer-out.c (DFS_write_tree_body): Drop
8065 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8066 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8067 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8068 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8069 Do not read DECL_ERROR_ISSUED.
8070 (unpack_ts_decl_with_vis_value_fields): Do not read
8071 DECL_DEFER_OUTPUT.
48a3fa69
UB		 8072 (lto_input_ts_binfo_tree_pointers): Do not read
8073 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH		 8074 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8075 write DECL_ERROR_ISSUED..
8076 (pack_ts_decl_with_vis_value_fields): Do not write
8077 DECL_DEFER_OUTPUT.
8078 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 8079 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH		 8080 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8081 * tree.h (tree_decl_common): Update comment.
8082 (DECL_ERROR_ISSUED): Remove.
8083
4b128ece
JJ		 80842013-08-28 Jakub Jelinek <jakub@redhat.com>
8085
8086 PR middle-end/58257
8087 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8088
12211b99 80892013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH		 8090
8091 * builtins.def (free): Declare leaf.
8092
8dce4dbc
DM		 80932013-08-27 David Malcolm <dmalcolm@redhat.com>
8094
8095 * gdbhooks.py: New.
8096 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8097 * configure: Regenerate.
8098
b8f6e610
MJ		 80992013-08-27 Martin Jambor <mjambor@suse.cz>
8100
8101 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8102 (ipa_ancestor_jf_data): Likewise.
8103 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8104 (ipa_get_jf_pass_through_type_preserved): New function.
8105 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8106 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 8107 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ		 8108 (ipa_get_jf_ancestor_result): Likewise.
8109 (propagate_vals_accross_pass_through): Use
8110 ipa_get_jf_pass_through_result to do all the value mappings.
8111 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8112 type_preserved flag.
8113 (ipa_set_jf_cst_copy): New function.
8114 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8115 (ipa_set_jf_arith_pass_through): Likewise.
8116 (ipa_set_ancestor_jf): Likewise.
8117 (compute_complex_assign_jump_func): Set type_preserved instead of
8118 punting.
8119 (ipa_compute_jump_functions_for_edge): Likewise.
8120 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8121 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 8122 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ		 8123 (ipa_write_jump_function): Stream the type_preserved flags.
8124 (ipa_read_jump_function): Likewise.
8125
74bf76ed
JJ		 81262013-08-27 Jakub Jelinek <jakub@redhat.com>
8127 Aldy Hernandez <aldyh@redhat.com>
8128
8129 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8130 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8131 * function.h (struct function): Add has_force_vect_loops and
8132 has_simduid_loops.
12211b99 8133 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ		 8134 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8135 handle it.
8136 * gimple.def: Update CLAUSES comments.
8137 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8138 (gimple_build_omp_for): Add argument to prototype.
8139 (gimple_omp_for_kind): New.
8140 (gimple_omp_for_set_kind): New.
8141 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8142 GOVD_DATA_SHARE_CLASS.
8143 (enum omp_region_type): Add ORT_SIMD.
8144 (gimple_add_tmp_var): Handle ORT_SIMD.
8145 (gimplify_var_or_parm_decl): Same.
8146 (is_gimple_stmt): Same.
8147 (omp_firstprivatize_variable): Same.
8148 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8149 (omp_notice_variable): Handle ORT_SIMD.
12211b99 8150 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ		 8151 (omp_check_private): Handle ORT_SIMD.
8152 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8153 OMP_CLAUSE_SAFELEN.
8154 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8155 Handle OMP_CLAUSE_LASTPRIVATE.
8156 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8157 OMP_CLAUSE_SAFELEN.
8158 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8159 (gimplify_expr): Handle OMP_SIMD.
8160 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8161 (expand_GOMP_SIMD_VF): New.
8162 (expand_GOMP_SIMD_LAST_LANE): New.
8163 * internal-fn.def (GOMP_SIMD_LANE): New.
8164 (GOMP_SIMD_VF): New.
8165 (GOMP_SIMD_LAST_LANE): New.
8166 * omp-low.c: Include target.h.
8167 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8168 OMP_CLAUSE_SAFELEN.
8169 (check_omp_nesting_restrictions): Same.
8170 (omp_max_vf): New.
8171 (lower_rec_simd_input_clauses): New.
8172 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8173 OMP_CLAUSE_LINEAR.
8174 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8175 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8176 (expand_omp_build_assign): New.
8177 (expand_omp_for_init_counts): New.
8178 (expand_omp_for_init_vars): New.
8179 (extract_omp_for_update_vars): New.
8180 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8181 and rewrite accordingly.
8182 (expand_omp_simd): New.
8183 (expand_omp_for): Use expand_omp_simd.
8184 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8185 (lower_omp_for): Do not lower the body.
8186 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8187 in their own loops.
8188 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB		 8189 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8190 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ		 8191 (gate_tree_if_conversion): Similarly.
8192 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8193 gimple_build_omp_for.
8194 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8195 * tree-parloops (create_parallel_loop): Pass kind argument to
8196 gimple_build_omp_for.
8197 * tree-pretty-print.c (dump_omp_clause): Add cases for
8198 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8199 OMP_CLAUSE__SIMDUID_.
8200 (dump_generic_node): Handle OMP_SIMD.
8201 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8202 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8203 unroll OMP_SIMD loops here.
12211b99 8204 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 8205 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 8206 loop->safelen.
74bf76ed
JJ		 8207 (vect_analyze_data_refs): Handle simd loops.
8208 * tree-vect-loop.c (vectorizable_live_operation): Handle
8209 IFN_GOMP_SIMD*.
12211b99 8210 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ		 8211 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8212 (vectorizable_load): Same.
12211b99 8213 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ		 8214 (struct simduid_to_vf): New.
8215 (simduid_to_vf::hash): New.
8216 (simduid_to-vf::equal): New.
8217 (struct simd_array_to_simduid): New.
8218 (simd_array_to_simduid::hash): New.
8219 (simd_array_to_simduid::equal): New.
8220 (adjust_simduid_builtins): New.
8221 (struct note_simd_array_uses_struct): New.
8222 (note_simd_array_uses_cb): New.
8223 (note_simd_array_uses): New.
8224 (vectorize_loops): Handle simd hints and adjust simd builtins
8225 accordingly.
8226 * tree-vectorizer.h (struct _stmt_vec_info): Add
8227 simd_lane_access_p field.
8228 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8229 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8230 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8231 (omp_clause_code_name): Same.
8232 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8233 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8234 * tree.def (OMP_SIMD): New entry.
12211b99
UB		 8235 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8236 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ		 8237 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8238 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8239 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8240 (OMP_CLAUSE_LINEAR_STEP): New.
8241 (OMP_CLAUSE_SAFELEN_EXPR): New.
8242 (OMP_CLAUSE__SIMDUID__DECL): New.
8243 (find_omp_clause): New prototype.
8244
d0cf4e84
L		 82452013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8246
8247 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8248 Haswell processor detection.
8249
992592ec
CW		 82502013-08-27 Christian Widmer <shadow@umbrox.de>
8251
8252 PR target/57927
8253 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8254 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8255 AVX2 capable processors.
8256
9116eb22
TB		 82572013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8258
8259 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8260 of vget_low_* with implementations in terms of other intrinsics.
8261
1c5abb53
MG		 82622013-08-27 Marc Glisse <marc.glisse@inria.fr>
8263
8264 PR middle-end/57219
8265 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8266 values to -1, 0 and 1.
8267
2e100703
VP		 82682013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8269
8270 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8271 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8272 (<optab><mode>3_insn): Remove.
8273 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8274 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8275 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8276 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8277 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8278 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8279 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8280 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8281 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8282 (ror<mode>3_insn): Likewise.
8283 * config/aarch64/predicates.md (aarch64_simd_register): New.
8284
4ded8276
RB		 82852013-08-27 Richard Biener <rguenther@suse.de>
8286
8287 PR tree-optimization/57521
8288 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8289 one edge is non-critical.
8290 (find_phi_replacement_condition): Make sure to use a non-critical
8291 edge. Cleanup and remove old bug workarounds.
8292 (bb_postdominates_preds): Remove.
8293 (if_convertible_loop_p_1): Do not compute post-dominators.
8294 (combine_blocks): Do not free post-dominators.
8295 (main_tree_if_conversion): Likewise.
8296 (pass_data_if_conversion): Add TODO_verify_ssa.
8297
5aa11061
DD		 82982013-08-27 DJ Delorie <dj@redhat.com>
8299
8300 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8301
fd91cfe3
YZ		 83022013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8303
8304 * function.c (assign_parm_find_data_types): Set passed_mode and
8305 nominal_mode to the TYPE_MODE of nominal_type for the built
8306 pointer type in case of the struct-pass-by-reference.
8307
907555ce
JR		 83082013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8309
8310 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8311 (UINT16_TYPE): Change default to "unsigned int".
8312
67518c93
JR		 8313 * config/avr/avr.opt (mfract-convert-truncate): New option.
8314 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8315 is set, round negative fractional integers according to n1169
8316 when converting to integer types.
8317
d6d989fc
JH		 83182013-08-26 Jan Hubicka <jh@suse.cz>
8319
8320 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8321 methods can not be called indirectly when their address is not taken.
8322
2aa3da06
JH		 83232013-08-26 Jan Hubicka <jh@suse.cz>
8324
12211b99
UB		 8325 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8326 ctor_for_folding.
2aa3da06 8327
0987ffe7
JH		 83282013-08-26 Jan Hubicka <jh@suse.cz>
8329
8330 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8331 can be unshared.
8332
befe8647
JR		 83332013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8334
8335 * reload.c (find_valid_class): Allow classes that do not include
8336 FIRST_PSEUDO_REGISTER - 1.
8337
a21e735e
JH		 83382013-08-26 Jan Hubicka <jh@suse.cz>
8339
8340 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8341 fix edge count/frequency when speculation failed; fix type check
8342 for the direct call.
8343
e067bd43
JH		 83442013-08-26 Jan Hubicka <jh@suse.cz>
8345
8346 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8347
e8aec975
JH		 83482013-08-26 Jan Hubicka <jh@suse.cz>
8349
8350 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8351 to be changed by edge redirection.
8352
d0b66480
JH		 83532013-08-26 Jan Hubicka <jh@suse.cz>
8354
12211b99
UB		 8355 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8356 formating; add sanity check.
d0b66480
JH		 8357 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8358 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8359 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8360 (resolve_noninline_speculation): Update callee keys, too.
8361
0f9aaac7
JH		 83622013-08-26 Jan Hubicka <jh@suse.cz>
8363
8364 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8365 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8366
4b37444e
JR		 83672013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8368
8369 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8370 into proper place.
8371
5339fc0c
UB		 83722013-08-26 Uros Bizjak <ubizjak@gmail.com>
8373
8374 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8375 (x86_64_elf_select_section): Ditto.
8376 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8377 arguments.
8378 (ix86_pass_by_reference): Ditto.
5339fc0c
UB		 8379 (output_set_got): Ditto.
8380 (ix86_unary_operator_ok): Ditto.
8381 (ix86_expand_builtin): Ditto.
8382
d5c3d3ef
JH		 83832013-08-23 Jan Hubicka <jh@suse.cz>
8384
5339fc0c 8385 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 8386
5b1e7435
JH		 83872013-08-23 Jan Hubicka <jh@suse.cz>
8388
8389 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8390 (tree_decl_with_vis): Add FINAL field.
8391
ead69dac
JL		 83922013-08-23 Jeff Law <law@redhat.com>
8393
8394 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8395 the debugging dump when the expression is fully redundant.
8396
025311c4
GDR		 83972013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8398
8399 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8400 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8401 * pretty-print.c (pp_formatted_text_data): Likewise.
8402 (pp_write_text_to_stream): Likewise.
8403 (pp_write_text_as_dot_label_to_stream): Likewise.
8404 (pp_append_r): Likewise.
8405 (pp_format): Likewise.
8406 (pp_flush): Likewise.
8407 (pp_clear_output_area): Likewise.
8408 (pp_append_text): Likewise.
8409 (pp_formatted_text): Likewise.
8410 (pp_remaining_character_count_for_line): Likewise.
8411 (pp_newline): Likewise.
8412 (pp_character): Likewise.
8413 (output_buffer::~output_buffer): Define.
8414 (pretty_printer::~pretty_printer): Destruct output buffer.
8415 * pretty-print.h (output_buffer::~output_buffer): Declare.
8416 (pretty_printer::~pretty_printer): Declare virtual.
8417
0cadbfaa
MG		 84182013-08-24 Marc Glisse <marc.glisse@inria.fr>
8419
8420 PR other/57324
8421 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8422 HOST_WIDE_INT_M1U): New macros.
8423 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8424 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8425 unsigned -1 for lshift.
8426 * cse.c (cse_insn): Likewise.
8427 * double-int.c (rshift_double, lshift_double): Likewise.
8428 * builtins.c (fold_builtin_bitop): Likewise.
8429 * combine.c (force_to_mode): Likewise.
8430 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8431 * simplify-rtx.c (simplify_const_unary_operation,
8432 simplify_const_binary_operation): Likewise.
8433 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8434 check_va_list_escapes): Likewise.
8435 * rtlanal.c (nonzero_bits1): Likewise.
8436 * expmed.c (expand_smod_pow2): Likewise.
8437 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8438
009e5353
JH		 84392013-08-23 Jan Hubicka <jh@suse.cz>
8440
8441 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8442 as having address taken.
8443
c4be6568
JH		 84442013-08-23 Jan Hubicka <jh@suse.cz>
8445
64cbf23d
JH		 8446 * ipa-utils.h (method_class_type): Declare.
8447 * ipa-devirt.c (method_class_type): Export.
8448
c4be6568
JH		 8449 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8450 do not walk base classes of anonymous types.
8451
97aba8e9
KK		 84522013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8453
8454 PR rtl-optimization/58220
8455 PR regression/58221
8456 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8457 handle SEQUENCE insns properly.
8458
0fc80001
GDR		 84592013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8460
8461 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8462 definition.
8463 (pp_newline_and_indent): Likewise.
8464 (pp_separate_with): Likewise.
8465 * pretty-print.c (pp_newline_and_flush): Define.
8466 (pp_newline_and_indent): Likewise.
8467 (pp_separate_with): Likewise.
8468
520a5868
JJ		 84692013-08-23 Jakub Jelinek <jakub@redhat.com>
8470
8471 PR target/58218
8472 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8473 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8474
e500c62a
KY		 84752013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8476
f43245d1
UB		 8477 * config/i386/predicates.md (ext_sse_reg_operand): New.
8478 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY		 8479 predicate to determine if EVEX is needed.
8480 (*movsi_internal): Ditto.
8481 (*movdf_internal): Ditto.
8482 (*movsf_internal): Ditto.
f43245d1 8483 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 8484
6b00d7dd
JJ		 84852013-08-23 Jakub Jelinek <jakub@redhat.com>
8486
8487 PR tree-optimization/58209
8488 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8489 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8490 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8491 emit POINTER_PLUS_EXPR.
8492 (create_tailcall_accumulator): For pointer result type accumulate in
8493 sizetype type.
8494
4f2a9f90
PC		 84952013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8496
8497 * configure.ac: Add backslashes missing from the last change.
8498 * configure: Regenerate.
8499
7bf4274e
JH		 85002013-08-22 Jan Hubicka <jh@susue.cz>
8501
4f2a9f90
PC		 8502 * ipa.c (function_and_variable_visibility): First remember function
8503 was global and then make it local.
7bf4274e 8504
1bea243d
JB		 85052013-08-22 Julian Brown <julian@codesourcery.com>
8506
8507 * configure.ac: Add aarch64 to list of arches which use "nop" in
8508 debug_line test.
8509 * configure: Regenerate.
8510
bcba7ecc
AK		 85112013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8512
8513 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8514 gnu_libc_has_function.
8515 * config/s390/tpf.h: Likewise.
8516
4c495b0d
JH		 85172013-08-22 Jan Hubicka <jh@susue.cz>
8518
8519 * timevar.c (validate_phases): Add cast.
8520
5e302bca
JH		 85212013-08-22 Jan Hubicka <jh@susue.cz>
8522
8523 * timevar.c (validate_phases): Use size_t for memory.
8524 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8525
da6ca2b5
GDR		 85262013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8527
8528 * pretty-print.h (output_buffer::output_buffer): Declare.
8529 (pretty_printer::pretty_printer): Likewise.
8530 (pp_construct): Remove.
8531 * pretty-print.c (output_buffer::output_buffer): Define.
8532 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8533 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8534 (print_gimple_expr): Likewise.
8535 (print_gimple_seq): Likewise.
8536 (gimple_dump_bb): Likewise.
8537 * sched-vis.c (dump_value_slim): Likewise.
8538 (dump_insn_slim): Likewise.
8539 (dump_rtl_slim): Likewise.
8540 (str_pattern_slim): Likewise.
8541 * tree-mudflap.c (mf_varname_tree): Likewise.
8542 * graph.c (print_graph_cfg): Likewise.
8543 (start_graph_dump): Likewise.
8544 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8545 placement-new.
8546 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8547 pretty printer initialization.
8548 * coretypes.h (diagnostic_context): Remove superflous type alias
8549 declaration.
8550 (pretty_printer): Likewise. Declare directly as a class.
8551 (pretty_print_info): Remove declaration as class.
8552 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8553 and pp_clear_output_area.
8554 (asan_add_global): Likewise.
8555
0e1474e5
JH		 85562013-08-22 Jan Hubicka <jh@suse.cz>
8557
8558 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8559 * ipa-utils.h (update_type_inheritance_graph): Declare.
8560 (possible_polymorphic_call_target_p): Declare.
8561 (possible_polymorphic_call_target_p): New.
8562 * ipa-devirt.c: Update toplevel comments.
8563 (cached_polymorphic_call_targets): Move up.
8564 (odr_type_d): Move ID down.
8565 (polymorphic_type_binfo_p): Update comment.
8566 (odr_hasher::remove): Likewise;
8567 (get_odr_type): Set anonymous_namespace.
8568 (dump_odr_type): Dump it.
8569 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8570 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8571 (record_binfo): Add comment.
5339fc0c
UB		 8572 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8573 built.
0e1474e5
JH		 8574 (devirt_node_removal_hook): Do not iCE when cache is freed.
8575 (possible_polymorphic_call_target_p): New predicate.
8576 (update_type_inheritance_graph): New function.
8577
3f97cb0b
AI		 85782013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8579 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8580 Sergey Lega <sergey.s.lega@intel.com>
8581 Anna Tikhonova <anna.tikhonova@intel.com>
8582 Ilya Tocar <ilya.tocar@intel.com>
8583 Andrey Turetskiy <andrey.turetskiy@intel.com>
8584 Ilya Verbin <ilya.verbin@intel.com>
8585 Kirill Yukhin <kirill.yukhin@intel.com>
8586 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8587
8588 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8589 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8590 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8591 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8592 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8593 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8594 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8595 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8596 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8597 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8598 OPT_mavx512pf, OPT_mavx512er cases.
8599 * config/i386/constraints.md (v): New constraint.
8600 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8601 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8602 (bit_AVX512CD): New.
8603 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8604 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8605 * config/i386/i386-c.c (ix86_target_macros_internal):
8606 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8607 __AVX512PF__.
8608 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8609 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8610 * config/i386/i386.c (regclass_map, dbx_register_map)
8611 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8612 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8613 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8614 -mavx512pf options.
8615 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8616 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8617 -mavx512pf options. Fix formatting.
8618 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8619 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8620 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8621 -mavx512cd, -mavx512pf options.
8622 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8623 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8624 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8625 ALL_SSE_REGS.
8626 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8627 (ix86_set_reg_reg_cost): Ditto.
8628 (x86_order_regs_for_local_alloc): Ditto.
8629 (MAX_VECT_LEN): Extend to 64-byte.
8630 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8631 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8632 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8633 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8634 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8635 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8636 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8637 (SSE_REG_MODE_P): Support new modes.
8638 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8639 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8640 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8641 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8642 (REG_CLASS_CONTENTS): Add new registers.
8643 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8644 (EXT_REX_SSE_REGNO_P): New.
8645 (HI_REGISTER_NAMES): Add new registers.
8646 * config/i386/i386.md: Define constants for new registers.
8647 (mode): Add new 512-bit modes.
8648 (prefix): Support evex prefix.
8649 (isa): Support avx512f, noavx512f, fma_avx512f.
8650 (ssemodesuffix): Add new 512-bit modes.
8651 (movxi): New.
8652 (*movxi_internal_avx512f): Ditto.
8653 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8654 Support MODE_XI.
8655 (*movsi_internal): Likewise.
8656 (*movdf_internal): Likewise.
8657 (*movsf_internal): Likewise.
8658 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8659 (<code><mode>3): Likewise.
5339fc0c
UB		 8660 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8661 New.
3f97cb0b
AI		 8662 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8663 with the new constraint "v".
8664 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8665 modes.
8666 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8667 with the new constraint "v".
8668 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8669 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8670 (*<plusminus_insn><mode>3): Likewise.
8671 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8672 (*mul<mode>3): Likewise.
8673 (<sse>_vmmul<mode>3): Likewise.
8674 (<sse>_div<mode>3): Likewise.
8675 (<sse>_vmdiv<mode>3): Likewise.
8676 (<sse>_sqrt<mode>2): Likewise.
8677 (<sse>_vmsqrt<mode>2): Likewise.
8678 (*<code><mode>3_finite): Likewise.
8679 (*<code><mode>3) <smaxmin>: Likewise.
8680 (<sse>_vm<code><mode>3): Likewise.
8681 (*<code><mode>3) <any_logic>: Likewise.
8682 (*fma_fmadd_<mode>): Likewise.
8683 (*fma_fmsub_<mode>): Likewise.
8684 (*fma_fnmadd_<mode>): Likewise.
8685 (*fma_fnmsub_<mode>): Likewise.
8686 (*fma_fmaddsub_<mode>): Likewise.
8687 (*fma_fmsubadd_<mode>): Likewise.
8688 (*fmai_fmadd_<mode>): Likewise.
8689 (*fmai_fmsub_<mode>): Likewise.
8690 (*fmai_fnmadd_<mode>): Likewise.
8691 (*fmai_fnmsub_<mode>): Likewise.
8692 (sse_cvtsi2ss): Likewise.
8693 (sse_cvtsi2ssq): Likewise.
8694 (sse_cvtss2si): Likewise.
8695 (sse_cvtss2si_2): Likewise.
8696 (sse_cvtss2siq): Likewise.
8697 (sse_cvtss2siq_2): Likewise.
8698 (sse_cvttss2si): Likewise.
8699 (sse_cvtss2siq_2): Likewise.
8700 (float<sseintvecmodelower><mode>2): Likewise.
8701 (sse2_cvtsd2si_2): Likewise.
8702 (sse2_cvtsd2siq_2): Likewise.
8703 (*<plusminus_insn><mode>3): Likewise.
8704 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8705 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8706 (ashr<mode>3): Likewise.
8707 (<shift_insn><mode>3): Likewise.
8708 (avx2_<code><mode>3): Likewise.
8709 (*avx2_<code><mode>3): Likewise.
8710 (*andnot<mode>3): Likewise.
8711 (*<code><mode>3) <any_logic>: Likewise.
8712 (abs<mode>2): Likewise.
8713 (avx2_permvar<mode>): Likewise.
8714 (avx2_perm<mode>_1): Likewise.
8715 (*avx_vpermilp<mode>): Likewise.
8716 (avx_vpermilvar<mode>3): Likewise.
8717 (avx2_ashrv<mode>): Likewise.
8718 (avx2_<shift_insn>v<mode>): Likewise.
8719 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8720 -mavx512cd.
8721 * doc/rtl.texi: Document XImode.
8722
dd5e8423
JL		 87232013-08-21 Jeff Law <law@redhat.com>
8724
b9ebee5d
JL		 8725 * tree-flow.h (register_jump_thread): Pass vector of edges
8726 instead of each important edge.
8727 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8728 thread path into a vector and pass that to register_jump_thread.
8729 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8730 passed in edge vector to the current 3-edge form.
8731
dd5e8423
JL		 8732 Revert:
8733 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8734
9bb6628e 8735 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL		 8736 delete_unmarked_insns removed anything.
8737
445dc8df
JR		 87382013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8739
8740 * reload.h (struct reg_equivs): Rename to ..
8741 (struct reg_equivs_s): .. this.
8742
e0df53dd
ML		 87432013-08-20 Martin Liska <marxin.liska@gmail.com>
8744
8745 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8746
e7d1d3eb
RO		 87472013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8748
5339fc0c 8749 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 8750
5562e26e
JL		 87512013-08-21 Jeff Law <law@redhat.com>
8752
8753 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8754 simplify assignments too. If the RHS collapses to a singleton
8755 range, then return the value for the range.
8756
c7ecdec6
KY		 87572013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8758
7a5911d3
UB		 8759 * config/i386/sse.md (V16): Rename to...
8760 (VMOVE): this.
8761 (mov<mode>): Update iterator name.
8762 (*mov<mode>_internal): Ditto.
8763 (push<mode>1): Ditto.
8764 (movmisalign<mode>): Ditto.
c7ecdec6 8765
bfa3b50a
JH		 87662013-08-20 Jan Hubicka <jh@suse.cz>
8767
8768 PR bootstrap/58186
8769 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8770 entry for direct edges.
8771 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8772
6cd4d135
DM		 87732013-08-20 David Malcolm <dmalcolm@redhat.com>
8774
8775 Revert my last two changes, r201865 and r201864:
8776
8777 Revert r201865:
8778 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8779
8780 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8781 instances can own GC refs.
8782
8783 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8784 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8785 (gcc::context::gt_pch_nx): Likewise.
8786 (gcc::context::gt_pch_nx): Likewise.
8787 * ggc.h (gt_ggc_mx <T>): New.
8788 (gt_pch_nx_with_op <T>): New.
8789 (gt_pch_nx <T>): New.
8790 * passes.c (opt_pass::gt_ggc_mx): New.
8791 (opt_pass::gt_pch_nx): New.
8792 (opt_pass::gt_pch_nx_with_op): New.
8793 (pass_manager::gt_ggc_mx): New.
8794 (pass_manager::gt_pch_nx): New.
8795 (pass_manager::gt_pch_nx_with_op): New.
8796 (pass_manager::operator new): Use
8797 ggc_internal_cleared_alloc_stat rather than xcalloc.
8798 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8799 (pass_manager::gt_ggc_mx): New.
8800 (pass_manager::gt_pch_nx): New.
8801 (pass_manager::gt_pch_nx_with_op): New.
8802 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8803 (opt_pass::operator new): New.
8804 (opt_pass::gt_ggc_mx): New.
8805 (opt_pass::gt_pch_nx): New.
8806 (opt_pass::gt_pch_nx_with_op): New.
8807
8808 Revert r201864:
8809 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8810
8811 * Makefile.in (GTFILES): Add context.h.
8812 * context.c (gcc::context::operator new): New.
8813 (gcc::context::gt_ggc_mx): New.
8814 (gcc::context::gt_pch_nx): New.
8815 (gcc::context::gt_pch_nx): New.
8816 * context.h (gcc::context): Add GTY((user)) marking.
8817 (gcc::context::operator new): New.
8818 (gcc::context::gt_ggc_mx): New.
8819 (gcc::context::gt_pch_nx): New.
8820 (gcc::context::gt_pch_nx): New.
8821 (g): Add GTY marking.
8822 (gt_ggc_mx (gcc::context *)): New.
8823 (gt_pch_nx (gcc::context *)): New.
8824 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8825 void *cookie)): New.
8826 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8827
a046136a
AM		 88282013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8829
8830 * dce.c (fini_dce): Call df_analyze again just in case
8831 delete_unmarked_insns removed anything.
8832
67598720
TJ		 88332013-08-20 Teresa Johnson <tejohnson@google.com>
8834
8835 PR rtl-optimizations/57451
8836 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8837 from crossing split section boundaries.
8838
f8693faf
MGD		 88392013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8840
8841 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8842 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8843 with MULTLIB_DEFAULTS.
8844
7d0b9a9c
NC		 88452013-08-20 Nick Clifton <nickc@redhat.com>
8846
8847 * target.def (narrow_volatile_bitfield): Note that the default
8848 value is false, not !TARGET_STRICT_ALIGN.
8849 * doc/tm.texi: Regenerate.
8850
8edb8dc8
PC		 88512013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8852
8853 Fix LIB_SPEC for systems without libpthread.
8854
8855 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8856 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8857 for Android.
8858 * config/i386/linux-common.h: Likewise.
8859 * config/mips/linux-common.h: Likewise.
8860
12211b99 88612013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ		 8862
8863 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8864 checks.
8865
e42c64cb
DM		 88662013-08-20 David Malcolm <dmalcolm@redhat.com>
8867
8868 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8869 instances can own GC refs.
8870
8871 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8872 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8873 (gcc::context::gt_pch_nx): Likewise.
8874 (gcc::context::gt_pch_nx): Likewise.
8875 * ggc.h (gt_ggc_mx <T>): New.
8876 (gt_pch_nx_with_op <T>): New.
8877 (gt_pch_nx <T>): New.
8878 * passes.c (opt_pass::gt_ggc_mx): New.
8879 (opt_pass::gt_pch_nx): New.
8880 (opt_pass::gt_pch_nx_with_op): New.
8881 (pass_manager::gt_ggc_mx): New.
8882 (pass_manager::gt_pch_nx): New.
8883 (pass_manager::gt_pch_nx_with_op): New.
8884 (pass_manager::operator new): Use
8885 ggc_internal_cleared_alloc_stat rather than xcalloc.
8886 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8887 (pass_manager::gt_ggc_mx): New.
8888 (pass_manager::gt_pch_nx): New.
8889 (pass_manager::gt_pch_nx_with_op): New.
8890 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8891 (opt_pass::operator new): New.
8892 (opt_pass::gt_ggc_mx): New.
8893 (opt_pass::gt_pch_nx): New.
8894 (opt_pass::gt_pch_nx_with_op): New.
8895
910c02a0
DM		 88962013-08-20 David Malcolm <dmalcolm@redhat.com>
8897
8898 * Makefile.in (GTFILES): Add context.h.
8899 * context.c (gcc::context::operator new): New.
8900 (gcc::context::gt_ggc_mx): New.
8901 (gcc::context::gt_pch_nx): New.
8902 (gcc::context::gt_pch_nx): New.
8903 * context.h (gcc::context): Add GTY((user)) marking.
8904 (gcc::context::operator new): New.
8905 (gcc::context::gt_ggc_mx): New.
8906 (gcc::context::gt_pch_nx): New.
8907 (gcc::context::gt_pch_nx): New.
8908 (g): Add GTY marking.
8909 (gt_ggc_mx (gcc::context *)): New.
8910 (gt_pch_nx (gcc::context *)): New.
8911 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8912 void *cookie)): New.
8913 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8914
af4b4236
AM		 89152013-08-20 Alan Modra <amodra@gmail.com>
8916
8917 PR target/57865
8918 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8919 (rs6000_emit_epilogue): Likewise.
8920
12211b99 89212013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC		 8922
8923 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8924
e2323f5b
PB		 89252013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8926 Jakub Jelinek <jakub@redhat.com>
8927
8928 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8929 (BUILT_IN_FABSD64): Likewise.
8930 (BUILT_IN_FABSD128): Likewise.
8931 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8932 (fold_builtin_1): Likewise.
5339fc0c
UB		 8933 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8934 destination and source operands.
e2323f5b
PB		 8935 (*abstd2_fpr): Likewise.
8936 (*nabstd2_fpr): Likewise.
8937
cfbf3ee8
RS		 89382013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8939
8940 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8941 JUMP_P and INSN_P.
8942
d3136aeb
AH		 89432013-08-19 Aldy Hernandez <aldyh@redhat.com>
8944
8945 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8946 incomplete.
8947
d33d9e47
AI		 89482013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8949
8950 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8951 * builtins.c (default_libc_has_function): New.
8952 (gnu_libc_has_function): Ditto.
8953 (no_c99_libc_has_function): Ditto.
8954 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8955 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8956 (fold_builtin_sincos): Likewise.
8957 (fold_builtin_cexp): Likewise.
8958 * builtins.def (DEF_C94_BUILTIN): Likewise.
8959 (DEF_C99_BUILTIN): Likewise.
8960 (DEF_C99_C90RES_BUILTIN): Likewise.
8961 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8962 definitions to using this define.
8963 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 8964 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI		 8965 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8966 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8967 * config/darwin.h: Ditto.
8968 * config/elfos.h: Ditto.
8969 * config/freebsd.h: Ditto.
8970 * config/i386/cygming.h: Ditto.
8971 * config/i386/djgpp.h: Ditto.
8972 * config/i386/i386-interix.h: Ditto.
8973 * config/microblaze/microblaze.h: Ditto.
8974 * config/mmix/mmix.h: Ditto.
8975 * config/gnu-user.h: Ditto.
8976 * config/ia64/hpux.h: Ditto.
8977 * config/pa/pa-hpux.h: Ditto.
8978 * config/pdp11/pdp11.h: Ditto.
8979 * config/picochip/picochip.h: Ditto.
8980 * config/linux.h: Ditto.
8981 * config/netbsd.h: Ditto.
8982 * config/openbsd.h: Ditto.
8983 * config/rs6000/aix43.h: Ditto.
8984 * config/rs6000/aix51.h: Ditto.
8985 * config/rs6000/aix52.h: Ditto.
8986 * config/rs6000/aix53.h: Ditto.
8987 * config/rs6000/aix61.h: Ditto.
8988 * config/rs6000/darwin.h: Ditto.
8989 * config/rs6000/linux.h: Ditto.
8990 * config/rs6000/linux64.h: Ditto.
8991 * config/s390/tpf.h: Ditto.
8992 * config/sol2-10.h: Ditto.
8993 * config/sol2.h: Ditto.
8994 * config/vms/vms.h: Ditto.
8995 * config/vxworks.h: Ditto.
8996 * config/linux-android.c (linux_android_libc_has_function):
8997 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8998 * config/linux-protos.h (linux_android_libc_has_function):
8999 New declaration.
9000 * config/i386/i386.c (ix86_libc_has_function): New.
9001 * config/i386/i386-protos.h
9002 (ix86_libc_has_function): New declaration.
9003 * config/i386/i386.md
9004 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9005 ("isinf<mode>2): Likewise.
9006 * convert.c (convert_to_integer): Using new target hook
9007 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9008 TARGET_C99_FUNCTIONS.
9009 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9010 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9011 * coretypes.h (function_class): New enum for different
9012 classes of functions.
9013 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9014 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9015 (TARGET_HAS_SINCOS): Likewise.
9016 (TARGET_LIBC_HAS_FUNCTION): New.
9017 * doc/tm.texi: Regenerated.
9018 * targhooks.h (default_libc_has_function): New declaration.
9019 (no_c99_libc_has_function): Ditto.
9020 (gnu_libc_has_function): Ditto.
9021 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9022 and TARGET_HAS_SINCOS.
9023
eefe9a99
JH		 90242013-08-18 Jan Hubicka <jh@suse.cz>
9025
9026 * Makeifle-in (ipa-devirt.o): New.
9027 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9028 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9029 (analyze_functions): Look into possible targets of polymorphic call.
9030 * dumpfile.c (dump_files): Add type-inheritance dump.
9031 * dumpfile.h (TDI_inheritance): New.
9032 * ipa-devirt.c: New file.
9033 * ipa-utils.h (odr_type_d): Forward declare.
9034 (odr_type): New type.
9035 (build_type_inheritance_graph): Declare.
9036 (possible_polymorphic_call_targets): Declare and introduce inline
9037 variant when only edge is pased.
9038 (dump_possible_polymorphic_call_targets): Likewise.
9039 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9040 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9041 (types_same_for_odr): ... here.
9042 * tree.h (type_in_anonymous_namespace_p): Declare.
9043
29b89442
JJ		 90442013-08-18 Jakub Jelinek <jakub@redhat.com>
9045
9046 PR tree-optimization/58006
9047 * tree-parloops.c (take_address_of): Don't ICE if get_name
9048 returns NULL.
9049 (eliminate_local_variables_stmt): Remove clobber stmts.
9050
4f219961
EB		 90512013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9052
9053 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9054 error message is issued for an alias to undefined symbol.
9055
1d5755ef
JH		 90562013-08-18 Jan Hubicka <jh@suse.cz>
9057
9058 * cgraph.c (cgraph_create_indirect_edge): Discover
9059 polymorphic calls and record basic info into indirect_info.
9060 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9061 devirtualization, ignore objc function calls.
9062 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9063 call with no parm index info.
9064 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9065 * tree.c (virtual_method_call_p): New function.
9066 * tree.h (virtual_method_call_p): Declare.
9067
5945bebf
JH		 90682013-08-16 Jan Hubicka <jh@suse.cz>
9069
9070 PR middle-end/58179
9071 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9072
47fa5d23
DE		 90732013-08-16 David Edelsohn <dje.gcc@gmail.com>
9074
9075 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9076 attribute.
9077
c74f54a0
DM		 90782013-08-16 David Malcolm <dmalcolm@redhat.com>
9079
9080 * gengtype.c (type_for_name): Add special-case support for
9081 locating types within the "gcc::" namespace.
9082 (open_base_files): Emit a "using namespace gcc" directive.
9083
f758f299
MM		 90842013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9085
9086 PR target/58160
9087 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9088 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9089
9090 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9091 array instead of each individual operand as a separate argument.
9092 (emit_fusion_gpr_load): Likewise.
9093 (expand_fusion_gpr_load): Add new function declaration.
9094
9095 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9096 signature to have the operands passed as an array, instead of as
9097 separate arguments. Allow ZERO_EXTEND to be in the memory
9098 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9099 depend on the register live/dead flags when peepholes are run.
9100 (expand_fusion_gpr_load): New function to be called from the
9101 peephole2 pass, to change the register that addis sets to be the
9102 target register.
9103 (emit_fusion_gpr_load): Change the calling signature to have the
9104 operands passed as an array, instead of as separate arguments.
9105 Allow ZERO_EXTEND to be in the memory address, and also
9106 SIGN_EXTEND if -mpower8-fusion-sign.
9107
9108 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9109 unspec enumeration.
9110 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9111 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 9112 depend on the register live/dead state when the peephole pass is done.
f758f299 9113
158f4e4f
DM		 91142013-08-16 David Malcolm <dmalcolm@redhat.com>
9115
9116 * gengtype.c (create_user_defined_type): Ensure that the kind
9117 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9118 declaration is seen before the GTY((user)) marking.
9119
22f8cea5
BE		 91202013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9121
9122 PR target/58105
9123 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9124
c49bdb2e
JH		 91252013-08-16 Jan Hubicka <jh@suse.cz>
9126
9127 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9128 arugment expected_type.
9129 (gimple_fold_call): Use it.
9130 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9131 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 9132 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH		 9133 (try_make_edge_direct_virtual_call): Likewise.
9134 * tree.c (obj_type_ref_class): New.
9135 * tree.h (obj_type_ref_class): Use it.
9136
4042dca9
GDR		 91372013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9138
9139 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9140 (rtl_slim_pp): Likewise.
9141 (init_rtl_slim_pretty_print): Likewise.
9142 (dump_value_slim): Don't call it. Use local pretty printer.
9143 (dump_insn_slim): Likewise.
9144 (dump_rtl_slim): Likewise.
9145 (str_pattern_slim): Likewise.
9146 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9147 Simplify.
9148
7a460594
JJ		 91492013-08-16 Jakub Jelinek <jakub@redhat.com>
9150
639dc669
JJ		 9151 PR tree-optimization/58164
9152 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9153 walk gimple_goto_dest of GIMPLE_GOTO.
9154
7a460594
JJ		 9155 PR tree-optimization/58165
9156 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9157 bi_call must be the last stmt in a bb, don't split_block, instead
9158 use fallthru edge from it and give up if there is none.
9159 Release conds vector when returning early.
9160
e3c02cc7
XDL		 91612013-08-14 Xinliang David Li <davidxl@google.com>
9162
9163 * config/i386/i386.c (ix86_option_override_internal):
9164 Remove unused variable and field.
9165
6469da2c
BS		 91662013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9167
9168 PR target/57949
22f8cea5 9169 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS		 9170 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9171 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9172 and Linux, correct BLKmode alignment when 128-bit alignment is
9173 required and compatibility flag is not set.
22f8cea5
BE		 9174 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9175 for zero-size arguments when compatibility flag is not set.
6469da2c 9176
4ca890e2
JJ		 91772013-08-14 Jakub Jelinek <jakub@redhat.com>
9178
9179 PR tree-optimization/58145
9180 * tree-sra.c (build_ref_for_offset): If prev_base has
9181 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9182
daec1759 91832013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 9184
daec1759 9185 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 9186 Fix uninitialized variable error.
daec1759 9187
3ad20bd4
XDL		 91882013-08-14 Xinliang David Li <davidxl@google.com>
9189
9190 * config/i386/i386.opt: Define two new options.
9191 * config/i386/x86-tune.def: Add arch selector field in macros.
9192 * config/i386/i386.h: Adjust macro definition.
9193 * config/i386/i386.c (ix86_option_override_internal):
9194 Refactor the code.
9195 (parse_mtune_ctrl_str): New function.
9196 (set_ix86_tune_features): New function.
9197 (ix86_function_specific_restore): Call the new helper function.
9198
e839e2a9
AB		 91992013-08-14 Andrey Belevantsev <abel@ispras.ru>
9200
9201 PR rtl-optimization/57662
9202 * sel-sched.c (code_motion_process_successors): When the current insn
9203 is removed after the recursive traversal, break from the loop.
9204 Add comments and debug printouts.
9205
27ed665f
JJ		 92062013-08-14 Jakub Jelinek <jakub@redhat.com>
9207 Alexandre Oliva <aoliva@redhat.com>
9208
9209 PR target/58067
9210 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9211 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9212 there also UNSPEC_PLTOFF.
9213
61c1a609
MP		 92142013-08-14 Marek Polacek <polacek@redhat.com>
9215
9216 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9217 NUM_CONDITIONS bit positions.
9218
0e901c67
CC		 92192013-08-13 Cary Coutant <ccoutant@google.com>
9220
9221 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9222 (attr_checksum): Hash vector contents instead of pointer.
9223 (attr_checksum_ordered): Likewise.
9224
061eff6d
UB		 92252013-08-13 Uros Bizjak <ubizjak@gmail.com>
9226
9227 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9228 when Pmode != word_mode. Add length_address attribute.
9229 (sse3_monitor_<mode>): Merge from sse3_monitor and
9230 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9231 Pmode != word_mode. Update insn length attribute.
9232 * config/i386/i386.c (ix86_option_override_internal): Update
9233 ix86_gen_monitor selection for merged sse3_monitor insn.
9234
8ec9249b
JB		 92352013-08-13 Julian Brown <julian@codesourcery.com>
9236
9237 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9238 perform invalid legitimization on greater-than-word-size modes for
9239 TARGET_E500_DOUBLE.
9240
761a8eb7
VM		 92412013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9242
9243 * ira.c (setup_class_translate_array): Use aclass instead of cl
9244 for classes not fully covered by allocno classes.
9245
99c2bd54
JJ		 92462013-08-13 Jakub Jelinek <jakub@redhat.com>
9247
4029a5e0
JJ		 9248 PR tree-optimization/57661
9249 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9250 * tree-inline.c (tree_function_versioning): Initialize it.
9251 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9252 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9253 that is not being copied.
9254
99c2bd54
JJ		 9255 PR sanitizer/56417
9256 * asan.c (instrument_strlen_call): Fix typo in comment.
9257 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9258
b5f8f063
SE		 92592013-08-13 Steve Ellcey <sellcey@mips.com>
9260
9261 * config/mips/mips.md (prefetch): Use lw instead of ld on
9262 loongson in 32bit mode.
9263
4de80584
NC		 92642013-08-13 Nick Clifton <nickc@redhat.com>
9265
9266 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9267
09ce3660
JH		 92682013-08-13 Jan Hubicka <jh@suse.cz>
9269
9270 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9271 introduced edge; fix typo in sanity check.
9272 (cgraph_resolve_speculation): Export; improve diagnostic.
9273 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9274 speculation at type mismatch.
9275 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9276 (cgraph_resolve_speculation): Declare.
9277 (symtab_can_be_discarded): New function.
9278 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9279 * ipa-inline-transform.c (speculation_removed): New global var.
9280 (clone_inlined_nodes): See if speculation can be removed.
9281 (inline_call): If speculations was removed, we growths may not match.
9282 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9283 (speculation_useful_p): New function.
9284 (resolve_noninline_speculation): New function.
9285 (inline_small_functions): Resolve useless speculations.
9286 * ipa-inline.h (speculation_useful_p): Declare
9287 * ipa.c (can_replace_by_local_alias): Simplify.
9288 (ipa_profile): Produce speculative calls in non-lto, too;
9289 add simple cost model; produce local aliases.
9290
f971dc24
DM		 92912013-08-13 David Malcolm <dmalcolm@redhat.com>
9292
9293 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 9294 PASS_MANAGER_H.
f971dc24 9295
2e507b9a
PC		 92962013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9297
9298 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9299
c97fd4b8
UB		 93002013-08-12 Uros Bizjak <ubizjak@gmail.com>
9301
9302 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9303 iterator instead of X87MODEF.
9304
93052013-08-12 Perez Read <netfirewall@gmail.com>
9306
9307 PR target/58132
9308 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9309 operand 0 for intel asm alternative.
9310 (*movabs<mode>_2): Ditto for operand 1.
9311
f4f4204c
JG		 93122013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9313
9314 * config/aarch64/arm_none.h
9315 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9316
133b1a8e
NC		 93172013-08-12 Nick Clifton <nickc@redhat.com>
9318
9319 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9320
15dd8b3a
YR		 93212013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9322
9323 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9324 expand for QI/HImode operand to produce more effictive code for
9325 unsigned char(short) --> float(double) conversion.
9326
94109a6a
AM		 93272013-08-12 Alexander Monakov <amonakov@ispras.ru>
9328
9329 * doc/invoke.texi: Mention that -ftls-model does not force the final
9330 model.
9331
4e115102
MP		 93322013-08-12 Marek Polacek <polacek@redhat.com>
9333 Marc Glisse <marc.glisse@inria.fr>
9334
9335 PR tree-optimization/57980
9336 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9337 when creating -1 constant.
9338
748d71f3
JH		 93392013-08-10 Jan Hubicka <jh@suse.cz>
9340
1f97e193 9341 Workaround binutils PR14342.
748d71f3
JH		 9342 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9343 (gimple_init_edge_profiler): Likewise.
9344 (gimple_gen_ic_func_profiler): Likewise.
9345
5979aa54
JH		 93462013-08-09 Jan Hubicka <jh@suse.cz>
9347
9348 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9349
ad83025e
XDL		 93502013-08-09 Xinliang David Li <davidxl@google.com>
9351
9352 * config/i386/stringop.def: New file.
9353 * config/i386/stringop.opt: New file.
9354 * config/i386/i386-opts.h: Include stringopt.def.
9355 * config/i386/i386.opt: Include stringopt.opt.
9356 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 9357 Override default size based stringop inline strategies with options.
ad83025e
XDL		 9358 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9359 New function.
9360
e59df5fd
JH		 93612013-08-09 Jan Hubicka <jh@suse.cz>
9362
c97fd4b8 9363 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 9364
634ab819
JH		 93652013-08-09 Jan Hubicka <jh@suse.cz>
9366
9367 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9368 CGRAPH_FREQ_MAX.
9369 (dump_cgraph_node): Dump profile-id.
9370 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9371 and common_target_probability.
9372 * lto-cgraph.c (lto_output_edge): Stream common targets.
9373 (lto_output_node): Stream profile ids.
9374 (input_node): Stream profile ids.
9375 (input_edge): Stream common targets.
9376 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9377 * ipa.c: Include value-prof.h
9378 (ipa_profile_generate_summary): Turn indirect call statement histograms
9379 into common targets.
9380 (ipa_profile): Turn common targets into speculative edges.
9381
2fa3d31b
JH		 93822013-08-09 Jan Hubicka <jh@suse.cz>
9383
9384 * cgraph.h (cgraph_node): Add profile_id.
9385 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 9386 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH		 9387 (del_node_map): Update.
9388 (find_func_by_funcdef_no): Replace by ...
9389 (find_func_by_profile_id): ... this one.
9390 (gimple_ic_transform): Do not remove useful histograms when
9391 speculation is not done; dump info when indirect call removal
9392 can happen at LTO.
9393 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9394 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9395 (__gcov_indirect_call_profiler_v2): .. this one.
9396 * profile.h (init_node_map): Update.
9397 * coverage.c (coverage_compute_profile_id): New function.
9398 * coverage.h (coverage_compute_profile_id): Declare.
9399 * tree-profile.c (init_ic_make_global_vars): Make
9400 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9401 (gimple_init_edge_profiler): Update prototype of
9402 __gcov_indirect_call_profiler.
9403 (gimple_gen_ic_func_profiler): Simplify.
9404 (tree_profiling): Use init_node_map
9405
042ae7d2
JH		 94062013-08-09 Jan Hubicka <jh@suse.cz>
9407
c97fd4b8
UB		 9408 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9409 non-speculative refs.
042ae7d2
JH		 9410 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9411 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9412 (cgraph_set_call_stmt): Likewise.
9413 (cgraph_create_edge_1): Fix release checking compilatoin;
9414 clear lto_stmt_uid.
9415 (cgraph_free_edge): Free indirect info.
9416 (cgraph_turn_edge_to_speculative): New function.
9417 (cgraph_speculative_call_info): New function.
9418 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 9419 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH		 9420 (dump_cgraph_node): Dump speculation.
9421 (verify_edge_count_and_frequency): Accept speculative edges.
9422 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9423 (verify_cgraph_node): Handle speculation.
9424 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9425 (cgraph_set_call_stmt): Update prototype.
9426 (cgraph_make_edge_direct): Update prototype.
9427 (cgraph_speculative_call_info): Declare.
9428 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9429 to change; update call of ipa_find_references.
9430 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9431 lto_stmt_uid and speculative flags.
9432 (ipa_dump_references): Dump speculation.
9433 (ipa_clone_references): Clone speculative flag.
9434 (ipa_clone_referring): Likewise.
9435 (ipa_clone_ref): New function.
9436 (ipa_find_reference): Look into lto_stmt_uids
9437 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9438 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9439 (ipa_find_reference): Update declaration.
9440 (ipa_clone_ref): Declare.
9441 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9442 stream speculative flag.
9443 (lto_output_ref): Stream statements uids and speculation.
9444 (input_ref): Likewise.
9445 (input_edge): Stream speuclation.
9446 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9447 (cgraph_set_call_stmt_including_clones): Handle speculation.
9448 * ipa-inline.c (heap_edge_removal_hook): New function.
9449 (inline_small_functions): Register it.
9450 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9451 also initialize refs.
9452 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9453 edge to change.
9454 (try_make_edge_direct_simple_call): Likewise.
9455 (try_make_edge_direct_simple_call): Likewise.
9456 (update_indirect_edges_after_inlining): Likewise.
9457 (remove_described_reference): Look proper lto_stmt_uid.
9458 (propagate_controlled_uses): Likewise.
9459 (propagate_controlled_uses): Liekwise.
9460 * tree-inline.c (copy_bb): Copy speculative edges.
9461 (redirect_all_calls): New function.
c97fd4b8
UB		 9462 (copy_cfg_body): Do redirection after loop info is updated.
9463 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 9464
aa1e10cc
JH		 94652013-08-09 Jan Hubicka <jh@suse.cz>
9466
9467 * lto-streamer-out.c (output_function): Renumber PHIs.
9468 * lto-streamer-in.c (input_function): Likewise.
9469
66adb8eb
JG		 94702013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9471
9472 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9473 (get_lane_unsigned): Likewise.
9474 (dup_lane_scalar): Likewise.
9475 (get_lane): enable for VALL.
9476 * config/aarch64/aarch64-simd.md
9477 (aarch64_dup_lane_scalar<mode>): Remove.
9478 (aarch64_get_lane_signed<mode>): Likewise.
9479 (aarch64_get_lane_unsigned<mode>): Likewise.
9480 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9481 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9482 (aarch64_get_lane<mode>): Enable for all vector modes.
9483 (aarch64_get_lanedi): Remove misleading constraints.
9484 * config/aarch64/arm_neon.h
9485 (__aarch64_vget_lane_any): Define.
9486 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9487 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9488 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9489 * config/aarch64/iterators.md (VDQQH): New.
9490 (VDQQHS): Likewise.
9491 (vwcore): Likewise.
9492
1f65ae7a
EB		 94932013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9494
9495 * configure.ac: Add GAS check for LEON instructions on SPARC.
9496 * configure: Regenerate.
9497 * config.in: Likewise.
9498 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9499 sparc*-*-* block.
9500 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9501 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9502 for LEON or LEON3.
9503 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9504 (AS_LEON_FLAG): New macro.
9505 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9506 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9507 Deal with LEON and LEON3 for the memory model.
23a6cb78 9508 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB		 9509 (atomic_compare_and_swap<mode>_1): Likewise.
9510 (*atomic_compare_and_swap<mode>_1): Likewise.
9511
ff522f7f
ZC		 95122013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9513
9514 * config/arm/neon.md (vcond): Fix floating-point vector
9515 comparisons against 0.
9516
cb1cca12
VM		 95172013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9518
9519 * lra-constraints.c (emit_spill_move): Remove assert.
9520 (process_alt_operands): Add more debugging
9521 output. Increase reject for spilling into memory. Decrease
9522 reject for reloading scratch.
9523 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9524
a3719e31
SE		 95252013-08-08 Steve Ellcey <sellcey@mips.com>
9526
9527 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9528 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9529 micromips incompatible. Add nan2008.
9530 (MULTILIB_DIRNAMES): Add nan2008.
9531 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9532 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9533 and micromips incompatible. Add nan2008.
9534 (MULTILIB_DIRNAMES): Add nan2008.
9535 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9536
f40423e2
RS		 95372013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9538
9539 PR rtl-optimization/58079
9540 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9541 simplify_comparison has widened a comparison with an integer.
9542
869b9125
KT		 95432013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9544
9545 * config/arm/neon.md (movmisalign<mode>): Disable when we
9546 don't allow unaligned accesses.
9547 (*movmisalign<mode>_neon_store): Likewise.
9548 (*movmisalign<mode>_neon_load): Likewise.
9549 (*movmisalign<mode>_neon_store): Likewise.
9550 (*movmisalign<mode>_neon_load): Likewise.
9551
71cafea9
JH		 95522013-08-08 Jan Hubicka <jh@suse.cz>
9553
9554 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9555 (make_pass_rebuild_cgraph_edges): Also clear references.
9556 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9557 * ipa-inline-transform.c (inline_transform): Remove all references
9558 after inlining.
c97fd4b8
UB		 9559 * cgraphunit.c (expand_function): Remove all references after
9560 expansion.
71cafea9
JH		 9561 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9562 (ipa_find_reference): Rewrite to iterator.
9563 (remove_stmt_references): Likewise.
9564 (ipa_clear_stmts_in_references): New function.
9565 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB		 9566 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9567 clear references.
71cafea9
JH		 9568 * ipa-split.c (split_function): Remove references in split function.
9569
27d2e612
RE		 95702013-08-08 Richard Earnshaw <rearnsha@arm.com>
9571
9572 PR target/57431
1f65ae7a 9573 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE		 9574 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9575
5efc661c
RE		 95762013-08-08 Richard Earnshaw <rearnsha@arm.com>
9577
9578 PR target/56979
1f65ae7a 9579 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE		 9580 suggested mode for the assignment isn't compatible with the
9581 registers required.
9582
f276d31d
BE		 95832013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9584
9585 PR target/58065
9586 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9587
4b8bc035
XDL		 95882013-08-07 Xinliang David Li <davidxl@google.com>
9589
9590 * config/i386/i386.opt: New option -mtune-ctrl=.
9591 * config/i386/x86-tune.def: New file.
9592 * config/i386/i386.h: include x86-tune.def.
9593 * config/i386/i386.c (ix86_option_override_internal):
9594 Parsing -mtune-ctrl= option and set tune features.
9595
12211b99 95962013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE		 9597
9598 PR other/12081
03b0ee0a 9599 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE		 9600 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9601 to gen_2arg_fn_t.
9602
5cf6635b
EB		 96032013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9604
9605 * rtl.h (update_alignments): Declare.
9606 * final.c (grow_label_align): New function extracted from...
9607 (shorten_branches): ...here. Call it.
9608 (update_alignments): New function.
9609 * reorg.c (sibling_labels): New variable.
9610 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9611 the new label along with it onto the sibling_labels vector.
9612 (fill_simple_delay_slots): Adjust call to get_label_before.
9613 (fill_slots_from_thread): Likewise.
9614 (relax_delay_slots): Likewise.
9615 (make_return_insns): Likewise.
9616 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9617
3c8ca1ab
EB		 96182013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9619
9620 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9621 document its semantics.
9622 (diagnostic_report_diagnostic): Adjust accordingly.
9623
5ee5b32c
DM		 96242013-08-07 David Malcolm <dmalcolm@redhat.com>
9625
03b0ee0a 9626 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM		 9627 (sparc_option_override): ...and port to new C++ pass API.
9628 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9629
abd566fa
PB		 96302013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9631
9632 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9633
2077db1b
CT		 96342013-08-06 Caroline Tice <cmtice@google.com>
9635
9636 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9637 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9638 * tree-pass.h: Add pass_vtable_verify.
9639 * varasm.c (assemble_variable): Add code to properly set the comdat
9640 section and name for the .vtable_map_vars section.
9641 (assemble_vtyv_preinit_initializer): New function.
9642 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9643 LINK_ONCE flag.
9644 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9645 * vtable-verify.c: New file.
9646 * vtable-verify.h: New file.
9647 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9648 initialiation levels.
9649 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9650 * passes.def: Insert pass_vtable_verify.
9651 * aclocal.m4: Reorder includes.
9652 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9653 -fvtv-counts options.
9654 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9655 as appropriate, if -fvtable-verify=... is used.
9656 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9657 -fvtable-verify=... is used.
9658 * Makefile.in (OBJS): Add vtable-verify.o to list.
9659 (vtable-verify.o): Add new build rule.
9660 (GTFILES): Add vtable-verify.c to list.
9661 * common.opt (fvtable-verify=): New flag.
9662 (vtv_priority): Values for fvtable-verify= flag.
9663 (fvtv-counts): New flag.
9664 (fvtv-debug): New flag.
9665 * tree.h (save_vtable_map_decl): New extern function decl.
9666
03085d1c
DM		 96672013-08-07 David Malcolm <dmalcolm@redhat.com>
9668
9669 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9670 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9671 (pass_data_rl78_devirt): ...new pass_data instance and...
9672 (make_pass_rl78_devirt): ...new function.
9673 (rl78_asm_file_start): Port pass registration to new C++ API.
9674
05555c4a
DM		 96752013-08-07 David Malcolm <dmalcolm@redhat.com>
9676
9677 * coretypes.h (rtl_opt_pass): Add.
9678 (gcc::context): Add.
9679 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9680 (epiphany_init): Port to new C++ pass API.
9681 (epiphany_optimize_mode_switching): Likewise.
9682 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9683 (pass_manager::get_pass_mode_switching): New.
9684 (pass_manager::get_pass_peephole2): New.
9685 * mode-switching.c (pass_mode_switching): Add clone method.
9686 * recog.c (pass_peephole2): Add clone method.
9687 (pass_split_all_insns): Add clone method.
9688
c5a12e26
DM		 96892013-08-06 David Malcolm <dmalcolm@redhat.com>
9690
03b0ee0a
UB		 9691 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9692 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 9693
a2e2a668
JH		 96942013-08-06 Jan Hubicka <jh@suse.cz>
9695
9696 * cgraph.c (cgraph_get_body): New function based on lto.c
9697 implementation.
9698 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB		 9699 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9700 LTO paths.
a2e2a668
JH		 9701 * cgraphunit.c (expand_function): Get body prior expanding.
9702 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB		 9703 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9704 really need.
a2e2a668
JH		 9705 * passes.c (do_per_function_toporder): Get body.
9706 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9707 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9708
15682f24
MJ		 97092013-08-06 Martin Jambor <mjambor@suse.cz>
9710
9711 PR fortran/57987
9712 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9713 is not re-finalized. Rename second parameter to no_collect.
9714
78f6dd68
MJ		 97152013-08-06 Martin Jambor <mjambor@suse.cz>
9716
9717 PR middle-end/58041
9718 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9719 MEM_REF has proper alignment information.
9720
12211b99 97212013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE		 9722
9723 PR other/12081
9724 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9725 class insn_gen_fn.
9726 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9727 rtx (*) (rtx, ...) with insn_gen_fn.
9728 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9729 insn_gen_fn::stored_funcptr. Add initializer braces.
9730
8ac69a6c
DM		 97312013-08-05 David Malcolm <dmalcolm@redhat.com>
9732
9733 Rewrite how instances of passes are cloned to remove assumptions
9734 about their sizes (thus allowing pass subclasses to have
9735 additional data fields, albeit non-GC-managed ones at this point).
9736
9737 * passes.c (make_pass_instance): Now that passes have clone
9738 methods, rewrite this function to eliminate XNEW and memcpy
9739 calls that used hardcoded sizes. Since this function no longer
9740 creates pass instances, rename it to...
03b0ee0a
UB		 9741 (add_pass_instance): ...this. Document the old way that passes were
9742 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM		 9743 (next_pass_1): Add an initial_pass argument for use by
9744 add_pass_instance.
9745 (position_pass): When adding multiple instances of a pass, use
9746 the pass's clone method, rather than relying on the XNEW/memcpy
9747 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB		 9748 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9749 the initial instance of the current pass within the pass manager.
8ac69a6c 9750
27a4cd48
DM		 97512013-08-05 David Malcolm <dmalcolm@redhat.com>
9752
9753 This is the automated part of the conversion of passes from C
9754 structs to C++ classes.
9755
9756 Patch autogenerated by refactor_passes.py from
9757 https://github.com/davidmalcolm/gcc-refactoring-scripts
9758 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9759
9760 * asan.c (pass_asan): Convert from a global struct to a subclass of
9761 gimple_opt_pass along with...
9762 (pass_data_asan): ...new pass_data instance and...
9763 (make_pass_asan): ...new function.
9764 (pass_asan_O0): Convert from a global struct to a subclass of
9765 gimple_opt_pass along with...
9766 (pass_data_asan_O0): ...new pass_data instance and...
9767 (make_pass_asan_O0): ...new function.
9768 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9769 subclass of rtl_opt_pass along with...
9770 (pass_data_inc_dec): ...new pass_data instance and...
9771 (make_pass_inc_dec): ...new function.
9772 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9773 a subclass of rtl_opt_pass along with...
9774 (pass_data_reorder_blocks): ...new pass_data instance and...
9775 (make_pass_reorder_blocks): ...new function.
9776 (pass_duplicate_computed_gotos): Convert from a global struct to a
9777 subclass of rtl_opt_pass along with...
9778 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9779 (make_pass_duplicate_computed_gotos): ...new function.
9780 (pass_partition_blocks): Convert from a global struct to a subclass of
9781 rtl_opt_pass along with...
9782 (pass_data_partition_blocks): ...new pass_data instance and...
9783 (make_pass_partition_blocks): ...new function.
9784 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9785 struct to a subclass of rtl_opt_pass along with...
9786 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9787 and...
9788 (make_pass_branch_target_load_optimize1): ...new function.
9789 (pass_branch_target_load_optimize2): Convert from a global struct to a
9790 subclass of rtl_opt_pass along with...
9791 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9792 and...
9793 (make_pass_branch_target_load_optimize2): ...new function.
9794 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9795 of rtl_opt_pass along with...
9796 (pass_data_jump): ...new pass_data instance and...
9797 (make_pass_jump): ...new function.
9798 (pass_jump2): Convert from a global struct to a subclass of
9799 rtl_opt_pass along with...
9800 (pass_data_jump2): ...new pass_data instance and...
9801 (make_pass_jump2): ...new function.
9802 * cfgexpand.c (pass_expand): Convert from a global struct to a
9803 subclass of rtl_opt_pass along with...
9804 (pass_data_expand): ...new pass_data instance and...
9805 (make_pass_expand): ...new function.
9806 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9807 of rtl_opt_pass along with...
9808 (pass_data_free_cfg): ...new pass_data instance and...
9809 (make_pass_free_cfg): ...new function.
9810 (pass_into_cfg_layout_mode): Convert from a global struct to a
9811 subclass of rtl_opt_pass along with...
9812 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9813 (make_pass_into_cfg_layout_mode): ...new function.
9814 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9815 subclass of rtl_opt_pass along with...
9816 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9817 (make_pass_outof_cfg_layout_mode): ...new function.
9818 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9819 struct to a subclass of gimple_opt_pass along with...
9820 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9821 (make_pass_build_cgraph_edges): ...new function.
9822 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9823 subclass of gimple_opt_pass along with...
9824 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9825 (make_pass_rebuild_cgraph_edges): ...new function.
9826 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9827 subclass of gimple_opt_pass along with...
9828 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9829 and...
9830 (make_pass_remove_cgraph_callee_edges): ...new function.
9831 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9832 struct to a subclass of rtl_opt_pass along with...
9833 (pass_data_stack_adjustments): ...new pass_data instance and...
9834 (make_pass_stack_adjustments): ...new function.
9835 * combine.c (pass_combine): Convert from a global struct to a subclass
9836 of rtl_opt_pass along with...
9837 (pass_data_combine): ...new pass_data instance and...
9838 (make_pass_combine): ...new function.
9839 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9840 global struct to a subclass of rtl_opt_pass along with...
9841 (pass_data_compare_elim_after_reload): ...new pass_data instance
9842 and...
9843 (make_pass_compare_elim_after_reload): ...new function.
9844 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9845 of rtl_opt_pass along with...
9846 (pass_data_rtl_cprop): ...new pass_data instance and...
9847 (make_pass_rtl_cprop): ...new function.
9848 * cse.c (pass_cse): Convert from a global struct to a subclass of
9849 rtl_opt_pass along with...
9850 (pass_data_cse): ...new pass_data instance and...
9851 (make_pass_cse): ...new function.
9852 (pass_cse2): Convert from a global struct to a subclass of
9853 rtl_opt_pass along with...
9854 (pass_data_cse2): ...new pass_data instance and...
9855 (make_pass_cse2): ...new function.
9856 (pass_cse_after_global_opts): Convert from a global struct to a
9857 subclass of rtl_opt_pass along with...
9858 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9859 (make_pass_cse_after_global_opts): ...new function.
9860 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9861 of rtl_opt_pass along with...
9862 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9863 (make_pass_ud_rtl_dce): ...new function.
9864 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9865 rtl_opt_pass along with...
9866 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9867 (make_pass_fast_rtl_dce): ...new function.
9868 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9869 a subclass of rtl_opt_pass along with...
9870 (pass_data_df_initialize_opt): ...new pass_data instance and...
9871 (make_pass_df_initialize_opt): ...new function.
9872 (pass_df_initialize_no_opt): Convert from a global struct to a
9873 subclass of rtl_opt_pass along with...
9874 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9875 (make_pass_df_initialize_no_opt): ...new function.
9876 (pass_df_finish): Convert from a global struct to a subclass of
9877 rtl_opt_pass along with...
9878 (pass_data_df_finish): ...new pass_data instance and...
9879 (make_pass_df_finish): ...new function.
9880 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9881 rtl_opt_pass along with...
9882 (pass_data_rtl_dse1): ...new pass_data instance and...
9883 (make_pass_rtl_dse1): ...new function.
9884 (pass_rtl_dse2): Convert from a global struct to a subclass of
9885 rtl_opt_pass along with...
9886 (pass_data_rtl_dse2): ...new pass_data instance and...
9887 (make_pass_rtl_dse2): ...new function.
9888 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9889 subclass of rtl_opt_pass along with...
9890 (pass_data_dwarf2_frame): ...new pass_data instance and...
9891 (make_pass_dwarf2_frame): ...new function.
9892 * except.c (pass_set_nothrow_function_flags): Convert from a global
9893 struct to a subclass of rtl_opt_pass along with...
9894 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9895 and...
9896 (make_pass_set_nothrow_function_flags): ...new function.
9897 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9898 subclass of rtl_opt_pass along with...
9899 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9900 and...
9901 (make_pass_convert_to_eh_region_ranges): ...new function.
9902 * final.c (pass_compute_alignments): Convert from a global struct to a
9903 subclass of rtl_opt_pass along with...
9904 (pass_data_compute_alignments): ...new pass_data instance and...
9905 (make_pass_compute_alignments): ...new function.
9906 (pass_final): Convert from a global struct to a subclass of
9907 rtl_opt_pass along with...
9908 (pass_data_final): ...new pass_data instance and...
9909 (make_pass_final): ...new function.
9910 (pass_shorten_branches): Convert from a global struct to a subclass of
9911 rtl_opt_pass along with...
9912 (pass_data_shorten_branches): ...new pass_data instance and...
9913 (make_pass_shorten_branches): ...new function.
9914 (pass_clean_state): Convert from a global struct to a subclass of
9915 rtl_opt_pass along with...
9916 (pass_data_clean_state): ...new pass_data instance and...
9917 (make_pass_clean_state): ...new function.
9918 * function.c (pass_instantiate_virtual_regs): Convert from a global
9919 struct to a subclass of rtl_opt_pass along with...
9920 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9921 (make_pass_instantiate_virtual_regs): ...new function.
9922 (pass_leaf_regs): Convert from a global struct to a subclass of
9923 rtl_opt_pass along with...
9924 (pass_data_leaf_regs): ...new pass_data instance and...
9925 (make_pass_leaf_regs): ...new function.
9926 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9927 subclass of rtl_opt_pass along with...
9928 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9929 and...
9930 (make_pass_thread_prologue_and_epilogue): ...new function.
9931 (pass_match_asm_constraints): Convert from a global struct to a
9932 subclass of rtl_opt_pass along with...
9933 (pass_data_match_asm_constraints): ...new pass_data instance and...
9934 (make_pass_match_asm_constraints): ...new function.
9935 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9936 subclass of rtl_opt_pass along with...
9937 (pass_data_rtl_fwprop): ...new pass_data instance and...
9938 (make_pass_rtl_fwprop): ...new function.
9939 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9940 rtl_opt_pass along with...
9941 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9942 (make_pass_rtl_fwprop_addr): ...new function.
9943 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9944 rtl_opt_pass along with...
9945 (pass_data_rtl_pre): ...new pass_data instance and...
9946 (make_pass_rtl_pre): ...new function.
9947 (pass_rtl_hoist): Convert from a global struct to a subclass of
9948 rtl_opt_pass along with...
9949 (pass_data_rtl_hoist): ...new pass_data instance and...
9950 (make_pass_rtl_hoist): ...new function.
9951 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9952 subclass of gimple_opt_pass along with...
9953 (pass_data_lower_cf): ...new pass_data instance and...
9954 (make_pass_lower_cf): ...new function.
9955 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9956 from a global struct to a subclass of gimple_opt_pass along with...
9957 (pass_data_strength_reduction): ...new pass_data instance and...
9958 (make_pass_strength_reduction): ...new function.
9959 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9960 of rtl_opt_pass along with...
9961 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9962 (make_pass_rtl_ifcvt): ...new function.
9963 (pass_if_after_combine): Convert from a global struct to a subclass of
9964 rtl_opt_pass along with...
9965 (pass_data_if_after_combine): ...new pass_data instance and...
9966 (make_pass_if_after_combine): ...new function.
9967 (pass_if_after_reload): Convert from a global struct to a subclass of
9968 rtl_opt_pass along with...
9969 (pass_data_if_after_reload): ...new pass_data instance and...
9970 (make_pass_if_after_reload): ...new function.
9971 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9972 a subclass of rtl_opt_pass along with...
9973 (pass_data_initialize_regs): ...new pass_data instance and...
9974 (make_pass_initialize_regs): ...new function.
9975 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9976 of ipa_opt_pass_d along with...
9977 (pass_data_ipa_cp): ...new pass_data instance and...
9978 (make_pass_ipa_cp): ...new function.
9979 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9980 global struct to a subclass of gimple_opt_pass along with...
9981 (pass_data_inline_parameters): ...new pass_data instance and...
9982 (make_pass_inline_parameters): ...new function.
9983 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9984 subclass of gimple_opt_pass along with...
9985 (pass_data_early_inline): ...new pass_data instance and...
9986 (make_pass_early_inline): ...new function.
9987 (pass_ipa_inline): Convert from a global struct to a subclass of
9988 ipa_opt_pass_d along with...
9989 (pass_data_ipa_inline): ...new pass_data instance and...
9990 (make_pass_ipa_inline): ...new function.
9991 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9992 struct to a subclass of gimple_opt_pass along with...
9993 (pass_data_local_pure_const): ...new pass_data instance and...
9994 (make_pass_local_pure_const): ...new function.
9995 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9996 ipa_opt_pass_d along with...
9997 (pass_data_ipa_pure_const): ...new pass_data instance and...
9998 (make_pass_ipa_pure_const): ...new function.
9999 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10000 to a subclass of ipa_opt_pass_d along with...
10001 (pass_data_ipa_reference): ...new pass_data instance and...
10002 (make_pass_ipa_reference): ...new function.
10003 * ipa-split.c (pass_split_functions): Convert from a global struct to
10004 a subclass of gimple_opt_pass along with...
10005 (pass_data_split_functions): ...new pass_data instance and...
10006 (make_pass_split_functions): ...new function.
10007 (pass_feedback_split_functions): Convert from a global struct to a
10008 subclass of gimple_opt_pass along with...
10009 (pass_data_feedback_split_functions): ...new pass_data instance and...
10010 (make_pass_feedback_split_functions): ...new function.
10011 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10012 global struct to a subclass of simple_ipa_opt_pass along with...
10013 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10014 instance and...
10015 (make_pass_ipa_function_and_variable_visibility): ...new function.
10016 (pass_ipa_free_inline_summary): Convert from a global struct to a
10017 subclass of simple_ipa_opt_pass along with...
10018 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10019 (make_pass_ipa_free_inline_summary): ...new function.
10020 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10021 subclass of ipa_opt_pass_d along with...
10022 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10023 and...
10024 (make_pass_ipa_whole_program_visibility): ...new function.
10025 (pass_ipa_profile): Convert from a global struct to a subclass of
10026 ipa_opt_pass_d along with...
10027 (pass_data_ipa_profile): ...new pass_data instance and...
10028 (make_pass_ipa_profile): ...new function.
10029 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10030 ipa_opt_pass_d along with...
10031 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10032 (make_pass_ipa_cdtor_merge): ...new function.
10033 * ira.c (pass_ira): Convert from a global struct to a subclass of
10034 rtl_opt_pass along with...
10035 (pass_data_ira): ...new pass_data instance and...
10036 (make_pass_ira): ...new function.
10037 (pass_reload): Convert from a global struct to a subclass of
10038 rtl_opt_pass along with...
10039 (pass_data_reload): ...new pass_data instance and...
10040 (make_pass_reload): ...new function.
10041 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10042 subclass of rtl_opt_pass along with...
10043 (pass_data_cleanup_barriers): ...new pass_data instance and...
10044 (make_pass_cleanup_barriers): ...new function.
10045 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10046 of rtl_opt_pass along with...
10047 (pass_data_loop2): ...new pass_data instance and...
10048 (make_pass_loop2): ...new function.
10049 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10050 rtl_opt_pass along with...
10051 (pass_data_rtl_loop_init): ...new pass_data instance and...
10052 (make_pass_rtl_loop_init): ...new function.
10053 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10054 rtl_opt_pass along with...
10055 (pass_data_rtl_loop_done): ...new pass_data instance and...
10056 (make_pass_rtl_loop_done): ...new function.
10057 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10058 subclass of rtl_opt_pass along with...
10059 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10060 (make_pass_rtl_move_loop_invariants): ...new function.
10061 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10062 rtl_opt_pass along with...
10063 (pass_data_rtl_unswitch): ...new pass_data instance and...
10064 (make_pass_rtl_unswitch): ...new function.
10065 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10066 subclass of rtl_opt_pass along with...
10067 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10068 and...
10069 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10070 (pass_rtl_doloop): Convert from a global struct to a subclass of
10071 rtl_opt_pass along with...
10072 (pass_data_rtl_doloop): ...new pass_data instance and...
10073 (make_pass_rtl_doloop): ...new function.
10074 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10075 a subclass of rtl_opt_pass along with...
10076 (pass_data_lower_subreg): ...new pass_data instance and...
10077 (make_pass_lower_subreg): ...new function.
10078 (pass_lower_subreg2): Convert from a global struct to a subclass of
10079 rtl_opt_pass along with...
10080 (pass_data_lower_subreg2): ...new pass_data instance and...
10081 (make_pass_lower_subreg2): ...new function.
10082 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10083 struct to a subclass of ipa_opt_pass_d along with...
10084 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10085 (make_pass_ipa_lto_gimple_out): ...new function.
10086 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10087 of ipa_opt_pass_d along with...
10088 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10089 (make_pass_ipa_lto_finish_out): ...new function.
10090 * mode-switching.c (pass_mode_switching): Convert from a global struct
10091 to a subclass of rtl_opt_pass along with...
10092 (pass_data_mode_switching): ...new pass_data instance and...
10093 (make_pass_mode_switching): ...new function.
10094 * modulo-sched.c (pass_sms): Convert from a global struct to a
10095 subclass of rtl_opt_pass along with...
10096 (pass_data_sms): ...new pass_data instance and...
10097 (make_pass_sms): ...new function.
10098 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10099 subclass of gimple_opt_pass along with...
10100 (pass_data_expand_omp): ...new pass_data instance and...
10101 (make_pass_expand_omp): ...new function.
10102 (pass_lower_omp): Convert from a global struct to a subclass of
10103 gimple_opt_pass along with...
10104 (pass_data_lower_omp): ...new pass_data instance and...
10105 (make_pass_lower_omp): ...new function.
10106 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10107 of gimple_opt_pass along with...
10108 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10109 (make_pass_diagnose_omp_blocks): ...new function.
10110 * passes.c (pass_early_local_passes): Convert from a global struct to
10111 a subclass of simple_ipa_opt_pass along with...
10112 (pass_data_early_local_passes): ...new pass_data instance and...
10113 (make_pass_early_local_passes): ...new function.
10114 (pass_all_early_optimizations): Convert from a global struct to a
10115 subclass of gimple_opt_pass along with...
10116 (pass_data_all_early_optimizations): ...new pass_data instance and...
10117 (make_pass_all_early_optimizations): ...new function.
10118 (pass_all_optimizations): Convert from a global struct to a subclass
10119 of gimple_opt_pass along with...
10120 (pass_data_all_optimizations): ...new pass_data instance and...
10121 (make_pass_all_optimizations): ...new function.
10122 (pass_all_optimizations_g): Convert from a global struct to a subclass
10123 of gimple_opt_pass along with...
10124 (pass_data_all_optimizations_g): ...new pass_data instance and...
10125 (make_pass_all_optimizations_g): ...new function.
10126 (pass_rest_of_compilation): Convert from a global struct to a subclass
10127 of rtl_opt_pass along with...
10128 (pass_data_rest_of_compilation): ...new pass_data instance and...
10129 (make_pass_rest_of_compilation): ...new function.
10130 (pass_postreload): Convert from a global struct to a subclass of
10131 rtl_opt_pass along with...
10132 (pass_data_postreload): ...new pass_data instance and...
10133 (make_pass_postreload): ...new function.
10134 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10135 subclass of rtl_opt_pass along with...
10136 (pass_data_gcse2): ...new pass_data instance and...
10137 (make_pass_gcse2): ...new function.
10138 * postreload.c (pass_postreload_cse): Convert from a global struct to
10139 a subclass of rtl_opt_pass along with...
10140 (pass_data_postreload_cse): ...new pass_data instance and...
10141 (make_pass_postreload_cse): ...new function.
10142 * predict.c (pass_profile): Convert from a global struct to a subclass
10143 of gimple_opt_pass along with...
10144 (pass_data_profile): ...new pass_data instance and...
10145 (make_pass_profile): ...new function.
10146 (pass_strip_predict_hints): Convert from a global struct to a subclass
10147 of gimple_opt_pass along with...
10148 (pass_data_strip_predict_hints): ...new pass_data instance and...
10149 (make_pass_strip_predict_hints): ...new function.
10150 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10151 of rtl_opt_pass along with...
10152 (pass_data_peephole2): ...new pass_data instance and...
10153 (make_pass_peephole2): ...new function.
10154 (pass_split_all_insns): Convert from a global struct to a subclass of
10155 rtl_opt_pass along with...
10156 (pass_data_split_all_insns): ...new pass_data instance and...
10157 (make_pass_split_all_insns): ...new function.
10158 (pass_split_after_reload): Convert from a global struct to a subclass
10159 of rtl_opt_pass along with...
10160 (pass_data_split_after_reload): ...new pass_data instance and...
10161 (make_pass_split_after_reload): ...new function.
10162 (pass_split_before_regstack): Convert from a global struct to a
10163 subclass of rtl_opt_pass along with...
10164 (pass_data_split_before_regstack): ...new pass_data instance and...
10165 (make_pass_split_before_regstack): ...new function.
10166 (pass_split_before_sched2): Convert from a global struct to a subclass
10167 of rtl_opt_pass along with...
10168 (pass_data_split_before_sched2): ...new pass_data instance and...
10169 (make_pass_split_before_sched2): ...new function.
10170 (pass_split_for_shorten_branches): Convert from a global struct to a
10171 subclass of rtl_opt_pass along with...
10172 (pass_data_split_for_shorten_branches): ...new pass_data instance
10173 and...
10174 (make_pass_split_for_shorten_branches): ...new function.
10175 * ree.c (pass_ree): Convert from a global struct to a subclass of
10176 rtl_opt_pass along with...
10177 (pass_data_ree): ...new pass_data instance and...
10178 (make_pass_ree): ...new function.
10179 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10180 subclass of rtl_opt_pass along with...
10181 (pass_data_stack_regs): ...new pass_data instance and...
10182 (make_pass_stack_regs): ...new function.
10183 (pass_stack_regs_run): Convert from a global struct to a subclass of
10184 rtl_opt_pass along with...
10185 (pass_data_stack_regs_run): ...new pass_data instance and...
10186 (make_pass_stack_regs_run): ...new function.
10187 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10188 subclass of rtl_opt_pass along with...
10189 (pass_data_cprop_hardreg): ...new pass_data instance and...
10190 (make_pass_cprop_hardreg): ...new function.
10191 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10192 subclass of rtl_opt_pass along with...
10193 (pass_data_reginfo_init): ...new pass_data instance and...
10194 (make_pass_reginfo_init): ...new function.
10195 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10196 of rtl_opt_pass along with...
10197 (pass_data_regmove): ...new pass_data instance and...
10198 (make_pass_regmove): ...new function.
10199 * regrename.c (pass_regrename): Convert from a global struct to a
10200 subclass of rtl_opt_pass along with...
10201 (pass_data_regrename): ...new pass_data instance and...
10202 (make_pass_regrename): ...new function.
10203 * reorg.c (pass_delay_slots): Convert from a global struct to a
10204 subclass of rtl_opt_pass along with...
10205 (pass_data_delay_slots): ...new pass_data instance and...
10206 (make_pass_delay_slots): ...new function.
10207 (pass_machine_reorg): Convert from a global struct to a subclass of
10208 rtl_opt_pass along with...
10209 (pass_data_machine_reorg): ...new pass_data instance and...
10210 (make_pass_machine_reorg): ...new function.
10211 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10212 of rtl_opt_pass along with...
10213 (pass_data_sched): ...new pass_data instance and...
10214 (make_pass_sched): ...new function.
10215 (pass_sched2): Convert from a global struct to a subclass of
10216 rtl_opt_pass along with...
10217 (pass_data_sched2): ...new pass_data instance and...
10218 (make_pass_sched2): ...new function.
10219 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10220 to a subclass of rtl_opt_pass along with...
10221 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10222 (make_pass_stack_ptr_mod): ...new function.
10223 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10224 to a subclass of rtl_opt_pass along with...
10225 (pass_data_rtl_store_motion): ...new pass_data instance and...
10226 (make_pass_rtl_store_motion): ...new function.
10227 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10228 of gimple_opt_pass along with...
10229 (pass_data_tracer): ...new pass_data instance and...
10230 (make_pass_tracer): ...new function.
10231 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10232 to a subclass of gimple_opt_pass along with...
10233 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10234 (make_pass_diagnose_tm_blocks): ...new function.
10235 (pass_lower_tm): Convert from a global struct to a subclass of
10236 gimple_opt_pass along with...
10237 (pass_data_lower_tm): ...new pass_data instance and...
10238 (make_pass_lower_tm): ...new function.
10239 (pass_tm_init): Convert from a global struct to a subclass of
10240 gimple_opt_pass along with...
10241 (pass_data_tm_init): ...new pass_data instance and...
10242 (make_pass_tm_init): ...new function.
10243 (pass_tm_mark): Convert from a global struct to a subclass of
10244 gimple_opt_pass along with...
10245 (pass_data_tm_mark): ...new pass_data instance and...
10246 (make_pass_tm_mark): ...new function.
10247 (pass_tm_edges): Convert from a global struct to a subclass of
10248 gimple_opt_pass along with...
10249 (pass_data_tm_edges): ...new pass_data instance and...
10250 (make_pass_tm_edges): ...new function.
10251 (pass_tm_memopt): Convert from a global struct to a subclass of
10252 gimple_opt_pass along with...
10253 (pass_data_tm_memopt): ...new pass_data instance and...
10254 (make_pass_tm_memopt): ...new function.
10255 (pass_ipa_tm): Convert from a global struct to a subclass of
10256 simple_ipa_opt_pass along with...
10257 (pass_data_ipa_tm): ...new pass_data instance and...
10258 (make_pass_ipa_tm): ...new function.
10259 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10260 subclass of gimple_opt_pass along with...
10261 (pass_data_call_cdce): ...new pass_data instance and...
10262 (make_pass_call_cdce): ...new function.
10263 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10264 subclass of gimple_opt_pass along with...
10265 (pass_data_build_cfg): ...new pass_data instance and...
10266 (make_pass_build_cfg): ...new function.
10267 (pass_split_crit_edges): Convert from a global struct to a subclass of
10268 gimple_opt_pass along with...
10269 (pass_data_split_crit_edges): ...new pass_data instance and...
10270 (make_pass_split_crit_edges): ...new function.
10271 (pass_warn_function_return): Convert from a global struct to a
10272 subclass of gimple_opt_pass along with...
10273 (pass_data_warn_function_return): ...new pass_data instance and...
10274 (make_pass_warn_function_return): ...new function.
10275 (pass_warn_function_noreturn): Convert from a global struct to a
10276 subclass of gimple_opt_pass along with...
10277 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10278 (make_pass_warn_function_noreturn): ...new function.
10279 (pass_warn_unused_result): Convert from a global struct to a subclass
10280 of gimple_opt_pass along with...
10281 (pass_data_warn_unused_result): ...new pass_data instance and...
10282 (make_pass_warn_unused_result): ...new function.
10283 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10284 a subclass of gimple_opt_pass along with...
10285 (pass_data_merge_phi): ...new pass_data instance and...
10286 (make_pass_merge_phi): ...new function.
10287 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10288 a subclass of gimple_opt_pass along with...
10289 (pass_data_lower_complex): ...new pass_data instance and...
10290 (make_pass_lower_complex): ...new function.
10291 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10292 gimple_opt_pass along with...
10293 (pass_data_lower_complex_O0): ...new pass_data instance and...
10294 (make_pass_lower_complex_O0): ...new function.
10295 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10296 subclass of gimple_opt_pass along with...
10297 (pass_data_lower_eh): ...new pass_data instance and...
10298 (make_pass_lower_eh): ...new function.
10299 (pass_refactor_eh): Convert from a global struct to a subclass of
10300 gimple_opt_pass along with...
10301 (pass_data_refactor_eh): ...new pass_data instance and...
10302 (make_pass_refactor_eh): ...new function.
10303 (pass_lower_resx): Convert from a global struct to a subclass of
10304 gimple_opt_pass along with...
10305 (pass_data_lower_resx): ...new pass_data instance and...
10306 (make_pass_lower_resx): ...new function.
10307 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10308 of gimple_opt_pass along with...
10309 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10310 (make_pass_lower_eh_dispatch): ...new function.
10311 (pass_cleanup_eh): Convert from a global struct to a subclass of
10312 gimple_opt_pass along with...
10313 (pass_data_cleanup_eh): ...new pass_data instance and...
10314 (make_pass_cleanup_eh): ...new function.
10315 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10316 to a subclass of simple_ipa_opt_pass along with...
10317 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10318 (make_pass_ipa_lower_emutls): ...new function.
10319 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10320 a subclass of gimple_opt_pass along with...
10321 (pass_data_if_conversion): ...new pass_data instance and...
10322 (make_pass_if_conversion): ...new function.
10323 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10324 subclass of gimple_opt_pass along with...
10325 (pass_data_build_ssa): ...new pass_data instance and...
10326 (make_pass_build_ssa): ...new function.
10327 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10328 global struct to a subclass of gimple_opt_pass along with...
10329 (pass_data_loop_distribution): ...new pass_data instance and...
10330 (make_pass_loop_distribution): ...new function.
10331 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10332 subclass of gimple_opt_pass along with...
10333 (pass_data_mudflap_1): ...new pass_data instance and...
10334 (make_pass_mudflap_1): ...new function.
10335 (pass_mudflap_2): Convert from a global struct to a subclass of
10336 gimple_opt_pass along with...
10337 (pass_data_mudflap_2): ...new pass_data instance and...
10338 (make_pass_mudflap_2): ...new function.
10339 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10340 subclass of gimple_opt_pass along with...
10341 (pass_data_mudflap_1): ...new pass_data instance and...
10342 (make_pass_mudflap_1): ...new function.
10343 (pass_mudflap_2): Convert from a global struct to a subclass of
10344 gimple_opt_pass along with...
10345 (pass_data_mudflap_2): ...new pass_data instance and...
10346 (make_pass_mudflap_2): ...new function.
10347 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10348 gimple_opt_pass along with...
10349 (pass_data_nrv): ...new pass_data instance and...
10350 (make_pass_nrv): ...new function.
10351 (pass_return_slot): Convert from a global struct to a subclass of
10352 gimple_opt_pass along with...
10353 (pass_data_return_slot): ...new pass_data instance and...
10354 (make_pass_return_slot): ...new function.
10355 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10356 to a subclass of gimple_opt_pass along with...
10357 (pass_data_object_sizes): ...new pass_data instance and...
10358 (make_pass_object_sizes): ...new function.
10359 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10360 global struct to a subclass of gimple_opt_pass along with...
10361 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10362 and...
10363 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10364 (pass_fixup_cfg): Convert from a global struct to a subclass of
10365 gimple_opt_pass along with...
10366 (pass_data_fixup_cfg): ...new pass_data instance and...
10367 (make_pass_fixup_cfg): ...new function.
10368 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10369 (make_pass_mudflap_1): ...new function.
10370 (pass_mudflap_2): Replace declaration with that of...
10371 (make_pass_mudflap_2): ...new function.
10372 (pass_asan): Replace declaration with that of...
10373 (make_pass_asan): ...new function.
10374 (pass_asan_O0): Replace declaration with that of...
10375 (make_pass_asan_O0): ...new function.
10376 (pass_tsan): Replace declaration with that of...
10377 (make_pass_tsan): ...new function.
10378 (pass_tsan_O0): Replace declaration with that of...
10379 (make_pass_tsan_O0): ...new function.
10380 (pass_lower_cf): Replace declaration with that of...
10381 (make_pass_lower_cf): ...new function.
10382 (pass_refactor_eh): Replace declaration with that of...
10383 (make_pass_refactor_eh): ...new function.
10384 (pass_lower_eh): Replace declaration with that of...
10385 (make_pass_lower_eh): ...new function.
10386 (pass_lower_eh_dispatch): Replace declaration with that of...
10387 (make_pass_lower_eh_dispatch): ...new function.
10388 (pass_lower_resx): Replace declaration with that of...
10389 (make_pass_lower_resx): ...new function.
10390 (pass_build_cfg): Replace declaration with that of...
10391 (make_pass_build_cfg): ...new function.
10392 (pass_early_tree_profile): Replace declaration with that of...
10393 (make_pass_early_tree_profile): ...new function.
10394 (pass_cleanup_eh): Replace declaration with that of...
10395 (make_pass_cleanup_eh): ...new function.
10396 (pass_sra): Replace declaration with that of...
10397 (make_pass_sra): ...new function.
10398 (pass_sra_early): Replace declaration with that of...
10399 (make_pass_sra_early): ...new function.
10400 (pass_early_ipa_sra): Replace declaration with that of...
10401 (make_pass_early_ipa_sra): ...new function.
10402 (pass_tail_recursion): Replace declaration with that of...
10403 (make_pass_tail_recursion): ...new function.
10404 (pass_tail_calls): Replace declaration with that of...
10405 (make_pass_tail_calls): ...new function.
10406 (pass_tree_loop): Replace declaration with that of...
10407 (make_pass_tree_loop): ...new function.
10408 (pass_tree_loop_init): Replace declaration with that of...
10409 (make_pass_tree_loop_init): ...new function.
10410 (pass_lim): Replace declaration with that of...
10411 (make_pass_lim): ...new function.
10412 (pass_tree_unswitch): Replace declaration with that of...
10413 (make_pass_tree_unswitch): ...new function.
10414 (pass_predcom): Replace declaration with that of...
10415 (make_pass_predcom): ...new function.
10416 (pass_iv_canon): Replace declaration with that of...
10417 (make_pass_iv_canon): ...new function.
10418 (pass_scev_cprop): Replace declaration with that of...
10419 (make_pass_scev_cprop): ...new function.
10420 (pass_empty_loop): Replace declaration with that of...
10421 (make_pass_empty_loop): ...new function.
10422 (pass_record_bounds): Replace declaration with that of...
10423 (make_pass_record_bounds): ...new function.
10424 (pass_graphite): Replace declaration with that of...
10425 (make_pass_graphite): ...new function.
10426 (pass_graphite_transforms): Replace declaration with that of...
10427 (make_pass_graphite_transforms): ...new function.
10428 (pass_if_conversion): Replace declaration with that of...
10429 (make_pass_if_conversion): ...new function.
10430 (pass_loop_distribution): Replace declaration with that of...
10431 (make_pass_loop_distribution): ...new function.
10432 (pass_vectorize): Replace declaration with that of...
10433 (make_pass_vectorize): ...new function.
10434 (pass_slp_vectorize): Replace declaration with that of...
10435 (make_pass_slp_vectorize): ...new function.
10436 (pass_complete_unroll): Replace declaration with that of...
10437 (make_pass_complete_unroll): ...new function.
10438 (pass_complete_unrolli): Replace declaration with that of...
10439 (make_pass_complete_unrolli): ...new function.
10440 (pass_parallelize_loops): Replace declaration with that of...
10441 (make_pass_parallelize_loops): ...new function.
10442 (pass_loop_prefetch): Replace declaration with that of...
10443 (make_pass_loop_prefetch): ...new function.
10444 (pass_iv_optimize): Replace declaration with that of...
10445 (make_pass_iv_optimize): ...new function.
10446 (pass_tree_loop_done): Replace declaration with that of...
10447 (make_pass_tree_loop_done): ...new function.
10448 (pass_ch): Replace declaration with that of...
10449 (make_pass_ch): ...new function.
10450 (pass_ccp): Replace declaration with that of...
10451 (make_pass_ccp): ...new function.
10452 (pass_phi_only_cprop): Replace declaration with that of...
10453 (make_pass_phi_only_cprop): ...new function.
10454 (pass_build_ssa): Replace declaration with that of...
10455 (make_pass_build_ssa): ...new function.
10456 (pass_build_alias): Replace declaration with that of...
10457 (make_pass_build_alias): ...new function.
10458 (pass_build_ealias): Replace declaration with that of...
10459 (make_pass_build_ealias): ...new function.
10460 (pass_dominator): Replace declaration with that of...
10461 (make_pass_dominator): ...new function.
10462 (pass_dce): Replace declaration with that of...
10463 (make_pass_dce): ...new function.
10464 (pass_dce_loop): Replace declaration with that of...
10465 (make_pass_dce_loop): ...new function.
10466 (pass_cd_dce): Replace declaration with that of...
10467 (make_pass_cd_dce): ...new function.
10468 (pass_call_cdce): Replace declaration with that of...
10469 (make_pass_call_cdce): ...new function.
10470 (pass_merge_phi): Replace declaration with that of...
10471 (make_pass_merge_phi): ...new function.
10472 (pass_split_crit_edges): Replace declaration with that of...
10473 (make_pass_split_crit_edges): ...new function.
10474 (pass_pre): Replace declaration with that of...
10475 (make_pass_pre): ...new function.
10476 (pass_profile): Replace declaration with that of...
10477 (make_pass_profile): ...new function.
10478 (pass_strip_predict_hints): Replace declaration with that of...
10479 (make_pass_strip_predict_hints): ...new function.
10480 (pass_lower_complex_O0): Replace declaration with that of...
10481 (make_pass_lower_complex_O0): ...new function.
10482 (pass_lower_complex): Replace declaration with that of...
10483 (make_pass_lower_complex): ...new function.
10484 (pass_lower_vector): Replace declaration with that of...
10485 (make_pass_lower_vector): ...new function.
10486 (pass_lower_vector_ssa): Replace declaration with that of...
10487 (make_pass_lower_vector_ssa): ...new function.
10488 (pass_lower_omp): Replace declaration with that of...
10489 (make_pass_lower_omp): ...new function.
10490 (pass_diagnose_omp_blocks): Replace declaration with that of...
10491 (make_pass_diagnose_omp_blocks): ...new function.
10492 (pass_expand_omp): Replace declaration with that of...
10493 (make_pass_expand_omp): ...new function.
10494 (pass_expand_omp_ssa): Replace declaration with that of...
10495 (make_pass_expand_omp_ssa): ...new function.
10496 (pass_object_sizes): Replace declaration with that of...
10497 (make_pass_object_sizes): ...new function.
10498 (pass_strlen): Replace declaration with that of...
10499 (make_pass_strlen): ...new function.
10500 (pass_fold_builtins): Replace declaration with that of...
10501 (make_pass_fold_builtins): ...new function.
10502 (pass_stdarg): Replace declaration with that of...
10503 (make_pass_stdarg): ...new function.
10504 (pass_early_warn_uninitialized): Replace declaration with that of...
10505 (make_pass_early_warn_uninitialized): ...new function.
10506 (pass_late_warn_uninitialized): Replace declaration with that of...
10507 (make_pass_late_warn_uninitialized): ...new function.
10508 (pass_cse_reciprocals): Replace declaration with that of...
10509 (make_pass_cse_reciprocals): ...new function.
10510 (pass_cse_sincos): Replace declaration with that of...
10511 (make_pass_cse_sincos): ...new function.
10512 (pass_optimize_bswap): Replace declaration with that of...
10513 (make_pass_optimize_bswap): ...new function.
10514 (pass_optimize_widening_mul): Replace declaration with that of...
10515 (make_pass_optimize_widening_mul): ...new function.
10516 (pass_warn_function_return): Replace declaration with that of...
10517 (make_pass_warn_function_return): ...new function.
10518 (pass_warn_function_noreturn): Replace declaration with that of...
10519 (make_pass_warn_function_noreturn): ...new function.
10520 (pass_cselim): Replace declaration with that of...
10521 (make_pass_cselim): ...new function.
10522 (pass_phiopt): Replace declaration with that of...
10523 (make_pass_phiopt): ...new function.
10524 (pass_forwprop): Replace declaration with that of...
10525 (make_pass_forwprop): ...new function.
10526 (pass_phiprop): Replace declaration with that of...
10527 (make_pass_phiprop): ...new function.
10528 (pass_tree_ifcombine): Replace declaration with that of...
10529 (make_pass_tree_ifcombine): ...new function.
10530 (pass_dse): Replace declaration with that of...
10531 (make_pass_dse): ...new function.
10532 (pass_nrv): Replace declaration with that of...
10533 (make_pass_nrv): ...new function.
10534 (pass_rename_ssa_copies): Replace declaration with that of...
10535 (make_pass_rename_ssa_copies): ...new function.
10536 (pass_sink_code): Replace declaration with that of...
10537 (make_pass_sink_code): ...new function.
10538 (pass_fre): Replace declaration with that of...
10539 (make_pass_fre): ...new function.
10540 (pass_check_data_deps): Replace declaration with that of...
10541 (make_pass_check_data_deps): ...new function.
10542 (pass_copy_prop): Replace declaration with that of...
10543 (make_pass_copy_prop): ...new function.
10544 (pass_vrp): Replace declaration with that of...
10545 (make_pass_vrp): ...new function.
10546 (pass_uncprop): Replace declaration with that of...
10547 (make_pass_uncprop): ...new function.
10548 (pass_return_slot): Replace declaration with that of...
10549 (make_pass_return_slot): ...new function.
10550 (pass_reassoc): Replace declaration with that of...
10551 (make_pass_reassoc): ...new function.
10552 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10553 (make_pass_rebuild_cgraph_edges): ...new function.
10554 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10555 (make_pass_remove_cgraph_callee_edges): ...new function.
10556 (pass_build_cgraph_edges): Replace declaration with that of...
10557 (make_pass_build_cgraph_edges): ...new function.
10558 (pass_local_pure_const): Replace declaration with that of...
10559 (make_pass_local_pure_const): ...new function.
10560 (pass_tracer): Replace declaration with that of...
10561 (make_pass_tracer): ...new function.
10562 (pass_warn_unused_result): Replace declaration with that of...
10563 (make_pass_warn_unused_result): ...new function.
10564 (pass_diagnose_tm_blocks): Replace declaration with that of...
10565 (make_pass_diagnose_tm_blocks): ...new function.
10566 (pass_lower_tm): Replace declaration with that of...
10567 (make_pass_lower_tm): ...new function.
10568 (pass_tm_init): Replace declaration with that of...
10569 (make_pass_tm_init): ...new function.
10570 (pass_tm_mark): Replace declaration with that of...
10571 (make_pass_tm_mark): ...new function.
10572 (pass_tm_memopt): Replace declaration with that of...
10573 (make_pass_tm_memopt): ...new function.
10574 (pass_tm_edges): Replace declaration with that of...
10575 (make_pass_tm_edges): ...new function.
10576 (pass_split_functions): Replace declaration with that of...
10577 (make_pass_split_functions): ...new function.
10578 (pass_feedback_split_functions): Replace declaration with that of...
10579 (make_pass_feedback_split_functions): ...new function.
10580 (pass_strength_reduction): Replace declaration with that of...
10581 (make_pass_strength_reduction): ...new function.
10582 (pass_ipa_lower_emutls): Replace declaration with that of...
10583 (make_pass_ipa_lower_emutls): ...new function.
10584 (pass_ipa_function_and_variable_visibility): Replace declaration with
10585 that of...
10586 (make_pass_ipa_function_and_variable_visibility): ...new function.
10587 (pass_ipa_tree_profile): Replace declaration with that of...
10588 (make_pass_ipa_tree_profile): ...new function.
10589 (pass_early_local_passes): Replace declaration with that of...
10590 (make_pass_early_local_passes): ...new function.
10591 (pass_ipa_whole_program_visibility): Replace declaration with that
10592 of...
10593 (make_pass_ipa_whole_program_visibility): ...new function.
10594 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10595 (make_pass_ipa_lto_gimple_out): ...new function.
10596 (pass_ipa_increase_alignment): Replace declaration with that of...
10597 (make_pass_ipa_increase_alignment): ...new function.
10598 (pass_ipa_inline): Replace declaration with that of...
10599 (make_pass_ipa_inline): ...new function.
10600 (pass_ipa_free_lang_data): Replace declaration with that of...
10601 (make_pass_ipa_free_lang_data): ...new function.
10602 (pass_ipa_free_inline_summary): Replace declaration with that of...
10603 (make_pass_ipa_free_inline_summary): ...new function.
10604 (pass_ipa_cp): Replace declaration with that of...
10605 (make_pass_ipa_cp): ...new function.
10606 (pass_ipa_reference): Replace declaration with that of...
10607 (make_pass_ipa_reference): ...new function.
10608 (pass_ipa_pure_const): Replace declaration with that of...
10609 (make_pass_ipa_pure_const): ...new function.
10610 (pass_ipa_pta): Replace declaration with that of...
10611 (make_pass_ipa_pta): ...new function.
10612 (pass_ipa_lto_finish_out): Replace declaration with that of...
10613 (make_pass_ipa_lto_finish_out): ...new function.
10614 (pass_ipa_tm): Replace declaration with that of...
10615 (make_pass_ipa_tm): ...new function.
10616 (pass_ipa_profile): Replace declaration with that of...
10617 (make_pass_ipa_profile): ...new function.
10618 (pass_ipa_cdtor_merge): Replace declaration with that of...
10619 (make_pass_ipa_cdtor_merge): ...new function.
10620 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10621 of...
10622 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10623 (pass_init_datastructures): Replace declaration with that of...
10624 (make_pass_init_datastructures): ...new function.
10625 (pass_fixup_cfg): Replace declaration with that of...
10626 (make_pass_fixup_cfg): ...new function.
10627 (pass_expand): Replace declaration with that of...
10628 (make_pass_expand): ...new function.
10629 (pass_instantiate_virtual_regs): Replace declaration with that of...
10630 (make_pass_instantiate_virtual_regs): ...new function.
10631 (pass_rtl_fwprop): Replace declaration with that of...
10632 (make_pass_rtl_fwprop): ...new function.
10633 (pass_rtl_fwprop_addr): Replace declaration with that of...
10634 (make_pass_rtl_fwprop_addr): ...new function.
10635 (pass_jump): Replace declaration with that of...
10636 (make_pass_jump): ...new function.
10637 (pass_jump2): Replace declaration with that of...
10638 (make_pass_jump2): ...new function.
10639 (pass_lower_subreg): Replace declaration with that of...
10640 (make_pass_lower_subreg): ...new function.
10641 (pass_cse): Replace declaration with that of...
10642 (make_pass_cse): ...new function.
10643 (pass_fast_rtl_dce): Replace declaration with that of...
10644 (make_pass_fast_rtl_dce): ...new function.
10645 (pass_ud_rtl_dce): Replace declaration with that of...
10646 (make_pass_ud_rtl_dce): ...new function.
10647 (pass_rtl_dce): Replace declaration with that of...
10648 (make_pass_rtl_dce): ...new function.
10649 (pass_rtl_dse1): Replace declaration with that of...
10650 (make_pass_rtl_dse1): ...new function.
10651 (pass_rtl_dse2): Replace declaration with that of...
10652 (make_pass_rtl_dse2): ...new function.
10653 (pass_rtl_dse3): Replace declaration with that of...
10654 (make_pass_rtl_dse3): ...new function.
10655 (pass_rtl_cprop): Replace declaration with that of...
10656 (make_pass_rtl_cprop): ...new function.
10657 (pass_rtl_pre): Replace declaration with that of...
10658 (make_pass_rtl_pre): ...new function.
10659 (pass_rtl_hoist): Replace declaration with that of...
10660 (make_pass_rtl_hoist): ...new function.
10661 (pass_rtl_store_motion): Replace declaration with that of...
10662 (make_pass_rtl_store_motion): ...new function.
10663 (pass_cse_after_global_opts): Replace declaration with that of...
10664 (make_pass_cse_after_global_opts): ...new function.
10665 (pass_rtl_ifcvt): Replace declaration with that of...
10666 (make_pass_rtl_ifcvt): ...new function.
10667 (pass_into_cfg_layout_mode): Replace declaration with that of...
10668 (make_pass_into_cfg_layout_mode): ...new function.
10669 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10670 (make_pass_outof_cfg_layout_mode): ...new function.
10671 (pass_loop2): Replace declaration with that of...
10672 (make_pass_loop2): ...new function.
10673 (pass_rtl_loop_init): Replace declaration with that of...
10674 (make_pass_rtl_loop_init): ...new function.
10675 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10676 (make_pass_rtl_move_loop_invariants): ...new function.
10677 (pass_rtl_unswitch): Replace declaration with that of...
10678 (make_pass_rtl_unswitch): ...new function.
10679 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10680 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10681 (pass_rtl_doloop): Replace declaration with that of...
10682 (make_pass_rtl_doloop): ...new function.
10683 (pass_rtl_loop_done): Replace declaration with that of...
10684 (make_pass_rtl_loop_done): ...new function.
10685 (pass_web): Replace declaration with that of...
10686 (make_pass_web): ...new function.
10687 (pass_cse2): Replace declaration with that of...
10688 (make_pass_cse2): ...new function.
10689 (pass_df_initialize_opt): Replace declaration with that of...
10690 (make_pass_df_initialize_opt): ...new function.
10691 (pass_df_initialize_no_opt): Replace declaration with that of...
10692 (make_pass_df_initialize_no_opt): ...new function.
10693 (pass_reginfo_init): Replace declaration with that of...
10694 (make_pass_reginfo_init): ...new function.
10695 (pass_inc_dec): Replace declaration with that of...
10696 (make_pass_inc_dec): ...new function.
10697 (pass_stack_ptr_mod): Replace declaration with that of...
10698 (make_pass_stack_ptr_mod): ...new function.
10699 (pass_initialize_regs): Replace declaration with that of...
10700 (make_pass_initialize_regs): ...new function.
10701 (pass_combine): Replace declaration with that of...
10702 (make_pass_combine): ...new function.
10703 (pass_if_after_combine): Replace declaration with that of...
10704 (make_pass_if_after_combine): ...new function.
10705 (pass_ree): Replace declaration with that of...
10706 (make_pass_ree): ...new function.
10707 (pass_partition_blocks): Replace declaration with that of...
10708 (make_pass_partition_blocks): ...new function.
10709 (pass_match_asm_constraints): Replace declaration with that of...
10710 (make_pass_match_asm_constraints): ...new function.
10711 (pass_regmove): Replace declaration with that of...
10712 (make_pass_regmove): ...new function.
10713 (pass_split_all_insns): Replace declaration with that of...
10714 (make_pass_split_all_insns): ...new function.
10715 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10716 (make_pass_fast_rtl_byte_dce): ...new function.
10717 (pass_lower_subreg2): Replace declaration with that of...
10718 (make_pass_lower_subreg2): ...new function.
10719 (pass_mode_switching): Replace declaration with that of...
10720 (make_pass_mode_switching): ...new function.
10721 (pass_sms): Replace declaration with that of...
10722 (make_pass_sms): ...new function.
10723 (pass_sched): Replace declaration with that of...
10724 (make_pass_sched): ...new function.
10725 (pass_ira): Replace declaration with that of...
10726 (make_pass_ira): ...new function.
10727 (pass_reload): Replace declaration with that of...
10728 (make_pass_reload): ...new function.
10729 (pass_clean_state): Replace declaration with that of...
10730 (make_pass_clean_state): ...new function.
10731 (pass_branch_prob): Replace declaration with that of...
10732 (make_pass_branch_prob): ...new function.
10733 (pass_value_profile_transformations): Replace declaration with that
10734 of...
10735 (make_pass_value_profile_transformations): ...new function.
10736 (pass_postreload_cse): Replace declaration with that of...
10737 (make_pass_postreload_cse): ...new function.
10738 (pass_gcse2): Replace declaration with that of...
10739 (make_pass_gcse2): ...new function.
10740 (pass_split_after_reload): Replace declaration with that of...
10741 (make_pass_split_after_reload): ...new function.
10742 (pass_branch_target_load_optimize1): Replace declaration with that
10743 of...
10744 (make_pass_branch_target_load_optimize1): ...new function.
10745 (pass_thread_prologue_and_epilogue): Replace declaration with that
10746 of...
10747 (make_pass_thread_prologue_and_epilogue): ...new function.
10748 (pass_stack_adjustments): Replace declaration with that of...
10749 (make_pass_stack_adjustments): ...new function.
10750 (pass_peephole2): Replace declaration with that of...
10751 (make_pass_peephole2): ...new function.
10752 (pass_if_after_reload): Replace declaration with that of...
10753 (make_pass_if_after_reload): ...new function.
10754 (pass_regrename): Replace declaration with that of...
10755 (make_pass_regrename): ...new function.
10756 (pass_cprop_hardreg): Replace declaration with that of...
10757 (make_pass_cprop_hardreg): ...new function.
10758 (pass_reorder_blocks): Replace declaration with that of...
10759 (make_pass_reorder_blocks): ...new function.
10760 (pass_branch_target_load_optimize2): Replace declaration with that
10761 of...
10762 (make_pass_branch_target_load_optimize2): ...new function.
10763 (pass_leaf_regs): Replace declaration with that of...
10764 (make_pass_leaf_regs): ...new function.
10765 (pass_split_before_sched2): Replace declaration with that of...
10766 (make_pass_split_before_sched2): ...new function.
10767 (pass_compare_elim_after_reload): Replace declaration with that of...
10768 (make_pass_compare_elim_after_reload): ...new function.
10769 (pass_sched2): Replace declaration with that of...
10770 (make_pass_sched2): ...new function.
10771 (pass_stack_regs): Replace declaration with that of...
10772 (make_pass_stack_regs): ...new function.
10773 (pass_stack_regs_run): Replace declaration with that of...
10774 (make_pass_stack_regs_run): ...new function.
10775 (pass_df_finish): Replace declaration with that of...
10776 (make_pass_df_finish): ...new function.
10777 (pass_compute_alignments): Replace declaration with that of...
10778 (make_pass_compute_alignments): ...new function.
10779 (pass_duplicate_computed_gotos): Replace declaration with that of...
10780 (make_pass_duplicate_computed_gotos): ...new function.
10781 (pass_variable_tracking): Replace declaration with that of...
10782 (make_pass_variable_tracking): ...new function.
10783 (pass_free_cfg): Replace declaration with that of...
10784 (make_pass_free_cfg): ...new function.
10785 (pass_machine_reorg): Replace declaration with that of...
10786 (make_pass_machine_reorg): ...new function.
10787 (pass_cleanup_barriers): Replace declaration with that of...
10788 (make_pass_cleanup_barriers): ...new function.
10789 (pass_delay_slots): Replace declaration with that of...
10790 (make_pass_delay_slots): ...new function.
10791 (pass_split_for_shorten_branches): Replace declaration with that of...
10792 (make_pass_split_for_shorten_branches): ...new function.
10793 (pass_split_before_regstack): Replace declaration with that of...
10794 (make_pass_split_before_regstack): ...new function.
10795 (pass_convert_to_eh_region_ranges): Replace declaration with that
10796 of...
10797 (make_pass_convert_to_eh_region_ranges): ...new function.
10798 (pass_shorten_branches): Replace declaration with that of...
10799 (make_pass_shorten_branches): ...new function.
10800 (pass_set_nothrow_function_flags): Replace declaration with that of...
10801 (make_pass_set_nothrow_function_flags): ...new function.
10802 (pass_dwarf2_frame): Replace declaration with that of...
10803 (make_pass_dwarf2_frame): ...new function.
10804 (pass_final): Replace declaration with that of...
10805 (make_pass_final): ...new function.
10806 (pass_rtl_seqabstr): Replace declaration with that of...
10807 (make_pass_rtl_seqabstr): ...new function.
10808 (pass_release_ssa_names): Replace declaration with that of...
10809 (make_pass_release_ssa_names): ...new function.
10810 (pass_early_inline): Replace declaration with that of...
10811 (make_pass_early_inline): ...new function.
10812 (pass_inline_parameters): Replace declaration with that of...
10813 (make_pass_inline_parameters): ...new function.
10814 (pass_update_address_taken): Replace declaration with that of...
10815 (make_pass_update_address_taken): ...new function.
10816 (pass_convert_switch): Replace declaration with that of...
10817 (make_pass_convert_switch): ...new function.
10818 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10819 to a subclass of simple_ipa_opt_pass along with...
10820 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10821 (make_pass_ipa_tree_profile): ...new function.
10822 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10823 subclass of gimple_opt_pass along with...
10824 (pass_data_sra_early): ...new pass_data instance and...
10825 (make_pass_sra_early): ...new function.
10826 (pass_sra): Convert from a global struct to a subclass of
10827 gimple_opt_pass along with...
10828 (pass_data_sra): ...new pass_data instance and...
10829 (make_pass_sra): ...new function.
10830 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10831 gimple_opt_pass along with...
10832 (pass_data_early_ipa_sra): ...new pass_data instance and...
10833 (make_pass_early_ipa_sra): ...new function.
10834 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10835 subclass of gimple_opt_pass along with...
10836 (pass_data_ccp): ...new pass_data instance and...
10837 (make_pass_ccp): ...new function.
10838 (pass_fold_builtins): Convert from a global struct to a subclass of
10839 gimple_opt_pass along with...
10840 (pass_data_fold_builtins): ...new pass_data instance and...
10841 (make_pass_fold_builtins): ...new function.
10842 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10843 subclass of gimple_opt_pass along with...
10844 (pass_data_copy_prop): ...new pass_data instance and...
10845 (make_pass_copy_prop): ...new function.
10846 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10847 global struct to a subclass of gimple_opt_pass along with...
10848 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10849 (make_pass_rename_ssa_copies): ...new function.
10850 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10851 subclass of gimple_opt_pass along with...
10852 (pass_data_dce): ...new pass_data instance and...
10853 (make_pass_dce): ...new function.
10854 (pass_dce_loop): Convert from a global struct to a subclass of
10855 gimple_opt_pass along with...
10856 (pass_data_dce_loop): ...new pass_data instance and...
10857 (make_pass_dce_loop): ...new function.
10858 (pass_cd_dce): Convert from a global struct to a subclass of
10859 gimple_opt_pass along with...
10860 (pass_data_cd_dce): ...new pass_data instance and...
10861 (make_pass_cd_dce): ...new function.
10862 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10863 subclass of gimple_opt_pass along with...
10864 (pass_data_dominator): ...new pass_data instance and...
10865 (make_pass_dominator): ...new function.
10866 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10867 gimple_opt_pass along with...
10868 (pass_data_phi_only_cprop): ...new pass_data instance and...
10869 (make_pass_phi_only_cprop): ...new function.
10870 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10871 subclass of gimple_opt_pass along with...
10872 (pass_data_dse): ...new pass_data instance and...
10873 (make_pass_dse): ...new function.
10874 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10875 a subclass of gimple_opt_pass along with...
10876 (pass_data_forwprop): ...new pass_data instance and...
10877 (make_pass_forwprop): ...new function.
10878 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10879 struct to a subclass of gimple_opt_pass along with...
10880 (pass_data_tree_ifcombine): ...new pass_data instance and...
10881 (make_pass_tree_ifcombine): ...new function.
10882 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10883 subclass of gimple_opt_pass along with...
10884 (pass_data_ch): ...new pass_data instance and...
10885 (make_pass_ch): ...new function.
10886 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10887 subclass of gimple_opt_pass along with...
10888 (pass_data_tree_loop): ...new pass_data instance and...
10889 (make_pass_tree_loop): ...new function.
10890 (pass_tree_loop_init): Convert from a global struct to a subclass of
10891 gimple_opt_pass along with...
10892 (pass_data_tree_loop_init): ...new pass_data instance and...
10893 (make_pass_tree_loop_init): ...new function.
10894 (pass_lim): Convert from a global struct to a subclass of
10895 gimple_opt_pass along with...
10896 (pass_data_lim): ...new pass_data instance and...
10897 (make_pass_lim): ...new function.
10898 (pass_tree_unswitch): Convert from a global struct to a subclass of
10899 gimple_opt_pass along with...
10900 (pass_data_tree_unswitch): ...new pass_data instance and...
10901 (make_pass_tree_unswitch): ...new function.
10902 (pass_predcom): Convert from a global struct to a subclass of
10903 gimple_opt_pass along with...
10904 (pass_data_predcom): ...new pass_data instance and...
10905 (make_pass_predcom): ...new function.
10906 (pass_vectorize): Convert from a global struct to a subclass of
10907 gimple_opt_pass along with...
10908 (pass_data_vectorize): ...new pass_data instance and...
10909 (make_pass_vectorize): ...new function.
10910 (pass_graphite): Convert from a global struct to a subclass of
10911 gimple_opt_pass along with...
10912 (pass_data_graphite): ...new pass_data instance and...
10913 (make_pass_graphite): ...new function.
10914 (pass_graphite_transforms): Convert from a global struct to a subclass
10915 of gimple_opt_pass along with...
10916 (pass_data_graphite_transforms): ...new pass_data instance and...
10917 (make_pass_graphite_transforms): ...new function.
10918 (pass_check_data_deps): Convert from a global struct to a subclass of
10919 gimple_opt_pass along with...
10920 (pass_data_check_data_deps): ...new pass_data instance and...
10921 (make_pass_check_data_deps): ...new function.
10922 (pass_iv_canon): Convert from a global struct to a subclass of
10923 gimple_opt_pass along with...
10924 (pass_data_iv_canon): ...new pass_data instance and...
10925 (make_pass_iv_canon): ...new function.
10926 (pass_scev_cprop): Convert from a global struct to a subclass of
10927 gimple_opt_pass along with...
10928 (pass_data_scev_cprop): ...new pass_data instance and...
10929 (make_pass_scev_cprop): ...new function.
10930 (pass_record_bounds): Convert from a global struct to a subclass of
10931 gimple_opt_pass along with...
10932 (pass_data_record_bounds): ...new pass_data instance and...
10933 (make_pass_record_bounds): ...new function.
10934 (pass_complete_unroll): Convert from a global struct to a subclass of
10935 gimple_opt_pass along with...
10936 (pass_data_complete_unroll): ...new pass_data instance and...
10937 (make_pass_complete_unroll): ...new function.
10938 (pass_complete_unrolli): Convert from a global struct to a subclass of
10939 gimple_opt_pass along with...
10940 (pass_data_complete_unrolli): ...new pass_data instance and...
10941 (make_pass_complete_unrolli): ...new function.
10942 (pass_parallelize_loops): Convert from a global struct to a subclass
10943 of gimple_opt_pass along with...
10944 (pass_data_parallelize_loops): ...new pass_data instance and...
10945 (make_pass_parallelize_loops): ...new function.
10946 (pass_loop_prefetch): Convert from a global struct to a subclass of
10947 gimple_opt_pass along with...
10948 (pass_data_loop_prefetch): ...new pass_data instance and...
10949 (make_pass_loop_prefetch): ...new function.
10950 (pass_iv_optimize): Convert from a global struct to a subclass of
10951 gimple_opt_pass along with...
10952 (pass_data_iv_optimize): ...new pass_data instance and...
10953 (make_pass_iv_optimize): ...new function.
10954 (pass_tree_loop_done): Convert from a global struct to a subclass of
10955 gimple_opt_pass along with...
10956 (pass_data_tree_loop_done): ...new pass_data instance and...
10957 (make_pass_tree_loop_done): ...new function.
10958 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10959 struct to a subclass of gimple_opt_pass along with...
10960 (pass_data_cse_reciprocals): ...new pass_data instance and...
10961 (make_pass_cse_reciprocals): ...new function.
10962 (pass_cse_sincos): Convert from a global struct to a subclass of
10963 gimple_opt_pass along with...
10964 (pass_data_cse_sincos): ...new pass_data instance and...
10965 (make_pass_cse_sincos): ...new function.
10966 (pass_optimize_bswap): Convert from a global struct to a subclass of
10967 gimple_opt_pass along with...
10968 (pass_data_optimize_bswap): ...new pass_data instance and...
10969 (make_pass_optimize_bswap): ...new function.
10970 (pass_optimize_widening_mul): Convert from a global struct to a
10971 subclass of gimple_opt_pass along with...
10972 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10973 (make_pass_optimize_widening_mul): ...new function.
10974 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10975 subclass of gimple_opt_pass along with...
10976 (pass_data_phiopt): ...new pass_data instance and...
10977 (make_pass_phiopt): ...new function.
10978 (pass_cselim): Convert from a global struct to a subclass of
10979 gimple_opt_pass along with...
10980 (pass_data_cselim): ...new pass_data instance and...
10981 (make_pass_cselim): ...new function.
10982 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10983 subclass of gimple_opt_pass along with...
10984 (pass_data_phiprop): ...new pass_data instance and...
10985 (make_pass_phiprop): ...new function.
10986 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10987 subclass of gimple_opt_pass along with...
10988 (pass_data_pre): ...new pass_data instance and...
10989 (make_pass_pre): ...new function.
10990 (pass_fre): Convert from a global struct to a subclass of
10991 gimple_opt_pass along with...
10992 (pass_data_fre): ...new pass_data instance and...
10993 (make_pass_fre): ...new function.
10994 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10995 subclass of gimple_opt_pass along with...
10996 (pass_data_reassoc): ...new pass_data instance and...
10997 (make_pass_reassoc): ...new function.
10998 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10999 subclass of gimple_opt_pass along with...
11000 (pass_data_sink_code): ...new pass_data instance and...
11001 (make_pass_sink_code): ...new function.
11002 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11003 subclass of gimple_opt_pass along with...
11004 (pass_data_strlen): ...new pass_data instance and...
11005 (make_pass_strlen): ...new function.
11006 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11007 struct to a subclass of gimple_opt_pass along with...
11008 (pass_data_build_alias): ...new pass_data instance and...
11009 (make_pass_build_alias): ...new function.
11010 (pass_build_ealias): Convert from a global struct to a subclass of
11011 gimple_opt_pass along with...
11012 (pass_data_build_ealias): ...new pass_data instance and...
11013 (make_pass_build_ealias): ...new function.
11014 (pass_ipa_pta): Convert from a global struct to a subclass of
11015 simple_ipa_opt_pass along with...
11016 (pass_data_ipa_pta): ...new pass_data instance and...
11017 (make_pass_ipa_pta): ...new function.
11018 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11019 subclass of gimple_opt_pass along with...
11020 (pass_data_uncprop): ...new pass_data instance and...
11021 (make_pass_uncprop): ...new function.
11022 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11023 global struct to a subclass of gimple_opt_pass along with...
11024 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11025 (make_pass_late_warn_uninitialized): ...new function.
11026 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11027 to a subclass of gimple_opt_pass along with...
11028 (pass_data_init_datastructures): ...new pass_data instance and...
11029 (make_pass_init_datastructures): ...new function.
11030 (pass_early_warn_uninitialized): Convert from a global struct to a
11031 subclass of gimple_opt_pass along with...
11032 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11033 (make_pass_early_warn_uninitialized): ...new function.
11034 (pass_update_address_taken): Convert from a global struct to a
11035 subclass of gimple_opt_pass along with...
11036 (pass_data_update_address_taken): ...new pass_data instance and...
11037 (make_pass_update_address_taken): ...new function.
11038 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11039 struct to a subclass of gimple_opt_pass along with...
11040 (pass_data_release_ssa_names): ...new pass_data instance and...
11041 (make_pass_release_ssa_names): ...new function.
11042 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11043 subclass of gimple_opt_pass along with...
11044 (pass_data_stdarg): ...new pass_data instance and...
11045 (make_pass_stdarg): ...new function.
11046 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11047 global struct to a subclass of gimple_opt_pass along with...
11048 (pass_data_convert_switch): ...new pass_data instance and...
11049 (make_pass_convert_switch): ...new function.
11050 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11051 to a subclass of gimple_opt_pass along with...
11052 (pass_data_tail_recursion): ...new pass_data instance and...
11053 (make_pass_tail_recursion): ...new function.
11054 (pass_tail_calls): Convert from a global struct to a subclass of
11055 gimple_opt_pass along with...
11056 (pass_data_tail_calls): ...new pass_data instance and...
11057 (make_pass_tail_calls): ...new function.
11058 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11059 struct to a subclass of gimple_opt_pass along with...
11060 (pass_data_lower_vector): ...new pass_data instance and...
11061 (make_pass_lower_vector): ...new function.
11062 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11063 gimple_opt_pass along with...
11064 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11065 (make_pass_lower_vector_ssa): ...new function.
11066 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11067 to a subclass of gimple_opt_pass along with...
11068 (pass_data_slp_vectorize): ...new pass_data instance and...
11069 (make_pass_slp_vectorize): ...new function.
11070 (pass_ipa_increase_alignment): Convert from a global struct to a
11071 subclass of simple_ipa_opt_pass along with...
11072 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11073 (make_pass_ipa_increase_alignment): ...new function.
11074 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11075 gimple_opt_pass along with...
11076 (pass_data_vrp): ...new pass_data instance and...
11077 (make_pass_vrp): ...new function.
11078 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11079 subclass of simple_ipa_opt_pass along with...
11080 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11081 (make_pass_ipa_free_lang_data): ...new function.
11082 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11083 gimple_opt_pass along with...
11084 (pass_data_tsan): ...new pass_data instance and...
11085 (make_pass_tsan): ...new function.
11086 (pass_tsan_O0): Convert from a global struct to a subclass of
11087 gimple_opt_pass along with...
11088 (pass_data_tsan_O0): ...new pass_data instance and...
11089 (make_pass_tsan_O0): ...new function.
11090 * var-tracking.c (pass_variable_tracking): Convert from a global
11091 struct to a subclass of rtl_opt_pass along with...
11092 (pass_data_variable_tracking): ...new pass_data instance and...
11093 (make_pass_variable_tracking): ...new function.
11094 * web.c (pass_web): Convert from a global struct to a subclass of
11095 rtl_opt_pass along with...
11096 (pass_data_web): ...new pass_data instance and...
11097 (make_pass_web): ...new function.
11098 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11099 declaration with that of...
11100 (make_pass_mode_switch_use): ...new function.
11101 (pass_resolve_sw_modes): Replace declaration with that of...
11102 (make_pass_resolve_sw_modes): ...new function.
11103 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11104 from a global struct to a subclass of rtl_opt_pass along with...
11105 (pass_data_mode_switch_use): ...new pass_data instance and...
11106 (make_pass_mode_switch_use): ...new function.
11107 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11108 from a global struct to a subclass of rtl_opt_pass along with...
11109 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11110 (make_pass_resolve_sw_modes): ...new function.
11111 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11112 struct to a subclass of rtl_opt_pass along with...
11113 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11114 (make_pass_insert_vzeroupper): ...new function.
11115 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11116 global struct to a subclass of rtl_opt_pass along with...
11117 (pass_data_work_around_errata): ...new pass_data instance and...
11118 (make_pass_work_around_errata): ...new function.
11119 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11120 struct to a subclass of rtl_opt_pass along with...
11121 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11122 (make_pass_mips_machine_reorg2): ...new function.
11123
6a389ed5
DM		 111242013-08-05 David Malcolm <dmalcolm@redhat.com>
11125
11126 * passes.c (pass_manager::operator new): New.
11127
f7695dbf
DM		 111282013-08-05 David Malcolm <dmalcolm@redhat.com>
11129
11130 Handwritten part of conversion of passes to C++ classes.
11131
11132 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11133 (toplev.o): Add dep on PASS_MANAGER_H.
11134 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11135 of early local pases to reflect this moving from a global to a
11136 member of gcc::pass_manager.
11137 (cgraph_add_new_function): Likewise.
11138 * lto-cgraph.c (lto_output_node): Update for conversion of
11139 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11140 * passes.c (opt_pass::clone): New.
11141 (opt_pass::gate): New.
11142 (opt_pass::execute): New.
11143 (opt_pass::opt_pass): New.
11144 (pass_manager::execute_early_local_passes): New.
11145 (pass_manager::execute_pass_mode_switching): new.
11146 (finish_optimization_passes): Convert to...
11147 (pass_manager::finish_optimization_passes): ...this.
11148 (finish_optimization_passes): Update for conversion of passes to
11149 C++ classes.
11150 (register_dump_files_1): Use has_gate since we cannot portably
11151 check a vtable entry against NULL.
11152 (dump_one_pass): Likewise.
11153 (ipa_write_summaries_2): Likewise.
11154 (ipa_write_optimization_summaries_1): Likewise.
11155 (ipa_read_summaries_1): Likewise.
11156 (ipa_read_optimization_summaries_1): Likewise.
11157 (execute_ipa_stmt_fixups): Likewise.
11158 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11159 pass-creation functions rather than wiring up globals, and
11160 storing the results in fields of pass_manager generated using
11161 pass-instances.def.
11162 (pass_manager::dump_profile_report): Update for conversion of
11163 passes to C++ classes.
11164 (pass_manager::execute_ipa_summary_passes): Likewise.
11165 (execute_one_ipa_transform_pass): Likewise.
11166 (execute_one_pass): Use has_gate and has_execute since we cannot
11167 portably check a vtable entry against NULL.
11168 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11169 (pass_manager): Use pass-instances.def to add fields for the
11170 various pass instances.
11171 * toplev.c (finalize): Update for move of
11172 finish_optimization_passes to a method of gcc::pass_manager.
11173 * toplev.h (finish_optimization_passes): Move to method of class
11174 pass_manager.
11175 * tree-pass.h (struct pass_data): New.
03b0ee0a 11176 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM		 11177 (opt_pass::gate): Convert to virtual function.
11178 (opt_pass::~opt_pass): New.
11179 (opt_pass::clone): New.
11180 (opt_pass::execute): Convert to virtual function.
11181 (opt_pass::opt_pass): New.
11182 (opt_pass::ctxt_): new.
11183 (gimple_opt_pass): Convert to subclass of opt_pass.
11184 (gimple_opt_pass::gimple_opt_pass): New.
11185 (rtl_opt_pass): Convert to subclass of opt_pass.
11186 (rtl_opt_pass::rtl_opt_pass): New.
11187 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11188 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11189 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11190 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11191 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11192 invocation of pass_mode_switching to reflect this moving from a
11193 global to a member of gcc::pass_manager.
11194 (ix86_option_override): Rework how pass_insert_vzeroupper is
11195 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 11196 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 11197
d0b2f831
RE		 111982013-08-05 Richard Earnshaw <rearnsha@arm.com>
11199
11200 PR rtl-optimization/57708
11201 * recog.c (peep2_find_free_register): Validate all regs in a
11202 multi-reg mode.
11203
51a5c0c2
JH		 112042013-08-05 Jan Hubicka <jh@suse.cz>
11205
11206 PR lto/57602
03b0ee0a
UB		 11207 * cgraph.c (verify_cgraph_node): Accept local flags from other
11208 partitions.
51a5c0c2
JH		 11209 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11210 (function_and_variable_visibility): Likewise.
11211 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11212
65f0a120
GDR		 112132013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11214
11215 * graph.c (init_graph_slim_pretty_print): Remove.
11216 (print_graph_cfg): Do not call it. Use local pretty printer.
11217 (start_graph_dump): Likewise.
11218
b3f80694
GDR		 112192013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11220
11221 * gimple-pretty-print.c (buffer): Remove.
11222 (initialized): Likewise.
11223 (maybe_init_pretty_print): Likewise.
11224 (print_gimple_stmt): Do not call it. Use non-static local
11225 pretty_printer variable.
11226 (print_gimple_expr): Likewise.
11227 (print_gimple_seq): Likewise.
11228 (gimple_dump_bb): Likewise.
11229
11a877b3
GDR		 112302013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11231
11232 * asan.c (asan_pp): Remove.
11233 (asan_pp_initialized): Likewise.
11234 (asan_pp_initialize): Likewise.
11235 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11236 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11237 (asan_add_global): Likewise.
11238
b066401f
GDR		 112392013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11240
11241 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11242 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11243 * pretty-print.c (pp_formatted_text_data): Likewise.
11244 (pp_indent): Rename from pp_base_indent.
11245 (pp_format): Rename from pp_base_format.
11246 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11247 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11248 (pp_flush): Rename from pp_base_flush.
11249 (pp_set_line_maximum_length): Rename from
11250 pp_base_set_line_maximum_length.
11251 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11252 (pp_set_prefix): Rename from pp_base_set_prefix.
11253 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11254 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11255 (pp_append_text): Rename from pp_base_append_text.
11256 (pp_formatted_text): Rename from pp_base_formatted_text.
11257 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11258 (pp_remaining_character_count_for_line): Rename from
11259 pp_base_remaining_character_count_for_line.
11260 (pp_newline): Rename from pp_base_newline.
11261 (pp_character): Rename from pp_base_character.
11262 (pp_string): Rename from pp_base_string.
11263 (pp_maybe_space): Rename from pp_base_maybe_space.
11264 * asan.c (asan_pp_string): Adjust.
11265 (asan_emit_stack_protection): Likewise.
11266 (asan_add_global): Likewise.
11267 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11268 * tree-mudflap.c (mf_varname_tree): Likewise.
11269 * tree-pretty-print.c (pp_tree_identifier): Rename from
11270 pp_base_tree_identifier.
11271 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11272 Declare as function.
11273
137a1a27
GDR		 112742013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11275
11276 * pretty-print.h (pp_bar_bar): New.
11277 (pp_ampersand_ampersand): Likewise.
11278 (pp_less_equal): Likewise.
11279 (pp_greater_equal): Likewise.
11280 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11281 printer functions instead of pp_string or operators and punctuators.
11282 (dump_gimple_call): Likewise.
11283 (dump_gimple_omp_for): Likewise.
11284 (dump_gimple_transaction): Likewise.
11285 (dump_gimple_phi): Likewise.
11286 (pp_gimple_stmt_1): Likewise.
11287 * sched-vis.c (print_insn): Likewise.
11288 * tree-mudflap.c (mf_varname_tree): Likewise.
11289 * tree-pretty-print.c (dump_block_node): Likewise.
11290 (dump_generic_node): Likewise.
11291
815effe1
JH		 112922013-08-02 Jan Hubicka <jh@suse.cz>
11293
11294 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11295 boundaries.
11296 * lto-streamer-out.c (tree_is_indexable): Results decls and
11297 parm decls are not indexable.
11298 (DFS_write_tree_body): Do not follow args and results.
11299 (hash_tree): Likewise.
11300 (output_functions): Rearrange so struct function is needed
11301 only when real body is output; be able to also ouptut abstract
11302 functions; output DECL_ARGUMENTS and DECL_RESULT.
11303 (lto_output): When not in WPA, ale store abstract functions.
11304 (write_symbol): Do not care about RESULT_DECL.
11305 (output_symbol_p): Handle correctly sbtract decls.
11306 * lto-streamer-in.c (input_function): Rearrange so struct
11307 function can be NULL at entry; allow streaming of
11308 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11309 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11310 sanity check during LTO.
11311 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11312 RESULT_DECl and DECL_ARGUMENTS.
11313 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11314 Likewise.
11315
07838b13
GDR		 113162013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11317
11318 * pretty-print.h (pp_underscore): New.
11319 (pp_comma): Tidy.
11320 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11321 printer functions instead of pp_character.
11322 (dump_binary_rhs): Likewise.
11323 (dump_ternary_rhs): Likewise.
11324 (dump_gimple_call_args): Likewise.
11325 (pp_points_to_solution): Likewise.
11326 (dump_gimple_call): Likewise.
11327 (dump_gimple_switch): Likewise.
11328 (dump_gimple_cond): Likewise.
11329 (dump_gimple_bind): Likewise.
11330 (dump_gimple_try): Likewise.
11331 (dump_gimple_omp_for): Likewise.
11332 (dump_gimple_omp_continue): Likewise.
11333 (dump_gimple_omp_single): Likewise.
11334 (dump_gimple_omp_sections): Likewise.
11335 (dump_gimple_omp_block): Likewise.
11336 (dump_gimple_omp_critical): Likewise.
11337 (dump_gimple_transaction): Likewise.
11338 (dump_gimple_asm): Likewise.
11339 (dump_gimple_phi): Likewise.
11340 (dump_gimple_omp_parallel): Likewise.
11341 (dump_gimple_omp_task): Likewise.
11342 (dump_gimple_omp_atomic_load): Likewise.
11343 (dump_gimple_omp_atomic_store): Likewise.
11344 (dump_gimple_mem_ops): Likewise.
11345 (pp_gimple_stmt_1): Likewise.
11346 (pp_cfg_jump): Likewise.
11347 (dump_implicit_edges): Likewise.
11348 (gimple_dump_bb_for_graph): Likewise.
11349 * graph.c (draw_cfg_node): Likewise.
11350 * langhooks.c (lhd_print_error_function): Likewise.
11351 * sched-vis.c (print_exp): Likewise.
11352 (print_value): Likewise.
11353 (print_pattern): Likewise.
11354 (print_insn): Likewise.
11355 (rtl_dump_bb_for_graph): Likewise.
11356 * tree-pretty-print.c (dump_function_declaration): Likewise.
11357 (dump_array_domain): Likewise.
11358 (dump_omp_clause): Likewise.
11359 (dump_location): Likewise.
11360 (dump_generic_node): Likewise.
11361 (print_struct_decl): Likewise.
11362 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11363
bb0d2039
BS		 113642013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11365
11366 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11367 candidate table when replacing a candidate statement.
11368 (replace_rhs_if_not_dup): Likewise.
11369 (replace_one_candidate): Likewise.
11370
bc0ec027 113712013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 11372 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH		 11373
11374 * cgraphunit.c (add_new_function): Fix logic when adding from
11375 late IPA pass.
11376 (assemble_thunk): Rename to ...
03b0ee0a
UB		 11377 (expand_thunk); .. this one; export; get it working with
11378 general functions; make produced gimple valid.
bc0ec027
JH		 11379 * cgraph.h (expand_thunk): Declare.
11380
0e8853ee
JH		 113812013-08-02 Jan Hubicka <jh@suse.cz>
11382
03b0ee0a
UB		 11383 * ipa-cp.c (gather_context_independent_values): Use
11384 ipa_get_param_move_cost.
11385 (get_replacement_map): Remove PARAM; move parameter folding
11386 into tree-inline.c
0e8853ee
JH		 11387 (create_specialized_node): Update.
11388 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11389 assert that we have gimple body; update move_cost.
11390 (count_formal_params): Assert that we have gimple body.
11391 (ipa_dump_param): New function.
11392 (ipa_alloc_node_params): Break out from ...
11393 (ipa_initialize_node_params): ... here.
11394 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11395 (ipa_write_node_info): Stream move costs.
11396 (ipa_read_node_info): Read move costs.
11397 (ipa_update_after_lto_read): Do not recompute node params.
11398 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11399 (ipa_get_param): Check we are not in WPA.
11400 (ipa_get_param_move_cost): New.
11401 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11402 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11403 parm numbers to be present.
11404
ecee672b
VM		 114052013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11406
11407 PR rtl-optimization/58048
11408 * lra-constraints.c (process_alt_operands): Don't check asm
11409 operand on register.
11410
13ec0527
EB		 114112013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11412
11413 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11414 the implied StoreLoad barrier for atomic operations if before.
11415
af15184a
JH		 114162013-08-02 Jan Hubicka <jh@suse.cz>
11417 Martin Liska <marxin.liska@gmail.com>
11418
03b0ee0a
UB		 11419 * cgraph.c (cgraph_function_body_availability): Do not check
11420 cgraph flags.
af15184a
JH		 11421 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11422 symtab_node_availability): Declare.
11423 * ipa.c (can_replace_by_local_alias): New.
11424 (function_and_variable_visibility): Use it.
03b0ee0a
UB		 11425 * symtab.c (symtab_for_node_and_aliases,
11426 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 11427
01e54ef8
VM		 114282013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11429
11430 PR rtl-optimization/57963
03b0ee0a 11431 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM		 11432 (lra_constraints): Use them.
11433
e2fd7ca7
SN		 114342013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11435
03b0ee0a
UB		 11436 * config/arm/types.md (define_attr "type"): Add "load_acq"
11437 and "store_rel".
e2fd7ca7
SN		 11438 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11439 changes.
11440 (cortex_a53_store1): Likewise.
11441
0dcc26c3
JH		 114422013-08-01 Jan Hubicka <jh@suse.cz>
11443
03b0ee0a
UB		 11444 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11445 partitions are not needed.
0dcc26c3 11446
0bd72901
UB		 114472013-08-01 Uros Bizjak <ubizjak@gmail.com>
11448
11449 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11450 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11451 MAYBE_NON_Q_CLASS_P where appropriate.
11452
12123452
JH		 114532013-08-01 Jan Hubicka <jh@suse.cz>
11454
11455 * cgraph.h (release_function_body): Declare.
11456 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11457 of unused delcarations.
11458
ce852f9c
KT		 114592013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11460
03b0ee0a
UB		 11461 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11462 RTL form when subtracting a constant.
ce852f9c 11463
c743b246
KT		 114642013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11465
11466 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11467 Generate canonical plus rtx with negated immediate instead of minus
11468 where appropriate.
11469 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11470
c0c123ef
JH		 114712013-08-01 Jan Hubicka <jh@suse.cz>
11472
11473 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11474 (cgraph_release_function_body): Likewise.
11475 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11476 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11477 to used_as_abstract_origin.
11478 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11479 symbols used as abstract origins.
11480 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB		 11481 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11482 used_as_abstract_origin.
c0c123ef 11483 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB		 11484 used_as_abstract_origin; be ready for DECL_RESULT and
11485 DECL_ARGUMENTS to be NULL.
c0c123ef 11486
0bd72901
UB		 11487 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11488 for abstract functions.
11489 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11490 real symbols.
c0c123ef 11491
1f029433
JH		 114922013-08-01 Jan Hubicka <jh@suse.cz>
11493
11494 * profile.c (compute_value_histograms): Fix thinko.
11495
c451f4d6
SN		 114962013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11497
11498 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11499 aarch-common-protos.h to extra_headers.
11500 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11501 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11502 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11503
34b512f5
SN		 115042013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11505
11506 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11507 Include "../arm/types.md". Define "type" attribute for all patterns.
11508 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11509 attribute changes.
11510
d86e633a
MM		 115112013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11512
11513 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11514 to support power8 load fusion.
11515 (fusion_gpr_mem_load): Likewise.
11516
11517 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11518
11519 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11520 declarations for power8 load fusion.
11521 (emit_fusion_gpr_load): Likewise.
11522
11523 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11524 tuning for power8, turn on fusion mode by default. Turn on sign
11525 extending fusion mode if normal fusion mode is on, and we are at
11526 -O2 or -O3.
11527 (fusion_gpr_load_p): New function, return true if we can fuse an
11528 addis instruction with a dependent load to a GPR.
11529 (emit_fusion_gpr_load): Emit the instructions for power8 load
11530 fusion to GPRs.
11531
0bd72901 11532 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM		 11533 (VSX load fusion peepholes): New peepholes to fuse together an
11534 addi instruction with a VSX load instruction.
11535
11536 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11537 peepholes to fuse an addis instruction with a load to a GPR base
11538 register. If we are supporting sign extending fusions, convert
11539 sign extending loads to zero extending loads and add an explicit
11540 sign extension.
11541
c3f35647
SN		 115422013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11543
11544 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11545 aarch-common-protos.h to extra_headers.
11546 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11547 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11548 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11549 (arm_no_early_alu_shift_dep): Likewise.
11550 (arm_no_early_alu_shift_value_dep): Likewise.
11551 (arm_no_early_mul_dep): Likewise.
11552 (arm_no_early_store_addr_dep): Likewise.
11553 (arm_mac_accumulator_is_mul_result): Likewise.
11554 (arm_mac_accumulator_is_result): Likewise.
c3f35647 11555 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB		 11556 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11557 here to ...
c3f35647 11558 (arm_early_store_addr_dep): Likewise.
0bd72901
UB		 11559 (arm_no_early_alu_shift_dep): Likewise.
11560 (arm_no_early_alu_shift_value_dep): Likewise.
11561 (arm_no_early_mul_dep): Likewise.
11562 (arm_no_early_store_addr_dep): Likewise.
11563 (arm_mac_accumulator_is_mul_result): Likewise.
11564 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN		 11565 * config/arm/aarch-common-protos.h: ... here. New file.
11566 * config/arm/t-arm (aarch-common.o): Define.
11567
115682013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11569
11570 * config/arm/arm.md: Include new file "types.md".
11571 (define_attr "type"): Move from here to ...
11572 (define_attr "mul32"): Likewise.
11573 (define_attr "mul64"): Likewise.
11574 * config/arm/types.md: ... here. New file.
11575
228c1313
SH		 115762013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11577
11578 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11579 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11580
d6e9046d 115812013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG		 11582
11583 * gen-pass-instances.awk: Fix offset of substr().
11584
a167b052
DM		 115852013-07-31 David Malcolm <dmalcolm@redhat.com>
11586
11587 * Makefile.in (pass-instances.def): New.
11588 (passes.o): Replace dependency on passes.def with one on
11589 pass-instances.def
11590
11591 * gen-pass-instances.awk: New.
11592
11593 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11594 rather than passes.def, updating local definition of NEXT_PASS
11595 macro to add an extra NUM parameter (currently unused).
11596
315f8c0e
DM		 115972013-07-30 David Malcolm <dmalcolm@redhat.com>
11598
11599 * Makefile.in (PASS_MANAGER_H): New.
11600 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11601 (passes.o): Likewise.
11602 (statistics.o): Likewise.
11603 (cgraphunit.o): Likewise.
11604 (context.o): Depend on PASS_MANAGER_H.
11605
11606 * pass_manager.h: New.
11607
11608 * cgraphunit.c (cgraph_add_new_function): Update for moves
11609 of globals to fields of pass_manager.
11610 (analyze_function): Likewise.
11611 (expand_function): Likewise.
11612 (ipa_passes): Likewise.
11613 (compile): Likewise.
11614
11615 * context.c (context::context): New.
11616 * context.h (context::context): New.
11617 (context::get_passes): New.
11618 (context::passes_): New.
11619
11620 * lto-cgraph.c (input_node): Update for moves of globals to
11621 fields of pass_manager.
11622
11623 * passes.c (all_passes): Remove, in favor of a field of the
11624 same name within the new class pass_manager.
11625 (all_small_ipa_passes): Likewise.
11626 (all_lowering_passes): Likewise.
11627 (all_regular_ipa_passes): Likewise.
11628 (all_late_ipa_passes): Likewise.
11629 (all_lto_gen_passes): Likewise.
11630 (passes_by_id): Likewise.
11631 (passes_by_id_size): Likewise.
11632 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11633 the new class pass_manager.
11634 (set_pass_for_id): Convert to...
11635 (pass_manager::set_pass_for_id): ...method.
11636 (get_pass_for_id): Convert to...
11637 (pass_manager::get_pass_for_id): ...method.
11638 (register_one_dump_file): Move body of implementation into...
11639 (pass_manager::register_one_dump_file): ...here.
11640 (register_dump_files_1): Convert to...
11641 (pass_manager::register_dump_files_1): ...method.
11642 (register_dump_files): Convert to...
11643 (pass_manager::register_dump_files): ...method.
11644 (create_pass_tab): Update for moves of globals to fields of
11645 pass_manager.
11646 (dump_passes): Move body of implementation into...
11647 (pass_manager::dump_passes): ...here.
11648 (register_pass): Move body of implementation into...
11649 (pass_manager::register_pass): ...here.
11650 (init_optimization_passes): Convert into...
11651 (pass_manager::pass_manager): ...constructor for new
11652 pass_manager class, and initialize the pass_lists array.
11653 (check_profile_consistency): Update for moves of globals to
11654 fields of pass_manager.
11655 (dump_profile_report): Move body of implementation into...
11656 (pass_manager::dump_profile_report): ...here.
11657 (ipa_write_summaries_1): Update for moves of pass lists from
11658 being globals to fields of pass_manager.
11659 (ipa_write_optimization_summaries): Likewise.
11660 (ipa_read_summaries): Likewise.
11661 (ipa_read_optimization_summaries): Likewise.
11662 (execute_all_ipa_stmt_fixups): Likewise.
11663
11664 * statistics.c (statistics_fini): Update for moves of globals to
11665 fields of pass_manager.
11666
11667 * toplev.c (general_init): Replace call to
11668 init_optimization_passes with construction of the pass_manager
11669 instance.
11670
11671 * tree-pass.h (all_passes): Remove, in favor of a field of the
11672 same name within the new class pass_manager.
11673 (all_small_ipa_passes): Likewise.
11674 (all_lowering_passes): Likewise.
11675 (all_regular_ipa_passes): Likewise.
11676 (all_lto_gen_passes): Likewise.
11677 (all_late_ipa_passes): Likewise.
11678 (passes_by_id): Likewise.
11679 (passes_by_id_size): Likewise.
11680 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11681 the new class pass_manager.
11682 (get_pass_for_id): Remove.
11683
02cab1c1
RE		 116842013-07-30 Richard Earnshaw <rearnsha@arm.com>
11685
11686 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11687 configs.
11688
b2d3f886
RE		 116892013-07-30 Richard Earnshaw <rearnsha@arm.com>
11690
11691 * arm.md (mulhi3): New expand pattern.
11692
605e86fa
JH		 116932013-07-30 Jan Hubicka <jh@suse.cz>
11694 Martin Liska <marxin.liska@gmail.com>
11695
11696 * profile.c (compute_value_histograms): Do not ICE when
11697 there is mismatch only on some counters.
11698
14a87636
ZC		 116992013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11700
11701 PR rtl-optimization/57637
11702 * function.c (move_insn_for_shrink_wrap): Also check the
11703 GEN set of the LIVE problem for the liveness analysis
11704 if it exists, otherwise give up.
11705
0100cd3f
BS		 117062013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11707
11708 PR tree-optimization/57993
11709 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11710 replaced statement in the candidate table.
11711 (phi_add_costs): Return infinite cost when the hidden basis does
11712 not dominate all phis on which the candidate is dependent.
11713 (replace_one_candidate): Record replaced statement in the
11714 candidate table.
11715
f0d811f7
JR		 117162013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11717
11718 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11719 (ashlv2si3): New expander.
11720 (*ashlv2si3_i): New define_insn_and_split.
11721 * predicates.md (float_operation): Allow patterns with three
11722 basic sub-patterns.
11723
bba33211
JR		 11724 PR rtl-optimization/58021
11725 * mode-switching.c (create_pre_exit): Always split off preceding
11726 insns if we are not at the basic block head.
11727
ff3f3951
MR		 117282013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11729
11730 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11731 (UCLIBC_DYNAMIC_LINKER): New macro.
11732 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11733 `-mnan=2008'.
11734 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11735 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11736 `-mnan=2008'.
11737 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11738 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11739 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11740 for SF and DF modes. Use ieee_quad_format for TF mode.
11741 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11742 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11743 (mips_option_override): Handle `-mnan=legacy'.
11744 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11745 `-mabs=2008' and `-mnan=2008'.
11746 (OPTION_DEFAULT_SPECS): Add "nan" default.
11747 (ASM_SPEC): Handle `-mnan='.
11748 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11749 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11750 comment accordingly.
11751 (neg<mode>2): Likewise.
11752 * config/mips/mips.opt (mabs, mnan): New options.
11753 * doc/install.texi (Configuration): Document `--with-nan=' option.
11754 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11755 `-mnan=' options.
11756 (MIPS Options): Document them.
11757 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11758 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11759 * configure: Regenerate.
11760 * config.in: Regenerate.
11761
576e2f37
UB		 117622013-07-29 Uros Bizjak <ubizjak@gmail.com>
11763
11764 * config/i386/i386.md (float post-reload splitters): Do not check
11765 for subregs of SSE registers.
11766
117672013-07-29 Uros Bizjak <ubizjak@gmail.com>
11768 H.J. Lu <hongjiu.lu@intel.com>
11769
11770 PR target/57954
11771 PR target/57988
11772 * config/i386/i386.md (post-reload splitter
11773 to avoid partial SSE reg dependency stalls): New pattern.
11774
a71f0749
DV		 117752013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11776
11777 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11778
4c97f1cc
JR		 117792013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11780
11781 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11782 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11783
b0c733d8
CC		 117842013-07-26 Cary Coutant <ccoutant@google.com>
11785
11786 * dwarf2out.c (die_checksum_ordered): Don't include template
11787 instantiations in signature.
11788 (is_template_parameter): New function.
11789 (is_template_instantiation): New function.
11790 (generate_skeleton_bottom_up): Don't include template instantiations
11791 in type unit DIE.
11792 (generate_skeleton): Likewise.
11793 (break_out_comdat_types): Move recursive call to break out nested
11794 types earlier.
11795 (prune_unused_types_mark_generic_parms_dies): Call
11796 is_template_parameter.
11797
040d8a1c
IB		 117982013-07-26 Ian Bolton <ian.bolton@arm.com>
11799
11800 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11801 uses vector registers.
89fdc743 11802 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 11803
85bd4ac6 118042013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 11805 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT		 11806
11807 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 11808 where appropriate.
85bd4ac6
KT		 11809 * config/arm/arm.md (movcond_addsi): New splitter.
11810
b6051207
SA		 118112013-07-25 Sterling Augustine <saugustine@google.com>
11812
11813 * dwarf2out.c (size_of_pubnames): Move code to...
11814 (include_pubname_in_output): ...here. New.
11815 (want_pubnames): Rearrange.
11816 (output_pubnames): Call include_pubname_in_output. Move assertion.
11817
41a7c215
CM		 118182013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11819
11820 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11821
118222013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11823
11824 PR target/38836
11825 * doc/extend.texi: Remove obsolete builtins. Fix
11826 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11827
49bde175
JH		 118282013-07-25 Jan Hubicka <jh@suse.cz>
11829
11830 * cgraph.c (release_function_body): Break out from ...
11831 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11832 and DECL_ARGUMENTS.
11833 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11834 old_tree in the map.
11835 (create_specialized_node): Update.
11836 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11837 into index.
41a7c215
CM		 11838 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11839 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH		 11840 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11841 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11842 DECL_RESULT.
11843
4870352d
KT		 118442013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11845
11846 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11847 addsi3_carryin_alt2_<optab>): Correct output template.
11848
1d6aee1c
KT		 118492013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11850
11851 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11852 Adjust for arm_restrict_it.
11853 Remove trailing whitespace.
11854
c7cafd75
MK		 118552013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11856
017d38f5
MK		 11857 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11858 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11859
c7cafd75
MK		 11860 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11861
00b7527b
VM		 118622013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11863
11864 PR rtl-optimization/57960
11865 * lra-constraints.c (process_alt_operands): Use the right mode
11866 when checking strict_low.
11867
4f63dfc6
JH		 118682013-07-25 Jan Hubicka <jh@suse.cz>
11869
11870 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11871 * cgraph.c (cgraph_remove_node): Do not release function body
11872 when in cgraph streaming.
11873 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM		 11874 in other partitions are not considered reachable; fix handling of
11875 clones.
4f63dfc6 11876
e6dcfa18
RR		 118772013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11878
11879 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11880
0b93d3b6
RR		 118812013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11882
11883 PR target/19599
11884 PR target/57731
bb80c2eb 11885 PR target/57837
e6dcfa18 11886 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 11887 Ss with US. Adjust output for v5 and v4t.
41a7c215 11888 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 11889
0b93d3b6
RR		 11890 * config/arm/constraints.md ("Ss"): Rename to US.
11891
690688b3
TG		 118922013-07-25 Terry Guo <terry.guo@arm.com>
11893
11894 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11895 shift_add/shift_sub0/shift_sub1 RTXs.
11896
a4ad093b 118972013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 11898 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS		 11899
11900 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11901 (altivec_vpks<VI_char>ss): Likewise.
11902 (altivec_vpks<VI_char>us): Likewise.
11903 (altivec_vpku<VI_char>us): Likewise.
11904 (altivec_vpku<VI_char>um): Likewise.
11905
8c5005ce
DM		 119062013-07-24 David Malcolm <dmalcolm@redhat.com>
11907
11908 Introduce context class.
11909
11910 * Makefile.in (CONTEXT_H): New.
11911 (OBJS): Add context.o.
11912 (toplev.o): Add CONTEXT_H to dependencies.
11913 (context.o): New.
11914
41a7c215 11915 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM		 11916
11917 * context.c: New.
11918
11919 * context.h: New.
11920
ce4a9422
JR		 119212013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11922
11923 PR rtl-optimization/57968
11924 * mode-switching.c (create_pre_exit): Allow instructions that
11925 don't set a return register to need a non-exit mode.
11926
493f4c9e 119272013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11928 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 11929
b5860fd3
WS		 11930 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11931 operands to vperm for little endian.
11932 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11933 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 11934
73792b92 119352013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11936 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS		 11937
11938 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11939 two operands for little-endian.
11940
14c21302
SE		 119412013-07-23 Steve Ellcey <sellcey@mips.com>
11942
11943 * config/mips/mips.c (mips_case_values_threshold): New.
11944 (TARGET_CASE_VALUES_THRESHOLD): Define.
11945
a6056198 119462013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11947 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS		 11948
11949 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11950 selection of field for vector splat in little endian mode.
11951
dd7a40e1
MM		 119522013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11953
11954 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11955 expanders to rs6000.md.
11956 (ior<mode>3): Likewise.
11957 (and<mode>3): Likewise.
11958 (one_cmpl<mode>2): Likewise.
11959 (nor<mode>3): Likewise.
11960 (andc<mode>3): Likewise.
11961 (eqv<mode>3): Likewise.
11962 (nand<mode>3): Likewise.
11963 (orc<mode>3): Likewise.
11964
11965 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11966 declaration.
11967
11968 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11969 to split multi-word logical operations.
11970 (rs6000_split_logical_di): Likewise.
11971 (rs6000_split_logical): Likewise.
11972
11973 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11974 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11975 and allow TImode operations in 32-bit.
11976 (vsx_and<mode>3_64bit): Likewise.
11977 (vsx_ior<mode>3_32bit): Likewise.
11978 (vsx_ior<mode>3_64bit): Likewise.
11979 (vsx_xor<mode>3_32bit): Likewise.
11980 (vsx_xor<mode>3_64bit): Likewise.
11981 (vsx_one_cmpl<mode>2_32bit): Likewise.
11982 (vsx_one_cmpl<mode>2_64bit): Likewise.
11983 (vsx_nor<mode>3_32bit): Likewise.
11984 (vsx_nor<mode>3_64bit): Likewise.
11985 (vsx_andc<mode>3_32bit): Likewise.
11986 (vsx_andc<mode>3_64bit): Likewise.
11987 (vsx_eqv<mode>3_32bit): Likewise.
11988 (vsx_eqv<mode>3_64bit): Likewise.
11989 (vsx_nand<mode>3_32bit): Likewise.
11990 (vsx_nand<mode>3_64bit): Likewise.
11991 (vsx_orc<mode>3_32bit): Likewise.
11992 (vsx_orc<mode>3_64bit): Likewise.
11993
11994 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11995 logical types in GPRs.
11996
11997 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11998 logical insns to rs6000.md, and allow TImode operations in
11999 32-bit.
12000 (altivec_ior<mode>3): Likewise.
12001 (altivec_xor<mode>3): Likewise.
12002 (altivec_one_cmpl<mode>2): Likewise.
12003 (altivec_nor<mode>3): Likewise.
12004 (altivec_andc<mode>3): Likewise.
12005
12006 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12007 attributes for moving the 128-bit logical operations into
12008 rs6000.md.
12009 (BOOL_REGS_OUTPUT): Likewise.
12010 (BOOL_REGS_OP1): Likewise.
12011 (BOOL_REGS_OP2): Likewise.
12012 (BOOL_REGS_UNARY): Likewise.
12013 (BOOL_REGS_AND_CR0): Likewise.
12014 (one_cmpl<mode>2): Add support for DI logical operations on
12015 32-bit, splitting the operations to 32-bit.
12016 (anddi3): Likewise.
12017 (iordi3): Likewise.
12018 (xordi3): Likewise.
12019 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12020 changes to combine the 32/64-bit code, allow logical operations on
12021 TI mode in 32-bit, and to use similar match_operator patterns like
12022 scalar mode uses. Combine the Altivec and VSX code for logical
12023 operations, and move it here.
12024 (ior<mode>3, 128-bit types): Likewise.
12025 (xor<mode>3, 128-bit types): Likewise.
12026 (one_cmpl<mode>3, 128-bit types): Likewise.
12027 (nor<mode>3, 128-bit types): Likewise.
12028 (andc<mode>3, 128-bit types): Likewise.
12029 (eqv<mode>3, 128-bit types): Likewise.
12030 (nand<mode>3, 128-bit types): Likewise.
12031 (orc<mode>3, 128-bit types): Likewise.
12032 (and<mode>3_internal): Likewise.
12033 (bool<mode>3_internal): Likewise.
12034 (boolc<mode>3_internal1): Likewise.
12035 (boolc<mode>3_internal2): Likewise.
12036 (boolcc<mode>3_internal1): Likewise.
12037 (boolcc<mode>3_internal2): Likewise.
12038 (eqv<mode>3_internal1): Likewise.
12039 (eqv<mode>3_internal2): Likewise.
12040 (one_cmpl1<mode>3_internal): Likewise.
12041
12211b99 120422013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH		 12043
12044 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12045 Rename flag_stack_usage to flag_stack_usage_info.
12046
12211b99 120472013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 12048
a5f257fd
ME		 12049 * config/microblaze/sync.md: New file.
12050 * config/microblaze/microblaze.md: Include sync.md
12051 * config/microblaze/microblaze.c: Add print_operand 'y'.
12052 * config/microblaze/constraints.md: Add memory_contraint
12053 'Q' which is a single register.
450b0ebb 12054
49fba14f
EB		 120552013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12056
12057 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12058
c7b3b99f
PCC		 120592013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12060
12061 * reload.c (find_reloads): Exit loop once we find this operand
12062 cannot be reloaded somehow for this alternative.
12063
12064 * reload.c (find_reloads): Exit loop once we find a hard register.
12065
12066 * rtlanal.c (computed_jump_p): Exit loop once we find label
12067 reference is used.
12068
12069 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12070
12071 * cfgloopmanip.c (remove_path): Exit loop after setting
12072 irred_invalidated.
12073
12074 * gensupport.c (subst_dup): Avoid loop if code is not
12075 MATCH_DUP nor MATCH_OP_DUP.
12076
c8fbf1fa
NBJ		 120772013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12078
12079 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12080
9259db42
YZ		 120812013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12082
12083 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12084 true for SP_REGNUM if mode == ptr_mode.
12085 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12086 with value R0_REGNUM + 31.
12087
00edcfbe
YZ		 120882013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12089
12090 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12091 pad pointer-typed argument downward.
12092
43be9a95
YZ		 120932013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12094
12095 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12096 and __ILP32__ when the ILP32 model is in use.
12097
28514dda
YZ		 120982013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12099
12100 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12101 (aarch64_load_symref_appropriately): In the case of
12102 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12103 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12104 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12105 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12106 if 'mode' doesn't equal to 'ptr_mode'.
12107 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12108 'vcall_offset'; change to call aarch64_emit_move differently depending
12109 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12110 to calculate the upper bound of 'vcall_offset'.
12111 (aarch64_cannot_force_const_mem): Change to also return true if
12112 mode != ptr_mode.
12113 (aarch64_legitimize_reload_address): In the case of large
12114 displacements, add new local variable 'xmode' and an assertion
12115 based on it; change to use 'xmode' to generate the new rtx and
12116 reload.
12117 (aarch64_asm_trampoline_template): Change to generate the template
12118 differently depending on TARGET_ILP32 or not; change to use
12119 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12120 (aarch64_trampoline_size): Removed.
12121 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12122 and replace immediate literals with it. Change to use 'ptr_mode'
12123 instead of 'DImode' and call convert_memory_address if the mode
12124 of 'fnaddr' doesn't equal to 'ptr_mode'.
12125 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12126 to output symbol.
12127 (aarch64_elf_asm_destructor): Likewise.
12128 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12129 on TARGET_ILP32 instead of aarch64_trampoline_size.
12130 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12131 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12132 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12133 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12134 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12135 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12136 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12137 depending on the value of 'mode'.
12138 (add_losym_<mode>): New.
12139 (ldr_got_small_<mode>): New, based on ldr_got_small.
12140 (ldr_got_small): Remove.
12141 (ldr_got_small_sidi): New.
12142 * config/aarch64/iterators.md (P): New.
12143 (PTR): Change to 'ptr_mode' in the condition.
12144
17a819cb
YZ		 121452013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12146
12147 * config.gcc (aarch64*-*-*): Support --with-abi.
12148 (aarch64*-*-elf): Support --with-multilib-list.
12149 (aarch64*-*-linux*): Likewise.
12150 (supported_defaults): Add abi to aarch64*-*-*.
12151 * configure.ac: Mention AArch64 for --with-multilib-list.
12152 * configure: Re-generated.
12153 * config/aarch64/biarchilp32.h: New file.
12154 * config/aarch64/biarchlp64.h: New file.
12155 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12156 (ABI_SPEC): Ditto.
12157 (MULTILIB_DEFAULTS): Ditto.
12158 (DRIVER_SELF_SPECS): Ditto.
12159 (ASM_SPEC): Update to also substitute -mabi.
12160 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12161 file whose name depends on -mabi= and -mbig-endian.
12162 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12163 TARGET_ILP32.
12164 (POINTER_SIZE): New define.
12165 (POINTERS_EXTEND_UNSIGNED): Ditto.
12166 (enum aarch64_abi_type): New enumeration tag.
12167 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12168 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12169 (TARGET_ILP32): New define.
12170 * config/aarch64/aarch64.opt (mabi): New.
12171 (aarch64_abi): New.
12172 (ilp32, lp64): New values for -mabi.
12173 * config/aarch64/t-aarch64 (comma): New define.
12174 (MULTILIB_OPTIONS): Ditto.
12175 (MULTILIB_DIRNAMES): Ditto.
12176 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12177 * doc/invoke.texi: Document -mabi for AArch64.
12178
9f9cbdce
GJL		 121792013-07-23 Georg-Johann Lay <avr@gjlay.de>
12180
12181 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12182
7061977b 121832013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 12184 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS		 12185
12186 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12187 endianness when selecting field to splat.
12188
75a41878
EC		 121892013-07-22 Eric Christopher <echristo@gmail.com>
12190
12191 * dwarf2out.c (die_odr_checksum): New function to use
12192 CHECKSUM_ macros and ULEB128 for DIE tag.
12193 (generate_type_signature): Use.
41a7c215 12194
38ae58ca
EB		 121952013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12196
12197 * config.gcc (sparc*-*-*): Accept leon3 processor.
12198 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12199 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12200 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12201 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12202 (mfix-ut699): Adjust comment.
12203 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12204 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12205 (CPP_CPU_SPEC): Likewise.
12206 (ASM_CPU_SPEC): Likewise.
12207 * config/sparc/sparc.c (leon3_cost): New constant.
12208 (sparc_option_override): Add leon3 support.
12209 (mem_ref): New function.
12210 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12211 (sparc_do_work_around_errata): Look into the instruction in the delay
12212 slot and adjust accordingly. Add fix for the data cache nullify issues
12213 of the UT699. Change insertion position for the NOP.
12214 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12215 (leon3_load): New reservation.
12216 (leon_store): Bump latency to 2.
12217 (grfpu): New automaton.
12218 (grfpu_alu): New unit.
12219 (grfpu_ds): Likewise.
12220 (leon_fp_alu): Adjust.
12221 (leon_fp_mult): Delete.
12222 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12223 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12224 * config/sparc/sparc.md (cpu): Add leon3.
12225 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12226 (swapsi): Likewise.
12227 (atomic_test_and_set): Likewise.
12228 (ldstub): Likewise.
12229
04dfc6df
JU		 122302013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12231
12232 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12233 default for R5900 targets.
12234 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12235 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12236 * config/mips/mips.c (mips_option_override): Report an error for
12237 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12238 for -march=r5900 -mhard-float.
12239
c7b3b99f 122402013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC		 12241
12242 * df-problems.c (can_move_insns_across): Exit loop once we
12243 find a non-fixed, non-global register.
12244
12245 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12246 setting can_throw.
12247
12248 * omega.c (omega_eliminate_red): Break after setting red_found.
12249 (omega_problem_has_red_equations): Similarly after setting found.
12250 (omega_query_variable): Similarly after setting coupled.
12251
9b6e6981
MP		 122522013-07-22 Marek Polacek <polacek@redhat.com>
12253
12254 * gimplify.c: Don't include gimple.h twice.
12255
5e5f7673
KT		 122562013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12257
12258 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12259 instead of TARGET_THUMB1.
12260 (Pz): New constraint.
12261 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12262 encodings.
12263 (compare_negsi_si): Likewise.
12264 (compare_addsi2_op0): Likewise.
12265 (compare_addsi2_op1): Likewise.
12266 (addsi3_carryin_<optab>): Likewise.
12267 (addsi3_carryin_alt2_<optab>): Likewise.
12268 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12269 for arm_restrict_it.
12270 (subsi3_carryin): Likewise.
12271 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12272 (minmax_arithsi): Disable for arm_restrict_it.
12273 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12274 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12275 (satsi_<SAT:code>_shift): Likewise.
12276 (arm_shiftsi3): Add alternative for 16-bit encoding.
12277 (arm32_movhf): Disable for arm_restrict_it.
12278 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12279 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12280
859abddd
SN		 122812013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12282
12283 * config/arm/arm.md (attribute "insn"): Delete.
12284 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12285 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12286 (not_shiftsi): Update for attribute change.
12287 (not_shiftsi_compare0): Likewise.
12288 (not_shiftsi_compare0_scratch): Likewise.
12289 (arm_one_cmplsi2): Likewise.
12290 (thumb1_one_cmplsi2): Likewise.
12291 (notsi_compare0): Likewise.
12292 (notsi_compare0_scratch): Likewise.
12293 (thumb1_movdi_insn): Likewise.
12294 (arm_movsi_insn): Likewise.
12295 (movhi_insn_arch4): Likewise.
12296 (movhi_bytes): Likewise.
12297 (arm_movqi_insn): Likewise.
12298 (thumb1_movqi_insn): Likewise.
12299 (arm32_movhf): Likewise.
12300 (thumb1_movhf): Likewise.
12301 (arm_movsf_soft_insn): Likewise.
12302 (thumb1_movsf_insn): Likewise.
12303 (thumb_movdf_insn): Likewise.
12304 (movsicc_insn): Likewise.
12305 (movsfcc_soft_insn): Likewise.
12306 (and_scc): Likewise.
12307 (cond_move): Likewise.
12308 (if_move_not): Likewise.
12309 (if_not_move): Likewise.
12310 (if_shift_move): Likewise.
12311 (if_move_shift): Likewise.
12312 (if_shift_shift): Likewise.
12313 (if_not_arith): Likewise.
12314 (if_arith_not): Likewise.
12315 (cond_move_not): Likewise.
12316 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12317 (neon_mov<mode>): Likewise.
12318 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12319 (thumb2_movsi_vfp): Likewise.
12320 (movsf_vfp): Likewise.
12321 (thumb2_movsf_vfp): Likewise.
371e77e3
UB		 12322 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12323 change.
859abddd
SN		 12324 (cortexa7_older_only): Likewise.
12325 (cortexa7_younger): Likewise.
12326 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12327 (1020alu_shift_op): Likewise.
12328 (1020alu_shift_reg_op): Likewise.
12329 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12330 (alu_shift_op): Likewise.
12331 (alu_shift_reg_op): Likewise.
12332 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12333 (11_alu_shift_op): Likewise.
12334 (11_alu_shift_reg_op): Likewise.
12335 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12336 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12337 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12338 change.
859abddd
SN		 12339 (cortex_a15_alu_shift): Likewise.
12340 (cortex_a15_alu_shift_reg): Likewise.
12341 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12342 (cortex_a5_alu_shift): Likewise.
371e77e3
UB		 12343 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12344 change.
859abddd 12345 (cortex_a53_alu_shift): Likewise.
371e77e3
UB		 12346 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12347 change.
859abddd
SN		 12348 (cortex_a7_alu_reg): Likewise.
12349 (cortex_a7_alu_shift): Likewise.
12350 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12351 (cortex_a8_alu_shift): Likewise.
12352 (cortex_a8_alu_shift_reg): Likewise.
12353 (cortex_a8_mov): Likewise.
12354 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12355 (cortex_a9_dp_shift): Likewise.
12356 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12357 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12358 (cortex_r4_mov): Likewise.
12359 (cortex_r4_alu_shift): Likewise.
12360 (cortex_r4_alu_shift_reg): Likewise.
12361 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12362 (526_alu_shift_op): Likewise.
12363 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12364 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12365 (626te_alu_shift_op): Likewise.
12366 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12367 (726te_alu_op): Likewise.
12368 (726te_alu_shift_op): Likewise.
12369 (726te_alu_shift_reg_op): Likewise.
12370 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12371 (mp626_alu_shift_op): Likewise.
12372 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12373 (pj4_alu_e1_conds): Likewise.
12374 (pj4_alu): Likewise.
12375 (pj4_alu_conds): Likewise.
12376 (pj4_shift): Likewise.
12377 (pj4_shift_conds): Likewise.
12378 (pj4_alu_shift): Likewise.
12379 (pj4_alu_shift_conds): Likewise.
12380
95b97fac
KT		 123812013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12382
12383 * config/arm/predicates.md (shiftable_operator_strict_it):
12384 New predicate.
12385 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12386 Disable cond_exec version for arm_restrict_it.
12387 (thumb2_smaxsi3): Convert to generate cond_exec.
12388 (thumb2_sminsi3): Likewise.
12389 (thumb32_umaxsi3): Likewise.
12390 (thumb2_uminsi3): Likewise.
12391 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12392 (thumb2_neg_abssi2): Likewise.
12393 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12394 (thumb2_movsicc_insn): Adjust alternatives.
12395 (thumb2_mov_negscc): Disable for arm_restrict_it.
12396 (thumb2_mov_negscc_strict_it): New pattern.
12397 (thumb2_mov_notscc_strict_it): New pattern.
12398 (thumb2_mov_notscc): Disable for arm_restrict_it.
12399 (thumb2_ior_scc): Likewise.
12400 (thumb2_ior_scc_strict_it): New pattern.
12401 (thumb2_cond_move): Adjust for arm_restrict_it.
12402 (thumb2_cond_arith): Disable for arm_restrict_it.
12403 (thumb2_cond_arith_strict_it): New pattern.
12404 (thumb2_cond_sub): Adjust for arm_restrict_it.
12405 (thumb2_movcond): Likewise.
12406 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12407 (thumb2_zero_extendhisi2_v6): Likewise.
12408 (thumb2_zero_extendqisi2_v6): Likewise.
12409 (orsi_notsi_si): Likewise.
12410 (orsi_not_shiftsi_si): Likewise.
12411
d5a6ef82
GJL		 124122013-07-22 Georg-Johann Lay <avr@gjlay.de>
12413
12414 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12415 instruction sequence is 1 byte shorter.
12416
d9e71154
UB		 124172013-07-22 Uros Bizjak <ubizjak@gmail.com>
12418
12419 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12420 it is not needed after split.
12421
8c6381ec
IS		 124222013-07-20 Iain Sandoe <iain@codesourcery.com>
12423
12424 PR target/51784
12425 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12426 second label for nonlocal goto receivers. Don't output pic base labels
12427 unless we're producing PIC; mark that action unreachable().
12428 (ix86_save_reg): If the function contains a nonlocal label, save the
12429 PIC base reg.
12430 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 12431 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS		 12432 (update_pic_label_number_if_needed): New.
12433 (machopic_output_function_base_name): Adjust for nonlocal receiver
12434 case.
12435 (machopic_should_output_picbase_label): New.
12436 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12437 (nonlocal_goto_receiver): New insn and split.
12438
d05d0709
JG		 124392013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12440
12441 * config/aarch64/aarch64-builtins.c
12442 (aarch64_fold_builtin): Fold abs in all modes.
12443 * config/aarch64/aarch64-simd-builtins.def
12444 (abs): Enable for all modes.
12445 * config/aarch64/arm_neon.h
12446 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12447 (vabs_f64): Add missing intrinsic.
12448
ba590e63
IB		 124492013-07-19 Ian Bolton <ian.bolton@arm.com>
12450
12451 * config/aarch64/arm_neon.h (vabs_s64): New function
12452
02371798
GJL		 124532013-07-19 Georg-Johann Lay <avr@gjlay.de>
12454
12455 PR target/57516
12456 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12457 * config/avr/avr.md (adjust_len): Add `round'.
12458 * config/avr/avr-protos.h (avr_out_round): New prototype.
12459 (avr_out_plus): Add `out_label' argument.
12460 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12461 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12462 Handle the case where `insn' is just a pattern.
12463 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12464 (avr_out_round): New function.
12465 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12466
12211b99 124672013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH		 12468
12469 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12470 Add check for flag_stack_usage to handle -fstack-usage support
12471
12211b99 124722013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH		 12473
12474 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12475 interaction for new Power8 flags and VSX.
12476
df9dda2d
ST		 124772013-07-18 Sriraman Tallam <tmsriram@google.com>
12478
12479 PR middle-end/57698
12480 * tree-inline.c (expand_call_inline): Emit errors during
12481 early_inlining only if optimization is not turned on.
12482
0f9cdef4
DM		 124832013-07-18 David Malcolm <dmalcolm@redhat.com>
12484
12485 * passes.def: New.
12486
12487 * passes.c (init_optimization_passes): Move the construction of
12488 the pass hierarchy into a new passes.def file.
12489
12490 * Makefile.in (passes.o): Add dependency on passes.def.
12491
2efa4087
DM		 124922013-07-18 David Malcolm <dmalcolm@redhat.com>
12493
12494 * passes.c (init_optimization_passes): Introduce macros for
12495 constructing the tree of passes (INSERT_PASSES_AFTER,
12496 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12497 TERMINATE_PASS_LIST).
12498
bc404e1b
VM		 124992013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12500 Wei Mi <wmi@google.com>
12501
12502 PR rtl-optimization/57878
12503 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12504 top.
12505 (reload_pseudo_compare_func): Check nregs first for reload
12506 pseudos.
12507
d675843f
DM		 125082013-07-18 David Malcolm <dmalcolm@redhat.com>
12509
12510 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12511
c7b3b99f 125122013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC		 12513
12514 * read-rtl.c (validate_const_int): Once an invalid character is
12515 seen, quit the loop.
12516
12517 * gengtype.c (write_roots): Similarly once we find the "deletable"
12518 or "if_marked" option.
12519
52fceb44
SN		 125202013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12521
12522 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12523 "xtab" and "sat". Move value "clz" from here to ...
12524 (attriubte "type"): ... here.
12525 (satsi_<SAT:code>): Delete "insn" attribute.
12526 (satsi_<SAT:code>_shift): Likewise.
12527 (arm_zero_extendqisi2addsi): Likewise.
12528 (arm_extendqisi2addsi): Likewise.
12529 (clzsi2): Update for attribute changes.
12530 (rbitsi2): Likewise.
371e77e3
UB		 12531 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12532 attribute.
52fceb44
SN		 12533 (arm_usatsihi): Likewise.
12534 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12535
006bd006
SN		 125362013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12537
12538 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB		 12539 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12540 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN		 12541 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12542 in alphabetical order.
12543 (attribute "core_cycles"): Update for attribute changes.
12544 (arm_addsi3): Likewise.
12545 (addsi3_compare0): Likewise.
12546 (addsi3_compare0_scratch): Likewise.
12547 (addsi3_compare_op1): Likewise.
12548 (addsi3_compare_op2): Likewise.
12549 (compare_addsi2_op0): Likewise.
12550 (compare_addsi2_op1): Likewise.
12551 (addsi3_carryin_shift_<optab>): Likewise.
12552 (subsi3_carryin_shift): Likewise.
12553 (rsbsi3_carryin_shift): Likewise.
12554 (arm_subsi3_insn): Likewise.
12555 (subsi3_compare0): Likewise.
12556 (subsi3_compare): Likewise.
12557 (arm_andsi3_insn): Likewise.
12558 (thumb1_andsi3_insn): Likewise.
12559 (andsi3_compare0): Likewise.
12560 (andsi3_compare0_scratch): Likewise.
12561 (zeroextractsi_compare0_scratch
12562 (andsi_not_shiftsi_si): Likewise.
12563 (iorsi3_insn): Likewise.
12564 (iorsi3_compare0): Likewise.
12565 (iorsi3_compare0_scratch): Likewise.
12566 (arm_xorsi3): Likewise.
12567 (thumb1_xorsi3_insn): Likewise.
12568 (xorsi3_compare0): Likewise.
12569 (xorsi3_compare0_scratch): Likewise.
12570 (satsi_<SAT:code>_shift): Likewise.
12571 (rrx): Likewise.
12572 (arm_shiftsi3): Likewise.
12573 (shiftsi3_compare0): Likewise.
12574 (not_shiftsi): Likewise.
12575 (not_shiftsi_compare0): Likewise.
12576 (not_shiftsi_compare0_scratch): Likewise.
12577 (arm_one_cmplsi2): Likewise.
12578 (thumb_one_complsi2): Likewise.
12579 (notsi_compare0): Likewise.
12580 (notsi_compare0_scratch): Likewise.
12581 (thumb1_zero_extendhisi2): Likewise.
12582 (arm_zero_extendhisi2): Likewise.
12583 (arm_zero_extendhisi2_v6): Likewise.
12584 (arm_zero_extendhisi2addsi): Likewise.
12585 (thumb1_zero_extendqisi2): Likewise.
12586 (thumb1_zero_extendqisi2_v6): Likewise.
12587 (arm_zero_extendqisi2): Likewise.
12588 (arm_zero_extendqisi2_v6): Likewise.
12589 (arm_zero_extendqisi2addsi): Likewise.
12590 (thumb1_extendhisi2): Likewise.
12591 (arm_extendhisi2): Likewise.
12592 (arm_extendhisi2_v6): Likewise.
12593 (arm_extendqisi): Likewise.
12594 (arm_extendqisi_v6): Likewise.
12595 (arm_extendqisi2addsi): Likewise.
12596 (thumb1_extendqisi2): Likewise.
12597 (thumb1_movdi_insn): Likewise.
12598 (arm_movsi_insn): Likewise.
12599 (movsi_compare0): Likewise.
12600 (movhi_insn_arch4): Likewise.
12601 (movhi_bytes): Likewise.
12602 (arm_movqi_insn): Likewise.
12603 (thumb1_movqi_insn): Likewise.
12604 (arm32_movhf): Likewise.
12605 (thumb1_movhf): Likewise.
12606 (arm_movsf_soft_insn): Likewise.
12607 (thumb1_movsf_insn): Likewise.
12608 (movdf_soft_insn): Likewise.
12609 (thumb_movdf_insn): Likewise.
12610 (arm_cmpsi_insn): Likewise.
12611 (cmpsi_shiftsi): Likewise.
12612 (cmpsi_shiftsi_swp): Likewise.
12613 (arm_cmpsi_negshiftsi_si): Likewise.
12614 (movsicc_insn): Likewise.
12615 (movsfcc_soft_insn): Likewise.
12616 (arith_shiftsi): Likewise.
12617 (arith_shiftsi_compare0
12618 (arith_shiftsi_compare0_scratch
12619 (sub_shiftsi): Likewise.
12620 (sub_shiftsi_compare0
12621 (sub_shiftsi_compare0_scratch
12622 (and_scc): Likewise.
12623 (cond_move): Likewise.
12624 (if_plus_move): Likewise.
12625 (if_move_plus): Likewise.
12626 (if_move_not): Likewise.
12627 (if_not_move): Likewise.
12628 (if_shift_move): Likewise.
12629 (if_move_shift): Likewise.
12630 (if_shift_shift): Likewise.
12631 (if_not_arith): Likewise.
12632 (if_arith_not): Likewise.
12633 (cond_move_not): Likewise.
12634 (thumb1_ashlsi3): Set type attribute.
12635 (thumb1_ashrsi3): Likewise.
12636 (thumb1_lshrsi3): Likewise.
12637 (thumb1_rotrsi3): Likewise.
12638 (shiftsi3_compare0_scratch): Likewise.
12639 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12640 (neon_mov<mode>): Likewise.
371e77e3
UB		 12641 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12642 attribute changes.
006bd006
SN		 12643 (thumb2_movsi_insn): Likewise.
12644 (thumb2_cmpsi_neg_shiftsi): Likewise.
12645 (thumb2_extendqisi_v6): Likewise.
12646 (thumb2_zero_extendhisi2_v6): Likewise.
12647 (thumb2_zero_extendqisi2_v6): Likewise.
12648 (thumb2_shiftsi3_short): Likewise.
12649 (thumb2_addsi3_compare0_scratch): Likewise.
12650 (orsi_not_shiftsi_si): Likewise.
12651 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12652 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12653 changes.
12654 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12655 (1020alu_shift_op): Likewise.
12656 (1020alu_shift_reg_op): Likewise.
12657 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12658 (alu_shift_op): Likewise.
12659 (alu_shift_reg_op): Likewise.
12660 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12661 (11_alu_shift_op): Likewise.
12662 (11_alu_shift_reg_op): Likewise.
12663 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12664 (9_alu_shift_reg_op): Likewise.
371e77e3
UB		 12665 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12666 changes.
006bd006
SN		 12667 (cortex_a15_alu_shift): Likewise.
12668 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB		 12669 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12670 changes.
006bd006
SN		 12671 (cortex_a5_alu_shift): Likewise.
12672 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12673 changes.
12674 (cortex_a53_alu_shift): Likewise.
12675 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12676 changes.
12677 (cortex_a7_alu_reg): Likewise.
12678 (cortex_a7_alu_shift): Likewise.
371e77e3
UB		 12679 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12680 changes.
006bd006
SN		 12681 (cortex_a8_alu_shift): Likewise.
12682 (cortex_a8_alu_shift_reg): Likewise.
12683 (cortex_a8_mov): Likewise.
12684 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12685 (cortex_a9_dp_shift): Likewise.
371e77e3
UB		 12686 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12687 changes.
12688 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12689 changes.
006bd006
SN		 12690 (cortex_r4_mov): Likewise.
12691 (cortex_r4_alu_shift): Likewise.
12692 (cortex_r4_alu_shift_reg): Likewise.
12693 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12694 (526_alu_shift_op): Likewise.
12695 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12696 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12697 (626te_alu_shift_op): Likewise.
12698 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12699 (726te_alu_op): Likewise.
12700 (726te_alu_shift_op): Likewise.
12701 (726te_alu_shift_reg_op): Likewise.
12702 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12703 (mp626_alu_shift_op): Likewise.
12704 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12705 (pj4_alu_e1_conds): Likewise.
12706 (pj4_alu): Likewise.
12707 (pj4_alu_conds): Likewise.
12708 (pj4_shift): Likewise.
12709 (pj4_shift_conds): Likewise.
12710 (pj4_alu_shift): Likewise.
12711 (pj4_alu_shift_conds): Likewise.
371e77e3
UB		 12712 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12713 changes.
006bd006
SN		 12714 (cortexa7_older_only): Likewise.
12715 (cortexa7_younger): Likewise.
12716
651df1b2
DM		 127172013-07-18 David Malcolm <dmalcolm@redhat.com>
12718
12719 * ipa-pure-const.c (generate_summary): Rename to...
12720 (pure_const_generate_summary): ... this.
12721
ea717bd7
IS		 127222013-07-17 Iain Sandoe <iain@codesourcery.com>
12723
12724 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12725
a1259a13
YR		 127262013-07-17 Yvan Roux <yvan.roux@linaro.org>
12727
12728 PR target/57909
12729 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12730 usage in HI mode.
12731
167f68ed
AK		 127322013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12733
12734 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12735 enabled without -march=zEC12.
12736 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12737 flags to be set.
12738
f900a982
MR		 127392013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12740
12741 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12742 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12743 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12744 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12745 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12746 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12747 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12748 ISA_HAS_FP4.
12749 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12750 and ISA_HAS_NMADD3_NMSUB3.
12751 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12752 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12753 (nmsub4<mode>, nmsub3<mode>): Likewise.
12754 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12755
d4ed27eb
MR		 127562013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12757
12758 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12759 TARGET_MIPS5400 checking.
12760
7a5add18
PB		 127612013-07-16 Jakub Jelinek <jakub@redhat.com>
12762 Peter Bergner <bergner@vnet.ibm.com>
12763
12764 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12765 registers in the comment.
12766 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12767 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12768 rather than FIRST_PSEUDO_REGISTERS.
12769
127702013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB		 12771
12772 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12773 enable extra ISA flags with TARGET_HTM.
12774
d2ab0929
MR		 127752013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12776
12777 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12778 Fix comment typos.
12779
839c74bc
CH		 127802013-07-15 Cong Hou <congh@google.com>
12781
12782 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12783 in compare function for sorting.
12784
7a5add18 127852013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB		 12786
12787 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12788 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12789 * config/rs6000/rs6000.opt: Add -mhtm option.
12790 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12791 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12792 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12793 __HTM__ if the HTM instructions are available.
12794 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12795 htm_spr_reg_operand): New define_predicates.
12796 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12797 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12798 Include htm.md.
12799 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12800 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12801 HTM builtin functions.
12802 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12803 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12804 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12805 (rs6000_builtin_mask_calculate): Likewise.
12806 (rs6000_option_override_internal): Likewise.
12807 (bdesc_htm): Add new HTM builtin support.
12808 (htm_spr_num): New function.
12809 (htm_spr_regno): Likewise.
12810 (rs6000_htm_spr_icode): Likewise.
12811 (htm_expand_builtin): Likewise.
12812 (htm_init_builtins): Likewise.
12813 (rs6000_expand_builtin): Add support for HTM builtin functions.
12814 (rs6000_init_builtins): Likewise.
371e77e3
UB		 12815 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12816 option.
0258b6e4
PB		 12817 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12818 (TARGET_HTM, MASK_HTM): Define macros.
12819 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12820 (FIXED_REGISTERS): Likewise.
12821 (CALL_USED_REGISTERS): Likewise.
12822 (CALL_REALLY_USED_REGISTERS): Likewise.
12823 (REG_ALLOC_ORDER): Likewise.
12824 (enum reg_class): Likewise.
12825 (REG_CLASS_NAMES): Likewise.
12826 (REG_CLASS_CONTENTS): Likewise.
12827 (REGISTER_NAMES): Likewise.
12828 (ADDITIONAL_REGISTER_NAMES): Likewise.
12829 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12830 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12831 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12832 * config/rs6000/htm.md: New file.
12833 * config/rs6000/htmintrin.h: New file.
12834 * config/rs6000/htmxlintrin.h: New file.
12835
87dd8ab0
MS		 128362013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12837
12838 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12839 Define SYMBOL_TINY_GOT, update comment.
12840 * config/aarch64/aarch64.c
12841 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12842 (aarch64_expand_mov_immediate): Likewise.
12843 (aarch64_print_operand): Likewise.
12844 (aarch64_classify_symbol): Likewise.
12845 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12846 (ldr_got_tiny): Define.
12847
52d676b6
TG		 128482013-07-13 Tobias Grosser <tobias@grosser.es>
12849
12850 PR tree-optimization/54094
12851 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12852 scheduling dimension for the parallelism check from the polyhedral
12853 information in the AST.
12854 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12855 in 2D + 1 form.
12856
286e8fc1
JM		 128572013-07-13 Jason Merrill <jason@redhat.com>
12858
12859 * print-tree.c (debug_vec_tree): Use debug_raw.
12860 (debug_raw (vec<tree, va_gc> &)): New.
12861 (debug_raw (vec<tree, va_gc> *)): New.
12862 * tree.h: Declare them.
12863
e43257e8
BC		 128642013-07-13 Bin Cheng <bin.cheng@arm.com>
12865
12866 * ifcvt.c (ifcvt_after_combine): New static variable.
12867 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12868 for size.
12869 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12870 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12871 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12872
da582d46
MR		 128732013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12874
12875 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12876
18b0ea8f
MM		 128772013-07-12 Michael Matz <matz@suse.de>
12878
12879 PR middle-end/55771
12880 * convert.c (convert_to_real): Reject non-float inner types.
12881
e4f0f84d
TB		 128822013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12883
12884 * config/aarch64/aarch64-protos.h
12885 (aarch64_simd_immediate_valid_for_move): Remove.
12886 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12887 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12888 cases.
12889 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12890
76500022
SE		 128912013-07-11 Steve Ellcey <sellcey@mips.com>
12892
12893 * config/mips/mips.c (mips_conditional_register_usage): Do not
12894 use t[0-7] registers in MIPS16 mode when optimizing for size.
12895
78039734
ST		 128962013-07-11 Sriraman Tallam <tmsriram@google.com>
12897
12898 * config/i386/i386.c (dispatch_function_versions): Fix array
12899 indexing of function_version_info to match actual_versions.
12900
7b24b675
TJ		 129012013-07-11 Teresa Johnson <tejohnson@google.com>
12902
12903 * vec.h (struct va_gc): Move release out-of-line.
12904 (va_gc::release): Call ggc_free on released vec.
12905
85d4cbb8
UW		 129062013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12907
12908 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12909 Require GOT register as additional operand in UNSPEC.
12910 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12911 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12912 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12913 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12914 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12915 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12916 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12917
52b3f9e2
GJL		 129182013-07-11 Georg-Johann Lay <avr@gjlay.de>
12919
12920 PR target/57631
12921 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12922 name seen by assembler/linker rather if available.
12923
3f3e266e
AS		 129242013-07-11 Andreas Schwab <schwab@suse.de>
12925
12926 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12927
21231ca6
VM		 129282013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12929
371e77e3 12930 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 12931
d331d374
JM		 129322013-07-10 Joseph Myers <joseph@codesourcery.com>
12933
673c2f63
JM		 12934 * doc/tm.texi.in: Move hook documentation to ....
12935 * target.def: ... here.
12936
d331d374
JM		 12937 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12938 text on @hook line.
12939 * doc/tm.texi: Regenerate.
12940
dfeadaa0
PC		 129412013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12942
12943 PR c++/57869
12944 * doc/invoke.texi: Document -Wconditionally-supported.
12945
bb6f2bac
GJL		 129462013-07-10 Georg-Johann Lay <avr@gjlay.de>
12947
12948 PR target/57844
12949 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12950 of my_fp.
12951
5219b232
GJL		 129522013-07-10 Georg-Johann Lay <avr@gjlay.de>
12953
12954 PR target/57506
12955 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12956 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12957 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12958 Remove duplicate devices.
12959 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12960 * config/avr/t-multilib: Regenerate.
12961 * config/avr/avr-tables.opt: Regenerate.
12962 * doc/avr-mmcu.texi: Regenerate.
12963
283045c0
GJL		 129642013-07-10 Georg-Johann Lay <avr@gjlay.de>
12965
12966 PR target/56987
12967 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12968
ceb2d59f 129692013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 12970
ceb2d59f
GS		 12971 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12972 the cost of MULT when optimizing for size.
12973
055e0a99
JBG		 129742013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12975
12976 * config/cr16/cr16-protos.h: Don't include target.h.
12977
ef43e868
JM		 129782013-07-09 Joseph Myers <joseph@codesourcery.com>
12979
12980 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12981 adjust register size for TDmode and TFmode for VSX registers.
12982
4a283fdf
KT		 129832013-07-08 Kai Tietz <ktietz@redhat.com>
12984
12985 PR target/56892
12986 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12987 hook_bool_const_tree_true.
12988
2cf4c39e
AK		 129892013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12990
12991 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12992 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12993 * config/s390/s390.md: Define FPR*_REGNUM constants.
12994 Fix FPR2_REGNUM constant (18 -> 17).
12995 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12996 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12997 ("trunc<BFP:mode><DFP_ALL:mode>2")
12998 ("trunc<DFP_ALL:mode><BFP:mode>2")
12999 ("*extend<BFP:mode><DFP_ALL:mode>2")
13000 ("*extend<DFP_ALL:mode><BFP:mode>2")
13001 ("extend<BFP:mode><DFP_ALL:mode>2")
13002 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13003 FPR4_REGNUM.
13004
e6ac0270 130052013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 13006
6a2fa4b2 13007 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 13008
b89b22fc
AK		 130092013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13010
13011 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13012 and cfun_fpr_bit_p to cfun_fpr_save_p.
13013 (s390_frame_area, s390_register_info, s390_frame_info)
13014 (s390_emit_prologue, s390_emit_epilogue)
13015 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13016 register numbers.
13017 * config/s390/s390.h: Define *_REGNUM macros for floating point
13018 register numbers.
13019
07f398aa
EB		 130202013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13021
13022 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13023
75fefa91
PCC		 130242013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13025
13026 PR rtl-optimization/57786
13027 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13028 and break out of the loop when it is set to false.
13029
046f1eee
JJ		 130302013-07-08 Jakub Jelinek <jakub@redhat.com>
13031
8140c065
JJ		 13032 PR target/57819
13033 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13034 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13035 (const_int 63)) 0)).
13036 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13037 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13038 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13039
046f1eee
JJ		 13040 PR rtl-optimization/57829
13041 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13042 mask bits outside of mode are just sign-extension from mode to HWI.
13043
930b700b
MZ		 130442013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13045
13046 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13047 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13048 adjust_address instead of change_address to keep info about alignment.
13049 (emit_strmov): Remove.
13050 (emit_memmov): New function.
13051 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13052 (expand_movmem_epilogue): Likewise and return updated rtx for
13053 destination.
13054 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13055 destination and source.
13056 (decide_alignment): Refactor, handle vector_loop.
13057 (ix86_expand_movmem): Likewise.
13058 (ix86_expand_setmem): Likewise.
13059 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13060
7b9d1bd8
UB		 130612013-07-07 Uros Bizjak <ubizjak@gmail.com>
13062
13063 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13064 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13065
eabb5f48
UB		 130662013-07-06 Uros Bizjak <ubizjak@gmail.com>
13067
13068 * config/i386/sse.md (sse_movlhps): Change alternative 3
13069 of operand 2 to "m".
13070
130712013-07-06 Uros Bizjak <ubizjak@gmail.com>
13072
13073 PR target/57807
13074 * config/i386/sse.md (iptr): New mode attribute.
13075 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13076 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13077 (<sse>_vmmul<mode>3): Ditto.
13078 (<sse>_vmdiv<mode>3): Ditto.
13079 (sse_vmrcpv4sf2): Ditto.
13080 (<sse>_vmsqrt<mode>2): Ditto.
13081 (sse_vmrsqrtv4sf2): Ditto.
13082 (<sse>_vm<code><mode>3): Ditto.
13083 (avx_vmcmp<mode>3): Ditto.
13084 (<sse>_vmmaskcmp<mode>3): Ditto.
13085 (<sse>_comi): Ditto.
13086 (<sse>_ucomi): Ditto.
13087 (*xop_vmfrcz_<mode>): Ditto.
13088 (*fmai_fmadd_<mode>): Ditto.
13089 (*fmai_fmsub_<mode>): Ditto.
13090 (*fmai_fnmadd_<mode>): Ditto.
13091 (*fmai_fnmsub_<mode>): Ditto.
13092 (*fma4i_vmfmadd_<mode>): Ditto.
13093 (*fma4i_vmfmsub_<mode>): Ditto.
13094 (*fma4i_vmfnmadd_<mode>): Ditto.
13095 (*fma4i_vmfnmsub_<mode>): Ditto.
13096 (*xop_vmfrcz_<mode>): Ditto.
13097 (sse_cvtps2pi): Ditto.
13098 (sse_cvttps2pi): Ditto.
13099 (sse_cvtss2si): Ditto.
13100 (sse_cvtss2si_2): Ditto.
13101 (sse_cvtss2siq_2): Ditto.
13102 (sse_cvttss2si): Ditto.
13103 (sse_cvttss2siq): Ditto.
13104 (sse_cvtsd2si): Ditto.
13105 (sse_cvtsd2si_2): Ditto.
13106 (sse_cvtsd2siq_2): Ditto.
13107 (sse_cvttsd2si): Ditto.
13108 (sse_cvttsd2siq): Ditto.
13109 (sse_cvtsd2ss): Ditto.
13110 (sse_cvtss2sd): Ditto.
13111 (avx2_pbroadcast<mode>): Ditto.
13112 (avx2_pbroadcast<mode>_1): Ditto.
13113 (*avx_vperm_broadcast_v4sf): Ditto.
13114
13115 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13116 (sse_movlhps): Ditto.
13117 (sse_storehps): Ditto.
13118 (sse_loadhps): Ditto.
13119 (sse_storelps): Ditto.
13120 (sse_loadlps): Ditto.
13121 (*vec_concatv4sf): Ditto.
13122 (*vec_interleave_highv2df): Ditto.
13123 (*vec_interleave_lowv2df): Ditto.
13124 (*vec_extractv2df_1_sse): Ditto.
13125 (*vec_extractv2df_0_sse): Ditto.
13126 (sse2_storelpd): Ditto.
13127 (sse2_loadlpd): Ditto.
13128 (sse2_movsd): Ditto.
13129 (*vec_concatv4si): Ditto.
13130 (vec_concatv2di): Ditto.
13131
13132 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13133 for Intel asm dialect.
13134 (mmx_punpcklwd): Ditto.
13135 (mmx_punpckldq): Ditto.
13136
13137 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13138 for intel assembler dialect.
13139
1f6eac90
JJ		 131402013-07-06 Jakub Jelinek <jakub@redhat.com>
13141
13142 PR target/29776
13143 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13144 for BUILT_IN_C{LZ,LRSB}*.
13145 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13146 * tree-vrp.c (extract_range_basic): Handle
13147 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13148 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13149 fall thru to code calling set_value*.
13150 * builtins.c (expand_builtin): Remove *IMAX cases.
13151 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13152 if width is bigger than 2*HWI.
13153
2b778c9d
VM		 131542013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13155
13156 PR rtl-optimization/55342
13157 * lra-int.h (lra_subreg_reload_pseudos): New.
13158 * lra.c: Add undoing optional reloads to the block diagram.
13159 (lra_subreg_reload_pseudos): New.
13160 (lra_optional_reload_pseudos): Change comments.
13161 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13162 lra_optional_reload_pseudos after undo transformations.
13163 * lra-assigns.c (pseudo_prefix_title): New.
13164 (lra_setup_reg_renumber): Use it.
13165 (spill_for): Ditto. Check subreg reload pseudos too.
13166 (assign_by_spills): Consider subreg reload pseudos too.
13167 * lra-constraints.c (simplify_operand_subreg): Use
13168 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13169 (curr_insn_transform): Recognize and do optional reloads.
13170 (undo_optional_reloads): New.
13171 (lra_undo_inheritance): Call undo_optional_reloads.
13172
207156e4
TQ		 131732013-07-05 Thomas Quinot <quinot@adacore.com>
13174
13175 * tree-complex.c (expand_complex_operations_1): Fix typo.
13176
8990e73a
TB		 131772013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13178
13179 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13180 (tune_params): New member 'const vec_costs'.
13181 * config/aarch64/aarch64.c (generic_vector_cost): New.
13182 (generic_tunings): New member 'generic_vector_cost'.
13183 (aarch64_builtin_vectorization_cost): New.
13184 (aarch64_add_stmt_cost): New.
13185 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13186 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13187
f7bc421d
JJ		 131882013-07-03 Jakub Jelinek <jakub@redhat.com>
13189
13190 PR target/57777
13191 * config/i386/predicates.md (vsib_address_operand): Disallow
13192 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13193
e90d1568
HPN		 131942013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13195
13196 PR middle-end/55030
13197 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13198 expand_builtin_setjmp_receiver.
13199 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13200 with NULL for the label parameter.
13201 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13202 the frame-pointer. Adjust comments.
13203 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13204 only if LABEL is non-NULL.
13205
cbd7413b
YZ		 132062013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13207
13208 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13209 (ARM_ABI_AAPCS64): Ditto.
13210 (arm_abi): Ditto.
13211 (ARM_DEFAULT_ABI): Ditto.
13212
dec11868
JG		 132132013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13214
13215 * config/aarch64/aarch64-builtins.c
13216 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13217 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13218 (st1): Likewise.
13219 * config/aarch64/aarch64-simd.md
13220 (aarch64_ld1<VALL:mode>): New.
13221 (aarch64_st1<VALL:mode>): Likewise.
13222 * config/aarch64/arm_neon.h
13223 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13224
786fb9b6
ST		 132252013-07-02 Sriraman Tallam <tmsriram@google.com>
13226
0ba6aff3
UB		 13227 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13228 target ISA is AVX.
786fb9b6
ST		 13229 (ix86_option_override_internal):Turn on all -mavx target flags by
13230 default as they are dependent on AVX anyway.
13231
4b043553
CC		 132322013-07-02 Cary Coutant <ccoutant@google.com>
13233
13234 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13235 deterministic hash.
13236 (loc_checksum_ordered): Likewise.
13237 (hash_loc_operands): Remove inline keyword.
13238
c4e87a13
JJ		 132392013-07-02 Jakub Jelinek <jakub@redhat.com>
13240
13241 PR tree-optimization/57741
13242 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13243 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13244 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13245 Allow REAL_CST step_exprs if flag_associative_math.
13246 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13247
26366d28
IB		 132482013-07-02 Ian Bolton <ian.bolton@arm.com>
13249
0ba6aff3 13250 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 13251
2879bb2b
IB		 132522013-07-02 Ian Bolton <ian.bolton@arm.com>
13253
13254 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13255
afc5e8a6
KT		 132562013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13257
13258 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13259 encoding.
13260 (iorsi3_insn): Likewise.
13261 (arm_xorsi3): Likewise.
13262
31f8442b
SN		 132632013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13264
13265 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13266 to ...
13267 (attribute "type"): ... here, and prefix with "wmmx_".
13268 (attribute "core_cycles"): Update for attribute changes.
13269 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13270 (tbcstv4hi): Likewise.
13271 (tbcstv2si): Likewise.
13272 (iwmmxt_iordi3): Likewise.
13273 (iwmmxt_xordi3): Likewise.
13274 (iwmmxt_anddi3): Likewise.
13275 (iwmmxt_nanddi3): Likewise.
13276 (iwmmxt_arm_movdi): Likewise.
13277 (iwmmxt_movsi_insn): Likewise.
13278 (mov<mode>_internal): Likewise.
13279 (and<mode>3_iwmmxt): Likewise.
13280 (ior<mode>3_iwmmxt): Likewise.
13281 (xor<mode>3_iwmmxt): Likewise.
13282 (add<mode>3_iwmmxt): Likewise.
13283 (ssaddv8qi3): Likewise.
13284 (ssaddv4hi3): Likewise.
13285 (ssaddv2si3): Likewise.
13286 (usaddv8qi3): Likewise.
13287 (usaddv4hi3): Likewise.
13288 (usaddv2si3): Likewise.
13289 (sub<mode>3_iwmmxt): Likewise.
13290 (sssubv8qi3): Likewise.
13291 (sssubv4hi3): Likewise.
13292 (sssubv2si3): Likewise.
13293 (ussubv8qi3): Likewise.
13294 (ussubv4hi3): Likewise.
13295 (ussubv2si3): Likewise.
13296 (mulv4hi3_iwmmxt): Likewise.
13297 (smulv4hi3_highpart): Likewise.
13298 (umulv4hi3_highpart): Likewise.
13299 (iwmmxt_wmacs): Likewise.
13300 (iwmmxt_wmacsz): Likewise.
13301 (iwmmxt_wmacu): Likewise.
13302 (iwmmxt_wmacuz): Likewise.
13303 (iwmmxt_clrdi): Likewise.
13304 (iwmmxt_clrv8qi): Likewise.
13305 (iwmmxt_clr4hi): Likewise.
13306 (iwmmxt_clr2si): Likewise.
13307 (iwmmxt_uavgrndv8qi3): Likewise.
13308 (iwmmxt_uavgrndv4hi3): Likewise.
13309 (iwmmxt_uavgv8qi3): Likewise.
13310 (iwmmxt_uavgv4hi3): Likewise.
13311 (iwmmxt_tinsrb): Likewise.
13312 (iwmmxt_tinsrh): Likewise.
13313 (iwmmxt_tinsrw): Likewise.
13314 (iwmmxt_textrmub): Likewise.
13315 (iwmmxt_textrmsb): Likewise.
13316 (iwmmxt_textrmuh): Likewise.
13317 (iwmmxt_textrmsh): Likewise.
13318 (iwmmxt_textrmw): Likewise.
13319 (iwmxxt_wshufh): Likewise.
13320 (eqv8qi3): Likewise.
13321 (eqv4hi3): Likewise.
13322 (eqv2si3): Likewise.
13323 (gtuv8qi3): Likewise.
13324 (gtuv4hi3): Likewise.
13325 (gtuv2si3): Likewise.
13326 (gtv8qi3): Likewise.
13327 (gtv4hi3): Likewise.
13328 (gtv2si3): Likewise.
13329 (smax<mode>3_iwmmxt): Likewise.
13330 (umax<mode>3_iwmmxt): Likewise.
13331 (smin<mode>3_iwmmxt): Likewise.
13332 (umin<mode>3_iwmmxt): Likewise.
13333 (iwmmxt_wpackhss): Likewise.
13334 (iwmmxt_wpackwss): Likewise.
13335 (iwmmxt_wpackdss): Likewise.
13336 (iwmmxt_wpackhus): Likewise.
13337 (iwmmxt_wpackwus): Likewise.
13338 (iwmmxt_wpackdus): Likewise.
13339 (iwmmxt_wunpckihb): Likewise.
13340 (iwmmxt_wunpckihh): Likewise.
13341 (iwmmxt_wunpckihw): Likewise.
13342 (iwmmxt_wunpckilb): Likewise.
13343 (iwmmxt_wunpckilh): Likewise.
13344 (iwmmxt_wunpckilw): Likewise.
13345 (iwmmxt_wunpckehub): Likewise.
13346 (iwmmxt_wunpckehuh): Likewise.
13347 (iwmmxt_wunpckehuw): Likewise.
13348 (iwmmxt_wunpckehsb): Likewise.
13349 (iwmmxt_wunpckehsh): Likewise.
13350 (iwmmxt_wunpckehsw): Likewise.
13351 (iwmmxt_wunpckelub): Likewise.
13352 (iwmmxt_wunpckeluh): Likewise.
13353 (iwmmxt_wunpckeluw): Likewise.
13354 (iwmmxt_wunpckelsb): Likewise.
13355 (iwmmxt_wunpckelsh): Likewise.
13356 (iwmmxt_wunpckelsw): Likewise.
13357 (ror<mode>3): Likewise.
13358 (ashr<mode>3_iwmmxt): Likewise.
13359 (lshr<mode>3_iwmmxt): Likewise.
13360 (ashl<mode>3_iwmmxt): Likewise.
13361 (ror<mode>3_di): Likewise.
13362 (ashr<mode>3_di): Likewise.
13363 (lshr<mode>3_di): Likewise.
13364 (ashl<mode>3_di): Likewise.
13365 (iwmmxt_wmadds): Likewise.
13366 (iwmmxt_wmaddu): Likewise.
13367 (iwmmxt_tmia): Likewise.
13368 (iwmmxt_tmiaph): Likewise.
13369 (iwmmxt_tmiabb): Likewise.
13370 (iwmmxt_tmiatb): Likewise.
13371 (iwmmxt_tmiabt): Likewise.
13372 (iwmmxt_tmiatt): Likewise.
13373 (iwmmxt_tmovmskb): Likewise.
13374 (iwmmxt_tmovmskh): Likewise.
13375 (iwmmxt_tmovmskw): Likewise.
13376 (iwmmxt_waccb): Likewise.
13377 (iwmmxt_wacch): Likewise.
13378 (iwmmxt_waccw): Likewise.
13379 (iwmmxt_waligni): Likewise.
13380 (iwmmxt_walignr): Likewise.
13381 (iwmmxt_walignr0): Likewise.
13382 (iwmmxt_walignr1): Likewise.
13383 (iwmmxt_walignr2): Likewise.
13384 (iwmmxt_walignr3): Likewise.
13385 (iwmmxt_wsadb): Likewise.
13386 (iwmmxt_wsadh): Likewise.
13387 (iwmmxt_wsadbz): Likewise.
13388 (iwmmxt_wsadhz): Likewise.
13389 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13390 (iwmmxt_wabsdiffb): Likewise.
13391 (iwmmxt_wabsdiffh): Likewise.
13392 (iwmmxt_wabsdiffw): Likewise.
13393 (iwmmxt_waddsubhx): Likewise
13394 (iwmmxt_wsubaddhx): Likewise.
13395 (addc<mode>3): Likewise.
13396 (iwmmxt_avg4): Likewise.
13397 (iwmmxt_avg4r): Likewise.
13398 (iwmmxt_wmaddsx): Likewise.
13399 (iwmmxt_wmaddux): Likewise.
13400 (iwmmxt_wmaddsn): Likewise.
13401 (iwmmxt_wmaddun): Likewise.
13402 (iwmmxt_wmulwsm): Likewise.
13403 (iwmmxt_wmulwum): Likewise.
13404 (iwmmxt_wmulsmr): Likewise.
13405 (iwmmxt_wmulumr): Likewise.
13406 (iwmmxt_wmulwsmr): Likewise.
13407 (iwmmxt_wmulwumr): Likewise.
13408 (iwmmxt_wmulwl): Likewise.
13409 (iwmmxt_wqmulm): Likewise.
13410 (iwmmxt_wqmulwm): Likewise.
13411 (iwmmxt_wqmulmr): Likewise.
13412 (iwmmxt_wqmulwmr): Likewise.
13413 (iwmmxt_waddbhusm): Likewise.
13414 (iwmmxt_waddbhusl): Likewise.
13415 (iwmmxt_wqmiabb): Likewise.
13416 (iwmmxt_wqmiabt): Likewise.
13417 (iwmmxt_wqmiatb): Likewise.
13418 (iwmmxt_wqmiatt): Likewise.
13419 (iwmmxt_wqmiabbn): Likewise.
13420 (iwmmxt_wqmiabtn): Likewise.
13421 (iwmmxt_wqmiatbn): Likewise.
13422 (iwmmxt_wqmiattn): Likewise.
13423 (iwmmxt_wmiabb): Likewise.
13424 (iwmmxt_wmiabt): Likewise.
13425 (iwmmxt_wmiatb): Likewise.
13426 (iwmmxt_wmiatt): Likewise.
13427 (iwmmxt_wmiabbn): Likewise.
13428 (iwmmxt_wmiabtn): Likewise.
13429 (iwmmxt_wmiatbn): Likewise.
13430 (iwmmxt_wmiattn): Likewise.
13431 (iwmmxt_wmiawbb): Likewise.
13432 (iwmmxt_wmiawbt): Likewise.
13433 (iwmmxt_wmiawtb): Likewise.
13434 (iwmmxt_wmiawtt): Likewise.
13435 (iwmmxt_wmiawbbn): Likewise.
13436 (iwmmxt_wmiawbtn): Likewise.
13437 (iwmmxt_wmiawtbn): Likewise.
13438 (iwmmxt_wmiawttn): Likewise.
13439 (iwmmxt_wmerge): Likewise.
13440 (iwmmxt_tandc<mode>3): Likewise.
13441 (iwmmxt_torc<mode>3): Likewise.
13442 (iwmmxt_torvsc<mode>3): Likewise.
13443 (iwmmxt_textrc<mode>3): Likewise.
13444 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13445 (wmmxt_pack): Likewise.
13446 (wmmxt_mult_c1): Likewise.
13447 (wmmxt_mult_c2): Likewise.
13448 (wmmxt_alu_c1): Likewise.
13449 (wmmxt_alu_c2): Likewise.
13450 (wmmxt_alu_c3): Likewise.
13451 (wmmxt_transfer_c1): Likewise.
13452 (wmmxt_transfer_c2): Likewise.
13453 (wmmxt_transfer_c3): Likewise.
13454 (marvell_f_iwmmxt_wstr): Likewise.
13455 (marvell_f_iwmmxt_wldr): Likewise.
13456
bdb7bf8a
YZ		 134572013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13458
0ba6aff3 13459 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 13460
e6631d38
VM		 134612013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13462
13463 Revert:
13464 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13465 * lra-constraints.c (need_for_split_p): Check call used hard regs
13466 living through calls.
13467
13468 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13469 call used regs for call insn.
13470
53c1275b
JJ		 134712013-06-28 Jakub Jelinek <jakub@redhat.com>
13472
13473 PR target/57736
0ba6aff3
UB		 13474 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13475 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 13476
1141ed3f
BI		 134772013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13478
13479 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 13480
e32e4c4a
VM		 134812013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13482
13483 * lra-constraints.c (need_for_split_p): Check call used hard regs
13484 living through calls.
13485
f161bfd3
MM		 134862013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13487
13488 PR target/57744
13489 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13490 to tie with any other modes. Eliminate Altivec vector mode tests,
13491 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13492 code, to return 0 if testing MODE2 for a condition, if we've
13493 already tested MODE1 for the same condition.
13494
7eda14e1
MS		 134952013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13496
13497 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13498 layout.
13499
c822f852
MS		 135002013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13501
13502 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13503 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13504
da4f13a4
MS		 135052013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13506
0ba6aff3
UB		 13507 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13508 Define.
da4f13a4
MS		 13509 (aarch64_symbolic_constant_p): Remove.
13510 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13511 static. Fix line length and white space.
13512 (aarch64_symbolic_constant_p): Remove.
13513 * config/aarch64/predicates.md (aarch64_valid_symref):
13514 Use aarch64_classify_symbol_expression.
13515
956a95a5
KT		 135162013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13517
13518 * config/arm/constraints.md (Ts): New constraint.
13519 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13520 16-bit encodings.
13521 (compare_scc): Use "Ts" constraint for operand 0.
13522 (ior_scc_scc): Likewise.
13523 (and_scc_scc): Likewise.
13524 (and_scc_scc_nodom): Likewise.
13525 (ior_scc_scc_cmp): Likewise for operand 7.
13526 (and_scc_scc_cmp): Likewise.
13527 * config/arm/thumb2.md (thumb2_movsi_insn):
13528 Add alternatives for 16-bit encodings.
13529 (thumb2_movhi_insn): Likewise.
13530 (thumb2_movsicc_insn): Likewise.
13531 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13532 (thumb2_negscc): Use "Ts" constraint.
13533 Move mvn instruction outside cond_exec block.
13534 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13535 for 16-bit encodings.
13536
31dfce10
KT		 135372013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13538
13539 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13540 encoding.
13541 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13542 (mulsi3subsi): Likewise.
13543 (mulsidi3adddi): Likewise.
13544 (mulsidi3_v6): Likewise.
13545 (umulsidi3_v6): Likewise.
13546 (umulsidi3adddi_v6): Likewise.
13547 (smulsi3_highpart_v6): Likewise.
13548 (umulsi3_highpart_v6): Likewise.
13549 (mulhisi3tb): Likewise.
13550 (mulhisi3bt): Likewise.
13551 (mulhisi3tt): Likewise.
13552 (maddhisi4): Likewise.
13553 (maddhisi4tb): Likewise.
13554 (maddhisi4tt): Likewise.
13555 (maddhidi4): Likewise.
13556 (maddhidi4tb): Likewise.
13557 (maddhidi4tt): Likewise.
13558 (zeroextractsi_compare0_scratch): Likewise.
13559 (insv_zero): Likewise.
13560 (insv_t2): Likewise.
13561 (anddi_notzesidi_di): Likewise.
13562 (anddi_notsesidi_di): Likewise.
13563 (andsi_notsi_si): Likewise.
13564 (iordi_zesidi_di): Likewise.
13565 (xordi_zesidi_di): Likewise.
13566 (andsi_iorsi3_notsi): Likewise.
13567 (smax_0): Likewise.
13568 (smax_m1): Likewise.
13569 (smin_0): Likewise.
13570 (not_shiftsi): Likewise.
13571 (unaligned_loadsi): Likewise.
13572 (unaligned_loadhis): Likewise.
13573 (unaligned_loadhiu): Likewise.
13574 (unaligned_storesi): Likewise.
13575 (unaligned_storehi): Likewise.
13576 (extv_reg): Likewise.
13577 (extzv_t2): Likewise.
13578 (divsi3): Likewise.
13579 (udivsi3): Likewise.
13580 (arm_zero_extendhisi2addsi): Likewise.
13581 (arm_zero_extendqisi2addsi): Likewise.
13582 (compareqi_eq0): Likewise.
13583 (arm_extendhisi2_v6): Likewise.
13584 (arm_extendqisi2addsi): Likewise.
13585 (arm_movt): Likewise.
13586 (thumb2_ldrd): Likewise.
13587 (thumb2_ldrd_base): Likewise.
13588 (thumb2_ldrd_base_neg): Likewise.
13589 (thumb2_strd): Likewise.
13590 (thumb2_strd_base): Likewise.
13591 (thumb2_strd_base_neg): Likewise.
13592 (arm_negsi2): Add alternative for 16-bit encoding.
13593 (arm_one_cmplsi2): Likewise.
13594
1572e697
KT		 135952013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13596
13597 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13598 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13599 (movdfcc): Likewise.
13600 * config/arm/vfp.md (*thumb2_movsf_vfp):
13601 Disable predication for arm_restrict_it.
13602 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13603 (*thumb2_movdfcc_vfp): Likewise.
13604 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13605 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13606 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13607 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13608 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13609 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13610 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13611 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13612 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13613 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13614 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13615 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13616 Disable predication for arm_restrict_it.
13617
ebf8f0ea
KY		 136182013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13619
13620 * config/i386/bmiintrin.h (_bextr_u32): New.
13621 (_bextr_u64): Ditto.
13622
a290fcda
RS		 136232013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13624
13625 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13626 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13627 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13628 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13629 * config/mips/n32-elf.h: ...this new file.
13630
ea169996
MG		 136312013-06-27 Marc Glisse <marc.glisse@inria.fr>
13632
13633 PR target/57224
13634 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13635 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13636
12211b99 136372013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM		 13638
13639 * config/mips/mips-tables.opt: Regenerate.
13640 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 13641 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 13642 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 13643
2444244e
JJ		 136442013-06-27 Jakub Jelinek <jakub@redhat.com>
13645
cd06d2a2
JJ		 13646 PR target/57623
13647 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13648 constraints of operand 1 and 2.
13649
2444244e
JJ		 13650 PR target/57623
13651 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13652 to match RTL canonicalization. Swap predicates and
13653 constraints of operand 1 and 2.
13654
2f259720
VM		 136552013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13656
0ba6aff3
UB		 13657 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13658 Process OP_INOUT regs for splitting too.
2f259720 13659
fef4d2b3
JJ		 136602013-06-27 Jakub Jelinek <jakub@redhat.com>
13661
13662 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13663 decl before the loop, initialize to NULL.
13664 (vectorizable_load): Initialize ptr_incr to NULL.
13665
79ee9826
MJ		 136662013-06-27 Martin Jambor <mjambor@suse.cz>
13667
13668 PR lto/57208
13669 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13670 * ipa-ref.c (ipa_maybe_record_reference): New function.
13671 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13672 * ipa-cp.c (create_specialized_node): Record potential references from
13673 aggvals.
13674 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13675
e18b4a81
YZ		 136762013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13677
13678 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13679 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13680 'mode' to force_reg.
13681 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13682 (aarch64_expand_mov_immediate): Likewise.
13683
9c023bf0
YZ		 136842013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13685
13686 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13687 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13688
5a3fe9b6
AK		 136892013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13690
13691 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13692 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13693 (struct machine_function): Add tbegin_p.
13694 (s390_canonicalize_comparison): Fold CC mode compares to
13695 conditional jump if possible.
13696 (s390_emit_jump): Return the emitted jump.
13697 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13698 Handle CCRAWmode compares.
13699 (s390_option_override): Default to -mhtm if available.
13700 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13701 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13702 FPRs instead of df_regs_ever_live_p.
13703 (s390_optimize_nonescaping_tx): New function.
13704 (s390_init_frame_layout): Extend clobbered_regs array to cover
13705 FPRs as well.
13706 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13707 (s390_expand_tbegin): New function.
13708 (enum s390_builtin): New enum definition.
13709 (code_for_builtin): New array definition.
13710 (s390_init_builtins): New function.
13711 (s390_expand_builtin): New function.
13712 (TARGET_INIT_BUILTINS): Define.
13713 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 13714 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK		 13715 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13716 (s390_alc_comparison): Likewise.
13717 * config/s390/s390-modes.def: Add CCRAWmode.
13718 * config/s390/s390.h (processor_flags): Add PF_TX.
13719 (TARGET_CPU_HTM): Define macro.
13720 (TARGET_HTM): Define macro.
13721 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13722 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13723 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13724 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 13725 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK		 13726 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13727 ("*cc_to_int"): Move up.
13728 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13729 constants other than 0.
13730 ("*ccraw_to_int"): New insn and splitter definition.
13731 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13732 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13733 ("tx_assist"): New expander.
13734 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13735 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13736 * config/s390/s390.opt: Add -mhtm option.
13737 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13738 * config/s390/htmxlintrin.h: New file.
13739 * config/s390/htmintrin.h: New file.
13740 * config/s390/s390intrin.h: New file.
13741 * doc/extend.texi: Document htm builtins.
13742 * config.gcc: Add the new header files to extra_headers.
13743
9e216629
TS		 137442013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13745
13746 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 13747 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 13748
fd8c65e7 137492013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 13750 Pat Haugen <pthaugen@us.ibm.com>
13751 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM		 13752
13753 * config/rs6000/power8.md: New.
13754 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13755 setting for power8 entry.
13756 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13757 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13758 test for Power4/Power5 only.
13759 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13760 support.
13761 (force_new_group): Adjust comment.
13762 * config/rs6000/rs6000.md: Include power8.md.
13763
b24a2ce5
GY		 137642013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13765
13766 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13767 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13768 declaration.
13769 (tune_params): Update comment.
13770 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13771 (arm_max_conditional_execute): New function.
13772 (thumb2_final_prescan_insn): Use max_insn_skipped and
13773 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13774
f6fdeb62
JJ		 137752013-06-25 Jakub Jelinek <jakub@redhat.com>
13776
13777 PR tree-optimization/57705
13778 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13779 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 13780 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ		 13781 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13782
c13bc3d9
MJ		 137832013-06-25 Martin Jambor <mjambor@suse.cz>
13784
13785 PR middle-end/57670
13786 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13787 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13788 calls in the dump.
13789 (ipa_note_param_call): Initialize member_ptr flag.
13790 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13791 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13792 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13793 (ipa_read_indirect_edge_info): Likewise.
13794
8a845901
RB		 137952013-06-25 Richard Biener <rguenther@suse.de>
13796
13797 PR middle-end/56977
13798 * passes.c (init_optimization_passes): Move pass_fold_builtins
13799 and pass_dce earlier with -Og.
13800
33be0bec
EB		 138012013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13802
13803 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13804 <BIT_FIELD_REF>: Remove trailing TAB.
13805 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13806 remove blank line.
13807
5fe8e757
MJ		 138082013-06-24 Martin Jambor <mjambor@suse.cz>
13809
13810 PR tree-optimization/57358
13811 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13812 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13813 (ipa_analyze_params_uses): Generate pessimistic info when true.
13814
44a60244
MJ		 138152013-06-24 Martin Jambor <mjambor@suse.cz>
13816
13817 PR tree-optimization/57539
13818 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13819 global.inlined_to of the new node to it. All callers changed.
13820 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13821 inlining_into, pass it to cgraph_clone_node.
13822 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13823 ipa_free_edge_args_substructures.
13824 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13825 rdesc linked list. Do not assert rdesc edges have inlined caller.
13826 Assert we have found an rdesc in the rdesc list.
13827
7c5848b8
RB		 138282013-06-24 Richard Biener <rguenther@suse.de>
13829
0ba6aff3 13830 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB		 13831 (pointer_set_lookup): Declare.
13832 (class pointer_map): New template class implementing a
13833 generic pointer to T map.
13834 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13835 pointer_map<T>::contains, pointer_map<T>::insert,
13836 pointer_map<T>::traverse): New functions.
13837 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13838 (pointer_set_lookup): New function.
13839 (pointer_set_contains): Use pointer_set_lookup.
13840 (pointer_set_insert): Likewise.
13841 (insert_aux): Remove.
13842 (struct pointer_map_t): Embed a pointer_set_t.
13843 (pointer_map_create): Adjust.
13844 (pointer_map_destroy): Likewise.
13845 (pointer_map_contains): Likewise.
13846 (pointer_map_insert): Likewise.
13847 (pointer_map_traverse): Likewise.
13848 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13849 pointer_map<unsigned> instead of a pointer_map_t.
13850 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13851 (streamer_tree_cache_lookup): Likewise.
13852 (streamer_tree_cache_create): Likewise.
13853 (streamer_tree_cache_delete): Likewise.
13854 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13855 pointer_map<unsigned> instead of a pointer_map_t.
13856 (lto_init_tree_ref_encoder): Adjust.
13857 (lto_destroy_tree_ref_encoder): Likewise.
13858 * lto-section-out.c (lto_output_decl_index): Likewise.
13859 (lto_record_function_out_decl_state): Likewise.
13860 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13861
3dbc97a9
RB		 138622013-06-24 Richard Biener <rguenther@suse.de>
13863
13864 PR tree-optimization/57488
13865 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13866
8345965a
AM		 138672013-06-24 Alan Modra <amodra@gmail.com>
13868
13869 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13870 (gen_easy_altivec_constant): Likewise.
13871 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13872 easy_vector_constant_msb): Likewise.
13873
b7df379f
JJ		 138742013-06-23 Jakub Jelinek <jakub@redhat.com>
13875
13876 PR target/57688
13877 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13878 add missing return true.
13879
12211b99 138802013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE		 13881
13882 PR target/52483
13883 * config/sh/predicates.md (general_extend_operand): Invoke
13884 general_movsrc_operand for memory operands.
13885 (general_movsrc_operand): Allow reg+reg addressing, do not use
13886 general_operand for memory operands.
13887
97db2bf7
ST		 138882013-06-23 Sriraman Tallam <tmsriram@google.com>
13889
13890 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13891 when current target options does not apply.
13892 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13893 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB		 13894 * config/i386/bmiintrin.h: Pass appropriate target
13895 attributes to header.
97db2bf7
ST		 13896 * config/i386/mmintrin.h: Ditto.
13897 * config/i386/nmmintrin.h: Ditto.
13898 * config/i386/avx2intrin.h: Ditto.
13899 * config/i386/fxsrintrin.h: Ditto.
13900 * config/i386/tbmintrin.h: Ditto.
13901 * config/i386/xsaveintrin.h: Ditto.
13902 * config/i386/f16cintrin.h: Ditto.
13903 * config/i386/xtestintrin.h: Ditto.
13904 * config/i386/xsaveoptintrin.h: Ditto.
13905 * config/i386/bmi2intrin.h: Ditto.
13906 * config/i386/lzcntintrin.h: Ditto.
13907 * config/i386/smmintrin.h: Ditto.
13908 * config/i386/wmmintrin.h: Ditto.
13909 * config/i386/x86intrin.h: Remove all header include guards.
13910 * config/i386/prfchwintrin.h: Ditto.
13911 * config/i386/pmmintrin.h: Ditto.
13912 * config/i386/tmmintrin.h: Ditto.
13913 * config/i386/xmmintrin.h: Ditto.
13914 * config/i386/popcntintrin.h: Ditto.
13915 * config/i386/rdseedintrin.h: Ditto.
13916 * config/i386/ammintrin.h: Ditto.
13917 * config/i386/emmintrin.h: Ditto.
13918 * config/i386/immintrin.h: Remove all header include guards.
13919 * config/i386/fma4intrin.h: Ditto.
13920 * config/i386/lwpintrin.h: Ditto.
13921 * config/i386/xopintrin.h: Ditto.
13922 * config/i386/ia32intrin.h: Ditto.
13923 * config/i386/avxintrin.h: Ditto.
13924 * config/i386/rtmintrin.h: Ditto.
13925 * config/i386/fmaintrin.h: Ditto.
13926 * config/i386/mm3dnow.h: Ditto.
13927
495e6879
ST		 139282013-06-22 Sriraman Tallam <tmsriram@google.com>
13929
13930 * common/config/i386/i386-common.c: Handle LZCNT.
13931
11e8f687
AK		 139322013-06-22 Andi Kleen <ak@linux.intel.com>
13933
13934 * doc/extend.texi: Use __atomic_store_n instead of
13935 __atomic_store in HLE example.
13936
12211b99 139372013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE		 13938
13939 * config/sh/sh.c: Remove <cstdlib> workaround.
13940
927f908b
AK		 139412013-06-21 Andi Kleen <ak@linux.intel.com>
13942
0ba6aff3 13943 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 13944
cca410b6
AK		 139452013-06-21 Andi Kleen <ak@linux.intel.com>
13946
13947 * doc/extend.texi: Document that __atomic_clear and
13948 __atomic_test_and_set should only be used with bool.
13949
d84db1ed
JH		 139502013-06-20 Jan Hubicka <jh@suse.cz>
13951
13952 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13953 types_same_for_odr.
13954 * tree.c (decls_same_for_odr): New function.
13955 (same_for_edr): New function.
13956 (types_same_for_odr): New function.
13957 (get_binfo_at_offset): Use it.
13958 * tree.h (types_same_for_odr): Declare.
13959
12211b99 139602013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE		 13961 Jason Merrill <jason@redhat.com>
13962
13963 * system.h: Include <cstdlib> as well as <stdlib.h>.
13964
49b4ff63
UB		 139652013-06-20 Uros Bizjak <ubizjak@gmail.com>
13966
13967 PR target/57655
13968 * config/i386/i386.c (construct_container): Report error if
13969 long double is used with disabled x87 float returns.
13970
7a89ad00
JH		 139712013-06-20 Jan Hubicka <jh@suse.cz>
13972
13973 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13974
12211b99 139752013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR		 13976
13977 PR rtl-optimization/57425
13978 PR rtl-optimization/57569
13979 * alias.c (write_dependence_p): Remove parameters mem_mode and
13980 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13981 Changed all callers.
13982 (canon_anti_dependence): Get comments and semantics in sync.
13983 Add parameter mem_canonicalized. Changed all callers.
13984 * rtl.h (canon_anti_dependence): Update prototype.
13985
d16e9a99
RB		 139862013-06-20 Richard Biener <rguenther@suse.de>
13987
13988 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13989 byte case, inline streamer_read_uchar and defer section
13990 overrun check.
13991
c3a9b91b
RB		 139922013-06-20 Richard Biener <rguenther@suse.de>
13993
13994 PR tree-optimization/57584
13995 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13996 SSA names into the expanded expression that take part in
13997 abnormal coalescing.
13998
371e77e3 139992013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS		 14000
14001 * gcov.c (print_usage): Handle new option.
14002 (process_args): Ditto.
14003 (get_gcov_intermediate_filename): New function.
14004 (output_intermediate_file): New function.
14005 (output_gcov_file): New function
14006 (generate_results): Handle new option.
14007 (release_function): Relase demangled name.
14008 (read_graph_file): Handle demangled name.
14009 (output_lines): Ditto.
14010 * doc/gcov.texi: Document gcov intermediate format.
14011
11f5b71a
VM		 140122013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14013
14014 PR bootstrap/57604
14015 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14016 (lra_emit_add): Use the functions. Add comment about Y as an
14017 address segment.
14018
e1f9340b
DE		 140192013-06-19 David Edelsohn <dje.gcc@gmail.com>
14020
14021 PR driver/57652
14022 * collect2.c (collect_atexit): New.
14023 (collect_exit): Delete.
14024 (main): Register collect_atexit with atexit.
14025 (collect_wait): Change collect_exit to exit.
14026 (do_wait): Same.
14027 * collect2.h (collect_exit): Delete.
14028 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14029
e3f9e0ac
WM		 140302013-06-19 Wei Mi <wmi@google.com>
14031
14032 PR rtl-optimization/57518
14033 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14034 if regno is used in paradoxical subreg.
14035 (update_equiv_regs): Check pdx_subregs[regno] before
14036 set a reg to be equivalent with a mem.
14037
a09f1a76
MK		 140382013-06-19 Matthias Klose <doko@ubuntu.com>
14039
14040 PR driver/57651
14041 * file-find.h (find_a_file): Add a mode parameter.
14042 * file-find.c (find_a_file): Likewise.
14043 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14044 with X_OK for the executables.
14045 * collect2.c (main): Call find_a_file with X_OK.
14046
7d18b0ad
SE		 140472013-06-19 Steve Ellcey <sellcey@mips.com>
14048
14049 PR target/56942
14050 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14051 Use NEXT_INSN instead of next_real_insn.
14052
6a6dac52
JH		 140532013-06-19 Jan Hubicka <jh@suse.cz>
14054
14055 * cgraph.h (const_value_known_p): Replace by ...
14056 (ctor_for_folding): .. this one.
14057 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14058 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14059 * expr.c (expand_expr_real_1): Likewise.
14060 (string_constant): Likewise.
14061 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14062 * ipa.c (process_references): Likewise.
14063 (symtab_remove_unreachable_nodes): Likewise.
14064 * ipa-inline-analysis.c (param_change_prob): Likewise.
14065 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14066 (get_base_constructor): Likwise.
14067 * varpool.c (varpool_remove_node): Likewise.
14068 (varpool_remove_initializer): LIkewise.
14069 (dump_varpool_node): LIkwise.
14070 (const_value_known_p): Rewrite to ...
14071 (ctor_for_folding): ... this one.
14072
216c12ab
JJ		 140732013-06-19 Jakub Jelinek <jakub@redhat.com>
14074
14075 PR driver/57651
14076 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14077 PERSONALITY in $PATH derived prefixes.
14078
d53e2f99
JL		 140792013-06-19 Jeff Law <law@redhat.com>
14080
ecdbf306
JL		 14081 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14082 in comment.
14083
d53e2f99
JL		 14084 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14085 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14086 booleans.
14087
89e7fc04
SN		 140882013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14089
14090 * config/arm/vfp.md: Move VFP instruction classification documentation
14091 to ...
14092 * config/arm/arm.md: ... here. Update instruction classification
14093 documentation.
14094
5e7f6aaa
RE		 140952013-06-19 Richard Earnshaw <rearnsha@arm.com>
14096
14097 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14098 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14099 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14100
5fdc1e5d
SB		 141012013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14102
14103 PR target/57609
14104 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14105 with NEXT_INSN. Use tablejump_p to check for jump table data
14106 insns.
14107
e952afba
PC		 141082013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14109
14110 PR c++/56544
14111 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14112 that now in C++ the value is correct per the C++ standards.
14113
3fd0303a
RB		 141142013-06-19 Richard Biener <rguenther@suse.de>
14115
14116 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14117 for global context.
14118
b04e4b08
AK		 141192013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14120
14121 Revert:
14122 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14123
14124 PR target/57609
14125 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14126 with next_active_insn.
14127
bef8491a
ST		 141282013-06-18 Sriraman Tallam <tmsriram@google.com>
14129
14130 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14131 functions are inlined during failures to flag an error.
14132 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14133 in early inline pass.
22fc7d3d 14134
c21b257d
L		 141352013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14136
14137 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14138 in comments.
14139
33255ae3
JB		 141402013-06-18 Julian Brown <julian@codesourcery.com>
14141
14142 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14143 Permit virtual register pre-reload if !strict.
14144 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14145 change.
14146 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14147 prototype.
14148 * config/arm/neon.md (movmisalign<mode>): Use
14149 neon_perm_struct_or_reg_operand instead of
14150 neon_struct_or_register_operand.
14151 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14152 neon_permissive_struct_operand instead of neon_struct_operand.
14153 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14154 neon_vector_mem_operand.
14155 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14156 neon_vector_mem_operand.
14157 (neon_permissive_struct_operand): New.
14158 (neon_struct_or_register_operand): Rename to...
14159 (neon_perm_struct_or_reg_operand): This. Adjust call to
14160 neon_vector_mem_operand.
14161
d579fcda
RB		 141622013-06-18 Richard Biener <rguenther@suse.de>
14163
14164 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14165 * lto-streamer.h: Include pointer-set.h.
14166 (struct lto_decl_slot): Remove.
14167 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14168 Remove next_index entry.
14169 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14170 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14171 (lto_init_tree_ref_encoder): Adjust.
14172 (lto_destroy_tree_ref_encoder): Likewise.
14173 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14174 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14175 (lto_output_decl_index): Adjust.
14176 (lto_new_out_decl_state): Likewise.
14177 (lto_record_function_out_decl_state): Likewise.
14178 * lto-streamer-out.c (copy_function): Likewise.
14179
31ee20ba
RB		 141802013-06-18 Richard Biener <rguenther@suse.de>
14181
14182 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14183 * cgraphunit.c: Include cfgloop.h.
14184 (init_lowered_empty_function): Initialize the loop tree.
14185 (assemble_thunk): Insert new BBs into loops.
14186
ac0511f2
RB		 141872013-06-18 Richard Biener <rguenther@suse.de>
14188
14189 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14190 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14191 the map from cache entry to cache index optional.
14192 (streamer_tree_cache_replace_tree): Adjust accordingly.
14193 (streamer_tree_cache_append): Likewise.
14194 (streamer_tree_cache_delete): Likewise.
14195 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14196 streamer cache map from cache entry to cache index.
14197 * lto-streamer-out.c (create_output_block): Adjust.
14198
09485a08
SN		 141992013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14200
22fc7d3d
UB		 14201 * config/arm/arm.md (attribute "insn"): Move multiplication and
14202 division attributes to...
09485a08
SN		 14203 (attribute "type"): ... here. Remove mult.
14204 (attribute "mul32"): New attribute.
14205 (attribute "mul64"): Add umaal.
14206 (*arm_mulsi3): Update attributes.
14207 (*arm_mulsi3_v6): Likewise.
14208 (*thumb_mulsi3): Likewise.
14209 (*thumb_mulsi3_v6): Likewise.
14210 (*mulsi3_compare0): Likewise.
14211 (*mulsi3_compare0_v6): Likewise.
14212 (*mulsi_compare0_scratch): Likewise.
14213 (*mulsi_compare0_scratch_v6): Likewise.
14214 (*mulsi3addsi): Likewise.
14215 (*mulsi3addsi_v6): Likewise.
14216 (*mulsi3addsi_compare0): Likewise.
14217 (*mulsi3addsi_compare0_v6): Likewise.
14218 (*mulsi3addsi_compare0_scratch): Likewise.
14219 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14220 (*mulsi3subsi): Likewise.
14221 (*mulsidi3adddi): Likewise.
14222 (*mulsi3addsi_v6): Likewise.
14223 (*mulsidi3adddi_v6): Likewise.
14224 (*mulsidi3_nov6): Likewise.
14225 (*mulsidi3_v6): Likewise.
14226 (*umulsidi3_nov6): Likewise.
14227 (*umulsidi3_v6): Likewise.
14228 (*umulsidi3adddi): Likewise.
14229 (*umulsidi3adddi_v6): Likewise.
14230 (*smulsi3_highpart_nov6): Likewise.
14231 (*smulsi3_highpart_v6): Likewise.
14232 (*umulsi3_highpart_nov6): Likewise.
14233 (*umulsi3_highpart_v6): Likewise.
14234 (mulhisi3): Likewise.
14235 (*mulhisi3tb): Likewise.
14236 (*mulhisi3bt): Likewise.
14237 (*mulhisi3tt): Likewise.
14238 (maddhisi4): Likewise.
14239 (*maddhisi4tb): Likewise.
14240 (*maddhisi4tt): Likewise.
14241 (maddhidi4): Likewise.
14242 (*maddhidi4tb): Likewise.
14243 (*maddhidi4tt): Likewise.
14244 (divsi3): Likewise.
14245 (udivsi3): Likewise.
14246 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14247 (thumb2_mulsi_short_compare0): Likewise.
14248 (thumb2_mulsi_short_compare0_scratch): Likewise.
14249 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14250 (1020mult2): Likewise.
14251 (1020mult3): Likewise.
14252 (1020mult4): Likewise.
14253 (1020mult5): Likewise.
14254 (1020mult6): Likewise.
22fc7d3d
UB		 14255 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14256 change.
09485a08
SN		 14257 (cortex_a15_mult64): Likewise.
14258 (cortex_a15_sdiv): Likewise.
14259 (cortex_a15_udiv): Likewise.
14260 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14261 (mult2): Likewise.
14262 (mult3): Likewise.
14263 (mult4): Likewise.
14264 (mult5): Likewise.
14265 (mult6): Likewise.
14266 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14267 (pj4_ir_div): Likewise.
14268 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14269 (11_mult2): Likewise.
14270 (11_mult3): Likewise.
14271 (11_mult4): Likewise.
14272 (11_mult5): Likewise.
14273 (11_mult6): Likewise.
14274 (11_mult7): Likewise.
14275 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14276 (cortex_a8_mla): Likewise.
14277 (cortex_a8_mull): Likewise.
14278 (cortex_a8_smulwy): Likewise.
14279 (cortex_a8_smlald): Likewise.
14280 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14281 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14282 (cortex_r4_mul_3): Likewise.
14283 (cortex_r4_mla_4): Likewise.
14284 (cortex_r4_mla_3): Likewise.
14285 (cortex_r4_smlald): Likewise.
14286 (cortex_r4_mull): Likewise.
14287 (cortex_r4_sdiv): Likewise.
14288 (cortex_r4_udiv): Likewise.
14289 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14290 (cortex_a7_idiv): Likewise.
14291 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14292 (9_mult2): Likewise.
14293 (9_mult3): Likewise.
14294 (9_mult4): Likewise.
14295 (9_mult5): Likewise.
14296 (9_mult6): Likewise.
14297 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14298 (cortex_a53_sdiv): Likewise.
14299 (cortex_a53_udiv): Likewise.
14300 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14301 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14302 (mp626_mult2): Likewise.
14303 (mp626_mult3): Likewise.
14304 (mp626_mult4): Likewise.
14305 * config/arm/fa526.md (526_mult1): Update attribute change.
14306 (526_mult2): Likewise.
14307 * config/arm/arm-generic.md (mult): Update attribute change.
14308 (mult_ldsched_strongarm): Likewise.
14309 (mult_ldsched): Likewise.
14310 (multi_cycle): Likewise.
14311 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14312 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14313 (606te_mult2): Likewise.
14314 (606te_mult3): Likewise.
14315 (606te_mult4): Likewise.
14316 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14317 (cortex_a9_mac16): Likewise.
14318 (cortex_a9_multiply): Likewise.
14319 (cortex_a9_mac): Likewise.
14320 (cortex_a9_multiply_long): Likewise.
14321 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14322 (626te_mult2): Likewise.
14323 (626te_mult3): Likewise.
14324 (626te_mult4): Likewise.
14325
8951f345
RB		 143262013-06-18 Richard Biener <rguenther@suse.de>
14327
14328 PR lto/57334
14329 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14330
d8101d05
AK		 143312013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14332
14333 PR target/57609
14334 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14335 with next_active_insn.
14336
69eff9da
AM		 143372013-06-18 Alan Modra <amodra@gmail.com>
14338
14339 * config/rs6000/rs6000.h (enum data_align): New.
14340 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14341 (DATA_ABI_ALIGNMENT): Define.
14342 (CONSTANT_ALIGNMENT): Correct comment.
14343 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14344 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14345
8ed00d76
DM		 143462013-06-17 David Malcolm <dmalcolm@redhat.com>
14347
14348 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14349 ATTRIBUTE_UNUSED marking.
14350
3d33d06b
SN		 143512013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14352
14353 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14354 alternative and update.
14355 (aarch64_dup_lanedi): Delete.
14356 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14357 * config/aarch64/aarch64-simd-builtins.def: Update.
14358
ee03e71d
RB		 143592013-06-17 Richard Biener <rguenther@suse.de>
14360
14361 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14362 (lto_input_scc): Declare.
14363 (lto_input_tree_1): Likewise.
14364 (struct lto_stats_d): Add num_tree_bodies_output and
14365 num_pickle_refs_output.
14366 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14367 (lto_read_tree_1): Split out from ...
14368 (lto_read_tree): ... this.
14369 (lto_input_scc): New function.
14370 (lto_input_tree_1): Split out from ...
14371 (lto_input_tree): ... this. Handle LTO_tree_scc.
14372 (lto_data_in_create): Create the streamer cache without hashes.
14373 * lto-streamer-out.c (create_output_block): Create the streamer
14374 cache with hashes when not doing WPA.
14375 (lto_write_tree_1): Split out from ...
14376 (lto_write_tree): ... this.
14377 (get_symbol_initial_value): New function.
14378 (lto_output_tree_1): Split out from ...
14379 (lto_output_tree): ... this. Write trees as series of SCCs
14380 using a DFS walk via DFS_write_tree.
14381 (struct sccs, struct scc_entry): New types.
14382 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14383 (DFS_write_tree_body): New function.
14384 (DFS_write_tree): Likewise.
14385 (hash_tree): Likewise.
14386 (scc_entry_compare): Likewise.
14387 (hash_scc): Likewise.
14388 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14389 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14390 TREE_CHAIN as regular reference.
14391 (streamer_read_integer_cst): Remove.
14392 (streamer_get_pickled_tree): Adjust.
14393 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14394 of DECL_EXTERNALs in BLOCK_VARS for now.
14395 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14396 reference.
14397 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14398 Add hash value argument and record that if hashes are recorded
14399 in the cache.
14400 (streamer_tree_cache_insert_1): Adjust.
14401 (streamer_tree_cache_insert): Likewise.
14402 (streamer_tree_cache_insert_at): Rename to ...
14403 (streamer_tree_cache_replace_tree): ... this and adjust.
14404 (streamer_tree_cache_append): Adjust.
14405 (record_common_node): Likewise.
14406 (streamer_tree_cache_create): Add argument whether to
14407 record hash values together with trees.
14408 (streamer_tree_cache_delete): Adjust.
14409 * tree-streamer.h (struct streamer_tree_cache_d): Add
14410 vector of hashes.
14411 (streamer_read_integer_cst): Remove.
14412 (streamer_tree_cache_insert): Adjust.
14413 (streamer_tree_cache_append): Likewise.
14414 (streamer_tree_cache_insert_at): Rename to ...
14415 (streamer_tree_cache_replace_tree): ... this and adjust.
14416 (streamer_tree_cache_create): Add argument whether to record hashes.
14417 (streamer_tree_cache_get): Rename to ...
14418 (streamer_tree_cache_get_tree): ... this.
14419 (streamer_tree_cache_get_hash): New function.
14420 * tree.c (cache_integer_cst): New function.
14421 * tree.h (cache_integer_cst): Declare.
14422 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14423 * lto-symtab.c (lto_varpool_replace_node): Only release
14424 DECL_INITIAL of non-prevailing decls.
14425 * varpool.c (varpool_remove_initializer): Do not release
14426 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14427
0de86a92
JU		 144282013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14429
14430 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14431 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14432 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14433 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14434 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14435 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14436 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14437 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14438 instead of TARGET_64BIT.
14439 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14440 Require ISA_HAS_<D>DIV.
14441
74240413
RS		 144422013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14443
14444 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14445 (mips*-*-linux*): Move default with_llsc setting to where other
14446 defaults are set.
14447 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14448 with_arch block.
14449 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14450 Likewise. Remove default with_tune setting. Move default float
14451 setting to its own block. Handle with_llsc in the same block as above.
14452
12211b99 144532013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR		 14454
14455 PR rtl-optimization/57425
14456 PR rtl-optimization/57569
299a5f6a 14457 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR		 14458 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14459 Changed all callers.
14460 (canon_anti_dependence): New function.
14461 * cse.c (check_dependence): Use canon_anti_dependence.
14462 * cselib.c (cselib_invalidate_mem): Likewise.
14463 * rtl.h (canon_anti_dependence): Declare.
14464
b259d352
JU		 144652013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14466
14467 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14468 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14469 ".set mips3" for 64-bit targets.
14470
4de09b85
DC		 144712013-06-15 Dehao Chen <dehao@google.com>
14472
14473 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14474 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14475 (gimple_check_call_args): Likewise.
14476 * value-prof.c (check_ic_target): Likewise.
14477 * ipa-inline.c (early_inliner): Likewise.
14478 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14479 * cgraph.c (cgraph_create_edge_1): Likewise.
14480 (cgraph_make_edge_direct): Likewise.
14481
d1aa4a76
MM		 144822013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14483
14484 PR target/57615
14485 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14486 rs6000_output_move_128bit to handle emitting quad memory
14487 operations. Set attribute length to 8 bytes.
14488
12211b99 144892013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP		 14490
14491 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14492 New pattern.
14493 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14494 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14495 (aarch64_<su>mlsl<mode>): Likewise.
14496
f5273574
MS		 144972013-06-14 Mike Stump <mikestump@comcast.net>
14498
14499 * Makefile.in (TARGET_H): Add insn-codes.h.
14500
ef7e7ab2
AM		 145012013-06-14 Alan Modra <amodra@gmail.com>
14502
14503 PR middle-end/57134
14504 PR middle-end/57586
14505 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14506 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14507 bitfield expansion when EXPAND_MEMORY.
14508 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14509
135faab6
MM		 145102013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14511
14512 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14513 test for clearing quad memory on 32-bit later.
14514
948a1fd9
MG		 145152013-06-13 Marc Glisse <marc.glisse@inria.fr>
14516
14517 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14518 (fold_negate_expr): Likewise.
14519 (fold_real_zero_addition_p): Handle vectors.
14520 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14521
0597fb9c
AM		 145222013-06-14 Alan Modra <amodra@gmail.com>
14523
14524 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14525
e5b962d0
JH		 145262013-06-13 Jan Hubicka <jh@suse.cz>
14527
14528 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14529 Local comdats are not externally visible.
14530 * symtab.c (dump_symtab_base): Dump externally visible.
14531 (verify_symtab_base): Verify back links in the symtab hash.
14532
a8355e51
BC		 145332013-06-13 Bin Cheng <bin.cheng@arm.com>
14534
14535 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14536 CONVERT_EXPR as equal nodes.
14537
8f7e6e33
BC		 145382013-06-13 Bin Cheng <bin.cheng@arm.com>
14539
22fc7d3d 14540 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 14541
5d418483
MG		 145422013-06-13 Marc Glisse <marc.glisse@inria.fr>
14543
14544 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14545 Generalize to complex and vector.
14546 * tree.c (build_all_ones_cst): New function.
14547 * tree.h (build_all_ones_cst): Declare it.
14548
50751417
AM		 145492013-06-13 Alan Modra <amodra@gmail.com>
14550
14551 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14552 * config/rs6000/rs6000.md (signbittf2): New insn.
14553 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14554 (abstf2_internal, cmptf_internal2): Likewise.
14555 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14556
4b02c962 145572013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14558 Pat Haugen <pthaugen@us.ibm.com>
14559 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM		 14560
14561 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14562 power8 byte, half-word, and quad-word atomic instructions.
14563 (emit_store_conditional): Likewise.
14564 (rs6000_expand_atomic_compare_and_swap): Likewise.
14565 (rs6000_expand_atomic_op): Likewise.
14566
14567 * config/rs6000/sync.md (larx): Add new modes for power8.
14568 (stcx): Likewise.
14569 (AINT): New mode iterator to include TImode as well as normal
14570 integer modes on power8.
14571 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14572 that VSX registers are not considered. Use AINT mode iterator
14573 instead of INT1 to allow inclusion of quad word atomic operations
14574 on power8.
14575 (load_locked<mode>): Likewise.
14576 (store_conditional<mode>): Likewise.
14577 (atomic_compare_and_swap<mode>): Likewise.
14578 (atomic_exchange<mode>): Likewise.
14579 (atomic_nand<mode>): Likewise.
14580 (atomic_fetch_<fetchop_name><mode>): Likewise.
14581 (atomic_nand_fetch<mode>): Likewise.
14582 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14583 each type.
14584 (ATOMIC): On power8, add QImode, HImode modes.
14585 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14586 modes that promote to SImode.
14587 (load_lockedti): Convert TImode arguments to PTImode, so that we
14588 get a guaranteed even/odd register pair.
14589 (load_lockedpti): Likewise.
14590 (store_conditionalti): Likewise.
14591 (store_conditionalpti): Likewise.
14592
14593 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14594 atomic load/store instructions.
14595 (HSI): Likewise.
14596
710b6cc1
RS		 145972013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14598
14599 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14600 loads.
14601 (insn_count): New attribute, with most cases extracted from...
14602 (length): ...here. Redefine most cases in terms of insn_count.
14603 (single_insn): Delete.
14604 (can_delay): Use insn_count to check for single instructions.
14605 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14606 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14607 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14608 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14609 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14610 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14611 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14612 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14613 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14614 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14615 rather than "length".
14616 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14617 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14618 Use "insn_count" rather than "length".
14619 * config/mips/mips-dsp.md
14620 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14621 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14622 length attributes.
14623
1951f101
MG		 146242013-06-12 Marc Glisse <marc.glisse@inria.fr>
14625
14626 PR tree-optimization/57361
14627 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14628
8b033a8a
SN		 146292013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14630
22fc7d3d
UB		 14631 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14632 to split.
8b033a8a
SN		 14633 (aarch64_simd_combine<mode>): New instruction expansion.
14634 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14635 function prototype.
14636 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14637 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14638
ca0f62a8
JH		 146392013-06-12 Jan Hubicka <jh@suse.cz>
14640
14641 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14642 decl has when in streaming stage.
14643 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14644 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14645
12211b99 146462013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS		 14647
14648 PR target/57578
14649 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14650
bfdeda2c
JJ		 146512013-06-12 Jakub Jelinek <jakub@redhat.com>
14652
14653 PR tree-optimization/57537
14654 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14655 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14656
a4fa02d1
RB		 146572013-06-12 Richard Biener <rguenther@suse.de>
14658
14659 * data-streamer.h (streamer_write_char_stream): CSE
14660 obs->current_pointer.
14661 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14662 streamer_write_char_stream manually and optimize the resulting loop.
14663 (streamer_write_hwi_stream): Likewise.
14664
bbf9ad07
JH		 146652013-06-12 Jan Hubicka <jh@suse.cz>
14666
14667 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14668 * cgraph.h (varpool_create_empty_node): Declare.
14669 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14670 duplicated nodes.
14671 * symtab.c (symtab_unregister_node): Be lax about missin entries
14672 in node hash.
14673 (symtab_get_node): Update comment.
14674 * varpool.c (varpool_create_empty_node): Break out from ...
14675 (varpool_node_for_decl): ... here.
14676 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14677
a53f90ad
EB		 146782013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14679
14680 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14681 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14682 part. Use straight-line flow at the end.
14683 <COMPONENT_REF>: Remove superfluous else.
14684 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14685
946f9306
JJ		 146862013-06-12 Jakub Jelinek <jakub@redhat.com>
14687
14688 PR target/56564
14689 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14690 target hook even for !TREE_PUBLIC decls. If no resolution info
14691 is available, return false for common and external decls.
14692
a7fca6f0
KP		 146932013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14694
14695 * config/rl78/constraints.md (U): New constraint.
14696 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14697 valloc attribute.
14698
c7d8f446
MM		 146992013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14700
14701 PR target/57589
14702 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14703 to allow returning address to AT_PLATFORM name.
14704
08346abd
JH		 147052013-06-11 Jan Hubicka <jh@suse.cz>
14706
14707 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14708 * cgraph.h (symtab_node_base): Add weakref flag.
14709 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14710 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14711 (output_weakrefs): Use weakref flag.
14712 * fold-const.c (simple_operand_p): Handle WEAK.
14713 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14714 * ipa.c (varpool_externally_visible_p): Drop weakref.
14715 (function_and_variable_visibility): Update comment; fix weakref
14716 sanity checks; do not clear DECL_WEAK on them.
14717 * lto-cgraph.c (lto_output_node): update.
14718 (lto_output_varpool_node): Update.
14719 (input_overwrite_node): Update.
14720 (input_node): Update.
14721 (input_varpool_node): Update.
14722 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14723 (lto_symtab_merge_symbols): Add sanity check.
14724 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14725 * passes.c (rest_of_decl_compilation): Set static flag, too.
14726 * symtab.c (dump_symtab_base): Dump weakref.
14727 (verify_symtab_base): Sanity check weakrefs.
14728 (symtab_make_decl_local): Remove duplicated code.
14729 (symtab_alias_ultimate_target): Simplify.
14730 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 14731
2e6a2f88
TV		 147322013-06-11 Tom de Vries <tom@codesourcery.com>
14733
14734 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14735 sequence_vect == NULL.
14736
d9af3652
DD		 147372013-06-11 DJ Delorie <dj@redhat.com>
14738
14739 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14740 (rl78_unwind_word_mode): New.
14741
ffa66012
DM		 147422013-06-11 David Malcolm <dmalcolm@redhat.com>
14743
14744 * final.c (debug_prefix_maps): Make static.
14745
783f2d35
DM		 147462013-06-11 David Malcolm <dmalcolm@redhat.com>
14747
14748 * function.c (initial_trampoline): Remove stray copy.
14749
7af26def
SN		 147502013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14751
14752 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14753
97756c0e
MJ		 147542013-06-11 Martin Jambor <mjambor@suse.cz>
14755
14756 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14757 within bounds at the beginning of the function.
14758
2c7b8bf6
AM		 147592013-06-11 Alan Modra <amodra@gmail.com>
14760
14761 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14762 reporting.
14763 (get_named_section): Don't NULL !DECL_P decl.
14764
77cea46e
IZ		 147652013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14766
14767 * doc/invoke.texi (core-avx2): Document.
14768 (slm): Likewise.
14769 (atom): Updated with MOVBE.
14770
888f0920
RB		 147712013-06-11 Richard Biener <rguenther@suse.de>
14772
22fc7d3d 14773 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 14774
6b438d58
AB		 147752013-06-11 Anton Blanchard <anton@samba.org>
14776
14777 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14778 correct shift value in little-endian mode.
14779
3bc66938
JJ		 147802013-06-11 Jakub Jelinek <jakub@redhat.com>
14781
14782 PR target/56564
14783 * varasm.c (get_variable_align): Move #endif to the right place.
14784
a5dc7f88
CC		 147852013-06-10 Cary Coutant <ccoutant@google.com>
14786
14787 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14788 for hash so that hash table traversal order is deterministic.
14789
19be72ab 147902013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 14791 Pat Haugen <pthaugen@us.ibm.com>
14792 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM		 14793
14794 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14795 of vectors in GPRS if they are direct moves or quad word load or
14796 store moves.
14797
14798 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14799 declaration.
14800 (direct_move_p): Likewise.
14801 (quad_load_store_p): Likewise.
14802
14803 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14804 classes into bins based on the physical register type.
14805 (reg_class_to_reg_type): Likewise.
14806 (IS_STD_REG_TYPE): Likewise.
14807 (IS_FP_VECT_REG_TYPE): Likewise.
14808 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14809 use direct move instructions.
14810 (reload_gpr_vsx): Likewise.
14811 (reload_vsx_gpr): Likewise.
14812 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14813 information that is a simplification of register classes. Also
14814 precalculate direct move reload helpers.
14815 (direct_move_p): New function to return true if the operation can
14816 be done as a direct move instruciton.
14817 (quad_load_store_p): New function to return true if the operation
14818 is a quad memory operation.
14819 (rs6000_legitimize_address): If quad memory, only allow register
14820 indirect for TImode addresses.
14821 (rs6000_legitimate_address_p): Likewise.
14822 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14823 (rs6000_reload_register_type): Likewise.
14824 (register_to_reg_type): Return register type.
14825 (rs6000_secondary_reload_simple_move): New helper function for
14826 secondary reload and secondary memory needed to identify anything
14827 that is a simple move, and does not need reloading.
14828 (rs6000_secondary_reload_direct_move): New helper function for
14829 secondary reload to identify cases that can be done with several
14830 instructions via the direct move instructions.
14831 (rs6000_secondary_reload_move): New helper function for secondary
14832 reload to identify moves between register types that can be done.
14833 (rs6000_secondary_reload): Add support for quad memory operations
14834 and for direct move.
14835 (rs6000_secondary_memory_needed): Likewise.
14836 (rs6000_debug_secondary_memory_needed): Change argument names.
14837 (rs6000_output_move_128bit): New function to return the move to
14838 use for 128-bit moves, including knowing about the various
14839 limitations of quad memory operations.
14840
14841 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14842 memory operations. call rs6000_output_move_128bit for the actual
14843 instruciton(s) to generate.
14844 (vsx_movti_64bit): Likewise.
14845
14846 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14847 (UNSPEC_P8V_MTVSRWZ): Likewise.
14848 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14849 (UNSPEC_P8V_MTVSRD): Likewise.
14850 (UNSPEC_P8V_XXPERMDI): Likewise.
14851 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14852 (UNSPEC_FUSION_GPR): Likewise.
14853 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 14854 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM		 14855 (f32_sv): Likewise.
14856 (f32_dm): Likewise.
14857 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14858 loads and direct move instructions.
14859 (zero_extendsidi2_lfiwzx): Likewise.
14860 (extendsidi2_lfiwax): Likewise.
14861 (extendsidi2_nocell): Likewise.
14862 (floatsi<mode>2_lfiwax): Likewise.
14863 (lfiwax): Likewise.
14864 (floatunssi<mode>2_lfiwzx): Likewise.
14865 (lfiwzx): Likewise.
14866 (fix_trunc<mode>_stfiwx): Likewise.
14867 (fixuns_trunc<mode>_stfiwx): Likewise.
14868 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14869 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14870 (parity<mode>2_cmpb): Set length/type attr.
14871 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14872 for 'mr.' to fast_compare.
14873 (bpermd_<mode>): Change type attr to popcnt.
14874 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14875 (p8_mtvsrwz_1): Likewise.
14876 (p8_mtvsrwz_2): Likewise.
14877 (reload_fpr_from_gpr<mode>): Likewise.
14878 (p8_mtvsrd_1): Likewise.
14879 (p8_mtvsrd_2): Likewise.
14880 (p8_xxpermdi_<mode>): Likewise.
14881 (reload_vsx_from_gpr<mode>): Likewise.
14882 (reload_vsx_from_gprsf): Likewise.
14883 (p8_mfvsrd_3_<mode>): LIkewise.
14884 (reload_gpr_from_vsx<mode>): Likewise.
14885 (reload_gpr_from_vsxsf): Likewise.
14886 (p8_mfvsrd_4_disf): Likewise.
14887 (multi-word GPR splits): Do not split direct moves or quad memory
14888 operations.
14889
16876bdc
DM		 148902013-06-10 David Malcolm <dmalcolm@redhat.com>
14891
14892 * tree-into-ssa.c (interesting_blocks): Make static.
14893
df8a1d28
JJ		 148942013-06-10 Jakub Jelinek <jakub@redhat.com>
14895
14896 PR target/56564
14897 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14898 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14899 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14900 (get_variable_align): New function.
14901 (get_variable_section, emit_bss, emit_common,
14902 assemble_variable_contents, place_block_symbol): Use
14903 get_variable_align instead of DECL_ALIGN.
14904 (assemble_noswitch_variable): Add align argument, use it
14905 instead of DECL_ALIGN.
14906 (assemble_variable): Adjust caller. Use get_variable_align
14907 instead of DECL_ALIGN.
14908 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14909 caller.
14910 (DATA_ABI_ALIGNMENT): Define.
14911 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14912 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14913 opt is false, only return the psABI mandated alignment increase.
14914 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14915 (DATA_ABI_ALIGNMENT): ... this.
14916 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14917 (DATA_ABI_ALIGNMENT): ... this.
14918 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14919 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14920 (DATA_ABI_ALIGNMENT): ... this.
14921 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14922 * doc/tm.texi: Regenerated.
14923
57b29ca6
UB		 149242013-06-10 Uros Bizjak <ubizjak@gmail.com>
14925
14926 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14927 cmp_code to construct REG_EQUAL note.
14928
3e56ed50
JJ		 149292013-06-09 Jakub Jelinek <jakub@redhat.com>
14930
14931 PR target/57568
14932 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14933 that operands[2] doesn't overlap with operands[0].
14934
74fe2a1d
DE		 149352013-06-09 David Edelsohn <dje.gcc@gmail.com>
14936 Jan Hubicka <jh@suse.cz>
14937
14938 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14939 hack to mark symbols as used.
14940
ed52a84e
VM		 149412013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14942
14943 PR rtl-optimization/57559
14944 * lra-constraints.c (process_alt_operands): Don't discourage
14945 memory with known offset for offsetable memory constraint.
14946 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14947
299316ed
EB		 149482013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14949
14950 * varasm.c (struct oc_local_state): Reorder fields.
14951 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14952 and adjust accordingly.
14953 (output_constructor): Reorder initialization code and adjust call to
14954 output_constructor_bitfield.
14955
88ac60d3
JH		 149562013-06-07 Jan Hubicka <jh@suse.cz>
14957
14958 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14959
4cdce1a8
DM		 149602013-06-07 David Malcolm <dmalcolm@redhat.com>
14961
14962 * tree-object-size.c (unknown): Make const.
14963
b5e0425c
AK		 149642013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14965
14966 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14967 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14968 for last alternative in the cpu_facility attribute.
14969
79678d04
KT		 149702013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14971
14972 PR target/56315
14973 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14974 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14975 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14976 * config/arm/constraints.md (Dg): New constraint.
14977 * config/arm/neon.md (xordi3_neon): Remove.
14978 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14979 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14980
b31ddbdb
KT		 149812013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14982
14983 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14984 Clean up alternatives.
14985
8f90b7d4
AM		 149862013-06-07 Alan Modra <amodra@gmail.com>
14987
14988 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14989 va_list_gpr_size.
14990
961ce119
AM		 149912013-06-07 Alan Modra <amodra@gmail.com>
14992
14993 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14994
33e49835
KT		 149952013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14996
14997 * config/arm/constraints.md (Df): New constraint.
14998 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14999 Correct length attribute for last two alternatives.
15000
acd17ae6
AM		 150012013-06-07 Alan Modra <amodra@gmail.com>
15002
15003 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15004 override user -mfp-in-toc.
15005 (offsettable_ok_by_alignment): Consider just the current access
15006 rather than the whole object, unless BLKmode. Handle
15007 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15008 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15009 for -mcmodel=medium.
15010 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15011 override user -mfp-in-toc or -msum-in-toc. Default to
15012 -mno-fp-in-toc for -mcmodel=medium.
15013
73310b0e
DD		 150142013-06-06 DJ Delorie <dj@redhat.com>
15015
15016 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15017 TARGET_VALID_POINTER_MODE.
15018
0bd62dca 150192013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15020 Pat Haugen <pthaugen@us.ibm.com>
15021 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM		 15022
15023 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15024 Document new power8 builtins.
15025
15026 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15027 condition code register, to allow 128-bit logical operations to be
15028 done in the VSX or GPR registers.
15029 (nor<mode>3): Use the canonical form for nor.
15030 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15031 vclz*, and vpopcnt* vector instructions.
15032 (nand<mode>3): Likewise.
15033 (orc<mode>3): Likewise.
15034 (clz<mode>2): LIkewise.
15035 (popcount<mode>2): Likewise.
15036
15037 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15038 that only the GPRs are recognized.
15039
15040 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15041 support for new power8 builtins.
15042
15043 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15044 builtin functions.
15045 (xscvdpspn): Likewise.
15046 (vclz): Likewise.
15047 (vclzb): Likewise.
15048 (vclzh): Likewise.
15049 (vclzw): Likewise.
15050 (vclzd): Likewise.
15051 (vpopcnt): Likewise.
15052 (vpopcntb): Likewise.
15053 (vpopcnth): Likewise.
15054 (vpopcntw): Likewise.
15055 (vpopcntd): Likewise.
15056 (vgbbd): Likewise.
15057 (vmrgew): Likewise.
15058 (vmrgow): Likewise.
15059 (eqv): Likewise.
15060 (eqv_v16qi3): Likewise.
15061 (eqv_v8hi3): Likewise.
15062 (eqv_v4si3): Likewise.
15063 (eqv_v2di3): Likewise.
15064 (eqv_v4sf3): Likewise.
15065 (eqv_v2df3): Likewise.
15066 (nand): Likewise.
15067 (nand_v16qi3): Likewise.
15068 (nand_v8hi3): Likewise.
15069 (nand_v4si3): Likewise.
15070 (nand_v2di3): Likewise.
15071 (nand_v4sf3): Likewise.
15072 (nand_v2df3): Likewise.
15073 (orc): Likewise.
15074 (orc_v16qi3): Likewise.
15075 (orc_v8hi3): Likewise.
15076 (orc_v4si3): Likewise.
15077 (orc_v2di3): Likewise.
15078 (orc_v4sf3): Likewise.
15079 (orc_v2df3): Likewise.
15080
15081 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15082 allow power8 quad mode in 64-bit.
15083 (rs6000_builtin_vectorized_function): Add support to vectorize
15084 ISA 2.07 count leading zeros, population count builtins.
15085 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15086 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15087 (builtin_function_type): Add vgbbd builtin function which takes an
15088 unsigned argument.
15089 (altivec_expand_vec_perm_const): Add support for new power8 merge
15090 instructions.
15091
15092 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15093 that does not include TImdoe for use with 32-bit.
15094 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15095 instructions.
15096 (UNSPEC_VSX_CVDPSPN): Likewise.
15097 (vsx_xscvdpspn): Likewise.
15098 (vsx_xscvspdpn): Likewise.
15099 (vsx_xscvdpspn_scalar): Likewise.
15100 (vsx_xscvspdpn_directmove): Likewise.
15101 (vsx_and<mode>3): Split logical operations into 32-bit and
15102 64-bit. Add support to do logical operations on TImode as well as
15103 VSX vector types. Allow logical operations to be done in either
15104 VSX registers or in general purpose registers in 64-bit mode. Add
15105 splitters if GPRs were used. For AND, add clobber of CCmode to
15106 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15107 encoding.
15108 (vsx_and<mode>3_32bit): Likewise.
15109 (vsx_and<mode>3_64bit): Likewise.
15110 (vsx_ior<mode>3): Likewise.
15111 (vsx_ior<mode>3_32bit): Likewise.
15112 (vsx_ior<mode>3_64bit): Likewise.
15113 (vsx_xor<mode>3): Likewise.
15114 (vsx_xor<mode>3_32bit): Likewise.
15115 (vsx_xor<mode>3_64bit): Likewise.
15116 (vsx_one_cmpl<mode>2): Likewise.
15117 (vsx_one_cmpl<mode>2_32bit): Likewise.
15118 (vsx_one_cmpl<mode>2_64bit): Likewise.
15119 (vsx_nor<mode>3): Likewise.
15120 (vsx_nor<mode>3_32bit): Likewise.
15121 (vsx_nor<mode>3_64bit): Likewise.
15122 (vsx_andc<mode>3): Likewise.
15123 (vsx_andc<mode>3_32bit): Likewise.
15124 (vsx_andc<mode>3_64bit): Likewise.
15125 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15126 and xxlorc instructions.
15127 (vsx_eqv<mode>3_64bit): Likewise.
15128 (vsx_nand<mode>3_32bit): Likewise.
15129 (vsx_nand<mode>3_64bit): Likewise.
15130 (vsx_orc<mode>3_32bit): Likewise.
15131 (vsx_orc<mode>3_64bit): Likewise.
15132
15133 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15134
15135 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15136 instruction.
15137 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15138 (p8_vmrgow): Likewise.
15139 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15140 GPRs to be split under VSX.
15141 (p8v_clz<mode>2): Add power8 count leading zero support.
15142 (p8v_popcount<mode>2): Add power8 population count support.
15143 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15144 support.
15145
15146 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15147 instruction.
15148
15149 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15150 builtin functions.
15151 (vec_nand): Likewise.
15152 (vec_vclz): Likewise.
15153 (vec_vclzb): Likewise.
15154 (vec_vclzd): Likewise.
15155 (vec_vclzh): Likewise.
15156 (vec_vclzw): Likewise.
15157 (vec_vgbbd): Likewise.
15158 (vec_vmrgew): Likewise.
15159 (vec_vmrgow): Likewise.
15160 (vec_vpopcnt): Likewise.
15161 (vec_vpopcntb): Likewise.
15162 (vec_vpopcntd): Likewise.
15163 (vec_vpopcnth): Likewise.
15164 (vec_vpopcntw): Likewise.
15165
37684c46
VM		 151662013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15167
15168 PR rtl-optimization/57468
15169 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15170 spilled pseudos.
15171
d89ae27c
VM		 151722013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15173
15174 PR rtl-optimization/57459
15175 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15176 type when setting live regs.
15177
3597e113
VM		 151782013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15179
15180 * config/s390/s390.opt (mlra): New option.
15181 * config/s390/s390.c (s390_decompose_address): Check displacement
15182 for all registers for LRA.
15183 (s390_secondary_reload): Don't used secondary reloads for LRA.
15184 (s390_lra_p): New function.
15185 (TARGET_LRA_P): Define.
15186 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15187 of attribute cpu_facility to zarch for the last alternative.
15188 (*cmpmem_short): Ditto.
15189
01037aeb
EB		 151902013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15191
15192 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15193 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15194 (arm_expand_prologue): Likewise.
15195
3371a64f
TJ		 151962013-06-06 Teresa Johnson <tejohnson@google.com>
15197
15198 PR c++/53743
15199 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15200 as this is now done by redirect_edge_and_branch_force.
15201 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15202 barriers, and fix interaction with splitting.
15203 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15204 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15205 reflect changes made in the routine.
15206 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15207 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15208 since this is called in cfglayout mode, and replace partition fixup
15209 with assert as that is now done by force_nonfallthru_and_redirect.
15210 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15211 already be marked with region crossing note.
15212 (insert_section_boundary_note): Make non-static, gate on flag
15213 has_bb_partition, rewrite to also check for multiple partitions.
15214 (rest_of_handle_reorder_blocks): Remove call to
15215 insert_section_boundary_note, now done later during free_cfg.
15216 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15217 * bb-reorder.h (insert_section_boundary_note): Declare.
15218 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15219 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15220 invoke insert_section_boundary_note.
15221 (try_redirect_by_replacing_jump): Remove unnecessary
15222 check for region crossing note.
15223 (fixup_partition_crossing): New function.
15224 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15225 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15226 in non-cfglayout mode.
15227 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15228 remove old code that tried to do this. Emit barrier correctly
15229 when we are in cfglayout mode.
15230 (last_bb_in_partition): New function.
15231 (rtl_split_edge): Correctly fixup partition boundaries.
15232 (commit_one_edge_insertion): Remove old code that tried to
15233 fixup region crossing edge since this is now handled in
15234 split_block, and set up insertion point correctly since
15235 block may now end in a jump.
15236 (verify_hot_cold_block_grouping): Guard against checking when not in
15237 linearized RTL mode.
15238 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15239 notes.
15240 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15241 rtl_verify_flow_info, so not called in cfglayout mode.
15242 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15243 (fixup_reorder_chain): Remove old code that attempted to fixup region
15244 crossing note as this is now handled in force_nonfallthru_and_redirect.
15245 (duplicate_insn_chain): Don't duplicate switch section notes.
15246 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15247 note.
15248 * basic-block.h (emit_barrier_after_bb): Declare.
15249
66071e10
KT		 152502013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15251
15252 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15253 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15254 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15255
0e26bf3d
KT		 152562013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15257
15258 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15259 where appropriate.
15260 * config/arm/ldmstm.md: Regenerate.
15261
12b4e7ef
KT		 152622013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15263
15264 * config/arm/sync.md (atomic_loaddi_1):
15265 Disable predication for arm_restrict_it.
15266 (arm_load_exclusive<mode>): Likewise.
15267 (arm_load_exclusivesi): Likewise.
15268 (arm_load_exclusivedi): Likewise.
15269 (arm_load_acquire_exclusive<mode>): Likewise.
15270 (arm_load_acquire_exclusivesi): Likewise.
15271 (arm_load_acquire_exclusivedi): Likewise.
15272 (arm_store_exclusive<mode>): Likewise.
15273 (arm_store_exclusive<mode>): Likewise.
15274 (arm_store_release_exclusivedi): Likewise.
15275 (arm_store_release_exclusive<mode>): Likewise.
15276
15d16c8a
RB		 152772013-06-06 Richard Biener <rguenther@suse.de>
15278
15279 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15280 after LTO_null.
15281 (lto_tag_is_tree_code_p): Adjust.
15282 (lto_tag_is_gimple_code_p): Likewise.
15283 (lto_gimple_code_to_tag): Likewise.
15284 (lto_tag_to_gimple_code): Likewise.
15285 (lto_tree_code_to_tag): Likewise.
15286 (lto_tag_to_tree_code): Likewise.
15287 * data-streamer.h (streamer_write_hwi_in_range): Use
15288 uhwi streaming to stream the normalized range.
15289 (streamer_read_hwi_in_range): Likewise.
15290
17a7fc37
KT		 152912013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15292
15293 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15294 (predicable_short_it): Likewise.
15295 (predicated): Likewise.
15296 (enabled): Handle above.
15297 (define_cond_exec): Set predicated attribute to yes.
15298
b57ca59b
MS		 152992013-06-05 Mike Stump <mikestump@comcast.net>
15300
15301 * gdbinit.in (__FUNCTION__): Add.
15302
c1e183a9
UB		 153032013-06-05 Uros Bizjak <ubizjak@gmail.com>
15304
15305 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15306 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15307
0378bddb
JH		 153082013-06-05 Jan Hubicka <jh@suse.cz>
15309
15310 * varasm.c (mark_decl_referenced): Revert the removal until targets
15311 are fixed.
15312
8f8a2057
DE		 153132013-06-05 David Edelsohn <dje.gcc@gmail.com>
15314
15315 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15316 instead of mark_decl_referenced.
15317
edb983b2
JH		 153182013-06-05 Jan Hubicka <jh@suse.cz>
15319
15320 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15321 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15322 and symtab_used_from_object_file_p.
15323 (cgraph_make_node_local_1): Clear forced_by_abi.
15324 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15325 * cgraph.h (symtab_node_base): Add forced_by_abi.
15326 (decide_is_variable_needed): Remove.
15327 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15328 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15329 (decide_is_symbol_needed): ... this one; handle symbols in general;
15330 always analyze virtuals; honnor forced_by_abi.
15331 (cgraph_finalize_function): Update.
15332 (varpool_finalize_decl): Update.
15333 (symbol_defined_and_needed): Remove.
15334 (analyze_functions): Update.
15335 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15336 output_refs, input_overwrite_node): Handle forced_by_abi.
15337 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15338 (address_taken_from_non_vtable_p): ... this one.
15339 (comdat_can_be_unshared_p_1): New function.
15340 (cgraph_comdat_can_be_unshared_p): Rename to ...
15341 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15342 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15343 (function_and_variable_visibility): Clear forced_by_abi as needed.
15344 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15345 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH		 15346 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15347 * varpool.c (decide_is_variable_needed): Remove.
15348
9912dbe5
KT		 153492013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15350
15351 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15352 (arm_option_override): Override arm_restrict_it where appropriate.
15353 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15354 * config/arm/arm.opt (mrestrict-it): New command-line option.
15355 * doc/invoke.texi: Document -mrestrict-it.
15356
34c136b6
DM		 153572013-06-05 David Malcolm <dmalcolm@redhat.com>
15358
15359 * tsan.c (tsan_atomic_table): Make const.
15360
a367df53
RB		 153612013-06-05 Richard Biener <rguenther@suse.de>
15362
15363 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15364 index associated with the tree we are supposed to replace.
15365 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15366 TREE_ASM_WRITTEN as zero for everything but SSA names.
15367
70d51a19
DM		 153682013-06-05 David Malcolm <dmalcolm@redhat.com>
15369
15370 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15371
b9bd2075
JH		 153722013-06-04 Jan Hubicka <jh@suse.cz>
15373
15374 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15375 (input_node, input_varpool_node): Handle correctly external same
15376 body aliases.
15377 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15378 nodes at ltrans stage.
15379
aaae719d
JH		 153802013-06-04 Jan Hubicka <jh@suse.cz>
15381
15382 * ipa-inline.c (update_caller_keys): Fix availability test.
15383 (update_callee_keys): Likewise.
15384 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15385 to follow ELF standard.
15386
107eea2c
JU		 153872013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15388
15389 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15390 (mips64r5900el-*-elf*): New configurations.
15391 * config/mips/mips-cpus.def (r5900): New processor.
15392 * config/mips/mips-tables.opt: Regenerate.
15393 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15394 (mips_issue_rate): Handle PROCESSOR_R5900.
15395 (mips_reorg_process_insns): Force reorder mode for the R5900.
15396 * config/mips/mips.h (TARGET_MIPS5900): Define.
15397 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15398 TARGET_MIPS5900.
15399 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15400 TARGET_MIPS5900.
15401 * config/mips/mips.md (processor): Add r5900.
15402 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15403
b7342d25
IB		 154042013-06-04 Ian Bolton <ian.bolton@arm.com>
15405
15406 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15407 into function to generate MOVI instruction.
c1e183a9 15408 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB		 15409 (aarch64_preferred_simd_mode): Turn into wrapper.
15410 (aarch64_output_scalar_simd_mov_immediate): New function.
15411 * config/aarch64/aarch64-protos.h: Add prototype for above.
15412
81c2dfb9
IB		 154132013-06-04 Ian Bolton <ian.bolton@arm.com>
15414
15415 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15416 element_char member.
15417 (sizetochar): Return signed char.
15418 (aarch64_simd_valid_immediate): Remove elchar and other
15419 unnecessary variables.
15420 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15421 Calculate element_char as required.
15422 * config/aarch64/aarch64-protos.h: Update and move prototype
15423 for aarch64_output_simd_mov_immediate.
15424 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15425 Update arguments.
15426
48063b9d
IB		 154272013-06-04 Ian Bolton <ian.bolton@arm.com>
15428
15429 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15430 information completed by aarch64_simd_valid_immediate.
15431 (aarch64_legitimate_constant_p): Update arguments.
15432 (aarch64_simd_valid_immediate): Work with struct rather than many
15433 pointers.
15434 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15435 (aarch64_simd_make_constant): Update arguments.
15436 (aarch64_output_simd_mov_immediate): Work with struct rather than
15437 many pointers. Output immediate directly rather than as operand.
15438 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15439 Update prototype.
15440 * config/aarch64/constraints.md (Dn): Update arguments.
15441
3ea63f60
IB		 154422013-06-04 Ian Bolton <ian.bolton@arm.com>
15443
15444 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15445 longer static.
15446 (aarch64_simd_immediate_valid_for_move): Remove.
15447 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15448 (aarch64_simd_make_constant): Update call.
15449 (aarch64_output_simd_mov_immediate): Update call.
15450 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15451 Add prototype.
15452 * config/aarch64/constraints.md (Dn): Update call.
15453
d8edd899
IB		 154542013-06-04 Ian Bolton <ian.bolton@arm.com>
15455
15456 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15457 return type to bool for prototype.
15458 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15459 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15460 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15461
44b20bb8
CM		 154622013-06-04 Catherine Moore <clm@codesourcery.com>
15463
15464 * config/mips/mips.opt (meva): New.
c1e183a9 15465 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM		 15466 (ASM_SPEC): Handle -meva.
15467 * doc/invoke.texi (meva): Document.
15468
52befbd8
AM		 154692013-06-04 Alan Modra <amodra@gmail.com>
15470
15471 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15472 constant output.
15473
aadaf24e
KT		 154742013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15475
15476 * rtl.def: Add extra fourth optional field to define_cond_exec.
15477 * gensupport.c (process_one_cond_exec): Process attributes from
15478 define_cond_exec.
15479 * doc/md.texi: Document fourth field in define_cond_exec.
15480
0cd9e9ee
EB		 154812013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15482
15483 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15484 out the processing order as in store_bit_field_1.
15485
cb0f665d
JH		 154862013-06-04 Jan Hubicka <jh@suse.cz>
15487
15488 PR middle-end/57500
15489 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15490 non-VAR_DECL node if it does not exist yet.
15491
53984b9b
RS		 154922013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15493
15494 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15495 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15496 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15497 target_cpu_default setting.
15498
5d318fd4
TJ		 154992013-06-03 Teresa Johnson <tejohnson@google.com>
15500
15501 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15502 default to -fopt-info=optimized instead of all.
15503 * doc/invoke.texi: Ditto.
15504 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15505 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15506 (execute_vect_slp): Emit BB vectorization success under
15507 MSG_OPTIMIZED_LOCATIONS.
15508 * tree-vect-slp.c (vect_slp_transform_bb): Change
15509 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15510 * tree-vect-loop.c (vect_transform_loop): Ditto.
15511
ec3af349
JM		 155122013-06-03 Jason Merrill <jason@redhat.com>
15513
15514 PR c++/57415
15515 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15516 Use TARGET_EXPR for C++.
15517
1baf9159
JJ		 155182013-06-03 Jakub Jelinek <jakub@redhat.com>
15519
15520 PR rtl-optimization/57268
0cd9e9ee 15521 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ		 15522 if DEBUG_INSN_P (insn).
15523
15524 Reapply
15525 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15526
15527 PR rtl-optimization/57268
15528 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15529 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15530
0f1d3965
YR		 155312013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15532
15533 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15534 (ix86_avoid_lea_for_addr): Likewise.
15535 (exact_dependency_1): Likewise.
15536 (ix86_adjust_cost): Likewise.
15537 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15538 removed.
15539 (do_reorder_for_imul): Fix typo, formatting and
15540 !reload_completed check removed.
15541 (ix86_sched_reorder): Fix typo and formatting.
15542 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15543 list.
15544
3bc30db4
SN		 155452013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15546
15547 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15548
19cc6fac
EB		 155492013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15550
15551 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15552 <STRING_CST>: Likewise.
15553 <VECTOR_CST>: Likewise.
15554
c68a6e08
JW		 155552013-06-01 Janus Weil <janus@gcc.gnu.org>
15556 Mikael Morin <mikael@gcc.gnu.org>
15557
15558 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15559 * config.in: Regenerated.
15560 * configure: Regenerated.
15561
38e55e5c
JH		 155622013-06-01 Jan Hubicka <jh@suse.cz>
15563
15564 PR middle-end/57366
15565 * cgraphunit.c (compile): When weakref is not supported,
15566 set up transparent aliases before final output pass.
15567 * varasm.c (assemble_alias): Do not try to do it here.
15568
eb51d2ff
JH		 155692013-06-01 Jan Hubicka <jh@suse.cz>
15570
15571 PR middle-end/57467
15572 * passes.c (for_per_function): Skip unanalyzed functions.
15573
40a7fe1e
JH		 155742013-06-01 Jan Hubicka <jh@suse.cz>
15575
15576 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15577 (lto_symtab_merge_symbols_1): ... this one.
15578 (lto_symtab_merge_cgraph_nodes): Rename to ...
15579 (lto_symtab_merge_symbols): ... this one; simplify.
15580 * cgraph.c (same_body_aliases_done): Rename to ...
15581 (cpp_implicit_aliases_done): ... this one.
15582 (cgraph_create_function_alias): Update.
15583 (cgraph_same_body_alias): Update.
c1e183a9 15584 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 15585 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 15586 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH		 15587 (cgraph_node): Remove same_body_alias.
15588 (varpool_node): Remove alias_of and extra_name_alias.
15589 (same_body_aliases_done): Rename to ..
15590 (cpp_implicit_aliases_done): ... this one.
15591 (symtab_alias_ultimate_target): Add default parameter.
15592 (symtab_resolve_alias): New function.
15593 (fixup_same_cpp_alias_visibility): Declare.
15594 (cgraph_function_node): Add default parameter.
15595 (cgraph_node_asm_name): Likewise.
15596 (cgraph_function_or_thunk_node): Add default parameter; do
15597 not ICE when it is NULL.
15598 (varpool_variable_node): Likewise.
15599 * tree-emutls.c (create_emultls_var): Update.
15600 (ipa_lower_emutls): Update.
15601 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15602 (cgraph_reset_node): Reset alias info.
15603 (cgraph_finalize_function): Update.
15604 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15605 (analyze_function): Simplify.
15606 (cgraph_process_same_body_aliases): Simplify.
15607 (analyze_functions): Fixup same body aliases.
15608 (handle_alias_pairs): Simplify.
15609 (assemble_thunk): Update.
15610 (assemble_thunks_and_aliases): Update.
15611 (output_weakrefs): Rewrite.
15612 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15613 (lto_output_varpool_node): Likewise.
15614 (compute_ltrans_boundary): Remve assert.
15615 (get_alias_symbol): New functoin.
15616 (input_node): Rewrite alias handling.
15617 (input_varpool_node): Likewise.
15618 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15619 * ipa.c (process_references): Handle weakrefs correctly.
15620 (symtab_remove_unreachable_nodes): Likewise.
15621 * trans-mem.c (get_cg_data): Update.
15622 (ipa_tm_create_version_alias): Update.
15623 (ipa_tm_execute): Update.
15624 * symtab.c (dump_symtab_base): Dump aliases.
15625 (verify_symtab_base): Verify aliases.
15626 (symtab_node_availability): New function.
15627 (symtab_alias_ultimate_target): Simplify.
15628 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15629 handle all the fixup cases.
15630 (symtab_resolve_alias): New function.
15631 * passes.c (ipa_write_summaries): Handle weakrefs.
15632 * varpool.c (varpool_analyze_node): Simplify.
15633 (assemble_aliases): Update.
15634 (varpool_create_variable_alias): Simplify.
15635 (varpool_extra_name_alias): Simplify.
15636 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15637 (lto_symtab_merge_symbols): ... this one.
15638
21759881
DT		 156392013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15640
15641 Revert
15642 PR rtl-optimization/57268
15643 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15644 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15645
daf4e940
TB		 156462013-06-01 Tobias Burnus <burnus@net-b.de>
15647
15648 Partially reverted:
15649 2013-05-31 Tobias Burnus <burnus@net-b.de>
15650
15651 PR middle-end/57073
15652 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15653 further up.
15654
c359d8d0
DT		 156552013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15656
15657 PR rtl-optimization/57268
15658 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15659 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15660
4172215d
EB		 156612013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15662
15663 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15664 unordered comparison operators when -fno-trapping-math is in effect
15665 on the e500.
15666 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15667 and implement unordered comparison operators properly on the e500.
15668
a8c50132
EB		 156692013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15670
15671 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15672 for constant scalar integers.
15673 (simplify_relational_operation_1): Likewise.
15674
44626634
SB		 156752013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15676
15677 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15678 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15679 Fix comment.
15680
55805e54
YR		 156812013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15682 Igor Zamyatin <igor.zamyatin@intel.com>
15683
15684 Silvermont (SLM) architecture performance tuning.
15685 * config/i386/i386.h (enum ix86_tune_indices): Add
15686 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15687 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15688
15689 * config/i386/i386.c (initial_ix86_tune_features)
15690 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15691 (ix86_lea_outperforms): Handle Silvermont tuning.
15692 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15693 call.
15694 (ix86_use_lea_for_mov): Likewise.
15695 (ix86_avoid_lea_for_addr): Likewise.
15696 (ix86_lea_for_add_ok): Likewise.
15697 (exact_dependency_1): New function.
15698 (exact_store_load_dependency): Likewise.
15699 (ix86_adjust_cost): Handle Silvermont tuning.
15700 (do_reoder_for_imul): Likewise.
15701 (swap_top_of_ready_list): New function.
15702 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15703
15704 * config/i386/i386.md (peepholes that split memory operand in fp
15705 converts): New.
15706
e19c9de2
MS		 157072013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15708
15709 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15710 Remove un-necessary braces.
15711
38e6c9a6
MS		 157122013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15713
15714 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15715 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15716
e3530904
TB		 157172013-05-31 Tobias Burnus <burnus@net-b.de>
15718
15719 PR middle-end/57073
c1e183a9 15720 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 15721
b6af05a9
KT		 157222013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15723
15724 PR target/56315
15725 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15726 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15727 * config/arm/neon.md (iordi3_neon): Remove.
15728 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15729 * config/arm/predicates.md (imm_for_neon_logic_operand):
15730 Move to earlier in the file.
15731 (neon_logic_op2): Likewise.
15732 (arm_iordi_operand_neon): New predicate.
15733
f800c166
RB		 157342013-05-31 Richard Biener <rguenther@suse.de>
15735
15736 PR tree-optimization/57478
15737 PR tree-optimization/57453
15738 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15739 are life as well.
15740
e571fa59
KP		 157412013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15742
15743 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15744 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15745
0fa6e0ef
TB		 157462013-05-30 Tobias Burnus <burnus@net-b.de>
15747 Thomas Koenig <tkoenig@gcc.gnu.org>
15748
15749 PR middle-end/57073
15750 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15751 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15752
4688ddf5
SB		 157532013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15754
0fa6e0ef 15755 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 15756
3b9ceb4b
VM		 157572013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15758
15759 * target.def (register_usage_leveling_p): New hook.
15760 * targhooks.c (default_register_usage_leveling_p): New.
15761 * targhooks.h (default_register_usage_leveling_p): New prototype.
15762 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15763 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15764 * doc/tm.texi: Update.
15765 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15766
65f9e789
IB		 157672013-05-30 Ian Bolton <ian.bolton@arm.com>
15768
15769 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15770 (*insv_reg<mode>): New define_insn.
15771
12211b99 157722013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR		 15773
15774 PR rtl-optimization/57439
15775 * postreload.c (move2add_valid_value_p): Check that we have
15776 a zero subreg_regno_offset when accessing the register in
15777 the requested mode.
15778
0b871ccf
YR		 157792013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15780 Igor Zamyatin <igor.zamyatin@intel.com>
15781
15782 Silvermont (SLM) architecture pipeline model, tuning and
15783 insn selection.
15784 * config.gcc: Add slm config options and target.
15785
15786 * config/i386/slm.md: New.
15787
15788 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15789
f43245d1 15790 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR		 15791 PROCESSOR_SLM.
15792 (ix86_target_macros_internal): Likewise.
15793
f43245d1 15794 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR		 15795 (m_SLM): New macro flag.
15796 (initial_ix86_tune_features): Set m_SLM.
15797 (x86_accumulate_outgoing_args): Likewise.
15798 (x86_arch_always_fancy_math_387): Likewise.
15799 (processor_target_table): Add slm cost.
15800 (cpu_names): Add slm cpu name.
15801 (x86_option_override_internal): Set SLM ISA.
15802 (ix86_issue_rate): New case PROCESSOR_SLM.
15803 (ia32_multipass_dfa_lookahead): Likewise.
15804 (fold_builtin_cpu): Add slm.
15805
15806 * config/i386/i386.h (TARGET_SLM): New target macro.
15807 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15808 (processor_type): Add PROCESSOR_SLM.
15809
15810 * config/i386/i386.md (cpu): Add new value "slm".
15811 (slm.md): Include slm.md.
15812
24d5b097
XG		 158132013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15814 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15815
15816 * config/arm/arm-protos.h: Add and update function protos.
15817 * config/arm/arm.c (use_simple_return_p): New added.
15818 (thumb2_expand_return): Check simple_return flag.
15819 * config/arm/arm.md: Add simple_return and conditional simple_return.
15820 * config/arm/iterators.md: Add iterator for return and simple_return.
15821
c1cccc15
ZC		 158222013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15823
15824 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15825 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15826 (arm_emit_vfp_multi_reg_pop): Likewise.
15827 (thumb2_emit_ldrd_pop): Likewise.
15828 (arm_expand_epilogue): Add misc REG_CFA notes.
15829 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15830
3a4f280b
LC		 158312013-05-29 Lawrence Crowl <crowl@google.com>
15832
15833 * config/arm/t-arm: Update for below.
15834
15835 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15836 Change type to hash_table. Update dependent calls and types.
15837
15838 * config/i386/t-cygming: Update for below.
15839
15840 * config/i386/t-interix: Update for below.
15841
15842 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15843 Change type to hash_table. Update dependent calls and types.
15844 (i386_find_on_wrapper_list::wrappers): Likewise.
15845
15846 * config/ia64/t-ia64: Update for below.
15847
15848 * config/ia64/ia64.c (bundle_state_table):
15849 Change type to hash_table. Update dependent calls and types.
15850
15851 * config/mips/mips.c (mips_reorg_process_insns::htab):
15852 Change type to hash_table. Update dependent calls and types.
15853
15854 * config/sol2.c (solaris_comdat_htab):
15855 Change type to hash_table. Update dependent calls and types.
15856
15857 * config/t-sol2: Update for above.
15858
1388a0e3
TJ		 158592013-05-29 Teresa Johnson <tejohnson@google.com>
15860
15861 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15862 functions are not yet marked as defined.
15863
a5965b52 158642013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 15865 Pat Haugen <pthaugen@us.ibm.com>
15866 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM		 15867
15868 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15869 instructions.
15870 (VEC_A): Likewise.
15871 (VEC_C): Likewise.
15872 (vrotl<mode>3): Likewise.
15873 (vashl<mode>3): Likewise.
15874 (vlshr<mode>3): Likewise.
15875 (vashr<mode>3): Likewise.
15876
15877 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15878 support for power8 V2DI builtins.
15879
15880 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15881 power8 V2DI builtins.
15882 (vupkhsw): Likewise.
15883 (vupklsw): Likewise.
15884 (vaddudm): Likewise.
15885 (vminsd): Likewise.
15886 (vmaxsd): Likewise.
15887 (vminud): Likewise.
15888 (vmaxud): Likewise.
15889 (vpkudum): Likewise.
15890 (vpksdss): Likewise.
15891 (vpkudus): Likewise.
15892 (vpksdus): Likewise.
15893 (vrld): Likewise.
15894 (vsld): Likewise.
15895 (vsrd): Likewise.
15896 (vsrad): Likewise.
15897 (vsubudm): Likewise.
15898 (vcmpequd): Likewise.
15899 (vcmpgtsd): Likewise.
15900 (vcmpgtud): Likewise.
15901 (vcmpequd_p): Likewise.
15902 (vcmpgtsd_p): Likewise.
15903 (vcmpgtud_p): Likewise.
15904 (vupkhsw): Likewise.
15905 (vupklsw): Likewise.
15906 (vaddudm): Likewise.
15907 (vmaxsd): Likewise.
15908 (vmaxud): Likewise.
15909 (vminsd): Likewise.
15910 (vminud): Likewise.
15911 (vpksdss): Likewise.
15912 (vpksdus): Likewise.
15913 (vpkudum): Likewise.
15914 (vpkudus): Likewise.
15915 (vrld): Likewise.
15916 (vsld): Likewise.
15917 (vsrad): Likewise.
15918 (vsrd): Likewise.
15919 (vsubudm): Likewise.
15920
15921 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15922 support for power8 V2DI instructions.
15923
15924 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15925 power8 V2DI instructions. Combine pack and unpack insns to use an
15926 iterator for each mode. Check whether a particular mode supports
15927 Altivec instructions instead of just checking TARGET_ALTIVEC.
15928 (UNSPEC_VPKUWUM): Likewise.
15929 (UNSPEC_VPKSHSS): Likewise.
15930 (UNSPEC_VPKSWSS): Likewise.
15931 (UNSPEC_VPKUHUS): Likewise.
15932 (UNSPEC_VPKSHUS): Likewise.
15933 (UNSPEC_VPKUWUS): Likewise.
15934 (UNSPEC_VPKSWUS): Likewise.
15935 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15936 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15937 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15938 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15939 (UNSPEC_VUPKHSB): Likewise.
15940 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15941 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15942 (UNSPEC_VUPKHSH): Likewise.
15943 (UNSPEC_VUPKLSB): Likewise.
15944 (UNSPEC_VUPKLSH): Likewise.
15945 (VI2): Likewise.
15946 (VI_char): Likewise.
15947 (VI_scalar): Likewise.
15948 (VI_unit): Likewise.
15949 (VP): Likewise.
15950 (VP_small): Likewise.
15951 (VP_small_lc): Likewise.
15952 (VU_char): Likewise.
15953 (add<mode>3): Likewise.
15954 (altivec_vaddcuw): Likewise.
15955 (altivec_vaddu<VI_char>s): Likewise.
15956 (altivec_vadds<VI_char>s): Likewise.
15957 (sub<mode>3): Likewise.
15958 (altivec_vsubcuw): Likewise.
15959 (altivec_vsubu<VI_char>s): Likewise.
15960 (altivec_vsubs<VI_char>s): Likewise.
15961 (altivec_vavgs<VI_char>): Likewise.
15962 (altivec_vcmpbfp): Likewise.
15963 (altivec_eq<mode>): Likewise.
15964 (altivec_gt<mode>): Likewise.
15965 (altivec_gtu<mode>): Likewise.
15966 (umax<mode>3): Likewise.
15967 (smax<mode>3): Likewise.
15968 (umin<mode>3): Likewise.
15969 (smin<mode>3): Likewise.
15970 (altivec_vpkuhum): Likewise.
15971 (altivec_vpkuwum): Likewise.
15972 (altivec_vpkshss): Likewise.
15973 (altivec_vpkswss): Likewise.
15974 (altivec_vpkuhus): Likewise.
15975 (altivec_vpkshus): Likewise.
15976 (altivec_vpkuwus): Likewise.
15977 (altivec_vpkswus): Likewise.
15978 (altivec_vpks<VI_char>ss): Likewise.
15979 (altivec_vpks<VI_char>us): Likewise.
15980 (altivec_vpku<VI_char>us): Likewise.
15981 (altivec_vpku<VI_char>um): Likewise.
15982 (altivec_vrl<VI_char>): Likewise.
15983 (altivec_vsl<VI_char>): Likewise.
15984 (altivec_vsr<VI_char>): Likewise.
15985 (altivec_vsra<VI_char>): Likewise.
15986 (altivec_vsldoi_<mode>): Likewise.
15987 (altivec_vupkhsb): Likewise.
15988 (altivec_vupkhs<VU_char>): Likewise.
15989 (altivec_vupkls<VU_char>): Likewise.
15990 (altivec_vupkhsh): Likewise.
15991 (altivec_vupklsb): Likewise.
15992 (altivec_vupklsh): Likewise.
15993 (altivec_vcmpequ<VI_char>_p): Likewise.
15994 (altivec_vcmpgts<VI_char>_p): Likewise.
15995 (altivec_vcmpgtu<VI_char>_p): Likewise.
15996 (abs<mode>2): Likewise.
15997 (vec_unpacks_hi_v16qi): Likewise.
15998 (vec_unpacks_hi_v8hi): Likewise.
15999 (vec_unpacks_lo_v16qi): Likewise.
16000 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16001 (vec_unpacks_lo_v8hi): Likewise.
16002 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16003 (vec_pack_trunc_v8h): Likewise.
16004 (vec_pack_trunc_v4si): Likewise.
16005 (vec_pack_trunc_<mode>): Likewise.
16006
16007 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16008 V2DI builtins.
16009 (vec_vmaxsd): Likewise.
16010 (vec_vmaxud): Likewise.
16011 (vec_vminsd): Likewise.
16012 (vec_vminud): Likewise.
16013 (vec_vpksdss): Likewise.
16014 (vec_vpksdus): Likewise.
16015 (vec_vpkudum): Likewise.
16016 (vec_vpkudus): Likewise.
16017 (vec_vrld): Likewise.
16018 (vec_vsld): Likewise.
16019 (vec_vsrad): Likewise.
16020 (vec_vsrd): Likewise.
16021 (vec_vsubudm): Likewise.
16022 (vec_vupkhsw): Likewise.
16023 (vec_vupklsw): Likewise.
16024
e70670cf
JH		 160252013-05-29 Jan Hubicka <jh@suse.cz>
16026
16027 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16028 flags; reorder rest of fields in more consistent way.
16029 (varpool_node): Remove analyzed, finalized and alias.
16030 (cgraph_ndoe): Likewise.
16031 (symtab_alias_ultimate_target): New function.
16032 (cgraph_function_node): Move offline.
16033 (cgraph_reset_node): Declare.
16034 (cgraph_comdat_can_be_unshared_p): Remove.
16035 (varpool_remove_initializer): Declare.
16036 (varpool_first_defined_variable, varpool_next_defined_variable
16037 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16038 (cgraph_function_with_gimple_body_p): Update.
16039 (varpool_all_refs_explicit_p): Update.
16040 (symtab_alias_target): New function.
16041 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16042 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB		 16043 (cgraph_function_or_thunk_node): Simplify using
16044 symtab_alias_ultimate_target.
e70670cf
JH		 16045 (varpool_variable_node): Likewise.
16046 * cgraph.c (cgraph_create_function_alias): Update.
16047 (cgraph_add_thunk): Update.
16048 (cgraph_remove_node): Update.
16049 (dump_cgraph_node): Do not dump removed flags.
16050 (cgraph_function_body_availability): Update.
16051 (cgraph_propagate_frequency): Update.
16052 (verify_cgraph_node): Check sanity of local flag.
16053 (cgraph_function_node): Move here from cgraph.h; revamp for
16054 cgraph_function_or_thunk_node.
16055 * lto-symtab.c (lto_varpool_replace_node): Update.
16056 (lto_symtab_resolve_can_prevail_p): Update.
16057 (lto_symtab_merge_cgraph_nodes): Update.
16058 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16059 propagate_constants_accross_call, devirtualization_time_bonus,
16060 ipcp_propagate_stage): Update.
16061 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB		 16062 * ipa-inline-transform.c (clone_inlined_nodes,
16063 preserve_function_body_p): Update.
e70670cf
JH		 16064 * ipa-reference.c (propagate): Update.
16065 (write_node_summary_p): Update.
16066 * toplev.c (wrapup_global_declaration_2): Update.
16067 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16068 (analyze_function) ... this one.
16069 (cgraph_process_new_functions): Update.
16070 (cgraph_reset_node): Export.
16071 (cgraph_finalize_function): Update.
16072 (cgraph_add_new_function): Update.
16073 (process_function_and_variable_attributes): Update.
16074 (varpool_finalize_decl): Update.
16075 (symbol_finalized): Remove.
16076 (symbol_finalized_and_needed): Rename to ...
16077 (symbol_defined_and_needed): ... update.
16078 (cgraph_analyze_functions): Update.
16079 (handle_alias_pairs): Update.
16080 (mark_functions_to_output): Update.
16081 (assemble_thunk): Update.
16082 (output_in_order): Update.
16083 (output_weakrefs): Update.
16084 (finalize_compilation_unit): Update.
16085 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16086 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16087 input_node, input_varpool_node): Update.
16088 * dbxout.c (dbxout_expand_expr): Update.
16089 * cgraphclones.c (cgraph_clone_node): Update.
16090 (cgraph_copy_node_for_versioning): Update.
16091 (cgraph_materialize_clone): Update.
16092 (cgraph_materialize_all_clones): Update.
16093 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16094 propagate_pure_const, propagate_nothrow): Update.
16095 * lto-streamer-out.c (lto_output, write_symbol): Update.
16096 * ipa-utils.c (ipa_reverse_postorder): Update.
16097 * ipa-inline.c (can_inline_edge_p): Update.
16098 (update_caller_keys, ipa_inline): Update.
16099 * dwarf2out.c (reference_to_unused,
16100 premark_types_used_by_global_vars_helper): Update.
16101 * tree-eh.c (tree_could_trap_p): Update.
16102 * ipa-split.c (consider_split, execute_split_functions): Update.
16103 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB		 16104 has_addr_references_p): Update; move ahead in file for better
16105 readability.
e70670cf
JH		 16106 (process_references): Simplify.
16107 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16108 bodies are removed.
16109 (cgraph_comdat_can_be_unshared_p): Make static.
16110 (cgraph_externally_visible_p): Update.
16111 (varpool_externally_visible_p): Update.
16112 (function_and_variable_visibility): Update.
16113 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16114 ipa_tm_mark_force_output_node): Update.
16115 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16116 estimate_edge_devirt_benefit, inline_generate_summary,
16117 inline_write_summary): Update.
16118 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16119 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB		 16120 (ipa_print_node_params, ipa_prop_read_section,
16121 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH		 16122 * varasm.c (mark_decl_referenced): Update.
16123 (assemble_alias, dump_tm_clone_pairs): Update.
16124 * tree-inline.c (copy_bb): Update.
16125 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16126 Update.
16127 * symtab.c (dump_symtab_base): Print new flags.
16128 (verify_symtab_base): Verify new flags.
16129 (symtab_alias_ultimate_target): New function.
16130 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16131 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16132 Update.
c1e183a9
UB		 16133 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16134 Update.
e70670cf
JH		 16135 * i386.c (ix86_get_function_versions_dispatcher,
16136 ix86_generate_version_dispatcher_body): Update.
16137 (fold_builtin_cpu): Use varpool_add_new_variable.
16138 * varpool.c (varpool_remove_initializer): Break out from ...
16139 (varpool_remove_node): ... this one.
16140 (dump_varpool_node, varpool_node_for_asm,
16141 cgraph_variable_initializer_availability, varpool_analyze_node,
16142 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB		 16143 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16144 Update.
e70670cf 16145
182802ad
JH		 161462013-05-29 Jan Hubicka <jh@suse.cz>
16147
16148 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16149
8190b609
ER		 161502013-05-29 Easwaran Raman <eraman@google.com>
16151
16152 PR tree-optimization/57442
16153 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16154 when control exits the main loop.
16155
69f5aa9b
SKS		 161562013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16157
16158 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16159 and RX600.
c1e183a9 16160 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS		 16161 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16162 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16163 option.
16164
4b847da9
BS		 161652013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16166
16167 PR tree-optimization/57441
16168 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16169 Don't limit size of incr_vec to number of candidates.
16170
4fc43c39
SE		 161712013-05-29 Steve Ellcey <sellcey@imgtec.com>
16172
16173 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16174 and mips16 directories.
c1e183a9 16175 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE		 16176 (MULTILIB_DIRNAMES): Ditto.
16177 (MULTILIB_EXCEPTIONS): Add new exceptions.
16178 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16179 (MULTILIB_DIRNAMES): Ditto.
16180 (MULTILIB_EXCEPTIONS): Add new exceptions.
16181
12211b99 161822012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS		 16183 Marcus Shawcroft <marcus.shawcroft@arm.com>
16184
16185 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16186 SYMBOL_TINY_ABSOLUTE.
16187 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16188 SYMBOL_TINY_ABSOLUTE.
16189 (aarch64_expand_mov_immediate): Likewise.
16190 (aarch64_classify_symbol): Likewise.
16191 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16192 Permit SYMBOL_TINY_ABSOLUTE.
16193 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16194
12211b99 161952013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS		 16196 Marcus Shawcroft <marcus.shawcroft@arm.com>
16197
16198 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16199 Refactor if/switch. Replace gcc_assert with if.
16200
c0186656
GG		 162012013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16202
16203 * config/i386/i386.c (initial_ix86_tune_features): Enable
16204 FP Reassociation for AMD bdver1 and bdver2.
16205
d20188f3
MJ		 162062013-05-29 Martin Jambor <mjambor@suse.cz>
16207
16208 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16209 and IMAGPART_EXPR do not occur within other handled_components.
16210
292cba13
RB		 162112013-05-29 Richard Biener <rguenther@suse.de>
16212
16213 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16214 access on whether the use is in the BB we currently try to
16215 vectorize.
16216 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16217 vectorize to vect_bb_slp_scalar_cost.
16218
6eddf228
RB		 162192013-05-29 Richard Biener <rguenther@suse.de>
16220
16221 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16222 computing scalar cost offsetted by stmts that are kept live
16223 by scalar uses.
16224 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16225 for computation of scalar cost.
16226
7df36117
SE		 162272013-05-28 Steve Ellcey <sellcey@mips.com>
16228
16229 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16230
36536d79
BI		 162312013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16232
16233 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16234 array notation built-in reduction functions.
16235 * doc/passes.texi (Passes): Added documentation about changes done
16236 for Cilk Plus.
16237 * doc/invoke.texi (C Dialect Options): Added documentation about
16238 the -fcilkplus flag.
16239 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16240 (BUILTINS_DEF): Depend on cilkplus.def.
16241 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16242 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16243 * cilkplus.def: New file.
16244
12211b99 162452013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR		 16246
16247 PR rtl-optimization/57439
16248 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16249
9daf6dbb
ER		 162502013-05-28 Easwaran Raman <eraman@google.com>
16251
16252 PR tree-optimization/57337
16253 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16254 (find_insert_point): Correctly identify the insertion point
16255 when two statements with the same UID is compared.
16256
fbd7e877
RB		 162572013-05-28 Richard Biener <rguenther@suse.de>
16258
16259 PR tree-optimization/56787
16260 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16261 from the list of data references.
16262 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16263 clobbers.
16264 (vect_analyze_loop_operations): Likewise.
16265 (vect_transform_loop): Remove clobbers.
16266
bbba1117
MJ		 162672013-05-28 Martin Jambor <mjambor@suse.cz>
16268
16269 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16270 and REALPART_EXPRs have scalar type.
16271
bd388c2a
RB		 162722013-05-28 Richard Biener <rguenther@suse.de>
16273
16274 PR tree-optimization/57411
16275 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16276 virtual operands.
16277 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16278 virtual operand propagation.
16279
2f56a311
EB		 162802013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16281
16282 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16283 destination register for bmasksi_vis.
16284 (vector_init_bshuffle): Likewise.
16285 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16286
5c3eacbb
EB		 162872013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16288
16289 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16290 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16291 mode if the instruction isn't available in the original mode.
16292 * config/sparc/sparc.opt (mfix-ut699): New option.
16293 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16294 (divdf3): Turn into expander.
16295 (divdf3_nofix): New insn.
16296 (divdf3_fix): Likewise.
16297 (divsf3): Disable if -mfix-ut699.
16298 (sqrtdf2): Turn into expander.
16299 (sqrtdf2_nofix): New insn.
16300 (sqrtdf2_fix): Likewise.
16301 (sqrtsf2): Disable if -mfix-ut699.
16302
a1756c0a
RB		 163032013-05-27 Richard Biener <rguenther@suse.de>
16304
16305 PR middle-end/57412
16306 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16307 block for the new loop.
16308
5a892248
RB		 163092013-05-27 Richard Biener <rguenther@suse.de>
16310
16311 PR tree-optimization/57343
16312 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16313 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16314 (number_of_iterations_cond): Do not build the folded tree.
16315
d1de852b
RB		 163162013-05-27 Richard Biener <rguenther@suse.de>
16317
16318 Revert
16319 PR middle-end/57381
16320 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16321 OEP_CONSTANT_ADDRESS_OF retained.
16322
16323 PR tree-optimization/57417
16324 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16325 for unchanged base.
16326 (set_ssa_val_to): Compare addresses using
16327 get_addr_base_and_unit_offset.
16328
12211b99 163292013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR		 16330
16331 PR rtl-optimization/56833
16332 * postreload.c (move2add_record_mode): New function.
16333 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16334 (move2add_use_add2_insn): Use move2add_record_sym_value.
16335 (move2add_use_add3_insn): Likewise.
16336 (reload_cse_move2add): Use move2add_valid_value_p and
16337 move2add_record_mode. Invalidate call-clobbered and REG_INC
16338 affected regs by setting reg_mode to VOIDmode.
16339 (move2add_note_store): Don't pretend the inside of a SUBREG is
16340 the actual destination. Invalidate single/leading registers by
16341 setting reg_mode to VOIDmode.
16342 Use move2add_record_sym_value, move2add_valid_value_p and
16343 move2add_record_mode.
16344
b03be25f
RB		 163452013-05-27 Richard Biener <rguenther@suse.de>
16346
16347 PR tree-optimization/57396
16348 * tree-affine.c (double_int_constant_multiple_p): Properly
16349 return false for val == 0 and div != 0.
16350
44e88db2
RS		 163512013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16352
16353 * config/mips/mips.h: Use #elif in preprocessor conditions.
16354
3b859704
RS		 163552013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16356
16357 PR target/53916
16358 * config/mips/constraints.md (kl): New constraint.
16359 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16360 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16361 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16362 from LO for MIPS16.
16363 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16364
c979d5f5
RS		 163652013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16366
16367 PR target/55777
16368 * config/mips/mips.c (mips_can_inline_p): New function.
16369 (TARGET_CAN_INLINE_P): Define.
16370
8e90de43
SB		 163712013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16372
16373 * sched-int.h (ds_t, dw_t): Make unsigned int.
16374 Fix documentation that describes how all the ds_t bits are used.
16375 Reserve the last bit for delayed-branch scheduling.
16376 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16377 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16378 (gen_dep_weak_1): Remove prototype.
16379 * sched-deps.c (get_dep_weak_1): Make static.
16380 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16381 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16382 * doc/tm.texi: Regenerate.
16383 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16384
cb5cb194
SB		 163852013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16386
16387 PR debug/56950
16388 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16389
5bd93ff6
NS		 163902013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16391 Sandra Loosemore <sandra@codesourcery.com>
16392
16393 * config.gcc (powerpc-*): Allow native for with-cpu.
16394
2343af65
JL		 163952013-05-24 Jeff Law <law@redhat.com>
16396
16397 PR tree-optimization/57124
16398 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16399 conversion feeding a condition if the range has an overflow
16400 if -fstrict-overflow. Add warnings for when we do make the
16401 transformation.
16402
3ad695b9
DC		 164032013-05-24 Dehao Chen <dehao@google.com>
16404
04960246 16405 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC		 16406 (locus_discrim_hasher): Likewise.
16407 (locus_discrim_hasher::hash): Likewise.
16408 (locus_discrim_hasher::equal): Likewise.
16409
82338059
MJ		 164102013-05-24 Martin Jambor <mjambor@suse.cz>
16411
16412 PR tree-optimization/57294
16413 * cgraph.h (ipa_record_stmt_references): Declare.
16414 * cgraphbuild.c (ipa_record_stmt_references): New function.
16415 (build_cgraph_edges): Use ipa_record_stmt_references.
16416 (rebuild_cgraph_edges): Likewise.
16417 (cgraph_rebuild_references): Likewise.
16418 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16419 associated with the old statement and build references from the
16420 newly built statements.
16421 * ipa-ref.c (ipa_remove_stmt_references): New function.
16422 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16423
1ccd4874
VM		 164242013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16425
55805e54 16426 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 16427 mem-mem moves.
55805e54 16428 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 16429 too.
55805e54 16430 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM		 16431 secondary memory moves.
16432 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16433 reg set up in the current insn.
16434
25e25c73
DC		 164352013-05-24 Dehao Chen <dehao@google.com>
16436
04960246 16437 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC		 16438 hash function.
16439 (locus_descrim_hasher::equal): Likewise.
04960246 16440 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC		 16441 (make_edges): Likewise.
16442 (next_discriminator_for_locus): Likewise.
16443 (same_line_p): Likewise.
16444 (assign_discriminators): Likewise.
16445 (make_cond_expr_edges): Likewise.
16446 (make_gimple_switch_edges): Likewise.
16447 (make_goto_expr_edges): Likewise.
16448 (make_gimple_asm_edges): Likewise.
16449
50d38551
IB		 164502013-05-24 Ian Bolton <ian.bolton@arm.com>
16451
16452 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16453 X format specifier to only display bottom 16 bits.
16454 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16455 immediate to match for operand 2, since it will be masked.
16456
aea0101d
RB		 164572013-05-24 Richard Biener <rguenther@suse.de>
16458
16459 PR tree-optimization/57287
16460 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16461 all SSA names that occur in abnormal PHIs.
16462
634e03d3
AI		 164632013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16464
16465 PR tree-ssa/57385
16466 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16467 that index is not negative.
16468
b17c024f
EB		 164692013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16470
16471 PR rtl-optimization/55177
16472 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16473 (simplify_byte_swapping_operation): New.
16474 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16475 (simplify_relational_operation_1): Deal with BSWAP.
16476
46aeac1b
RH		 164772013-05-23 Richard Henderson <rth@redhat.com>
16478
16479 PR target/56742
16480 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16481 (ix86_reorg): Call it.
16482
70cc1536
UB		 164832013-05-23 Uros Bizjak <ubizjak@gmail.com>
16484
16485 PR target/57379
16486 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16487 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16488 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16489
b3851501
CB		 164902013-05-23 Christian Bruel <christian.bruel@st.com>
16491
16492 PR debug/57351
16493 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16494
12211b99 164952013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS		 16496 Marcus Shawcroft <marcus.shawcroft@arm.com>
16497
16498 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16499 * config/aarch64/constraints.md (Usa): Remove.
16500 * doc/md.texi (AArch64 Usa): Remove.
16501
12211b99 165022013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS		 16503 Marcus Shawcroft <marcus.shawcroft@arm.com>
16504
16505 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16506 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16507 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16508 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16509
12211b99 165102013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP		 16511
16512 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16513 instruction (AdvSIMD).
16514 * config/aarch64/aarch64-builtins.c
16515 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16516 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16517
76c36cb1 165182013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ		 16519
16520 PR middle-end/57347
16521 * tree.h (contains_bitfld_component_ref_p): Declare.
16522 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB		 16523 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16524 caller.
7d2fb524
MJ		 16525 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16526 not access a bit-field. Assert all final offsets are byte-aligned.
16527
ce521ae6
RB		 165282013-05-23 Richard Biener <rguenther@suse.de>
16529
16530 PR tree-optimization/57380
16531 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16532 least one invariant or re-used load.
16533 * passes.c (init_optimization_passes): Move pass_phiprop before
16534 pass_forwprop.
16535
75c7257f
JG		 165362013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16537
16538 * config/aarch64/aarch64-simd.md
16539 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16540
af6d44b5
RB		 165412013-05-23 Richard Biener <rguenther@suse.de>
16542
16543 PR middle-end/57381
16544 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16545 OEP_CONSTANT_ADDRESS_OF retained.
16546
bd3647bf
JJ		 165472013-05-23 Jakub Jelinek <jakub@redhat.com>
16548
16549 PR middle-end/57344
70cc1536
UB		 16550 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16551 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 16552
a55757ea
RB		 165532013-05-23 Richard Biener <rguenther@suse.de>
16554
16555 PR rtl-optimization/57341
16556 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16557 instead of true_dependence.
16558
cfb00b41
DM		 165592013-05-22 David Malcolm <dmalcolm@redhat.com>
16560
dd1c676f
DM		 16561 * bb-reorder.c (branch_threshold): Make const.
16562 (exec_threshold): Ditto.
cfb00b41 16563
f62511da 165642013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16565 Pat Haugen <pthaugen@us.ibm.com>
16566 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16567
16568 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16569 documentation for the power8 crypto builtins.
16570
16571 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16572
16573 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16574 macros for defining power8 builtin functions.
16575 (BU_P8V_AV_2): Likewise.
16576 (BU_P8V_AV_P): Likewise.
16577 (BU_P8V_VSX_1): Likewise.
16578 (BU_P8V_OVERLOAD_1): Likewise.
16579 (BU_P8V_OVERLOAD_2): Likewise.
16580 (BU_CRYPTO_1): Likewise.
16581 (BU_CRYPTO_2): Likewise.
16582 (BU_CRYPTO_3): Likewise.
16583 (BU_CRYPTO_OVERLOAD_1): Likewise.
16584 (BU_CRYPTO_OVERLOAD_2): Likewise.
16585 (XSCVSPDP): Fix typo, point to the correct instruction.
16586 (VCIPHER): Add power8 crypto builtins.
16587 (VCIPHERLAST): Likewise.
16588 (VNCIPHER): Likewise.
16589 (VNCIPHERLAST): Likewise.
16590 (VPMSUMB): Likewise.
16591 (VPMSUMH): Likewise.
16592 (VPMSUMW): Likewise.
16593 (VPERMXOR_V2DI): Likewise.
16594 (VPERMXOR_V4SI: Likewise.
16595 (VPERMXOR_V8HI: Likewise.
16596 (VPERMXOR_V16QI: Likewise.
16597 (VSHASIGMAW): Likewise.
16598 (VSHASIGMAD): Likewise.
16599 (VPMSUM): Likewise.
16600 (VPERMXOR): Likewise.
16601 (VSHASIGMA): Likewise.
16602
16603 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16604 __CRYPTO__ if the crypto instructions are available.
16605 (altivec_overloaded_builtins): Add support for overloaded power8
16606 builtins.
16607
16608 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16609 support for power8 crypto builtins.
16610 (builtin_function_type): Likewise.
16611 (altivec_init_builtins): Add support for builtins that take vector
16612 long long (V2DI) arguments.
16613
16614 * config/rs6000/crypto.md: New file, define power8 crypto
16615 instructions.
16616
166172013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB		 16618 Pat Haugen <pthaugen@us.ibm.com>
16619 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM		 16620
16621 * doc/invoke.texi (Option Summary): Add power8 options.
16622 (RS/6000 and PowerPC Options): Likewise.
16623
16624 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16625 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16626 wm, wn, wr documentation.
16627
f43245d1 16628 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM		 16629 registers if direct move instructions are enabled.
16630 (wn): New constraint for no registers.
16631 (wq): New constraint for quad word even GPR registers.
16632 (wr): New constraint if 64-bit instructions are enabled.
16633 (wv): New constraint if power8 vector instructions are enabled.
16634 (wQ): New constraint for quad word memory locations.
16635
f43245d1 16636 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM		 16637 constraint for 0..15 for crypto instructions.
16638 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16639 as GPR and floating point registers.
16640 (int_reg_operand): New predicate to match only GPR registers.
16641 (base_reg_operand): New predicate to match base registers.
16642 (quad_int_reg_operand): New predicate to match even GPR registers
16643 for quad memory operations.
16644 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16645 operations in both GPR and VSX registers.
16646 (quad_memory_operand): New predicate for quad memory operations.
16647 (reg_or_indexed_operand): New predicate for direct move support.
16648
f43245d1 16649 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM		 16650 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16651 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16652 (POWERPC_MASKS): Add power8 options.
16653 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16654 various options.
16655
f43245d1 16656 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM		 16657 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16658
f43245d1 16659 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM		 16660 (-mpower8-fusion): New power8 options.
16661 (-mpower8-fusion-sign): Likewise.
16662 (-mpower8-vector): Likewise.
16663 (-mcrypto): Likewise.
16664 (-mdirect-move): Likewise.
16665 (-mquad-memory): Likewise.
16666
f43245d1 16667 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM		 16668 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16669 registers.
70cc1536 16670 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM		 16671 (rs6000_debug_vector_unit): Add p8_vector.
16672 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16673 definitions. Also print fusion state.
16674 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16675 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16676 (rs6000_option_override_internal): Add support for power8.
16677 (rs6000_common_init_builtins): Add debugging for skipped builtins
16678 if -mdebug=builtin.
16679 (rs6000_adjust_cost): Add power8 support.
16680 (rs6000_issue_rate): Likewise.
16681 (insn_must_be_first_in_group): Likewise.
16682 (insn_must_be_last_in_group): Likewise.
16683 (force_new_group): Likewise.
16684 (rs6000_register_move_cost): Likewise.
16685 (rs6000_opt_masks): Likewise.
16686
16687 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16688 power8 capable assembler, default to power7 options.
16689 (TARGET_DIRECT_MOVE): Likewise.
16690 (TARGET_CRYPTO): Likewise.
16691 (TARGET_P8_VECTOR): Likewise.
16692 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16693 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16694 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16695 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16696 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16697 (TARGET_XSCVDPSPN): Likewise.
16698 (TARGET_XSCVSPDPN): Likewsie.
16699 (TARGET_SYNC_HI_QI): Likewise.
16700 (TARGET_SYNC_TI): Likewise.
16701 (MASK_CRYPTO): Likewise.
16702 (MASK_DIRECT_MOVE): Likewise.
16703 (MASK_P8_FUSION): Likewise.
16704 (MASK_P8_VECTOR): Likewise.
70cc1536
UB		 16705 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16706 temporary used by some of the direct move instructions to get two FP
16707 temporary registers does not force creation of a stack frame.
f62511da
MM		 16708 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16709 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16710 that any VSX registers are tieable, even if they are also an
16711 Altivec vector mode.
16712 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16713 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16714 (RS6000_BTM_CRYPTO): Likewise.
16715 (RS6000_BTM_COMMON): Likewise.
16716
16717 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16718 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16719 (enum rs6000_vector): Add power8 vector support.
16720
73a1a707
RR		 167212013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16722
16723 PR target/19599
16724 PR target/57340
16725 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16726 (any_sibcall_could_use_r3): this and handle indirect calls.
16727 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16728
d28073d4
BS		 167292013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16730
16731 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16732
bee0b10c
RB		 167332013-05-22 Richard Biener <rguenther@suse.de>
16734
16735 PR middle-end/57349
16736 * profile.c (branch_prob): Do not split blocks that are
16737 abnormally receiving from ECF_RETURNS_TWICE functions.
16738
98409b51
RS		 167392013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16740
16741 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16742 address mode. Move pointer mode initialization to the same place.
16743
c0602ab8
MZ		 167442013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16745
16746 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16747 while it has any effect.
16748
4f7a634e
ER		 167492013-05-21 Easwaran Raman <eraman@google.com>
16750
16751 PR tree-optimization/57322
9daf6dbb
ER		 16752 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16753 UID of the statement added to the BB to be 1.
4f7a634e 16754
a4ce1258
JJ		 167552013-05-21 Jakub Jelinek <jakub@redhat.com>
16756
16757 PR tree-optimization/57331
70cc1536
UB		 16758 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16759 of conversion from pointer type to integral type with integer.
a4ce1258 16760
1b14621a
MJ		 167612013-05-21 Martin Jambor <mjambor@suse.cz>
16762
16763 PR lto/57289
16764 * ipa-prop.c (ipa_read_node_info): Process param_used and
16765 controlled_uses in the same order as when writing.
16766
e60661f0
MG		 167672013-05-21 Magnus Granberg <baldrick@free.fr>
16768
16769 PR plugins/56754
ee49aa34 16770 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 16771
4a61be9a
RB		 167722013-05-21 Richard Biener <rguenther@suse.de>
16773
16774 PR tree-optimization/57318
16775 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16776 estimate stmts with side-effects as likely eliminated.
16777
c52da5f7
RB		 167782013-05-21 Richard Biener <rguenther@suse.de>
16779
16780 PR tree-optimization/57330
16781 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16782 preserve the call stmts fntype.
16783
7ec67e2a
RB		 167842013-05-21 Richard Biener <rguenther@suse.de>
16785
16786 PR tree-optimization/57303
16787 * tree-ssa-sink.c (statement_sink_location): Improve killing
16788 stmt detection and properly handle self-assignments.
16789
b112d513
CB		 167902013-05-21 Christian Bruel <christian.bruel@st.com>
16791
55805e54
YR		 16792 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16793 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB		 16794 registers. Set register size out of the PARALLEL loop.
16795
14c2ec26
OE		 167962013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16797
16798 PR target/56547
16799 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16800 (*fmasf4, *fmasf4_media): New insns.
16801
da734fa1
RS		 168022013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16803
16804 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16805 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16806 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16807 (mips_idiv_insns): Update the comments to say that the returned
16808 instruction counts are in units of BASE_INSN_LENGTH.
16809 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16810 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16811 using 2 rather than 4 as the length of indirect MIPS16 and
16812 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16813 length of a NOP. Don't divide MIPS16 lengths by 2.
16814 (mips16_split_long_branches): Assume a branch is long if the
16815 length is greater than 4 rather than 8.
16816 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16817 rather than multiplying them by 2. Multiply instruction counts
16818 by BASE_INSN_LENGTH rather than 4.
16819 (*jump_mips16, tls_get_tp_mips16_<mode>)
16820 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16821
13719e8b
RS		 168222013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16823
16824 * config/mips/mips.md (extended_mips16): Remove branch case.
16825 (length): Remove duplicated extended_mips16 test.
16826
c3850d14
RS		 168272013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16828
16829 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16830
8da2e059
RS		 168312013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16832
16833 * recog.h (Recog_data): Rename to...
16834 (recog_data_d): ...this.
16835 (recog_data): Update accordingly.
16836 * recog.c (recog_data): Likewise.
16837 * reload.c (save_recog_data): Likewise.
16838 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16839 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16840
2df013f3
JB		 168412013-05-17 Julian Brown <julian@codesourcery.com>
16842
16843 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16844 found in a REG_EQUAL note, invalidate it.
16845
371e77e3 168462013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER		 16847
16848 * tree-ssa-reassoc.c (find_insert_point): New function.
16849 (insert_stmt_after): Likewise.
16850 (get_def_stmt): Likewise.
16851 (ensure_ops_are_available): Likewise.
16852 (not_dominated_by): Likewise.
16853 (rewrite_expr_tree): Do not move statements beyond what is
16854 necessary. Remove call to swap_ops_for_binary_stmt...
16855 (reassociate_bb): ... and move it here.
16856 (build_and_add_sum): Assign UIDs for new statements.
16857 (linearize_expr): Likewise.
16858 (do_reassoc): Renumber gimple statement UIDs.
16859
e01c7cca
JH		 168602013-05-17 Jan Hubicka <jh@suse.cz>
16861
16862 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16863 weakrefs.
16864 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB		 16865 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16866 weakrefs.
e01c7cca
JH		 16867 (output_weakrefs): Update.
16868
c3272a92
PCC		 168692013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16870 Martin Jambor <mjambor@suse.cz>
16871
16872 PR middle-end/57276
16873 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16874 value that corresponds to the given aggval is found in values vector.
16875
11c2aa39
UB		 168762013-05-17 Uros Bizjak <ubizjak@gmail.com>
16877
16878 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16879 sse, sse2, sse3, ssse3 and sse4a flags to options.
16880
abecc8c6
DM		 168812013-05-17 David Malcolm <dmalcolm@redhat.com>
16882
16883 * gengtype-state.c: (s_expr_writer): New class, to handle
16884 prettifying of output layout of s-expressions.
16885 (state_writer): New class, to write out gtype.state.
16886 (state_written_type_count): Move this variable into member data of
16887 state_writer.
16888 (s_expr_writer::s_expr_writer): New code: constructor for new class
16889 (state_writer::state_writer(): ditto
16890 (s_expr_writer::write_new_line): New function
16891 (s_expr_writer::write_any_indent): ditto
16892 (s_expr_writer::begin_s_expr): ditto
16893 (s_expr_writer::end_s_expr): ditto
16894 (write_state_fileloc): convert to method of state_writer...
16895 (state_writer:: write_state_fileloc): ...and use methods of
16896 s_expr_writer to write indentation into the gtype.state output file
16897 to visually represent the hierarchical structure of the list
16898 structures
16899 (write_state_fields): ditto, renaming to...
16900 (state_writer::write_state_fields)
16901 (write_state_a_string): ditto, renaming to...
16902 (state_writer::write_state_a_string)
16903 (write_state_string_option): ditto, renaming to...
16904 (state_writer::write_state_string_option)
16905 (write_state_type_option): ditto, renaming to...
16906 (state_writer::write_state_type_option)
16907 (write_state_nested_option): ditto, renaming to...
16908 (state_writer::write_state_nested_option)
16909 (write_state_option): ditto, renaming to...
16910 (state_writer::write_state_option)
16911 (write_state_options): ditto, renaming to...
16912 (state_writer::write_state_options)
16913 (write_state_lang_bitmap): ditto, renaming to...
16914 (state_writer::write_state_lang_bitmap)
16915 (write_state_version): ditto, renaming to...
16916 (state_writer::write_state_version)
16917 (write_state_scalar_type): ditto, renaming to...
16918 (state_writer::write_state_scalar_type)
16919 (write_state_string_type): ditto, renaming to...
16920 (state_writer::write_state_string_type)
16921 (write_state_undefined_type): ditto, renaming to...
16922 (state_writer::write_state_undefined_type)
16923 (write_state_struct_union_type): ditto, renaming to...
16924 (state_writer::write_state_struct_union_type)
16925 (write_state_struct_type): ditto, renaming to...
16926 (state_writer::write_state_struct_type)
16927 (write_state_user_struct_type): ditto, renaming to...
16928 (state_writer::write_state_user_struct_type)
16929 (write_state_lang_struct_type): ditto, renaming to...
16930 (state_writer::write_state_lang_struct_type)
16931 (write_state_param_struct_type): ditto, renaming to...
16932 (state_writer::write_state_param_struct_type)
16933 (write_state_pointer_type): ditto, renaming to...
16934 (state_writer::write_state_pointer_type)
16935 (write_state_array_type): ditto, renaming to...
16936 (state_writer::write_state_array_type)
16937 (write_state_gc_used): ditto, renaming to...
16938 (state_writer::write_state_gc_used)
16939 (write_state_common_type_content): ditto, renaming to...
16940 (state_writer::write_state_common_type_content)
16941 (write_state_type): ditto, renaming to...
16942 (state_writer::write_state_type)
16943 (write_state_pair_list): ditto, renaming to...
16944 (state_writer::write_state_pair_list)
16945 (write_state_pair): ditto, renaming to...
16946 (state_writer::write_state_pair)
16947 (write_state_typedefs): ditto, renaming to...
16948 (state_writer::write_state_typedefs)
16949 (write_state_structures): ditto, renaming to...
16950 (state_writer::write_state_structures)
16951 (write_state_param_structs): ditto, renaming to...
16952 (state_writer::write_state_param_structs)
16953 (write_state_variables): ditto, renaming to...
16954 (state_writer::write_state_variables)
16955 (write_state_srcdir): ditto, renaming to...
16956 (state_writer::write_state_srcdir)
16957 (write_state_files_list): ditto, renaming to...
16958 (state_writer::write_state_files_list)
16959 (write_state_languages): ditto, renaming to...
16960 (state_writer::write_state_languages)
16961 (write_state): create a state_writer instance and use it when
16962 writing out the state file
16963
d6545f29
MS		 169642013-05-17 Mike Stump <mikestump@comcast.net>
16965
816a3f73 16966 PR rtl-optimization/57304
d6545f29
MS		 16967 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16968 accessing DF_REF_REAL_LOC.
16969
38c821cf
JJ		 169702013-05-17 Jakub Jelinek <jakub@redhat.com>
16971
16972 PR rtl-optimization/57281
16973 PR rtl-optimization/57300
16974 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16975 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16976 what the other splitter did if the registers are dead.
16977
2a293391
RB		 169782013-05-17 Richard Biener <rguenther@suse.de>
16979
16980 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16981 MEM_REF offsets.
16982
17042d2b
JJ		 169832013-05-17 Jakub Jelinek <jakub@redhat.com>
16984
16985 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16986 linking.
16987
5b115c1f
MP		 169882013-05-17 Marek Polacek <polacek@redhat.com>
16989
11c2aa39
UB		 16990 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16991 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 16992
52d84413
JJ		 169932013-05-17 Jakub Jelinek <jakub@redhat.com>
16994
68119618
JJ		 16995 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16996 vect_external_def oprnd1 with loop_vinfo, try to emit
16997 optional cast, negation and and stmts on the loop preheader
16998 edge instead of into the pattern def seq.
16999
52d84413
JJ		 17000 PR tree-optimization/57051
17001 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17002 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17003
17b962bd
NC		 170042013-05-16 Nick Clifton <nickc@redhat.com>
17005
17006 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17007 (rl78_is_naked_func): New function.
56aefbf7
UB		 17008 (rl78_expand_prologue): Skip prologue generation for naked functions.
17009 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC		 17010 * doc/extend.texi (naked): Add RL78 to the list of processors
17011 that supports this attribute.
17012
b1a0f84e
JL		 170132013-05-16 Jeff Law <law@redhat.com>
17014
17015 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17016
af0e415b
UB		 170172013-05-16 Uros Bizjak <ubizjak@gmail.com>
17018
17019 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17020 cache parameters using detect_caches_amd also for CYRIX,
17021 NSC and TM2 signatures.
17022
19db293a
UB		 170232013-05-16 Uros Bizjak <ubizjak@gmail.com>
17024 Dzianis Kahanovich <mahatma@eu.by>
17025
17026 PR target/45359
17027 PR target/46396
17028 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17029 VIA/Centaur processors and determine their cache parameters
17030 using detect_caches_amd.
17031
251a41b9
TJ		 170322013-05-16 Teresa Johnson <tejohnson@google.com>
17033
17034 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17035 (rtl_verify_edges): New function.
17036 (rtl_verify_bb_insns): Ditto.
17037 (rtl_verify_bb_pointers): Ditto.
17038 (rtl_verify_bb_insn_chain): Ditto.
17039 (rtl_verify_fallthru): Ditto.
17040 (rtl_verify_bb_layout): Ditto.
17041 (rtl_verify_flow_info_1): Outline checks into new functions.
17042 (rtl_verify_flow_info): Ditto.
17043
f14540b6
SE		 170442013-05-16 Steve Ellcey <sellcey@imgtec.com>
17045
17046 * cfghooks.c (copy_bbs): Add update_dominance argument.
17047 * cfghooks.h (copy_bbs): Update prototype.
17048 * tree-cfg.c (gimple_duplicate_sese_region):
17049 Add update_dominance argument.
17050 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17051 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17052 gimple_duplicate_sese_region call.
17053 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17054 Update copy_bbs call.
17055 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17056 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17057
7e9a3abb
JJ		 170582013-05-16 Jakub Jelinek <jakub@redhat.com>
17059
17060 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17061 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17062 vect_recog_rotate_pattern.
17063 (vect_recog_rotate_pattern): New function.
17064
427b248d
JM		 170652013-05-16 Jason Merrill <jason@redhat.com>
17066
17067 * Makefile.in (LLINKER): New variable.
17068 (mostlyclean): Remove link mutex.
17069 * configure.ac: Handle --enable-link-mutex.
17070 * lock-and-run.sh: New script.
17071
b871e3d2
RR		 170722013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17073
17074 PR target/19599
17075 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17076 for NULL decl.
17077
ce858126
RO		 170782013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17079
17080 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17081
798d3d04
GY		 170822013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17083
17084 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17085 * config/arm/arm.c (next_consecutive_mem): New function.
17086 (gen_movmem_ldrd_strd): Likewise.
17087 * config/arm/arm.md (movmemqi): Update condition and code.
17088 (unaligned_loaddi, unaligned_storedi): New patterns.
17089
0baddc45
RO		 170902013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17091
17092 * config.gcc: Obsolete *-*-solaris2.9*.
17093 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17094
aa9d5bcf
RB		 170952013-05-16 Richard Biener <rguenther@suse.de>
17096
17097 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17098 earlier, after GRAPHITE transforms and IV canonicalization.
17099
5a0f4dd3
JJ		 171002013-05-16 Jakub Jelinek <jakub@redhat.com>
17101
17102 * omp-low.c (extract_omp_for_data): For collapsed loops,
17103 if at least one of the loops is known at compile time to
17104 iterate zero times, set count to 0.
17105 (expand_omp_regimplify_p): New function.
17106 (expand_omp_for_generic): For collapsed loops, if at least
17107 one of the loops isn't known to iterate at least once,
17108 add runtime check with setting count to 0.
17109 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17110 For unsigned types if it isn't known at compile time that
17111 the loop will iterate at least once, add runtime check to bypass
17112 the whole loop if initial condition isn't true.
17113
e3753785
NS		 171142013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17115
17116 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17117
43bb4dd1
MG		 171182013-05-16 Marc Glisse <marc.glisse@inria.fr>
17119
17120 PR middle-end/57286
17121 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17122 transformations to avoid an infinite loop.
17123
3571dde6
MP		 171242013-05-16 Marek Polacek <polacek@redhat.com>
17125
17126 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17127
b8b3f0ca
LE		 171282013-05-15 Leif Ekblad <leif@rdos.net>
17129
17130 * config/i386/i386.c (ix86_decompose_address): Use
17131 DEFAULT_TLS_SEG_REG to access TLS segment register.
17132 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17133 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17134 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17135
e299a383
RS		 171362013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17137
17138 PR target/57260
17139 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17140 sibling calls to functions that would normally be lazily bound,
17141 unless $gp is call-clobbered.
17142
e7208ea3
UB		 171432013-05-15 Uros Bizjak <ubizjak@gmail.com>
17144
19db293a 17145 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 17146 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB		 17147 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17148 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB		 17149 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17150 of TARGET_3DNOW.
17151 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17152
e7413f3d
AS		 171532013-05-15 Andreas Schwab <schwab@suse.de>
17154
17155 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17156 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17157 third operand.
17158
ad4db775
TJ		 171592013-05-15 Teresa Johnson <tejohnson@google.com>
17160
17161 * loop-unroll.c (report_unroll_peel): Check decision before
17162 emitting unroll/peel message.
17163
af205f67
TJ		 171642013-05-15 Teresa Johnson <tejohnson@google.com>
17165
17166 * function.h (has_bb_partition): New rtl_data flag.
17167 (bb_reorder_complete): Ditto.
17168 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17169 instead of flag_reorder_blocks_and_partition.
17170 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17171 with some enhancements.
17172 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17173 * bb-reorder.c (connect_traces): Check for has_bb_partition
17174 instead of flag_reorder_blocks_and_partition.
17175 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17176 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17177 verify_hot_cold_block_grouping.
17178 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17179
9adcfa3c
RR		 171802013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17181
17182 PR target/19599
17183 * config/arm/predicates.md (call_insn_operand): New predicate.
17184 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17185 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17186 if insn is not a tail call.
17187 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17188 registers.
17189 * config/arm/arm.h (enum reg_class): New caller save register class.
17190 (REG_CLASS_NAMES): Likewise.
17191 (REG_CLASS_CONTENTS): Likewise.
17192 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17193 without decls.
17194
ccb3ad87
RB		 171952013-05-15 Richard Biener <rguenther@suse.de>
17196
17197 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17198 of MSG_OPTIMIZED_LOCATIONS.
17199 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17200 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17201 message.
17202 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17203 of MSG_OPTIMIZED_LOCATIONS.
17204 (execute_vect_slp): Likewise.
17205 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17206 (vect_create_cond_for_alias_checks): Likewise.
17207 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17208 (vect_recog_widen_mult_pattern): Likewise.
17209 (vect_recog_widen_sum_pattern): Likewise.
17210 (vect_recog_over_widening_pattern): Likewise.
17211 (vect_recog_widen_shift_pattern): Likewise.
17212 (vect_recog_vector_vector_shift_pattern): Likewise.
17213 (vect_recog_divmod_pattern): Likewise.
17214 (vect_recog_mixed_size_cond_pattern): Likewise.
17215 (vect_recog_bool_pattern): Likewise.
17216 (vect_pattern_recog_1): Likewise.
17217
48b1474e
MJ		 172182013-05-15 Martin Jambor <mjambor@suse.cz>
17219
17220 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17221 non-functions to builtin_unreachable.
17222 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17223 correct when new direct edges were discovered.
17224
9de04252
MJ		 172252013-05-15 Martin Jambor <mjambor@suse.cz>
17226
17227 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17228 header, print symbol order instead of node uid, print more information
17229 about indirect edge targets.
17230 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17231 uids.
17232 (ipa_make_edge_direct_to_target): Likewise.
17233 (remove_described_reference): Likewise.
17234 (propagate_controlled_uses): Likewise.
17235 (ipa_print_node_params): Also print symbol order.
17236 (ipcp_transform_function): Print symbol order instead of node uids.
17237 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17238 (cgraph_get_create_real_symbol_node): Likewise.
17239 * ipa-cp.c (print_lattice): Likewise.
17240 (print_all_lattices): Likewise.
17241 (determine_versionability): Likewise.
17242 (initialize_node_lattices): Likewise.
17243 (estimate_local_effects): Likewise.
17244 (update_profiling_info): Likewise.
17245 (create_specialized_node): Likewise.
17246 (perhaps_add_new_callers): Likewise.
17247 (decide_about_value): Likewise.
17248 (decide_whether_version_node): Likewise.
17249 (identify_dead_nodes): Likewise.
17250 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17251 (dump_inline_summary): Likewise.
17252 (estimate_node_size_and_time): Likewise.
17253 (inline_analyze_function): Likewise.
17254 * ipa-inline.c (report_inline_failed_reason): Likewise.
17255 (want_early_inline_function_p): Likewise.
17256 (edge_badness): Likewise.
17257 (update_edge_key): Likewise.
17258 (inline_small_functions): Likewise. Add dumping of order to two other
17259 dumps.
17260 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17261 instead of node uids.
17262 (propagate_pure_const): Likewise.
17263 (propagate_pure_const): Likewise.
17264 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17265 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17266 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17267 of node uids.
17268 * tree-pretty-print.c (dump_function_header): Likewise.
17269 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17270 Print symbol order instead of node uids.
17271
1dd03b91
AK		 172722013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17273
17274 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17275 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17276
fdf6a7b9
RB		 172772013-05-15 Richard Biener <rguenther@suse.de>
17278
17279 PR tree-optimization/57275
17280 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17281 return value for fail to do runtime alias checks for gather loads.
17282
2d6e4603
JH		 172832013-05-15 Jan Hubicka <jh@suse.cz>
17284
17285 PR lto/57038
17286 PR lto/47375
e7208ea3
UB		 17287 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17288 weakrefs are not external.
17289 (lto_symtab_merge_decls): Fix thinko when dealing with
17290 non-lto_symtab decls.
2d6e4603
JH		 17291 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17292 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17293 * varpool.c (dump_varpool_node): Dump more flags.
17294
83f44b39
GG		 172952013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17296
17297 * config/i386/i386.c (processor_alias_table): Add instruction
17298 FSGSBASE for AMD bdver3 architecture.
17299
2e55d062
JJ		 173002013-05-14 Jakub Jelinek <jakub@redhat.com>
17301
17302 * tree.c (warn_deprecated_use): Print file:line using locus color.
17303 * diagnostic.c (diagnostic_report_current_module): Print file:line
17304 and file:line:column using locus color.
17305
0bfdb81e
MS		 173062013-05-14 Mike Stump <mikestump@comcast.net>
17307
17308 * gdbinit.in: Add __null.
17309
a508ef22
MS		 173102013-05-14 Mike Stump <mikestump@comcast.net>
17311
17312 * recog.h: Rename struct recog_data to Recog_data.
17313 * recog.c: Likewise.
17314 * reload.c (can_reload_into): Likewise.
17315 * config/picochip/picochip.c: Likewise.
17316
e7180acb
MS		 173172013-05-14 Mike Stump <mikestump@comcast.net>
17318
17319 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17320
61216c26
SB		 173212013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17322
fd6beed4
SB		 17323 * resource.h (struct resources): Remove unch_memory member.
17324 (CLEAR_RESOURCE): Don't clear unch_memory.
17325 * resource.c (mark_referenced_resources): Don't set it.
17326 (mark_set_resources): Likewise.
17327 (mark_target_live_regs): Don't clear it.
17328 (init_resource_info): Likewise.
17329 * reorg.c (resource_conflicts_p): Don't compare it.
17330 (redundant_insn): Don't set it.
17331
61216c26
SB		 17332 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17333 Remove prototypes.
17334 * emit-rtl.c (next_label): Remove unused function.
17335 (skip_consecutive_labels, link_cc0_insns): Move to ...
17336 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17337 only place where these functions are used, and make them static.
17338
418d1b87
MG		 173392013-05-14 Marc Glisse <marc.glisse@inria.fr>
17340
17341 * fold-const.c (fold_negate_expr): Handle vectors.
17342 (fold_truth_not_expr): Make it static.
17343 (fold_invert_truthvalue): New static function.
17344 (invert_truthvalue_loc): Handle vectors. Do not call
17345 fold_truth_not_expr directly.
17346 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17347 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17348 (fold_comparison): Handle vector constants.
17349 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17350 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17351 * tree.h (fold_truth_not_expr): Remove declaration.
17352
fc21784d
JG		 173532013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17354
17355 * config/aarch64/aarch64-simd.md
17356 (aarch64_vcond_internal<mode>): Rename to...
17357 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17358 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17359 float modes. Clarify all iterator modes.
17360 (vcond<mode><mode>): Use new name for vcond expanders.
17361 (vcond<v_cmp_result><mode>): Likewise.
17362 (vcondu<mode><mode>: Likewise.
17363 * config/aarch64/iterators.md (VDQF_COND): New.
17364
d4c52634
MG		 173652013-05-14 Marc Glisse <marc.glisse@inria.fr>
17366
17367 PR bootstrap/57266
17368 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17369 variable for the shift amount. Check that we shift by non-negative
17370 amounts.
17371
2b261262
CLT		 173722013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17373
17374 PR target/42017
17375 * config/arm/arm.h (EPILOGUE_USES): Only return true
17376 for LR_REGNUM after epilogue_completed.
17377
12211b99 173782013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR		 17379
17380 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 17381 is error_mark_node.
6e022d7b 17382
56cf7859
RO		 173832013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17384
17385 PR target/57261
17386 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17387 and Solaris 11+/x86 with gld.
17388 * configure: Regenerate.
17389
75776c6d
JJ		 173902013-05-14 Jakub Jelinek <jakub@redhat.com>
17391
17392 * expmed.c (expand_shift_1): Canonicalize rotates by
17393 constant bitsize / 2 to bitsize - 1.
7f998021 17394 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ		 17395 case ROTATERT>: Likewise.
17396
17397 Revert:
17398 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17399
17400 * config/i386/i386.md (rotateinv): New code attr.
17401 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17402 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17403 roll $31, %eax, etc.
17404
df35498a
RB		 174052013-05-14 Richard Biener <rguenther@suse.de>
17406
17407 PR middle-end/57235
17408 * tree-eh.c (sink_clobbers): Give up for successors with
17409 multiple predecessors and no virtual uses.
17410
cc6e7ece
EB		 174112013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17412
17413 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17414 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17415
f778dd4d
JJ		 174162013-05-14 Jakub Jelinek <jakub@redhat.com>
17417
17418 PR middle-end/57251
17419 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17420 the case when both op0 and op1 have VOIDmode.
17421
bad4df9b
KP		 174222013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17423
17424 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17425 in multiply-accumulate mode.
17426
56f3e9ac
GW		 174272013-05-13 Guozhi Wei <carrot@google.com>
17428
17429 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17430
a3d7ab92
KT		 174312013-05-13 Kai Tietz <ktietz@redhat.com>
17432
17433 PR target/56975
17434 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 17435 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 17436 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 17437 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT		 17438 (ix86_expand_prologue): Likewise.
17439 (ix86_expand_split_stack_prologue): Likewise.
17440 (legitimate_pic_address_disp_p): Likewise.
17441 (legitimize_pic_address): Likewise.
17442 (legitimize_tls_address): Likewise.
17443 (legitimize_pe_coff_symbol): Likewise.
17444 (output_pic_addr_const): Likewise.
17445 (construct_plt_address): Likewise.
17446 (ix86_expand_call): Likewise.
17447 (x86_output_mi_thunk): Likewise.
17448 (x86_function_profiler): Likewise.
17449
c59b7e28
SN		 174502013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17451
17452 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17453 similar switch cases.
17454 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17455 (aarch64_simd_mov_to_<mode>low): Delete.
17456 (aarch64_simd_mov_to_<mode>high): Delete.
17457 (move_lo_quad_<mode>): Add w<-r alternative.
17458 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17459 (aarch64_simd_mov_from_*): Update type attribute.
17460 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17461 statement.
17462
5f28524a
JH		 174632013-05-13 Jan Hubicka <jh@suse.cz>
17464
17465 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17466 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB		 17467 ix86_expand_epilogue, emit_i387_cw_initialization,
17468 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17469 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 17470
ae6fa899
JJ		 174712013-05-13 Jakub Jelinek <jakub@redhat.com>
17472
17473 PR tree-optimization/45216
17474 PR tree-optimization/57157
17475 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17476 the (-Y) & (B - 1) variant if OP is |.
17477 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17478 return shifted. Use (-op1) & (prec - 1) as other_amount
17479 instead of prec - op1.
17480
4502fe8d
MJ		 174812013-05-13 Martin Jambor <mjambor@suse.cz>
17482
17483 PR middle-end/42371
17484 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17485 (ipa_constant_data): New type.
17486 (ipa_jump_func): Use ipa_constant_data to hold information about
17487 constant jump functions.
17488 (ipa_get_jf_constant): Adjust to jump function type changes.
17489 (ipa_get_jf_constant_rdesc): New function.
17490 (ipa_param_descriptor): New field controlled_uses.
17491 (ipa_get_controlled_uses): New function.
17492 (ipa_set_controlled_uses): Likewise.
17493 * ipa-ref.h (ipa_find_reference): Declare.
17494 * ipa-prop.c (ipa_cst_ref_desc): New type.
17495 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17496 changes.
17497 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17498 New parameter cs. Adjust all callers.
17499 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17500 (remove_described_reference): New function.
17501 (jfunc_rdesc_usable): Likewise.
17502 (try_make_edge_direct_simple_call): Decrement controlled use count,
17503 attempt to remove reference if it hits zero.
17504 (combine_controlled_uses_counters): New function.
17505 (propagate_controlled_uses): Likewise.
17506 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17507 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17508 (ipa_print_node_params): Print described use counter.
17509 (ipa_write_jump_function): Adjust to jump function type changes.
17510 (ipa_read_jump_function): New parameter CS, pass it to
17511 ipa_set_jf_constant. Adjust caller.
17512 (ipa_write_node_info): Stream controlled use count
17513 (ipa_read_node_info): Likewise.
17514 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17515 asserting.
17516 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17517 count. Remove cloning-added reference if it reaches zero.
17518 * ipa-ref.c (ipa_find_reference): New function.
17519
0864bfc2
GG		 175202013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17521
e7208ea3 17522 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG		 17523 alignment values for AMD BD and BT architectures.
17524
640bfeb2
MG		 175252013-05-13 Marc Glisse <marc.glisse@inria.fr>
17526
17527 * tree-vect-generic.c (uniform_vector_p): Move ...
17528 * tree.c (uniform_vector_p): ... here.
17529 * tree.h (uniform_vector_p): Declare it.
17530 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17531 into a scalar.
17532
3a60f32b
JJ		 175332013-05-13 Jakub Jelinek <jakub@redhat.com>
17534
198fe1bf
JJ		 17535 PR tree-optimization/57230
17536 * tree-ssa-strlen.c (handle_char_store): Record length for
17537 array store from STRING_CST.
17538
3a60f32b
JJ		 17539 PR tree-optimization/57230
17540 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17541 check.
17542
566be57c
JR		 175432013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17544
17545 * config/epiphany/epiphany.c (epiphany_init): Check size of
17546 NUM_MODES_FOR_MODE_SWITCHING.
17547 (epiphany_expand_prologue):
17548 Remove CONFIG_REGNUM initial value handling code.
17549 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17550 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 17551 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR		 17552 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17553 Don't return 1 for FP_MODE_NONE.
17554 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17555 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17556 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17557 * config/epiphany/epiphany.md (save_config): New pattern.
17558
0f2c2331
UB		 175592013-05-12 Uros Bizjak <ubizjak@gmail.com>
17560
17561 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17562
5b3f93c7
UB		 175632013-05-10 Uros Bizjak <ubizjak@gmail.com>
17564
17565 * config/i386/i386.md (memory): Handle sseishft1.
17566 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17567 (*vec_extractv2di_1): Ditto.
17568
1f873f0e
VM		 175692013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17570
17571 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17572 saved registers.
17573
341427fa 175742013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH		 17575
17576 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17577 Add mthumb/march=armv7-a multilib.
17578 Add mthumb/march=armv7-r multilib.
17579 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17580
9e69bdde
RC		 175812013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17582
17583 * config/v850/t-rtems: Add more multilibs.
17584
9ff09a22
RB		 175852013-05-10 Richard Biener <rguenther@suse.de>
17586
17587 PR tree-optimization/57214
17588 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17589 not propagate from SSA names that occur in abnormal PHI nodes.
17590
a5e0cd1d
MG		 175912013-05-10 Marc Glisse <marc.glisse@inria.fr>
17592
17593 * stor-layout.c (element_precision): New function.
17594 * machmode.h (element_precision): Declare it.
17595 * tree.c (build_minus_one_cst): New function.
17596 (element_precision): Likewise.
17597 * tree.h (build_minus_one_cst): Declare new function.
17598 (element_precision): Likewise.
17599 * fold-const.c (operand_equal_p): Use element_precision.
17600 (fold_binary_loc): Handle vector types.
17601 * convert.c (convert_to_integer): Use element_precision.
17602 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17603 separately.
17604
cb2558bc
RS		 176052013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17606
17607 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17608 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17609 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17610 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17611 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17612 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17613 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17614 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17615 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17616 (Uuw8): New constraints.
17617 (Usb4): Move into alphabetical order.
17618 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17619 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17620 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17621 previously unnamed patterns.
17622 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17623 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17624 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17625 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17626 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17627 of specific lengths.
17628
cb3b8d33
JJ		 176292013-05-10 Jakub Jelinek <jakub@redhat.com>
17630
6f93c008
JJ		 17631 * config/i386/i386.md (rotateinv): New code attr.
17632 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17633 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17634 roll $31, %eax, etc.
17635
cb3b8d33
JJ		 17636 PR tree-optimization/45216
17637 PR tree-optimization/57157
17638 * tree-ssa-forwprop.c (simplify_rotate): New function.
17639 (ssa_forward_propagate_and_combine): Call it.
17640
afb119be
RB		 176412013-05-10 Richard Biener <rguenther@suse.de>
17642
17643 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17644 disable peeling when we version for aliasing.
17645 (vector_alignment_reachable_p): Honor explicit user alignment.
17646 (vect_supportable_dr_alignment): Likewise.
17647 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17648 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17649 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17650 then peeling to arrange for the cost-model check to come first.
17651
01ae4861
AM		 176522013-05-10 Alan Modra <amodra@gmail.com>
17653
17654 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17655 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17656 * configure: Regenerate.
17657
ebc9a431
AM		 176582013-05-10 Alan Modra <amodra@gmail.com>
17659
17660 PR target/55033
17661 * varasm.c (default_elf_select_section): Move !DECL_P check..
17662 (get_named_section): ..to here before calling get_section_name.
17663 Adjust assertion.
17664 (default_section_type_flags): Add DECL_P check.
17665 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17666 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17667
d4bca93c
JR		 176682013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17669
17670 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17671 When using gen_stack_adjust_str with a register offset, add a
17672 REG_FRAME_RELATED_EXPR note.
17673
60ca9a65
UB		 176742013-05-09 Uros Bizjak <ubizjak@gmail.com>
17675
17676 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17677 (*vec_extractv4si_zext_mem): Ditto.
17678 (*vec_extractv2di): Add 0->x and x->x alternatives.
17679 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17680 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17681
a3409c02
JM		 176822013-05-09 Jason Merrill <jason@redhat.com>
17683
0138d6b2
JM		 17684 N3639 C++1y VLA support
17685 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17686 DECL_VALUE_EXPR.
17687
a3409c02
JM		 17688 * tree.c (build_constructor_va): New.
17689 * tree.h: Declare it.
17690
66e6b990
MJ		 176912013-05-09 Martin Jambor <mjambor@suse.cz>
17692
17693 PR lto/57084
17694 * gimple-fold.c (canonicalize_constructor_val): Call
17695 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17696
64cfa6c0
JH		 176972013-05-09 Jan Hubicka <jh@suse.cz>
17698 Richard Biener <rguenther@suse.de>
17699
17700 PR lto/54095
17701 * symtab.c (symtab_make_decl_local): Do not add private names.
17702
c3167b00
JH		 177032013-05-09 Jan Hubicka <jh@suse.cz>
17704
17705 PR lto/54095
17706 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17707 (unlink_from_assembler_name_hash): Likewise.
17708 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17709 symtab_unregister_node, symtab_initialize_asm_name_hash,
17710 change_decl_assembler_name): Update.
17711
12dc6974
SN		 177122013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17713
17714 * config/aarch64/aarch64.md: New movtf split.
17715 (*movtf_aarch64): Update.
17716 (aarch64_movdi_tilow): Handle TF modes and rename to
17717 aarch64_movdi_<mode>low.
17718 (aarch64_movdi_tihigh): Handle TF modes and rename to
17719 aarch64_movdi_<mode>high
17720 (aarch64_movtihigh_di): Handle TF modes and rename to
17721 aarch64_mov<mode>high_di
17722 (aarch64_movtilow_di): Handle TF modes and rename to
17723 aarch64_mov<mode>low_di
17724 (aarch64_movtilow_tilow): Remove spurious whitespace.
17725 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17726 splits.
17727 (aarch64_print_operand): Update.
17728
227eb343
AM		 177292013-05-09 Alan Modra <amodra@gmail.com>
17730
17731 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17732 powerpc64le.
17733 * configure: Regenerate.
17734
0b013847
UB		 177352013-05-08 Uros Bizjak <ubizjak@gmail.com>
17736
17737 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17738 splitter preparation statements.
17739 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17740 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17741 adjust_address_nv.
17742
1dc3d6e9
BS		 177432013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17744
17745 * gimple-ssa-strength-reduction.c (count_candidates): Change
17746 return value to int.
17747 (analyze_candidates_and_replace): Change type of length to int.
17748
e61e7d28
UB		 177492013-05-08 Uros Bizjak <ubizjak@gmail.com>
17750
17751 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17752 (*vec_extract<mode>): Use VI12_128 mode iterator.
17753 (*vec_extract<mode>_mem): Ditto.
17754 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17755 attribute.
17756
4fbfcf44
DN		 177572013-05-08 Diego Novillo <dnovillo@google.com>
17758
17759 PR bootstrap/54659
17760
17761 Revert:
4fbfcf44
DN		 17762 2012-08-17 Diego Novillo <dnovillo@google.com>
17763
e61e7d28
UB		 17764 PR bootstrap/54281
17765 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17766 * config.in: Regenerate.
17767 * configure: Regenerate.
17768 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 17769
702d8703
JH		 177702013-05-08 Jan Hubicka <jh@suse.cz>
17771
17772 PR lto/54095
17773 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17774 * cgraph.h (symtab_node_base): Add unique_name.
17775 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17776 input_overwrite_node, input_varpool_node): Stream unique_name.
17777 * cgraphclones.c (cgraph_create_virtual_clone,
17778 cgraph_function_versioning): Set unique_name.
17779 * ipa.c (function_and_variable_visibility): Set unique_name.
17780
8b28cf47
BS		 177812013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17782
17783 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17784 (alloc_cand_and_find_basis): Restrict conditional candidate
17785 processing to CAND_MULTs.
17786
e86074fd
JH		 177872013-05-08 Jan Hubicka <jh@suse.cz>
17788
17789 PR lto/54095
17790 lto-symtab.c (lto_symtab_symbol_p): New function.
17791 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17792 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17793 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17794 Skip static symbols.
17795
44398cbe
PC		 177962013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17797
17798 PR tree-optimization/57200
17799 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17800 Only call inform if the preceding warning_at returns true.
17801
f6bc1c4a
HS		 178022013-05-07 Han Shen <shenhan@google.com>
17803
17804 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17805 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17806 * common.opt (fstack-protector-strong): New option.
17807 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17808 * doc/invoke.texi (Optimization Options): Document
17809 "-fstack-protector-strong".
17810 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17811
4ffecb1f
SB		 178122013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17813
17814 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17815
5a107a0f
VM		 178162013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17817
17818 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17819 it is not lhs of the insn.
17820 (setup_reg_equiv): Remove insn having equiv memory which it is not
17821 lhs of the insn.
17822 * lra-constraints.c (process_address): Try to improve generation
17823 code for address base + disp.
17824 (lra_constraints): Make correct the code for checking insn setting
17825 up backward equivalence. Remove insn only if it is in the init
17826 insn list.
17827 * lra-eliminations.c (update_reg_eliminate): Change return value.
17828 (lra_eliminate): Use the result.
17829
3f5783ea
UB		 178302013-05-07 Uros Bizjak <ubizjak@gmail.com>
17831
17832 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17833 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17834 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17835 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17836 register target operands.
17837 (*vec_extractv8hi_sse2): New pattern.
17838 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17839 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17840 (*vec_extract<mode>_mem): New insn and split pattern.
17841
8a5800b8
CL		 178422013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17843
17844 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17845 (TARGET_ASAN_SHADOW_OFFSET): Define.
17846 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17847 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17848
7bf55a70
BS		 178492013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17850
17851 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17852 (incr_vec_index): Return -1 if increment not found.
17853 (create_add_on_incoming_edge): Assert if increment not found.
17854 (record_increment): Limit number of increments recorded.
17855 (all_phi_incrs_profitable): Return false if an increment not found.
17856 (replace_profitable_candidates): Don't process increments that were
17857 not recorded.
17858 (analyze_candidates_and_replace): Limit size of incr_vec.
17859
3f8825c0
RB		 178602013-05-07 Richard Biener <rguenther@suse.de>
17861
17862 * calls.c (special_function_p): setjmp-like functions are leaf.
17863 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17864 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17865
fd4842cd
SN		 178662013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17867
17868 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17869 (aarch64_simd_mov<mode>): New expander.
17870 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17871 (aarch64_simd_mov_to_<mode>high): Likewise.
17872 (aarch64_simd_mov_from_<mode>low): Likewise.
17873 (aarch64_simd_mov_from_<mode>high): Likewise.
17874 (aarch64_dup_lane<mode>): Update.
17875 (aarch64_dup_lanedi): New instruction pattern.
17876 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17877 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17878
a7a7d10e
BS		 178792013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17880
17881 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17882 (replace_mult_candidate): Remove unnecessary argument; remove
17883 unnecessary parameter from call to introduce_cast_before_cand.
17884 (replace_unconditional_candidate): Remove unnecessary parameter
17885 from call to replace_mult_candidate.
17886 (replace_conditional_candidate): Likewise.
17887 (insert_initializers): Use make_temp_ssa_name.
17888 (introduce_cast_before_cand): Remove unnecessary argument; use
17889 make_temp_ssa_name.
17890 (replace_one_candidate): Remove unnecessary argument; remove
17891 unnecessary parameter from calls to introduce_cast_before_cand.
17892 (replace_profitable_candidates): Remove unnecessary parameters
17893 from calls to replace_one_candidate.
17894
29105868
BS		 178952013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17896
17897 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17898 phi def as possibly hiding a basis for a CAND_ADD whose operands
17899 have been commuted in the analysis.
17900 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17901
4095f9fa
N		 179022013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17903
17904 * config/aarch64/aarch64.md
17905 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17906 shift value between 0-4.
17907
39e843e8
RB		 179082013-05-07 Richard Biener <rguenther@suse.de>
17909
17910 * double-int.h (rshift): New overload.
17911 * double-int.c (rshift): New function.
17912 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17913 (create_reference_ops_from_ref): Remove.
17914 (vn_reference_insert): Use shared ops for constructing the
17915 reference and copy it.
17916
0a1a83cb
RB		 179172013-05-07 Richard Biener <rguenther@suse.de>
17918
17919 PR middle-end/57190
17920 * tree-eh.c (sink_clobbers): Properly propagate
17921 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17922
43303d6f
JJ		 179232013-05-07 Jakub Jelinek <jakub@redhat.com>
17924
ba7e83f8
JJ		 17925 PR tree-optimization/57149
17926 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17927 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17928 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17929 uninit_undefined_value_p instead of ssa_undefined_value_p.
17930
43303d6f
JJ		 17931 PR debug/57184
17932 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17933 for modifier == EXPAND_INITIALIZER.
17934
14523c25
AB		 179352013-05-07 Anton Blanchard <anton@samba.org>
17936
17937 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17938 for powerpc64 little endian.
17939 * configure: Regenerate.
17940
cb7c8be9
GS		 179412013-05-06 Graham Stott <grahams@btinternet.com>
17942
17943 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17944 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17945 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17946 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17947
179482013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS		 17949
17950 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17951 codes which allow non-lvalues.
17952
9a0ee7b0
MG		 179532013-05-06 Marc Glisse <marc.glisse@inria.fr>
17954
17955 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17956 components are all 1s.
17957 (integer_minus_onep): New function.
17958 * tree.h (integer_minus_onep): Declare it.
17959 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17960 integer_minus_onep instead of integer_all_onesp.
17961
f2c17ea9
OE		 179622013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17963
17964 PR target/52933
17965 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17966 variations of these patterns.
17967
f75e6a51
UB		 179682013-05-06 Uros Bizjak <ubizjak@gmail.com>
17969
17970 * config/i386/i386.md (isa): Add x64_sse4 member.
17971 (enabled): Handle x64_sse4.
17972 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17973 instruction for 64bit SSE4_1 targets. Update insn attributes.
17974 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17975 instruction for SSE4_1 targets. Update insn attributes.
17976 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17977 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17978 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17979 const_1 selector.
17980 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17981 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17982 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17983
a986d468
OE		 179842013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17985
17986 PR target/57108
17987 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17988
382522cb
MK		 179892013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17990
17991 * final.c (do_assembler_dialects): Don't handle curly braces and
17992 vertical bar escaped by % as dialect delimiters.
17993 (output_asm_insn): Print curly braces and vertical bar if escaped
17994 by % and ASSEMBLER_DIALECT defined.
17995 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17996 * doc/tm.texi: Regenerated.
17997
fb0d5c60
SB		 179982013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17999
fb0d5c60
SB		 18000 * config/mips/mips.c: Include tree-pass.h.
18001 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18002 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18003 (pass_mips_machine_reorg2): New machine specific pass.
18004 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18005 (mips_option_override): Register the new pass.
18006 * rtl.h (cleanup_barriers): Remove prototype.
18007 (dbr_schedule): Likewise.
18008 * jump.c (cleanup_barriers): Make static.
18009 * reorg.c (dbr_schedule): Likewise.
18010
aa06a978
RB		 180112013-05-06 Richard Biener <rguenther@suse.de>
18012
18013 PR tree-optimization/57185
18014 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18015 SSA names properly.
18016
0b953bec
UB		 180172013-05-06 Uros Bizjak <ubizjak@gmail.com>
18018
18019 PR target/57106
18020 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18021 between "first_arg" and "insn", not "last" and "insn".
18022
28708525
WS		 180232013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18024
18025 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18026 (find_candidates_in_block): Re-enable slsr_process_phi.
18027 (create_phi_basis): Fix double counting of candidate adjustment.
18028
0107dca2
RB		 180292013-05-06 Richard Biener <rguenther@suse.de>
18030
18031 PR middle-end/57147
18032 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18033 the edge is also fallthru, preserve it and just clear the
18034 abnormal flag.
18035 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18036 also complex, preserve that and just clear the fallthru flag.
18037 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18038 update virtual operands.
18039
470d4d13
AM		 180402013-05-06 Alan Modra <amodra@gmail.com>
18041
18042 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18043 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18044 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18045 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18046 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18047 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18048
8f1dbf8d
AM		 180492013-05-06 Alan Modra <amodra@gmail.com>
18050
18051 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18052 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18053 (DEFAULT_ASM_ENDIAN): Define.
18054 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18055 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18056 Update -K PIC clause from sysv4.h.
18057 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18058 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18059
54890767
AM		 180602013-05-06 Alan Modra <amodra@gmail.com>
18061
18062 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18063 twice for little-endian.
18064 (ashrdi3_no_power, ashrdi3): Support little-endian.
18065
2353515d
OE		 180662013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18067
18068 PR target/55303
18069 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18070 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18071 related expanders.
18072 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18073 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18074 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18075 New predicates.
18076
33e67557
SB		 180772013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18078 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18079
18080 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18081 * config/pa/pa.opt: Make mbig-switch a no-op.
18082 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18083 (CASE_VECTOR_MODE): Always return SImode.
18084 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18085 for the !TARGET_BIG_SWITCH case.
18086 * config/pa/pa-linux.h: Likewise.
18087 * config/pa/pa-openbsd.h: Likewise.
18088 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18089 * config/pa/pa.md (short_jump): Remove define_insn.
18090 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18091 (casesi0): Remove define_insn.
18092 (type): Remove btable_branch.
18093 (pa_combine_type): Likewise.
18094 (in_nullified_branch_delay): Likewise.
18095 (in_call_delay): Likewise.
18096 (define_delay): Likewise.
18097 (define_insn_reservation "Z3"): Likewise.
18098 (define_insn_reservation "Z4"): Likewise.
18099 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18100 (pa_adjust_insn_length): Remove adjustment for btable branches.
18101 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18102 and mno-big-switch
18103
3095685e
UB		 181042013-05-05 Uros Bizjak <ubizjak@gmail.com>
18105
18106 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18107 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18108 Add m->r,x alternatives.
18109 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18110 splitters using SWI48x mode iterator.
18111 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18112 TARGET_64BIT. Add m->x alternative.
18113 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18114 Add o->x alternative. Enable for TARGET_SSE.
18115 (sse_storeq): Remove expander.
18116 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18117 with memory input operand.
18118 (*vec_extractv2di_1 splitter): New.
18119 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18120 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18121
4b36ae28
SB		 181222013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18123
18124 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18125 (INT_LOWPART): Delete.
18126 (extract_MB): Adjust.
18127 (extract_ME): Adjust.
18128 (print_operand): Adjust.
18129
da226db2
SB		 181302013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18131
18132 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18133 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18134 (reg_or_logical_cint_operand, easy_fp_constant,
18135 logical_const_operand): Delete "CONST_DOUBLE" case.
18136 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18137 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 18138 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB		 18139 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18140 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18141 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18142 test.
18143 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18144 CONST_DOUBLE DImode/VOIDmode case.
18145 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18146 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18147 CONST_DOUBLE VOIDmode case.
18148 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18149 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18150 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18151 Delete CONST_DOUBLE case.
18152 (splitters for mov FMOVE64 const_double): Delete
18153 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18154 "HOST_BITS_PER_WIDE_INT >= 64" test.
18155 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18156 case.
18157 (mov DI const_double): Delete.
18158
40de22d6
JJ		 181592013-05-04 Jakub Jelinek <jakub@redhat.com>
18160
18161 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18162 on op shows all bits zero in mode of a lowpart subreg, return zero.
18163
5ec6aff2
MM		 181642013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18165
18166 PR target/57150
18167 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18168 to save TFmode registers and DImode to save TImode registers for
18169 caller save operations.
18170 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18171 mark being partially clobbered since they only use the first
18172 double word.
18173
18174 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18175 and TDmode only use the upper 64-bits of each VSX register.
18176
2cefad90
BS		 181772013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18178
18179 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18180 (find_candidates_in_block): Disable slsr_process_phi.
18181
d6d7eee1
GW		 181822013-05-03 Guozhi Wei <carrot@google.com>
18183
18184 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB		 18185 constructor to ...
18186 (build_init_ctor): ... here.
d6d7eee1 18187
9b92d12b
BS		 181882013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18189
18190 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18191 (slsr_cand_d): Redefine def_phi.
18192 (stride_status, phi_adjust_status, count_phis_status): New enums.
18193 (find_phi_def): New.
18194 (find_basis_for_base_expr): New.
18195 (find_basis_for_candidate): Handle hidden bases.
18196 (alloc_cand_and_find_basis): Handle phi candidates.
18197 (slsr_process_phi): New.
18198 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18199 (create_mul_imm_cand): Likewise.
18200 (create_add_ssa_cand): Exclude phi base candidates.
18201 (create_add_imm_cand): Likewise.
18202 (slsr_process_cast): Likewise.
18203 (slsr_process_copy): Likewise.
18204 (find_candidates_in_block): Handle phi candidates.
18205 (dump_candidate): Likewise.
18206 (unconditional_cands): Delete.
18207 (unconditional_cands_with_known_stride_p): Delete.
18208 (phi_dependent_cand_p): New.
18209 (cand_increment): Handle phi-dependent candidates.
18210 (replace_dependent): Delete.
18211 (replace_mult_candidate): New.
18212 (replace_unconditional_candidate): New.
18213 (incr_vec_index): Move to avoid forward reference.
18214 (create_add_on_incoming_edge): New.
18215 (create_phi_basis): New.
18216 (replace_dependents): Delete.
18217 (replace_conditional_candidate): New.
18218 (phi_add_costs): New.
18219 (replace_uncond_cands_and_profitable_phis): New.
18220 (record_increment): Handle phi adjustments.
18221 (record_phi_increments): New.
18222 (record_increments): Handle phi adjustments.
18223 (phi_incr_cost): New.
18224 (lowest_cost_path): Handle phis.
18225 (total_savings): Likewise.
18226 (analyze_increments): Likewise.
18227 (ncd_with_phi): New.
18228 (ncd_of_cand_and_phis): New.
18229 (nearest_common_dominator_for_cands): Handle phi increments.
18230 (all_phi_incrs_profitable): New.
18231 (replace_profitable_candidates): Handle phi-dependent candidates.
18232 (analyze_candidates_and_replace): Likewise.
18233
68f073d4
TJ		 182342013-05-03 Teresa Johnson <tejohnson@google.com>
18235
18236 PR bootstrap/57154
18237 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18238 do not exceed REG_BR_PROB_BASE.
18239
a4ee7cb9
JL		 182402013-05-03 Jeff Law <law@redhat.com>
18241
ade67f70 18242 PR tree-optimization/57144
a4ee7cb9
JL		 18243 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18244 operand of the condition will bit into the new type when eliminating
18245 a cast feeding a condition.
18246
47954c4e
JJ		 182472013-05-03 Jakub Jelinek <jakub@redhat.com>
18248
18249 PR rtl-optimization/57130
3095685e
UB		 18250 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18251 of COMPARE as in_code to the recursive call if needed.
47954c4e 18252
3c21604f
UB		 182532013-05-03 Uros Bizjak <ubizjak@gmail.com>
18254
18255 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18256 (enabled): Handle new members.
18257 * config/i386/sse.md (*vec_concatv2si): Merge from
18258 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18259 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18260
12211b99 182612013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR		 18262
18263 PR tree-optimization/57027
18264 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18265 for fnms opportunity, check we got the prerequisite kind
18266 of tree / gimple before using accessor functions.
18267
07bfc9ec
RB		 182682013-05-03 Richard Biener <rguenther@suse.de>
18269
18270 * double-int.h (lshift): New overload without precision
18271 and arith argument.
18272 (operator *=, operator +=, operator -=): Move ...
18273 * double-int.c (operator *=, operator +=, operator -=): ... here
18274 and implement more efficiently.
18275 (mul_double_with_sign): Remove.
18276 (lshift_double): Adjust to take unsinged shift argument, push
18277 dispatching code to callers.
18278 (mul_double_wide_with_sign): Add early out for callers that
18279 are not interested in high parts or overflow.
18280 (lshift): New function.
18281 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18282 dispatch code here.
18283 (lrotate, rrotate): Use logical shifts.
18284 * expr.c (get_inner_reference): Use lshift.
18285 * fixed-value.c (do_fixed_divide): Likewise.
18286 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18287 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18288 (indirect_refs_may_alias_p): Likewise.
18289 (stmt_kills_ref_p_1): Likewise.
18290
7769bb64
VP		 182912013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18292
18293 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18294
1d0c8e5c
VP		 182952013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18296
18297 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18298 scalar form of FABD instruction.
18299
f15feaf9
VM		 183002013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18301
18302 * lra-constraints.c (process_alt_operands): Add checking alt
18303 number to choose the best alternative.
18304
d90e76d4
RB		 183052013-05-02 Richard Biener <rguenther@suse.de>
18306
18307 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18308 bitmap and its handling.
18309 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18310
9f8e7a96
RB		 183112013-05-02 Richard Biener <rguenther@suse.de>
18312
18313 PR middle-end/57140
18314 * tree-inline.c (copy_loops): Properly handle removed loops.
18315 (copy_cfg_body): Mark destination loops for fixup if source
18316 loops needed fixup.
18317
f3a81b39
GY		 183182013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18319
18320 PR target/56732
18321 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18322 generating simple_return for naked functions.
18323
7b920a9a
MJ		 183242013-05-02 Martin Jambor <mjambor@suse.cz>
18325
18326 PR middle-end/56988
18327 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18328 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18329 flags match.
18330 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18331 ipa_agg_replacement_value structures.
18332 (known_aggs_to_agg_replacement_list): Likewise.
18333 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18334 (read_agg_replacement_chain): Likewise.
18335 (ipcp_transform_function): Also check that by_ref flags match.
18336
2c41c19d
RB		 183372013-05-02 Richard Biener <rguenther@suse.de>
18338
18339 * graphds.h (struct graph): Add obstack member.
18340 * graphds.c (new_graph): Initialize obstack and allocate
18341 vertices from it.
18342 (add_edge): Allocate edge from the obstack.
3c21604f 18343 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 18344
8b47039c
TJ		 183452013-05-02 Teresa Johnson <tejohnson@google.com>
18346
18347 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18348 divides.
18349 * cfg.c (update_bb_profile_for_threading): Ditto.
18350 * tree-inline.c (copy_bb): Ditto.
18351 (copy_edges_for_bb): Ditto.
18352 (initialize_cfun): Ditto.
18353 (copy_cfg_body): Ditto.
18354 (expand_call_inline): Ditto.
18355 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18356 (estimate_node_size_and_time): Ditto.
18357 (inline_merge_summary): Ditto.
18358 * cgraphclones.c (cgraph_clone_edge): Ditto.
18359 (cgraph_clone_node): Ditto.
18360 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18361 (compute_trg_info): Ditto.
18362
da65928c
IB		 183632013-05-02 Ian Bolton <ian.bolton@arm.com>
18364
18365 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18366 S reg when fp attribute set.
18367 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18368
add3c965
IB		 183692013-05-02 Ian Bolton <ian.bolton@arm.com>
18370
18371 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18372 New pattern.
18373 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18374 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18375 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18376
fdd43ac4
RB		 183772013-05-02 Richard Biener <rguenther@suse.de>
18378
18379 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18380 (struct instantiate_cache_entry): New type.
18381 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18382 (struct instantiate_cache_type): New type.
18383 (set_instantiated_value, get_instantiated_value): Remove.
18384 (get_instantiated_value_entry): New function.
18385 (instantiate_scev_name): Use the new cache and adjust.
18386 (instantiate_scev_poly): Adjust.
18387 (instantiate_scev_binary): Likewise.
18388 (instantiate_array_ref): Likewise.
18389 (instantiate_scev_convert): Likewise.
18390 (instantiate_scev_not): Likewise.
18391 (instantiate_scev_3): Likewise.
18392 (instantiate_scev_2): Likewise.
18393 (instantiate_scev_r): Likewise.
18394 (instantiate_scev): Likewise.
18395 (resolve_mixers): Likewise.
18396
36ff9dfb
VM		 183972013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18398
18399 PR target/57091
18400 * lra-constraints.c (best_small_class_operands_num): Remove.
18401 (process_alt_operands): Remove small_class_operands_num. Take
18402 small classes operands into losers and only if the operand is not
18403 matched. Modify debugging output.
18404 (curr_insn_transform): Remove best_small_class_operands_num.
18405 Print insn name.
18406
36054fab
JG		 184072013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18408
18409 * config/aarch64/aarch64-builtins.c
18410 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18411 * config/aarch64/aarch64-simd-builtins.def
18412 (reduc_splus_): Add new modes.
18413 (reduc_uplus_): New.
18414 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18415 (reduc_uplus_v4sf): Likewise.
18416 (reduc_splus_v4sf): Likewise.
18417 (aarch64_addv<mode>): Likewise.
18418 (reduc_uplus_<mode>): Likewise.
18419 (reduc_splus_<mode>): Likewise.
18420 (aarch64_addvv2di): Likewise.
18421 (reduc_uplus_v2di): Likewise.
18422 (reduc_splus_v2di): Likewise.
18423 (aarch64_addvv2si): Likewise.
18424 (reduc_uplus_v2si): Likewise.
18425 (reduc_splus_v2si): Likewise.
18426 (reduc_<sur>plus_<mode>): New.
18427 (reduc_<sur>plus_v2di): Likewise.
18428 (reduc_<sur>plus_v2si): Likewise.
18429 (reduc_<sur>plus_v4sf): Likewise.
18430 (aarch64_addpv4sf): Likewise.
18431 * config/aarch64/arm_neon.h
18432 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18433 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18434 add UNSPEC_SADDV, UNSPEC_UADDV.
18435 (SUADDV): New.
18436 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18437
6dce23a8
JG		 184382013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18439
18440 * config/aarch64/arm_neon.h
18441 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18442
1598945b
JG		 184432013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18444
18445 * config/aarch64/aarch64-builtins
18446 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18447
998eaf97
JG		 184482013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18449
18450 * config/aarch64/aarch64-simd-builtins.def
18451 (reduc_smax_): New.
18452 (reduc_smin_): Likewise.
18453 (reduc_umax_): Likewise.
18454 (reduc_umin_): Likewise.
18455 (reduc_smax_nan_): Likewise.
18456 (reduc_smin_nan_): Likewise.
18457 (fmax): Remove.
18458 (fmin): Likewise.
18459 (smax): Update for V2SF, V4SF and V2DF modes.
18460 (smin): Likewise.
18461 (smax_nan): New.
18462 (smin_nan): Likewise.
18463 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18464 (<su><maxmin><mode>3): ...This, refactor.
18465 (s<maxmin><mode>3): New.
18466 (<maxmin_uns><mode>3): Likewise.
18467 (reduc_<maxmin_uns>_<mode>): Refactor.
18468 (reduc_<maxmin_uns>_v4sf): Likewise.
18469 (reduc_<maxmin_uns>_v2si): Likewise.
18470 (aarch64_<fmaxmin><mode>: Remove.
18471 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18472 new builtin names.
18473 (vmin<q>_f<32,64>): Likewise.
18474 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18475 (FMAXMIN): New.
18476 (su): Add mappings for smax, smin, umax, umin.
18477 (maxmin): New.
18478 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18479 (FMAXMIN): Rename as...
18480 (FMAXMIN_UNS): ...This.
18481 (maxminv): Remove.
18482 (fmaxminv): Likewise.
18483 (fmaxmin): Likewise.
18484 (maxmin_uns): New.
18485 (maxmin_uns_op): Likewise.
18486
bd11644e
JG		 184872013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18488
18489 * config/aarch64/arm_neon.h
18490 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18491 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18492 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18493
75dd5ace
JG		 184942013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18495
18496 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18497 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18498
7c19979f
JG		 184992013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18500
18501 * config/aarch64/aarch64-simd.md
18502 (vcond<mode>_internal): Handle special cases for constant masks.
18503 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18504 (vcondu<mode><mode>): Likewise.
18505 (vcond<v_cmp_result><mode>): New.
18506
bb60efd9
JG		 185072013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18508
18509 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18510 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18511 * config/aarch64/aarch64-simd-builtins.def
18512 (cmeq): Update to BUILTIN_VALLDI.
18513 (cmgt): Likewise.
18514 (cmge): Likewise.
18515 (cmle): Likewise.
18516 (cmlt): Likewise.
18517 * config/aarch64/arm_neon.h
18518 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18519 to builtins or C as appropriate.
18520
889b9412
JG		 185212013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18522
18523 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18524 (cmgeu): ...This.
18525 (cmhi): Rename to...
18526 (cmgtu): ...This.
18527 * config/aarch64/aarch64-simd.md
18528 (simd_mode): Add SF.
18529 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18530 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18531 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18532 (cstore<mode>_neg): ...This.
18533 * config/aarch64/iterators.md
18534 (VALLF): new.
18535 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18536 (COMPARISONS): New.
18537 (UCOMPARISONS): Likewise.
18538 (optab): Add missing comparisons.
18539 (n_optab): New.
18540 (cmp_1): Likewise.
18541 (cmp_2): Likewise.
18542 (CMP): Likewise.
18543 (cmp): Remove.
18544 (VCMP_S): Likewise.
18545 (VCMP_U): Likewise.
18546 (V_cmp_result): Add DF, SF modes.
18547 (v_cmp_result): Likewise.
18548 (v): Likewise.
18549 (vmtype): Likewise.
18550 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18551
0a7dbb76
GY		 185522013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18553
18554 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18555 define_insn to define_insn_and_split.
18556 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18557 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18558 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18559 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18560 (thumb2_negscc): Likewise.
18561
fb614ca6
GY		 185622013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18563
18564 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18565
9e64a0bf
GY		 185662013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18567
18568 * config/arm/thumb2.md: Remove trailing whitespaces.
18569
d6b28156
RS		 185702013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18571
18572 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18573 Use gen_int_mode rather than GEN_INT.
18574
f91674c3
L		 185752013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18576
b0dec607 18577 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L		 18578 debug_gimple_stmt.
18579
3551257c
RB		 185802013-04-30 Richard Biener <rguenther@suse.de>
18581
18582 PR middle-end/57122
3c21604f 18583 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 18584
4c1aff1c
RB		 185852013-04-30 Richard Biener <rguenther@suse.de>
18586
18587 PR middle-end/57107
18588 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18589
54b8379a
AB		 185902013-04-30 Andrey Belevantsev <abel@ispras.ru>
18591
18592 PR rtl-optimization/56957
18593 PR rtl-optimization/57105
54b8379a
AB		 18594 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18595 variable. Use just INSN_UID for determining whether an insn
18596 should be only disconnected from the insn stream.
18597 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18598
abc27962
JJ		 185992013-04-30 Jakub Jelinek <jakub@redhat.com>
18600
18601 PR tree-optimization/57104
18602 * tsan.c (instrument_expr): Don't instrument accesses to
18603 DECL_HARD_REGISTER VAR_DECLs.
18604
0fc822d0
RB		 186052013-04-30 Richard Biener <rguenther@suse.de>
18606
18607 * function.h (loops_for_fn): New inline function.
18608 (set_loops_for_fn): Likewise.
18609 * cfgloop.h (place_new_loop): Add struct function parameter.
18610 (get_loop): Likewise.
18611 (get_loops): Likewise.
18612 (number_of_loops): Likewise.
18613 (fel_next): Adjust.
18614 (fel_init): Likewise.
18615 * cfg.c (get_loop_copy): Adjust.
18616 * cfgloop.c (flow_loops_dump): Likewise.
18617 (record_loop_exits): Likewise.
18618 (verify_loop_structure): Likewise.
18619 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18620 (estimate_reg_pressure_cost): Likewise.
18621 (mark_loop_exit_edges): Likewise.
18622 * cfgloopmanip.c (place_new_loop): Likewise.
18623 (add_loop): Likewise.
18624 (duplicate_loop): Likewise.
18625 * graph.c (draw_cfg_nodes): Likewise.
18626 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18627 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18628 (extract_affine_chrec): Likewise.
18629 (build_scop_iteration_domain): Likewise.
18630 * graphite.c (graphite_initialize): Likewise.
18631 * ira-build.c (create_loop_tree_nodes): Likewise.
18632 (more_one_region_p): Likewise.
18633 (rebuild_regno_allocno_maps): Likewise.
18634 (mark_loops_for_removal): Likewise.
18635 (mark_all_loops_for_removal): Likewise.
18636 (remove_unnecessary_regions): Likewise.
18637 (ira_build): Likewise.
18638 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18639 * loop-init.c (fix_loop_structure): Likewise.
18640 (gate_rtl_move_loop_invariants): Likewise.
18641 (gate_rtl_unswitch): Likewise.
18642 (gate_rtl_unroll_and_peel_loops): Likewise.
18643 (rtl_doloop): Likewise.
18644 * lto-streamer-in.c (input_cfg): Likewise.
18645 * lto-streamer-out.c (output_cfg): Likewise.
18646 * modulo-sched.c (sms_schedule): Likewise.
18647 * predict.c (tree_estimate_probability): Likewise.
18648 (tree_estimate_probability_driver): Likewise.
18649 (estimate_loops): Likewise.
18650 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18651 (move_sese_region_to_fn): Likewise.
18652 (debug_loop_num): Likewise.
18653 * tree-chrec.c (chrec_evaluate): Likewise.
18654 (hide_evolution_in_other_loops_than_loop): Likewise.
18655 (chrec_component_in_loop_num): Likewise.
18656 (reset_evolution_in_loop): Likewise.
18657 (evolution_function_is_invariant_rec_p): Likewise.
18658 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18659 * tree-inline.c (copy_loops): Likewise.
18660 (copy_cfg_body): Likewise.
18661 (tree_function_versioning): Likewise.
18662 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18663 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18664 Likewise.
18665 (add_to_evolution_1): Likewise.
18666 (scev_const_prop): Likewise.
18667 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18668 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18669 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18670 (tree_ssa_lim_initialize): Likewise.
18671 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18672 (verify_loop_closed_ssa): Likewise.
18673 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18674 (tree_ssa_loop_im): Likewise.
18675 (tree_ssa_loop_unswitch): Likewise.
18676 (tree_vectorize): Likewise.
18677 (check_data_deps): Likewise.
18678 (tree_ssa_loop_ivcanon): Likewise.
18679 (tree_ssa_loop_bounds): Likewise.
18680 (tree_complete_unroll): Likewise.
18681 (tree_complete_unroll_inner): Likewise.
18682 (tree_parallelize_loops): Likewise.
18683 (tree_ssa_loop_prefetch): Likewise.
18684 (tree_ssa_loop_ivopts): Likewise.
18685 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18686 * tree-vectorizer.c (vectorize_loops): Likewise.
18687
37953bd3
MF		 186882013-04-29 Mike Frysinger <vapier@gentoo.org>
18689
18690 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18691 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18692 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18693 with EABI_LINK_SPEC.
18694
f9ed28db
UB		 186952013-04-29 Uros Bizjak <ubizjak@gmail.com>
18696
18697 PR target/44578
18698 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18699 alternative.
18700
deca73f5
VM		 187012013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18702
18703 PR target/57097
37953bd3 18704 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM		 18705 using memory for pseudos. Print cost dump for alternatives.
18706 Modify cost values for conflicts with early clobbers.
18707 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18708
fdca7d03
UB		 187092013-04-29 Uros Bizjak <ubizjak@gmail.com>
18710
18711 PR target/57098
18712 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18713
0b064172
IB		 187142013-04-29 Ian Bolton <ian.bolton@arm.com>
18715
18716 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18717 from/to S register.
18718 (movdi_aarch64): Support LDR/STR from/to D register.
18719
473cec55
IB		 187202013-04-29 Ian Bolton <ian.bolton@arm.com>
18721
18722 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18723 or higher by default.
18724
a6f30e66
RB		 187252013-04-29 Richard Biener <rguenther@suse.de>
18726
18727 PR middle-end/57075
18728 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18729 even if not adding abnormal edges for calls that can make
18730 abnormal gotos.
18731
0c2b2040
RB		 187322013-04-29 Richard Biener <rguenther@suse.de>
18733
18734 PR middle-end/57103
18735 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18736 TREE_BLOCK.
18737 (move_stmt_r): Remove redundant checking.
18738
f41f80f9
TJ		 187392013-04-29 Teresa Johnson <tejohnson@google.com>
18740
18741 PR bootstrap/57077
18742 * basic-block.h (apply_scale): New function.
18743 (apply_probability): Use apply_scale.
18744 * gimple-streamer-in.c (input_bb): Ditto.
18745 * lto-streamer-in.c (input_cfg): Ditto.
18746 * lto-cgraph.c (merge_profile_summaries): Ditto.
18747 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 18748 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ		 18749 (copy_edges_for_bb): Ditto.
18750 (copy_cfg_body): Ditto.
18751
315bbd2e
TV		 187522013-04-29 Tom de Vries <tom@codesourcery.com>
18753
18754 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18755 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18756 (tail_merge_optimize): Handle current_loops == NULL.
18757
ebbd90d8
JL		 187582013-04-26 Jeff Law <law@redhat.com>
18759
18760 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18761 (simplify_cond_using_ranges): Generalize code to simplify
18762 COND_EXPRs where one argument is a constant and the other
18763 is an SSA_NAME created by an integral type conversion.
18764
8b9b57eb
KT		 187652013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18766
18767 * config/arm/arm.md (store_minmaxsi): Use only when
18768 optimize_insn_for_size_p.
18769
9498e5dc
CB		 187702013-04-29 Christian Bruel <christian.bruel@st.com>
18771
18772 PR target/57108
18773 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18774
6093bc06
RB		 187752013-04-29 Richard Biener <rguenther@suse.de>
18776
18777 PR middle-end/57089
fdca7d03
UB		 18778 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18779 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB		 18780 (expand_omp_for_generic): Properly add loops.
18781 (expand_omp_for_static_nochunk): Likewise.
18782 (expand_omp_for_static_chunk): Likewise.
18783 (expand_omp_for): For the degenerate case fixup loops.
18784 (expand_omp_sections): Fix default bb placement in loops.
18785 (expand_omp_atomic_pipeline): Properly add loops.
18786
84aacbfd
KT		 187872013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18788
18789 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18790
8dee4479
TV		 187912013-04-29 Tom de Vries <tom@codesourcery.com>
18792
18793 * tree-ssa-tail-merge.c: Update header comment.
18794
47934dc4
JG		 187952013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18796
18797 * config/aarch64/arm_neon.h
18798 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18799 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18800 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18801 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18802 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18803 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18804 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18805 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18806
384be29f
JG		 188072013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18808
18809 * config/aarch64/aarch64-simd.md
18810 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18811 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18812 fix_trunc, fixuns_trunc.
18813 (ftrunc<VDQF:mode>2): New.
18814 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18815 (fix_trunc_optab): New.
18816
0386b123
JG		 188172013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18818
18819 * config/aarch64/aarch64-builtins.c
18820 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18821 iceilf, lround, iroundf.
18822
00fcb892
UB		 188232013-04-29 Uros Bizjak <ubizjak@gmail.com>
18824
18825 PR target/54349
18826 * config/i386/i386.h (enum ix86_tune_indices)
18827 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18828 New, split from X86_TUNE_INTER_UNIT_MOVES.
18829 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18830 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18831 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18832 (TARGET_INTER_UNIT_MOVES): Remove.
18833 * config/i386/i386.c (initial_ix86_tune_features): Update.
18834 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18835 (ix86_expand_convert_uns_didf_sse): Use
18836 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18837 (ix86_expand_vector_init_one_nonzero): Ditto.
18838 (ix86_expand_vector_init_interleave): Ditto.
18839 (inline_secondary_memory_needed): Return true for moves from SSE class
18840 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18841 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18842 * config/i386/constraints.md (Yi, Ym): Depend on
18843 TARGET_INTER_UNIT_MOVES_TO_VEC.
18844 (Yj, Yn): New constraints.
18845 * config/i386/i386.md (*movdi_internal): Change constraints of
18846 operand 1 from Yi to Yj and from Ym to Yn.
18847 (*movsi_internal): Ditto.
18848 (*movdf_internal): Ditto.
18849 (*movsf_internal): Ditto.
18850 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18851 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18852 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18853 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18854 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18855 * config/i386/sse.md (movdi_to_sse): Ditto.
18856 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18857 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18858 TARGET_INTER_UNIT_MOVES.
18859 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18860 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18861 instead of TARGET_INTER_UNIT_MOVES.
18862 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18863 operand 1 from Yi to Yj and from Ym to Yn.
18864
4c871069
JG		 188652013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18866
18867 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18868 (float_truncate_hi_): Likewise.
18869 (float_extend_lo_): Likewise.
18870 (float_truncate_lo_): Likewise.
18871 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18872 (aarch64_float_extend_lo_v2df): Likewise.
18873 (vec_unpacks_hi_v4sf): Likewise.
18874 (aarch64_float_truncate_lo_v2sf): Likewise.
18875 (aarch64_float_truncate_hi_v4sf): Likewise.
18876 (vec_pack_trunc_v2df): Likewise.
18877 (vec_pack_trunc_df): Likewise.
18878
1709ff9b
JG		 188792013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18880
18881 * config/aarch64/aarch64-builtins.c
18882 (aarch64_fold_builtin): Fold float conversions.
18883 * config/aarch64/aarch64-simd-builtins.def
18884 (floatv2si, floatv4si, floatv2di): New.
18885 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18886 * config/aarch64/aarch64-simd.md
18887 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18888 * config/aarch64/iterators.md (FLOATUORS): New.
18889 (optab): Add float, floatuns.
18890 (su_optab): Likewise.
18891
ce966824
JG		 188922013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18893
18894 * config/aarch64/aarch64-builtins.c
18895 (aarch64_builtin_vectorized_function): Use new names for
18896 fcvt builtins.
18897 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18898 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18899 (fcvtzu): Split as...
18900 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18901 (fcvtas): Split as...
18902 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18903 (fcvtau): Split as...
18904 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18905 (fcvtps): Split as...
18906 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18907 (fcvtpu): Split as...
18908 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18909 (fcvtms): Split as...
18910 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18911 (fcvtmu): Split as...
18912 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18913 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18914 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18915 (lfrintnusf, lfrintnudf): Likewise.
18916 * config/aarch64/aarch64-simd.md
18917 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18918 define_insn.
18919 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18920 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18921 (fcvt_pattern): Likewise.
18922
b9de24fe
JG		 189232013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18924
18925 * config/aarch64/aarch64-simd.md
18926 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18927 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18928
77a205be
JG		 189292013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18930
18931 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18932 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18933 (vrnd<a,m,n,p>_f32): Implement using builtins.
18934 (vrnd<i,x><q>_f<32, 64>): New.
18935
0659ce6f
JG		 189362013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18937
18938 * config/aarch64/aarch64-builtins.c
18939 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18940 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18941 (frintz): Rename to...
18942 (btrunc): ...this.
18943 (frintp): Rename to...
18944 (ceil): ...this.
18945 (frintm): Rename to...
18946 (floor): ...this.
18947 (frinti): Rename to...
18948 (nearbyint): ...this.
18949 (frintx): Rename to...
18950 (rint): ...this.
18951 (frinta): Rename to...
18952 (round): ...this.
18953 * config/aarch64/aarch64-simd.md
18954 (aarch64_frint<frint_suffix><mode>): Delete.
18955 (<frint_pattern><mode>2): Convert to insn.
18956 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18957 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18958 (frint_pattern): Likewise.
18959 (frint_suffix): Likewise.
18960
ea78906a
RB		 189612013-04-29 Richard Biener <rguenther@suse.de>
18962
18963 PR tree-optimization/57081
18964 * loop-init.c: Include tree-flow.h.
18965 (loop_optimizer_finalize): Free number of iteration estimates.
18966 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18967
baee1763
JJ		 189682013-04-29 Jakub Jelinek <jakub@redhat.com>
18969
94dc5332
JJ		 18970 PR tree-optimization/57083
18971 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18972 non-singleton shift count range, zero extend low_bound for uns case.
18973
baee1763
JJ		 18974 * config/i386/predicates.md (general_vector_operand): New predicate.
18975 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18976 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18977 if they aren't nonimmediate operands. If their original values
18978 satisfy const_vector_equal_evenodd_p, don't shift them.
18979 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18980 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18981 if not nonimmediate_operand.
18982 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18983 instead of register_operand.
18984 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18985
a9073727 189862013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB		 18987
18988 * stor-layout.c (finalize_size_functions): Allocate a structure and
18989 reset cfun before dumping the functions.
18990
ba8011e6
JJ		 189912013-04-27 Jakub Jelinek <jakub@redhat.com>
18992
d6fde69e
JJ		 18993 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18994
ba8011e6
JJ		 18995 PR target/56866
18996 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18997 use xop_pmacsdqh if uns_p.
18998 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18999 the immediate rotate count.
19000
2c62cbaa
VM		 190012013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19002
19003 * rtl.h (struct rtx_def): Add comment for field jump.
19004 (LRA_SUBREG_P): New macro.
19005 * recog.c (register_operand): Check LRA_SUBREG_P.
19006 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19007 stack frame.
19008 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19009 (lra_final_code_change): Skip subreg change for operators.
19010 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19011 if there are no operand changes.
19012 * lra-constraints.c (curr_insn_set): New.
19013 (match_reload): Set LRA_SUBREG_P.
19014 (emit_spill_move): Ditto.
19015 (check_and_process_move): Use curr_insn_set. Process only single
19016 set insns. Don't initialize sec_mem_p and change_p.
19017 (simplify_operand_subreg): Use LRA_SUBREG_P.
19018 (reg_in_class_p): New function.
19019 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19020 of #ifdef. Add code to remove cycling.
19021 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19022 non-null disp. Reload inner instead of disp when base and index
19023 are null. Try to put lo_sum into register.
19024 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19025 (check_and_process_move): Move code for move cost check to
19026 simple_move_p. Remove equiv_substitution.
19027 (simple_move_p): New function.
19028 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19029 curr_insn_set. Call check_and_process_move only for single set
19030 insns. Use the new function. Move call of check_and_process_move
19031 after operand equiv substitution and address process.
19032
e7d764f3
JJ		 190332013-04-26 Jakub Jelinek <jakub@redhat.com>
19034
19035 PR go/57045
19036 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19037 with nonlocal goto receivers or returns twice calls, ignore
19038 unininitialized values from abnormal edges to nl goto receiver
19039 or returns twice call.
19040
41e10689
JJ		 190412013-04-26 Jakub Jelinek <jakub@redhat.com>
19042
19043 PR tree-optimization/57051
19044 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19045 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19046 bitsize.
19047
d7ed20db
RB		 190482013-04-26 Richard Biener <rguenther@suse.de>
19049
19050 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19051 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19052 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19053 (fixup_loop_arrays_after_move): New function.
19054 (move_sese_region_to_fn): Properly outline the loop tree parts
19055 of the SESE region.
19056
df93505e
UB		 190572013-04-26 Uros Bizjak <ubizjak@gmail.com>
19058
19059 * config/i386/i386.md (type, unit): Fix long lines.
19060
dd366ec3
RB		 190612013-04-26 Richard Biener <rguenther@suse.de>
19062
19063 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19064 (lto-streamer-out.o): Likewise.
19065 * cfgloop.c (init_loops_structure): Export, add struct function
19066 argument and adjust.
19067 (flow_loops_find): Adjust.
19068 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19069 (init_loops_structure): Declare.
19070 * lto-streamer-in.c: Include cfgloop.h.
19071 (input_cfg): Input the loop tree.
19072 * lto-streamer-out.c: Include cfgloop.h.
19073 (output_cfg): Output the loop tree.
19074 (output_struct_function_base): Do not drop PROP_loops.
19075
a9e0d843
RB		 190762013-03-26 Richard Biener <rguenther@suse.de>
19077
19078 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19079 (pass_build_cfg): Provide PROP_loops.
19080 (move_sese_region_to_fn): Remove loops that are outlined into fn
19081 for now.
19082 * tree-inline.c: Include cfgloop.h.
19083 (initialize_cfun): Do not drop PROP_loops.
19084 (copy_loops): New function.
19085 (copy_cfg_body): Copy loop structure.
19086 (tree_function_versioning): Initialize destination loop tree.
19087 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19088 (pass_parallelize_loops): Do IL verification.
19089 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19090 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19091 the CFG make sure we fixup loops as well.
19092 * tree-ssa-tail-merge.c: Include cfgloop.h.
19093 (replace_block_by): When merging loop latches mark loops for fixup.
19094 * lto-streamer-out.c (output_struct_function_base): Drop
19095 PROP_loops for now.
19096 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19097 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19098 * ipa-split.c: Include cfgloop.h.
19099 (split_function): Add the new return block to the loop tree root.
19100 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19101 whether we have removed the forwarder block.
19102 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19103 * cfgloop.h (place_new_loop): Declare.
19104 * cfgloopmanip.c (place_new_loop): Export.
19105 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19106 (tree-switch-conversion.o): Likewise.
19107 (tree-complex.o): Likewise.
19108 (tree-inline.o): Likewise.
19109 (tree-ssa-tailmerge.o): Likewise.
19110 (ipa-split.o): Likewise.
19111 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19112 (tree-ssa-copy.o): Likewise.
19113 * tree-switch-conversion.c: Include cfgloop.h
19114 (process_switch): If we emit a bit-test cascade, schedule loops
19115 for fixup.
19116 * tree-complex.c: Include cfgloop.h.
19117 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19118 * asan.c: Include cfgloop.h.
19119 (create_cond_insert_point): Properly add new basic-blocks to
19120 loops, schedule loop fixup.
19121 * cfgloop.c (verify_loop_structure): Check that looks are not
19122 marked for fixup.
19123 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19124 to loops.
19125 (expand_omp_for_generic): Likewise.
19126 (expand_omp_sections): Likewise.
19127 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19128 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19129 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19130 is initialized, not when loops are present.
19131 * tree-parloops.c (parallelize_loops): Remove checking here.
19132 * passes.c (init_optimization_passes): Schedule a copy-propagation
19133 pass before complete unrolling of inner loops.
19134
e78e8a0b
JJ		 191352013-04-26 Jakub Jelinek <jakub@redhat.com>
19136
a2e836b2
JJ		 19137 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19138 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19139 (colorize_init): Add argument to _WIN32 version.
19140 * toplev.c: Include diagnostic-color.h.
19141 (process_options): Default to -fdiagnostics-color=auto if
19142 GCC_COLORS env var is in the environment.
19143 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB		 19144 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19145 env var is in the environment, the default is auto rather than never.
a2e836b2 19146
e78e8a0b
JJ		 19147 * diagnostic.h (file_name_as_prefix): Add context argument.
19148 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19149 the string as locus.
19150 * langhooks.c (lhd_print_error_function): Adjust caller.
19151
013e5ef9
LC		 191522013-04-25 Lawrence Crowl <crowl@google.com>
19153
19154 * var-tracking.c (shared_hash_def::htab):
19155 Change type to hash_table. Update dependent calls and types.
19156
4a8fb1a1
LC		 191572013-04-25 Lawrence Crowl <crowl@google.com>
19158
19159 * Makefile.in: Update as needed below.
19160
19161 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19162 Move declaration to after the type's method definitons.
19163
19164 * attribs.c (htab_t scoped_attributes::attribute_hash):
19165 Change type to hash_table. Update dependent calls and types.
19166
19167 * bitmap.c (htab_t bitmap_desc_hash):
19168 Change type to hash_table. Update dependent calls and types.
19169
19170 * cselib.c (htab_t cselib_hash_table):
19171 Change type to hash_table. Update dependent calls and types.
19172
19173 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19174 (hash_string_slot_node): Move implementation into lto-streamer.h
19175 struct string_slot_hasher.
19176 (eq_string_slot_node): Likewise.
19177
19178 * data-streamer-out.c: Update output_block::string_hash_table
19179 dependent calls and types.
19180
19181 * dwarf2cfi.c (htab_t trace_index):
19182 Change type to hash_table. Update dependent calls and types.
19183
19184 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19185 Change type to hash_table. Update dependent calls and types.
19186 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19187 (htab_t optimize_external_refs::map): Likewise.
19188 (htab_t output_comp_unit::extern_map): Likewise.
19189 (htab_t output_comdat_type_unit::extern_map): Likewise.
19190 (htab_t output_macinfo::macinfo_htab): Likewise.
19191 (htab_t optimize_location_lists::htab): Likewise.
19192 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19193
19194 * except.c (htab_t ehspec_hash_type):
19195 Change type to hash_table. Update dependent calls and types.
19196 (assign_filter_values::ttypes): Likewise.
19197 (assign_filter_values::ehspec): Likewise.
19198 (sjlj_assign_call_site_values::ar_hash): Likewise.
19199 (convert_to_eh_region_ranges::ar_hash): Likewise.
19200
19201 * gcse.c (htab_t pre_ldst_table):
19202 Change type to hash_table. Update dependent calls and types.
19203
19204 * ggc-common.c (htab_t saving_htab):
19205 Change type to hash_table. Update dependent calls and types.
19206 (htab_t loc_hash): Likewise.
19207 (htab_t ptr_hash): Likewise.
19208 (call_count): Rename ggc_call_count.
19209 (call_alloc): Rename ggc_call_alloc.
19210 (loc_descriptor): Rename make_loc_descriptor.
19211 (add_statistics): Rename ggc_add_statistics.
19212
19213 * ggc-common.c (saving_htab):
19214 Change type to hash_table. Update dependent calls and types.
19215
19216 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19217 (push_gimplify_context): Likewise.
19218 (pop_gimplify_context): Likewise.
19219 (struct gimple_temp_hash_elt): Added.
19220 (struct gimplify_hasher): Likewise.
19221 (struct gimplify_ctx.temp_htab):
19222 Change type to hash_table. Update dependent calls and types.
19223
19224 * gimple-fold.c: Include gimplify-ctx.h.
19225
19226 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19227 Change type to hash_table. Update dependent calls and types.
19228 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19229 avoid potential global name collision.
19230
19231 * gimplify.c: Include gimplify-ctx.h.
19232 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19233 (htab_t gimplify_ctx::temp_htab):
19234 Update dependent calls and types for new type hash_table.
19235 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19236 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19237
19238 * gimplify-ctx.h: New.
19239 (struct gimple_temp_hash_elt): Move from gimplify.c.
19240 (class gimplify_hasher): New.
19241 (struct gimplify_ctx): Move from gimple.h.
19242 (htab_t gimplify_ctx::temp_htab):
19243 Change type to hash_table. Update dependent calls and types.
19244
19245 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19246 (htab_t ivs_params::newivs_index):
19247 Change type to hash_table. Update dependent calls and types.
19248 (htab_t ivs_params::params_index): Likewise.
19249 (htab_t print_generated_program::params_index): Likewise.
19250 (htab_t gloog::newivs_index): Likewise.
19251 (htab_t gloog::params_index): Likewise.
19252
19253 * graphite.c: Include graphite-htab.h.
19254 4htab_t graphite_transform_loops::bb_pbb_mapping):
19255 Change type to hash_table. Update dependent calls and types.
19256
19257 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19258 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19259 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19260
19261 * graphite-dependences.c: Include graphite-htab.h.
19262 (loop_is_parallel_p): Change hash table type of parameter.
19263
19264 * graphite-htab.h: New.
19265 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19266 (extern find_pbb_via_hash): Move from graphite-poly.h.
19267 (extern loop_is_parallel_p): Move from graphite-poly.h.
19268 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19269
19270 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19271 (extern loop_is_parallel_p): Move to graphite-htab.h.
19272 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19273
19274 * haifa-sched.c (htab_t delay_htab):
19275 Change type to hash_table. Update dependent calls and types.
19276 (htab_t delay_htab_i2): Likewise.
19277
19278 * ira-color.c (htab_t allocno_hard_regs_htab):
19279 Change type to hash_table. Update dependent calls and types.
19280
19281 * ira-costs.c (htab_t cost_classes_htab):
19282 Change type to hash_table. Update dependent calls and types.
19283
19284 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19285 Change type to hash_table. Update dependent calls and types.
19286
19287 * loop-iv.c (htab_t bivs):
19288 Change type to hash_table. Update dependent calls and types.
19289
19290 * loop-unroll.c (htab_t opt_info::insns_to_split):
19291 Change type to hash_table. Update dependent calls and types.
19292 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19293
19294 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19295 (struct string_slot_hasher): New.
19296 (htab_t output_block::string_hash_table):
19297 Change type to hash_table. Update dependent calls and types.
19298
19299 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19300 (htab_t file_name_hash_table):
19301 Change type to hash_table. Update dependent calls and types.
19302
19303 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19304 calls and types.
19305
19306 * lto-streamer.c (htab_t tree_htab):
19307 Change type to hash_table. Update dependent calls and types.
19308
19309 * omp-low.c: Include gimplify-ctx.h.
19310
19311 * passes.c (htab_t name_to_pass_map):
19312 Change type to hash_table. Update dependent calls and types.
19313 (pass_traverse): Rename to passes_pass_traverse.
19314
19315 * plugin.c (htab_t event_tab):
19316 Change type to hash_table. Update dependent calls and types.
19317
19318 * postreload-gcse.c (htab_t expr_table):
19319 Change type to hash_table. Update dependent calls and types.
19320 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19321
19322 * sese.c (debug_rename_map_1): Make extern.
19323 (htab_t copy_bb_and_scalar_dependences::rename_map):
19324 Change type to hash_table. Update dependent calls and types.
19325
19326 * sese.h (extern debug_rename_map): Move to .c file.
19327
19328 * store-motion.c (htab_t store_motion_mems_table):
19329 Change type to hash_table. Update dependent calls and types.
19330
19331 * trans-mem.c (htab_t tm_new_mem_hash):
19332 Change type to hash_table. Update dependent calls and types.
19333
19334 * tree-browser.c (htab_t TB_up_ht):
19335 Change type to hash_table. Update dependent calls and types.
19336
19337 * tree-cfg.c (htab_t discriminator_per_locus):
19338 Change type to hash_table. Update dependent calls and types.
19339
19340 * tree-complex.c: Include tree-hasher.h
19341 (htab_t complex_variable_components):
19342 Change type to hash_table. Update dependent calls and types.
19343
19344 * tree-eh.c (htab_t finally_tree):
19345 Change type to hash_table. Update dependent calls and types.
19346
19347 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19348 struct int_tree_hasher.
19349 (extern int_tree_map_eq): Likewise.
19350 (uid_decl_map_hash): Removed.
19351 (extern decl_tree_map_eq): Likewise.
19352
19353 * tree-hasher.h: New.
19354 (struct int_tree_hasher): New.
19355 (typedef int_tree_htab_type): New.
19356
19357 * tree-inline.c: Include gimplify-ctx.h.
19358
19359 * tree-mudflap.c: Include gimplify-ctx.h.
19360
19361 * tree-parloops.c: Include tree-hasher.h.
19362 (htab_t eliminate_local_variables_stmt::decl_address):
19363 Change type to hash_table. Update dependent calls and types.
19364 (htab_t separate_decls_in_region::decl_copies): Likewise.
19365
19366 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19367 Change type to hash_table. Update dependent calls and types.
19368
19369 * tree-sra.c (candidates):
19370 Change type to hash_table. Update dependent calls and types.
19371
19372 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19373 in tree-flow.h.
19374 (int_tree_map_hash): Likewise.
19375
19376 * tree-ssa-dom.c (htab_t avail_exprs):
19377 Change type to hash_table. Update dependent calls and types.
19378
19379 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19380 Change type to hash_table. Update dependent calls and types.
19381
19382 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19383 Change type to hash_table. Update dependent calls and types.
19384
19385 * tree-ssa-phiopt.c (seen_ssa_names):
19386 Change type to hash_table. Update dependent calls and types.
19387
19388 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19389 Change type to hash_table. Update dependent calls and types.
19390
19391 * tree-ssa-uncprop.c (equiv):
19392 Change type to hash_table. Update dependent calls and types.
19393
c5a44004
JJ		 193942013-04-25 Jakub Jelinek <jakub@redhat.com>
19395
19396 PR rtl-optimization/57003
19397 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19398 call note_stores with kill_clobbered_value callback again after
19399 killing regs_invalidated_by_call.
19400
09962a4a
JG		 194012013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19402
19403 * config/aarch64/aarch64-simd.md
19404 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19405 (aarch64_simd_bsl<mode>): Likewise.
19406 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19407
ea28bb0b
MP		 194082013-04-25 Marek Polacek <polacek@redhat.com>
19409
19410 PR tree-optimization/57066
3c21604f 19411 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 19412
96659611
JG		 194132013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19414
19415 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19416
9697e620
JG		 194172013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19418
19419 * config/aarch64/aarch64-builtins.c
19420 (aarch64_fold_builtin): New.
19421 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19422 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19423 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19424 * config/aarch64/arm_neon.h
19425 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19426
0ac198d3
JG		 194272013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19428 Tejas Belagod <tejas.belagod@arm.com>
19429
19430 * config/aarch64/aarch64-builtins.c
19431 (aarch64_gimple_fold_builtin): New.
19432 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19433 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19434 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19435 (addvv4sf): Update.
19436 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19437
58cff58c
N		 194382013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19439
df93505e 19440 * config/aarch64/aarch64.md
58cff58c
N		 19441 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19442
7e0228bf
N		 194432013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19444
19445 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19446 (*ngcsi_uxtw): New pattern.
19447
5819f96f 194482013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 19449 Julian Brown <julian@codesourcery.com>
5819f96f
KT		 19450
19451 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19452 (TB_DREG): Add T_V4HF.
19453 (v4hf_UP): New macro.
19454 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 19455 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19456 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19457 built-ins.
df93505e 19458 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT		 19459 (arm_vector_mode_supported_p): Handle V4HF.
19460 (arm_mangle_map): Handle V4HFmode.
19461 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19462 * config/arm/arm_neon_builtins.def: Add entries for
19463 vcvtv4hfv4sf, vcvtv4sfv4hf.
19464 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19465 (neon_vcvtv4hfv4sf): Likewise.
19466 * config/arm/neon-gen.ml: Handle half-precision floating point
19467 features.
19468 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19469 * config/arm/arm_neon.h: Regenerate.
19470 * config/arm/neon.ml (type elts): Add F16.
19471 (type vectype): Add T_float16x4, T_floatHF.
19472 (type vecmode): Add V4HF.
19473 (type features): Add Requires_FP_bit feature.
19474 (elt_width): Handle F16.
19475 (elt_class): Likewise.
19476 (elt_of_class_width): Likewise.
19477 (mode_of_elt): Refactor.
19478 (type_for_elt): Handle F16, fix error messages.
19479 (vectype_size): Handle T_float16x4.
19480 (vcvt_sh): New function.
19481 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19482 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19483 (string_of_mode): Handle V4HF.
19484 * doc/arm-neon-intrinsics.texi: Regenerate.
19485
1ef395e4
JG		 194862013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19487
19488 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19489 format specifier in 'X' case.
19490
41c34e94
AM		 194912013-04-25 Alan Modra <amodra@gmail.com>
19492
19493 PR target/57052
19494 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19495 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19496 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19497 Repeat for many other rotate/shift and mask patterns using subregs.
19498 Name lshiftrt insns.
19499 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19500 on WORDS_BIG_ENDIAN.
19501
b9a7eb5d
AM		 195022013-04-25 Alan Modra <amodra@gmail.com>
19503
19504 * config.gcc: Support little-endian powerpc-linux targets.
19505 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19506 (LINK_OS_LINUX_SPEC): Define.
19507 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19508 Preserve MASK_LITTLE_ENDIAN.
19509 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19510 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19511 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19512 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19513 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19514 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19515 Correct fp word order for little-endian. Don't shift toc entries
19516 smaller than a word for little-endian.
19517 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19518 (bswapdi2 splits): Correct low-part subreg for little-endian.
19519 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19520 low/high where such is correct only for be.
19521 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19522 little-endian for -mcall-aixdesc.
19523
87f73374
AM		 195242013-04-25 Alan Modra <amodra@gmail.com>
19525
19526 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19527 replace_equiv_address_nv.
19528
cabf91cd
AM		 195292013-04-25 Alan Modra <amodra@gmail.com>
19530
19531 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19532
0ae24cc8
VM		 195332013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19534
19535 Revert:
19536 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19537 * rtl.h (struct rtx_def): ...
cabf91cd 19538
77bce07c
VM		 195392013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19540
19541 PR rtl-optimizations/57046
19542 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19543 for multi-reg splits.
19544
0db63e7f
L		 195452013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19546
19547 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19548
3362b6b6
SA		 195492013-04-24 Sterling Augustine <saugustine@google.com>
19550
19551 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19552 (comp_dir_string, debug_str_dwo_section): New.
19553 (DEBUG_STR_DWO_SECTION): Rename to ...
19554 (DEBUG_DWO_STR_SECTION): ... this.
19555 (DEBUG_NORM_STR_SECTION): Delete.
19556 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19557 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19558 (find_AT_string): Move most logic to ...
19559 (find_AT_string_in_table): ... here. New.
19560 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19561 add_skeleton_AT_string. Delete logic.
19562 (output_skeleton_debug_sections): Remove call to
19563 add_top_level_skeleton_die_attrs.
19564 (add_comp_dir_attribute): Move logic to comp_dir_string.
19565 (dwarf2out_init): Initialize debug_str_dwo_section.
19566 (output_indirect_string): Call find_string_form.
19567 (output_indirect_strings): Rewrite.
19568 (prune_unused_types): Empty skeleton_debug_str_hash.
19569 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19570 (dwarf2out_finish): Call output_indirect_strings.
19571
e93e18e9
PC		 195722013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19573
19574 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19575
f6ce35ac
VM		 195762013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19577
cabf91cd 19578 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM		 19579 (LRA_SUBREG_P): New macro.
19580 * recog.c (register_operand): Check LRA_SUBREG_P.
19581 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19582 stack frame.
19583 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19584 (lra_final_code_change): Skip subreg change for operators.
19585 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19586 if there are no operand changes.
19587 * lra-constraints.c (curr_insn_set): New.
19588 (match_reload): Set LRA_SUBREG_P.
19589 (emit_spill_move): Ditto.
19590 (check_and_process_move): Use curr_insn_set. Process only single
19591 set insns. Don't initialize sec_mem_p and change_p.
19592 (simplify_operand_subreg): Use LRA_SUBREG_P.
19593 (reg_in_class_p): New function.
19594 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19595 of #ifdef. Add code to remove cycling.
19596 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19597 non-null disp. Reload inner instead of disp when base and index
19598 are null. Try to put lo_sum into register.
19599 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 19600 (check_and_process_move): Move code for move cost check to
f6ce35ac 19601 simple_move_p. Remove equiv_substitution.
cabf91cd 19602 (simple_move_p): New function.
f6ce35ac
VM		 19603 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19604 curr_insn_set. Call check_and_process_move only for single set
19605 insns. Use the new function. Move call of check_and_process_move
19606 after operand equiv substitution and address process.
19607
38047d90
JG		 196082013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19609
19610 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19611 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19612 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19613
13f39b2e
PC		 196142013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19615
19616 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19617
97c116dc
MP		 196182013-04-24 Marek Polacek <polacek@redhat.com>
19619
19620 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19621 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19622 (select_loops_exit_conditions): Likewise.
19623 (number_of_iterations_for_all_loops): Likewise.
19624 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19625 (scev_analysis): Likewise.
19626
83082391 196272013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 19628 Chao-ying Fu <fu@mips.com>
83082391 19629
cabf91cd
AM		 19630 * config/mips/micromips.md (jraddiusp): New pattern.
19631 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19632 instruction if possible.
83082391 19633
19e34aa2
AM		 196342013-04-24 Alan Modra <amodra@gmail.com>
19635
19636 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19637
fdb6603c
JB		 196382013-04-24 Julian Brown <julian@codesourcery.com>
19639 Chung-Lin Tang <cltang@codesourcery.com>
19640
19641 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB		 19642 dependency behavior in enumeration type DIE generation. Add TODO note
19643 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 19644
bf190e8d
LC		 196452013-04-23 Lawrence Crowl <crowl@google.com>
19646
19647 * Makefile.in: Update as needed below.
19648
19649 * hash-table.h (class hash_table):
19650 Correct many methods with parameter types compare_type to the correct
19651 value_type. (Correct code was unlikely to notice the change.)
19652 (hash_table::elements_with_deleted) New.
19653 (class hashtable::iterator): New.
19654 (hashtable::begin()): New.
19655 (hashtable::end()): New.
19656 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19657
19658 * statistics.c (statistics_hashes):
19659 Change type to hash_table. Update dependent calls and types.
19660
19661 * tree-into-ssa.c (var_infos):
19662 Change type to hash_table. Update dependent calls and types.
19663
19664 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19665 Change type to hash_table. Update dependent calls and types.
19666
19667 * tree-ssa-loop-im.c (struct mem_ref.refs):
19668 Change type to hash_table. Update dependent calls and types.
19669
19670 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19671 Change type to hash_table. Update dependent calls and types.
19672
19673 * tree-ssa-sccvn.c (vn_tables_s::nary):
19674 Change type to hash_table. Update dependent calls and types.
19675 (vn_tables_s::phis): Likewise.
19676 (vn_tables_s::references): Likewise.
19677
19678 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19679 (vn_reference_eq): Update parameter and return types.
19680
19681 * tree-ssa-structalias.c (pointer_equiv_class_table):
19682 Change type to hash_table. Update dependent calls and types.
19683 (location_equiv_class_table): Likewise.
19684
19685 * tree-vect-data-refs.c: Consequential changes for making
19686 peeling a hash_table.
19687
19688 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19689 (destroy_loop_vec_info): Dependent hash_table update.
19690
19691 * tree-vectorizer.h (peeling_htab):
19692 Change type to hash_table. Update dependent calls and types.
19693
d70a81dd
SC		 196942013-04-23 Shiva Chen <shiva0217@gmail.com>
19695
cabf91cd
AM		 19696 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19697 to check the register content is equal or not.
19698 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19699 to assign register content record.
19700 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 19701 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM		 19702 * lra-int.h (struct lra_reg): Add offset member.
19703 (lra_reg_val_equal_p): New static inline function.
19704 (lra_update_reg_val_offset): New static inline function.
19705 (lra_assign_reg_val): New static inline function.
19706 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19707 to assign register content record.
19708 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 19709
b894a1f3
CM		 197102013-04-23 Catherine Moore <clm@codesourcery.com>
19711
19712 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19713 operands. Record compression.
19714
ecd14de9
XDL		 197152013-04-23 Xinliang David Li <davidxl@google.com>
19716
19717 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19718
92e776e9
RB		 197192013-04-23 Richard Biener <rguenther@suse.de>
19720
19721 PR middle-end/57036
19722 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19723 parameter, only add abnormal goto edges from the copied body
19724 if the call could perform abnormal gotos.
19725 (copy_cfg_body): Adjust.
19726
a15ee567
SN		 197272013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19728
19729 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19730
08c52234
AS		 197312013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19732
19733 * coretypes.h (gimple_stmt_iterator): Add struct to make
19734 compatible with C.
19735
999c1171
RB		 197362013-04-23 Richard Biener <rguenther@suse.de>
19737
19738 PR tree-optimization/57026
19739 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19740 from SSA names occuring in abnormal PHI nodes.
19741
53669259
AK		 197422013-04-22 Andi Kleen <ak@linux.intel.com>
19743
19744 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19745
1a0ad150
AK		 197462013-04-22 Andi Kleen <ak@linux.intel.com>
19747
19748 * lto/lto.c (print_lto_report_1): Declare early.
19749 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19750
057f8f20
AK		 197512013-04-22 Andi Kleen <ak@linux.intel.com>
19752
19753 * common.opt (-flto-report-wpa): Add.
19754 * doc/invoke.texi (-flto-report-wpa): Add.
19755 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19756 (lto_main): dito.
19757
473b1e05
XDL		 197582013-04-22 Xinliang David Li <davidxl@google.com>
19759
19760 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19761 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19762 * Makefile.in: New dependency
19763
cabf91cd 19764 David Daney <ddaney.cavm@gmail.com>
b1485a33 19765
cabf91cd
AM		 19766 * configure.ac (gcc_cv_as_micromips_support): Use the
19767 --fatal-warnings option.
19768 * configure: Regenerate.
b1485a33 19769
829d0168
MP		 197702013-04-22 Marek Polacek <polacek@redhat.com>
19771
19772 PR sanitizer/56990
19773 * tsan.c (instrument_expr): Don't instrument expression
19774 in case its size is zero.
19775
6d9b7208
UB		 197762013-04-22 Uros Bizjak <ubizjak@gmail.com>
19777
19778 PR target/57032
19779 Revert:
19780 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19781
19782 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19783
ea679d55
JG		 197842013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19785
19786 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19787 (gimple_stmt_iterator): New typedef.
19788 * gimple.h (gimple_stmt_iterator): Rename to...
19789 (gimple_stmt_iterator_d): ... This.
19790 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19791 trees be valid for GIMPLE and GENERIC.
19792 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19793 * gimple-fold.c (gimple_fold_call): Call target hook
19794 gimple_fold_builtin.
19795 * hooks.c (hook_bool_gsiptr_false): New.
19796 * hooks.h (hook_bool_gsiptr_false): New.
19797 * target.def (fold_stmt): New.
19798 * doc/tm.texi: Regenerate.
19799
88a581da
VM		 198002013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19801
19802 PR target/57018
19803 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19804 a set sp if no stack realignment.
19805
92be22dc
NC		 198062013-04-22 Nick Clifton <nickc@redhat.com>
19807
19808 * config.gcc (tilegx-linux): Extend extra_objs rather than
19809 overwriting it.
19810 (tilepro-linux): Likewise.
19811
0ddec79f
JG		 198122013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19813
19814 * config/aarch64/aarch64-builtins.c
19815 (CF): Remove.
19816 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19817 (VAR<1-12>): Add MAP parameter.
19818 (BUILTIN_*): Likewise.
19819 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19820 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19821 (aarch64_ushl_n<mode>): Likewise.
19822 (aarch64_sshr_n<mode>): Likewise.
19823 (aarch64_ushr_n<mode>): Likewise.
19824 (aarch64_<maxmin><mode>): Likewise.
19825 (aarch64_sqrt<mode>): Likewise.
19826 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19827 (vshr<q>_n_*): Likewise.
19828
0050faf8
JG		 198292013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19830
19831 * config/aarch64/aarch64-builtins.c
19832 (aarch64_simd_builtin_type_mode): Handle SF types.
19833 (sf_UP): Define.
19834 (BUILTIN_GPF): Define.
19835 (aarch64_init_simd_builtins): Handle SF types.
19836 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19837 (frecps): Likewise.
19838 (frecpx): Likewise.
19839 * config/aarch64/aarch64-simd.md
19840 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19841 (aarch64_frecpe<mode>): New.
19842 (aarch64_frecps<mode>): Likewise.
19843 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19844 (v8type): Add frecp<esx>.
19845 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19846 (aarch64_frecps<mode>): Likewise.
19847 * config/aarch64/iterators.md (FRECP): New.
19848 (frecp_suffix): Likewise.
19849 * config/aarch64/arm_neon.h
19850 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19851
0fad3dbc 198522013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB		 19853
19854 PR target/56995
19855 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19856 (REG_CLASS_NAMES): Idem.
19857 (REG_CLASS_CONTENTS): Idem.
19858 (REGCLASS_HAS_FP_REG): Idem.
19859 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19860 (sh_conditional_register_usage): Idem.
19861
3e8a33f9
JL		 198622013-04-21 Jeff Law <law@redhat.com>
19863
19864 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19865 (ssa_forward_propagate_and_combine): Use it.
19866
f38e1b0a
VM		 198672013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19868
19869 * lra.c: Update the flow chart diagram.
19870
682303da
VM		 198712013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19872
19873 PR rtl-optimization/56847
19874 * lra-constraints.c (process_alt_operands): Discourage alternative
19875 with non-matche doffsettable memory constraint fro memory with
19876 known offset.
19877
f6b64c35
RB		 198782013-04-19 Richard Biener <rguenther@suse.de>
19879
19880 PR tree-optimization/56982
19881 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19882 function.
19883 * gimplify.c (gimplify_call_expr): Notice special calls.
19884 (gimplify_modify_expr): Likewise.
19885 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19886 abnormal control flow receivers.
19887 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19888 in the same way as cfun->has_nonlocal_labels.
19889 (gimple_purge_dead_abnormal_call_edges): Likewise.
19890 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19891 receivers start a basic-block.
19892
01d8bf07
RB		 198932013-04-19 Richard Biener <rguenther@suse.de>
19894
19895 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19896 member ...
19897 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19898 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19899 (SLP_TREE_LOAD_PERMUTATION): Add.
19900 (vect_transform_slp_perm_load): Adjust prototype.
19901 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19902 (vect_free_slp_instance): Likewise.
19903 (vect_create_new_slp_node): Likewise.
19904 (vect_supported_slp_permutation_p): Remove.
19905 (vect_slp_rearrange_stmts): Adjust.
19906 (vect_supported_load_permutation_p): Likewise. Inline
19907 vect_supported_slp_permutation_p here.
19908 (vect_analyze_slp_instance): Compute load permutations per
19909 slp node instead of per instance.
19910 (vect_get_slp_defs): Adjust.
19911 (vect_transform_slp_perm_load): Likewise.
19912 (vect_schedule_slp_instance): Remove redundant code.
19913 (vect_schedule_slp): Remove hack for PR56270, add it ...
19914 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19915 CSE loads for SLP. Adjust.
19916
ede22fc3
GY		 199172013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19918
19919 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19920 spelling in two comments.
19921
67bc84fb
GY		 199222013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19923
19924 PR target/56797
19925 * config/arm/arm.c (load_multiple_sequence): Require SP
19926 as base register for loads if SP is in the register list.
19927
e248d83f
MJ		 199282013-04-19 Martin Jambor <mjambor@suse.cz>
19929
19930 PR tree-optimization/56718
19931 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19932 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19933 and made public. Adjusted all callers.
19934 (ipa_intraprocedural_devirtualization): New function.
19935 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19936 (ipa_intraprocedural_devirtualization): Likewise.
19937 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19938
4891e8f8
RB		 199392013-04-19 Richard Biener <rguenther@suse.de>
19940
19941 PR tree-optimization/57000
19942 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19943
dad89f7c
TG		 199442013-04-19 Terry Guo <terry.guo@arm.com>
19945
19946 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19947 Replace with ...
19948 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19949 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19950 (cortex_m4_fmacs): Use new reservations.
19951 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19952
72ea0d47
VM		 199532013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19954
f1e6512c 19955 PR rtl-optimization/56999
72ea0d47
VM		 19956 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19957 related code.
19958 (lra_coalesce): Remove split_origin_bitmap and related code.
19959 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19960 ranges if necessary.
19961
780a5b71
UB		 199622013-04-18 Uros Bizjak <ubizjak@gmail.com>
19963
19964 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19965 New array.
19966 (ix86_expand_call): Remove clobbered_registers array and use
19967 x86_64_ms_sysv_extra_clobbered_registers instead.
19968 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19969 Declare here.
19970 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19971 predicate.
19972 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19973 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19974 (*call_value_rex64_ms_sysv): Ditto.
19975
6f5a366a
CC		 199762013-04-18 Cary Coutant <ccoutant@google.com>
19977
19978 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19979 parent instead of die_mark.
19980
475b8f37
DN		 199812013-04-18 Diego Novillo <dnovillo@google.com>
19982
19983 * gimple.c (create_gimple_tmp): New.
19984 (get_expr_type): New.
19985 (build_assign): New.
19986 (build_type_cast): New.
19987 * gimple.h (enum ssa_mode): Define.
19988 (gimple_seq_set_location): New.
19989 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 19990 to use build_assign and build_type_cast.
475b8f37 19991
08940f33
RB		 199922013-04-18 Richard Biener <rguenther@suse.de>
19993
19994 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19995 handle negative step. Remove redundant checks.
19996 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19997 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19998 for negative step and grouped loads fail to vectorize.
19999
0e0f87d4
SB		 200002013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20001
20002 * emit-rtl.c (reset_insn_used_flags): New function.
20003 (reset_all_used_flags): Use it.
20004 (verify_insn_sharing): New function.
20005 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20006
4c445590
JJ		 200072013-04-18 Jakub Jelinek <jakub@redhat.com>
20008
20009 PR tree-optimization/56984
20010 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20011 and (x >> M) >= N don't register any assertion if N << M is the
20012 minimum value.
20013
6873ecab
SB		 200142013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20015
20016 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20017 do not delete_insn insns that have not yet been emitted, only
20018 unlink them with remove_insn.
20019 * df-scan.c (df_insn_delete): Revert r197492.
20020
3ccb989e
SB		 200212013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20022
20023 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20024 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20025
200262013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY		 20027
20028 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20029 define_insn_and_split.
20030 (and_scc,ior_scc,negscc): Likewise.
20031 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20032
3ccb989e 200332013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY		 20034
20035 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20036 can benefit from using a sequence of LDRD instructions in epilogue
20037 instead of a single LDM instruction.
20038
6d10a203
MLI		 200392013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20040
20041 PR 45688
20042 * doc/extend.texi: Fix typo.
20043
6983e6b5
RB		 200442013-04-17 Richard Biener <rguenther@suse.de>
20045
20046 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20047 (vect_build_slp_tree): ... here.
20048 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20049 match. Remove special-casing of mismatched complex loads.
20050 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20051 re-try the match with swapped commutative operands.
20052 (vect_supported_load_permutation_p): Remove special-casing of
20053 mismatched complex loads.
20054 (vect_analyze_slp_instance): Adjust.
20055
ef23e6a2
RB		 200562013-04-17 Richard Biener <rguenther@suse.de>
20057
20058 PR rtl-optimization/56921
20059 * cfgloop.h (struct loop): Add simple_loop_desc member.
20060 (struct niter_desc): Mark with GTY(()).
20061 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20062 * loop-iv.c (get_simple_loop_desc): Likewise.
20063 (free_simple_loop_desc): Likewise.
20064
20065 Revert
20066 2013-04-16 Richard Biener <rguenther@suse.de>
20067
20068 PR rtl-optimization/56921
20069 * loop-init.c (pass_rtl_move_loop_invariants): Add
20070 TODO_do_not_ggc_collect to todo_flags_finish.
20071 (pass_rtl_unswitch): Same.
20072 (pass_rtl_unroll_and_peel_loops): Same.
20073 (pass_rtl_doloop): Same.
20074
fc6f94f5
EB		 200752013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20076
20077 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20078 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20079 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20080 references.
20081 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20082 * tree-streamer.c (record_common_node): Adjust reference in comment.
20083
10a88311
TG		 200842013-04-17 Terry Guo <terry.guo@arm.com>
20085
20086 * config/arm/cortex-m4.md: Add a new bypass.
20087
6d9b7208 200882013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N		 20089
20090 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20091 New pattern.
20092 (*subs_<optab><mode>_multp2): New pattern.
20093 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20094 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20095
6d9b7208 200962013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N		 20097
20098 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20099 (*subs_mul_imm_<mode>): New pattern.
20100
18a6701e
DE		 201012013-04-16 David Edelsohn <dje.gcc@gmail.com>
20102
20103 PR target/56948
20104 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20105 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20106 (vsx_movti_32bit): Change j->wa to O->wa.
20107
07c37b2f
RB		 201082013-04-16 Richard Biener <rguenther@suse.de>
20109
20110 PR rtl-optimization/56921
20111 * loop-init.c (pass_rtl_move_loop_invariants): Add
20112 TODO_do_not_ggc_collect to todo_flags_finish.
20113 (pass_rtl_unswitch): Same.
20114 (pass_rtl_unroll_and_peel_loops): Same.
20115 (pass_rtl_doloop): Same.
20116
0e0f87d4 201172013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY		 20118
20119 * config/arm/arm.c (emit_multi_reg_push): New declaration
20120 for an existing function.
20121 (arm_emit_strd_push): New function.
20122 (arm_expand_prologue): Used here.
20123 (arm_emit_ldrd_pop): New function.
20124 (arm_expand_epilogue): Used here.
20125 (arm_get_frame_offsets): Update condition.
20126 (arm_emit_multi_reg_pop): Add a special case for load of a single
20127 register with writeback.
20128
5e8e2af4
UB		 201292013-04-16 Uros Bizjak <ubizjak@gmail.com>
20130
20131 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20132 description.
20133
9fd9ccf7
RB		 201342013-04-16 Richard Biener <rguenther@suse.de>
20135
20136 PR tree-optimization/56756
20137 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20138 (first_mem_ref_loc): New.
20139 (execute_sm): Place the load temporarily before a previous
20140 access instead of in the latch edge to ensure its SSA dependencies
20141 are defined at points dominating the load.
20142
96fba521
SB		 201432013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20144
4c8af858
SB		 20145 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20146 correct fix by moving header and footer insn to the footer of
20147 the merged basic block. Clear BB_END of the merged-away block.
20148
96fba521
SB		 20149 PR middle-end/43631
20150 * emit-rtl.c (make_note_raw): New function.
20151 (link_insn_into_chain): New static inline function.
20152 (add_insn): Use it.
20153 (add_insn_before, add_insn_after): Factor insn chain linking code...
20154 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20155 using link_insn_into_chain.
20156 (note_outside_basic_block_p): New helper function for emit_note_after
20157 and emit_note_before.
20158 (emit_note_after): Use nobb variant of add_insn_after if the note
20159 should not be contained in a basic block.
20160 (emit_note_before): Use nobb variant of add_insn_before if the note
20161 should not be contained in a basic block.
20162 (emit_note_copy): Use make_note_raw.
20163 (emit_note): Likewise.
20164 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20165 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20166 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20167 the moved barrier the tail of the basic block it follows.
20168 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20169
7b8265ba
JJ		 201702013-04-15 Jakub Jelinek <jakub@redhat.com>
20171
20172 PR tree-optimization/56962
20173 * gimple-ssa-strength-reduction.c (record_increment): Only set
20174 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20175 either rhs1 or rhs2 is equal to c->base_expr.
20176
5185d248
RB		 201772013-04-15 Richard Biener <rguenther@suse.de>
20178
20179 PR tree-optimization/56933
20180 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20181 member.
20182 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20183 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20184 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20185 dependence check ...
20186 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20187 ... here.
20188 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20189 GROUP_READ_WRITE_DEPENDENCE.
20190
a24243a0
AK		 201912013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20192
20193 * emit-rtl.c (reset_all_used_flags): New function.
20194 (verify_rtl_sharing): Call reset_all_used_flags before and after
20195 performing the checks.
20196
1c50eada
KT		 201972013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20198
20199 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20200 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20201 * config/arm/constraints.md (De): New constraint.
20202 * config/arm/neon.md (anddi3_neon): Delete.
20203 (neon_vand<mode>): Expand to standard anddi3 pattern.
20204 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20205 Move earlier in the file.
20206 (neon_inv_logic_op2): Likewise.
20207 (arm_anddi_operand_neon): New predicate.
20208
e927b6ad
RO		 202092013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20210
20211 * configure.ac (gcc_cv_ld_as_needed): Set
20212 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20213 Use -z ignore, -z record on *-*-solaris2*.
20214 (HAVE_LD_AS_NEEDED): Update comment.
20215 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20216 * configure: Regenerate.
20217 * config.in: Regenerate.
20218 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20219 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20220 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20221 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20222 equivalents. Fix markup.
20223 * doc/tm.texi: Regenerate.
20224
e0ea8797
AH		 202252013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20226
20227 * config/i386/i386.opt: New option mstack-protector-guard=.
20228 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20229 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20230 TARGET_SSP_TLS_GUARD.
20231 * config/i386/i386.c (ix86_option_override_internal): Set
20232 ix86_stack_protector_guard.
20233 * config/i386/i386.md (stack_protect_set): Enable for
20234 TARGET_SSP_TLS_GUARD only.
20235 (stack_protect_set_<mode>): Ditto.
20236 (stack_protect_test): Ditto.
20237 (stack_protect_test_<mode>): Ditto.
20238 * doc/invoke.texi (i386 Option): Document.
20239
811b72f9
EB		 202402013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20241
20242 PR target/56890
20243 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20244 (S_MODES): Set H_MODE bit.
20245 (SF_MODES): Set only S_MODE and SF_MODE bits.
20246 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20247 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20248 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20249 <MODE_FLOAT>: Likewise.
20250
5529fdd6
JY		 202512013-04-15 Joey Ye <joey.ye@arm.com>
20252
20253 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20254
517b1da2
JY		 202552013-04-15 Joey Ye <joey.ye@arm.com>
20256
20257 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20258 for real far jump.
20259 (thumb_far_jump_used_p): Count instruction size and set
20260 far_jump_used.
20261
01007ae0
EB		 202622013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20263
20264 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20265 * resource.c (mark_target_live_regs): Reformat conditional block.
20266
c46f6580
SB		 202672013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20268
20269 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20270 notes, they are emitted only just before final.
20271 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20272
1f397f45
SB		 202732013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20274
20275 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20276 * cfgrtl.c (delete_insn): Call it here instead.
20277 * lra-spills.c (lra_final_code_change): Use delete_insn.
20278 * haifa-sched.c (sched_remove_insn): Likewise.
20279 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20280 returning to the nop pool.
20281 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20282 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20283
58a51369
SB		 202842013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20285
20286 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20287 * doc/tm.texi: Regenerated.
20288
33159866
UB		 202892013-04-12 Uros Bizjak <ubizjak@gmail.com>
20290
20291 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20292 QImode checks.
20293
226e378f
SB		 202942013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20295
20296 * df-core.c (df_find_def): Compare register numbers.
20297 (df_find_use): Likewise.
20298
fafb9b18
VM		 202992013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20300
20301 PR target/56903
20302 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20303 lra_in_progress for return.
20304
9a946fd6
GY		 203052013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20306
20307 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20308 define_insn into define_insn_and_split and emit movsicc patterns.
20309
41b83758
GY		 203102013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20311
20312 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20313
d3afd9aa
RB		 203142013-04-12 Richard Biener <rguenther@suse.de>
20315
20316 * tree-pass.h (TODO_do_not_ggc_collect): New.
20317 * passes.c (execute_one_ipa_transform_pass): Honor
20318 TODO_do_not_ggc_collect.
20319 (execute_one_pass): Likewise.
20320
20321 Revert
20322 2013-04-10 Richard Biener <rguenther@suse.de>
20323
20324 * passes.c (init_optimization_passes): Remove reload pass.
20325 * ira.c (do_reload): Merge into ...
20326 (ira): ... this.
20327 (rest_of_handle_reload): Remove.
20328 (pass_reload): Likewise.
20329 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20330 of reload for vzeroupper pass placement.
20331
06f9b387
JJ		 203322013-04-12 Jakub Jelinek <jakub@redhat.com>
20333
20334 PR tree-optimization/56918
20335 PR tree-optimization/56920
20336 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20337 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20338 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20339 use wide_mul_with_sign method.
20340
953094d2
RB		 203412013-04-12 Richard Biener <rguenther@suse.de>
20342
20343 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20344 not be considered a gimple constant.
20345
bb506982
MG		 203462013-04-12 Marc Glisse <marc.glisse@inria.fr>
20347
20348 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20349 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20350 operations.
20351
4b84d650
JJ		 203522013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20353 Jakub Jelinek <jakub@redhat.com>
20354
20355 * opts.c: Include diagnostic-color.h.
20356 (common_handle_option): Handle OPT_fdiagnostics_color_.
20357 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20358 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20359 (diagnostic-color.o): New.
20360 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20361 (diagnostic_color_rule): New enum.
20362 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20363 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20364 the location string.
20365 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20366 either NULL, or color kind.
20367 * diagnostic-color.c: New file.
20368 * diagnostic-color.h: New file.
20369 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20370 arguments.
20371 * doc/invoke.texi (-fdiagnostics-color): Document.
20372 * pretty-print.h (pp_show_color): Define.
20373 (struct pretty_print_info): Add show_color field.
20374 * diagnostic.c: Include diagnostic-color.h.
20375 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20376 macros. Colorize error:, warning: etc. strings and also the location
20377 string.
20378 (diagnostic_show_locus): Colorize the caret line.
20379 * pretty-print.c: Include diagnostic-color.h.
20380 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20381 inside of %< %> quotes or quoted through q format modifier.
20382
067a1e71
AK		 203832013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20384
33159866 20385 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 20386
33159866 203872013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N		 20388
20389 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20390 code in CC_NZ mode.
20391 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20392 pattern.
20393
7b55f98f
MP		 203942013-04-11 Marek Polacek <polacek@redhat.com>
20395
20396 PR tree-optimization/48184
33159866 20397 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 20398
966b587e
EB		 203992013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20400
20401 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20402 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20403 (skip_simple_arithmetic): Tidy up.
20404 * tree.h (skip_simple_constant_arithmetic): Declare.
20405
33159866 204062013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N		 20407
20408 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20409
1d42c1ec
RB		 204102013-04-11 Richard Biener <rguenther@suse.de>
20411
20412 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20413 generate vector constants.
20414
4ba5ea11
RB		 204152013-04-11 Richard Biener <rguenther@suse.de>
20416
20417 PR tree-optimization/56878
20418 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20419 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20420 New function.
20421 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20422 Prefer to align the DR with the most invariant base address.
20423
f0defe58
SKS		 204242013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20425
20426 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20427 comment.
20428
d07458be
JG		 204292013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20430
20431 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20432 floating-point vector comparisons against 0.
20433
146b8692
JJ		 204342013-04-11 Jakub Jelinek <jakub@redhat.com>
20435
20436 PR tree-optimization/56899
20437 * fold-const.c (extract_muldiv_1): Apply distributive law
20438 only if TYPE_OVERFLOW_WRAPS (ctype).
20439
b8578ff7
BC		 204402013-04-11 Bin Cheng <bin.cheng@arm.com>
20441
20442 PR target/56124
20443 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20444 loading has side effect.
20445
0ea8a6f9
SB		 204462013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20447
20448 * config/sparc/sparc.c: Include tree-pass.h.
20449 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20450 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20451 head of file. Change return type. Split off gate function.
20452 (sparc_gate_work_around_errata): New function.
20453 (pass_work_around_errata): New pass definition.
20454 (insert_pass_work_around_errata) New pass insert definition to
20455 insert pass_work_around_errata just after delayed-branch scheduling.
20456 (sparc_option_override): Insert the pass.
20457 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20458
42e37616
DM		 204592013-04-10 David S. Miller <davem@davemloft.net>
20460
89deeb3b
DM		 20461 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20462 or -mcpu=hypersparc.
20463
42e37616
DM		 20464 * target.def (cstore_mode): New hook.
20465 * target.h: Include insn-codes.h
20466 * targhooks.c: Likewise.
20467 (default_cstore_mode): New function.
20468 * targhooks.h: Declare it.
20469 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20470 * doc/tm.texi: Rebuild.
20471 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20472 target hook, rather than inspecting the insn_data.
20473 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20474 (TARGET_CSTORE_MODE): Redefine.
20475 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20476 result patterns.
20477 * config/sparc/predicates.md (cstore_result_operand): New special
20478 predicate.
20479 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20480 Use it for operand 0.
20481 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20482 (*snesi_special): Likewise.
20483 (*snesi_zero): Likewise.
20484 (*seqsi_zero): Likewise.
20485 (*sltu_insn): Likewise.
20486 (*sgeu_insn): Likewise.
20487 (*seqdi_special): Make operand 0 and comparison operation be of
20488 DImode.
20489 (*snedi_special): Likewise.
20490 (*snedi_special_vis3): Likewise.
20491 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20492 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20493 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20494 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20495 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20496 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20497 (*sltu_extend_sp64): Likewise.
20498 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20499 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20500 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20501 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20502 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20503
95ca411e
YZ		 205042013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20505
20506 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20507 (aarch64_start_file): Use the new function.
20508
6782438d 205092013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 20510 Jason Merrill <jason@redhat.com>
6782438d
SKS		 20511
20512 * common.opt: Add -gdwarf.
20513 * opts.c (common_handle_option): Handle it.
20514 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20515
bb313b93
RB		 205162013-04-10 Richard Biener <rguenther@suse.de>
20517
20518 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20519 (execute_one_ipa_transform_pass): But unconditionally here.
20520 (execute_one_pass): And here.
20521 (init_optimization_passes): Remove reload pass.
20522 * tree-pass.h (TODO_ggc_collect): Remove.
20523 (pass_reload): Likewise.
20524 * ira.c (do_reload): Merge into ...
20525 (ira): ... this.
20526 (rest_of_handle_reload): Remove.
20527 (pass_reload): Likewise.
20528 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20529 of reload for vzeroupper pass placement.
20530 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20531 and todo_flags_finish of all passes.
20532
793d9a16
RB		 205332013-04-10 Richard Biener <rguenther@suse.de>
20534
20535 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20536 first_const_oprnd field, rename first_def_type to first_op_type.
20537 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20538 (vect_get_and_check_slp_defs): Always use the type of the
20539 operand. Allow mixed vect_external_def, vect_constant_def types.
20540 (vect_get_constant_vectors): Handle mixed vect_external_def,
20541 vect_constant_def types.
20542
12211b99 205432013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR		 20544
20545 PR tree-optimization/55524
20546 * tree-ssa-math-opts.c
20547 (convert_mult_to_fma): Don't use an fms construct
20548 when we don't have an fms operation, but fnma, and it looks
20549 likely that we'll be able to use the latter.
20550
12211b99 205512013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ		 20552
20553 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20554 function.
20555 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20556 inline fail caused by overwritable functions.
20557
34ab4a5b
CJW		 205582013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20559
20560 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20561 unnecessary bits in the constant power of two case.
20562
abf9bfbc
RB		 205632013-04-10 Richard Biener <rguenther@suse.de>
20564
20565 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20566 broken code swapping operands.
20567 (vect_build_slp_tree): Do not compute load permutations here.
20568 (vect_analyze_slp_instance): Compute load permutations here,
20569 after building the SLP tree.
20570
f408477e
CB		 205712013-04-09 Christian Bruel <christian.bruel@st.com>
20572
20573 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20574 of next/prev_real_insn.
20575
5ac42672
JH		 205762013-04-09 Jan Hubicka <jh@suse.cz>
20577
abf9bfbc
RB		 20578 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20579 Drop aliased parameter.
5ac42672
JH		 20580 (function_and_variable_visibility): Do not handle alias pairs.
20581 * cgraph.c (varpool_externally_visible_p): Update prototype.
20582 * varpool.c (varpool_add_new_variable): Update.
20583
5017f1d2
KT		 205842013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20585
20586 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20587
48eecbee
SB		 205882013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20589
a949cf1c
SB		 20590 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20591
48eecbee
SB		 20592 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20593
75ef8e3d
MP		 205942013-04-09 Marek Polacek <polacek@redhat.com>
20595
20596 PR tree-optimization/48762
33159866 20597 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 20598
23847df4
RB		 205992013-04-09 Richard Biener <rguenther@suse.de>
20600
20601 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20602 dealing with cost.
20603 (vect_build_slp_tree): Likewise.
20604 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20605 calculating the cost of a SLP instance.
20606 (vect_analyze_slp_instance): Use it from here, after building
20607 the SLP tree.
20608
ea3a0fde
JJ		 206092013-04-09 Jakub Jelinek <jakub@redhat.com>
20610
20611 PR middle-end/56883
20612 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20613 expand_omp_for_static_chunk): Use simple_p = true in
20614 force_gimple_operand_gsi calls when assigning to addressable decls.
20615
a32dfe9d
JL		 206162013-04-09 Jeff Law <law@redhat.com>
20617
20618 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20619 when the boolean was created by converting a wider object which
20620 had a boolean range.
20621
d755c7ef
RB		 206222013-04-09 Richard Biener <rguenther@suse.de>
20623
20624 * tree-vectorizer.h (slp_void_p): Remove.
20625 (slp_tree): Typedef before _slp_tree declaration.
20626 (struct _slp_tree): Use a vector of slp_tree as children.
20627 (vect_get_place_in_interleaving_chain): Remove.
20628 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20629 Move ...
20630 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20631 and make static.
20632 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20633 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20634 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20635 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20636 Use slp_node instead of slp_void_p and adjust.
20637
3d741091
RB		 206382013-04-09 Richard Biener <rguenther@suse.de>
20639
20640 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20641 work that is not necessary.
20642
39307ba7
JJ		 206432013-04-09 Jakub Jelinek <jakub@redhat.com>
20644
20645 PR tree-optimization/56854
20646 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20647 forward into clobber stmts if it would change MEM_REF lhs into
20648 non-MEM_REF.
20649
343881fd
MK		 206502013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20651
20652 * tree.c (type_hash_lookup, type_hash_add): Make static.
20653 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20654
3922658a
RB		 206552013-04-09 Richard Biener <rguenther@suse.de>
20656
20657 * tree.h (unsave_expr_now): Remove.
20658 * tree-inline.c (mark_local_for_remap_r): Remove.
20659 (unsave_expr_1): Likewise.
20660 (unsave_r): Likewise.
20661 (unsave_expr_now): Likewise.
20662 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20663 (propagate_tree_value): Likewise.
20664
9fb6b620
SB		 206652013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20666
20667 * doc/rtl.texi (sequence): Rewrite documentation to match the
20668 current use of SEQUENCE rtl objects.
20669 * rtl.def (SEQUENCE): Likewise.
20670
20671 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20672 Update documentation.
20673 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20674 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20675
20676 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20677
8ddb5a29
TJ		 206782013-04-08 Teresa Johnson <tejohnson@google.com>
20679
20680 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20681 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 20682 methods.
8ddb5a29
TJ		 20683 (estimate_edge_size_and_time): Add comment to suggest using rounding
20684 methods.
20685 (estimate_node_size_and_time): Ditto.
20686 (remap_edge_change_prob): Use helper rounding divide methods.
20687 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20688 (gimple_mod_pow2_value_transform): Ditto.
20689 (gimple_mod_subtract_transform): Ditto.
20690 (gimple_ic_transform): Ditto.
20691 (gimple_stringops_transform): Ditto.
20692 * stmt.c (conditional_probability): Ditto.
20693 (emit_case_dispatch_table): Ditto.
20694 * lto-cgraph.c (merge_profile_summaries): Ditto.
20695 * tree-optimize.c (execute_fixup_cfg): Ditto.
20696 * cfgcleanup.c (try_forward_edges): Ditto.
20697 * cfgloopmanip.c (scale_loop_profile): Ditto.
20698 (loopify): Ditto.
20699 (duplicate_loop_to_header_edge): Ditto.
20700 (lv_adjust_loop_entry_edge): Ditto.
20701 * tree-vect-loop.c (vect_transform_loop): Ditto.
20702 * profile.c (compute_branch_probabilities): Ditto.
20703 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20704 * lto-streamer-in.c (input_cfg): Ditto.
20705 * gimple-streamer-in.c (input_bb): Ditto.
20706 * ipa-cp.c (update_profiling_info): Ditto.
20707 (update_specialized_profile): Ditto.
20708 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20709 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 20710 rounding methods.
8ddb5a29
TJ		 20711 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20712 (compute_trg_info): Ditto.
20713 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20714 (purge_dead_edges): Ditto.
20715 * loop-unswitch.c (unswitch_loop): Ditto.
20716 * cgraphclones.c (cgraph_clone_edge): Ditto.
20717 (cgraph_clone_node): Ditto.
20718 * tree-inline.c (copy_bb): Ditto.
20719 (copy_edges_for_bb): Ditto.
20720 (initialize_cfun): Ditto.
20721 (copy_cfg_body): Ditto.
20722 (expand_call_inline): Ditto.
20723
661e6bd7
KT		 207242013-04-08 Kai Tietz <ktietz@redhat.com>
20725
20726 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20727 TARGET_CYGWIN64 by TARGET_64BIT.
20728
105766f3
JR		 207292013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20730
20731 * config/epiphany/epiphany.md (GPR_1): New constant.
20732 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20733 * config/epiphany/epiphany.c (gen_compare_reg):
20734 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20735 is already in place.
20736 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20737 Don't require being called during rtl expansion; If y operlaps r0,
20738 return 0.
20739 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20740 (epiphany_expand_epilogue): Likewise.
20741
8afab237
JR		 20742 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20743 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 20744 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 20745
fa7707d6
JR		 20746 * config/epiphany/constraints.md (CnL): New constraint.
20747 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20748 * config/epiphany/predicates.md (add_operand): Allow 1024.
20749
5138e18d
JR		 20750 * config/epiphany/epiphany.md (logical_op): New code iterator.
20751 (op_mnc): New code attribute.
20752 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20753 (mov_f+1, mov_f+2): New peephole2 patterns.
20754
2ccc703d
JR		 20755 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20756 (cstoresi4): Also allow re-use of zero result when doing a NE
20757 comparison to a non-zero operand.
aefb0819 20758 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 20759
093ac0a5
JR		 20760 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20761 Use gen_addsi3_i / gen_subsi3_i.
20762
f223bb13
JJ		 207632013-04-08 Jakub Jelinek <jakub@redhat.com>
20764
20765 PR c++/34949
20766 PR c++/50243
20767 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20768 contain anything but clobbers, at most one __builtin_stack_restore,
20769 optionally debug stmts and final resx, and if it has at least one
20770 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20771 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20772 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20773 which isn't defaut definition, remove them.
20774 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20775 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20776 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20777 with MEM_REF LHS with SSA_NAME address.
20778
4481581f
JL		 207792013-04-08 Jeff Law <law@redhat.com>
20780
20781 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 20782
451dabda
RB		 207832013-04-08 Richard Biener <rguenther@suse.de>
20784
20785 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20786 extra newline.
20787 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20788 determined vector type.
20789 (vect_analyze_data_refs): Likewise.
20790 (vect_get_new_vect_var): Adjust.
20791 (vect_create_destination_var): Preserve SSA name versions.
20792 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20793 not dump anything here.
20794
3b088b47
JR		 207952013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20796
20797 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20798 Add member lr_slot_known.
20799 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20800 if necessary.
20801 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20802 Remove code that sets lr_slot_offset according to what a previous
20803 version of epiphany_emit_save_restore used to do.
20804 (epiphany_emit_save_restore): When doing an lr save or restore,
20805 set/verify lr_slot_known and lr_slot_offset.
20806
d8484d41
XQ		 208072013-04-08 Xinyu Qi <xyqi@marvell.com>
20808
33159866
UB		 20809 PR target/54338
20810 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ		 20811 in ALL_REGS.
20812
481be1c4
RB		 208132013-04-08 Richard Biener <rguenther@suse.de>
20814
20815 * alias.c (find_base_term): Fix thinko in previous change.
20816
401f3a81
JJ		 208172013-04-08 Jakub Jelinek <jakub@redhat.com>
20818
20819 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20820 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20821 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20822 if possible to compute val.
20823 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20824 For QImode integers don't require anything about precision. Use
20825 const_with_all_bytes_same to find out if the constant doesn't have
20826 repeated bytes in it.
20827
03ed99a8
AK		 208282013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20829
20830 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20831 within mode size.
20832
781b2e62
MP		 208332013-04-08 Marek Polacek <polacek@redhat.com>
20834
20835 PR rtl-optimization/48182
20836 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20837 value to 1.
20838
27e430a2
JDA		 208392013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20840
20841 PR target/55487
20842 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20843 nuses, make sure we have a label.
20844
4902aa64
BS		 208452013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20846
20847 PR target/56843
20848 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20849 (rs6000_emit_swdiv_low_precision): Remove.
20850 (rs6000_emit_swdiv): Rewrite to handle between one and four
20851 iterations of Newton-Raphson generally; modify required number of
20852 iterations for some cases.
20853 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20854
7bca81dc
SB		 208552013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20856
20857 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20858 set-but-unused variable.
20859
20860 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20861 basic blocks of released function bodies garbage-collectable.
20862
20863 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20864 (struct rtl_opt_pass): Add TODO_df_finish.
20865
20866 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20867
4542a38a
GY		 208682013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20869
20870 * config/arm/constraints.md (q): New constraint.
20871 * config/arm/ldrdstrd.md: New file.
20872 * config/arm/arm.md (ldrdstrd.md) New include.
20873 (arm_movdi): Use "q" instead of "r" constraint
20874 for double-word memory access.
20875 (movdf_soft_insn): Likewise.
20876 * config/arm/vfp.md (movdi_vfp): Likewise.
20877 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 20878 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY		 20879 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20880 (mem_ok_for_ldrd_strd): Likewise.
20881 (output_move_double): Update assertion.
20882
2385b218
GY		 208832013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20884
20885 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20886
75fe1cb5
GY		 208872013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20888
20889 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20890 define_insn_and_split.
20891 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20892
dd660e8e
GY		 208932013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20894
20895 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20896 define_insn_and_split.
33159866 20897 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY		 20898 (shiftsi3_compare): New pattern.
20899 (rrx): New pattern.
20900 * config/arm/unspecs.md (UNSPEC_RRX): New.
20901
045e472c
GY		 209022013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20903
20904 * config/arm/arm.md (negdi_extendsidi): New pattern.
20905 (negdi_zero_extendsidi): Likewise.
20906
3f3bf1a8
GY		 209072013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20908
20909 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20910 define_insn_and_split.
20911 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20912 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20913
b0b49556
GY		 209142013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20915
20916 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20917 define_insn_and_split.
20918 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20919 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20920
d633dd84
GY		 209212013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20922
20923 * config/arm/arm.md (subsi3_carryin): New pattern.
20924 (subsi3_carryin_const): Likewise.
20925 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20926 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20927
f4499066
GY		 209282013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20929
20930 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20931
ceef6fd9
GY		 209322013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20933
20934 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 20935 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 20936
ddbdd8a7
KT		 209372013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20938
20939 * config/arm/arm.c (arm_expand_builtin): Change fcode
20940 type to unsigned int.
20941
8456d78a
RR		 209422013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20943
20944 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20945
526c230f
ILT		 209462013-04-04 Ian Lance Taylor <iant@google.com>
20947
33159866
UB		 20948 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20949 language standard.
526c230f 20950
3cfbe04d
SB		 209512013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20952
20953 PR middle-end/56729
20954 * df-scan.c (df_insn_delete): Disable failing assert.
20955
dfa3f8d0
KT		 209562013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20957
20958 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20959 New function prototype.
20960 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20961 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20962 (arm_builtin_vectorized_function): New function.
20963
f7837758
KT		 209642013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20965
20966 * config/arm/arm_neon_builtins.def: New file.
20967 * config/arm/arm.c (neon_builtin_data): Move contents to
20968 arm_neon_builtins.def.
20969 (enum arm_builtins): Include neon builtin definitions.
20970 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 20971 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 20972
39c1b6db
MP		 209732013-04-04 Marek Polacek <polacek@redhat.com>
20974
20975 PR tree-optimization/48186
20976 * predict.c (maybe_hot_frequency_p): Return false if
20977 HOT_BB_FREQUENCY_FRACTION is 0.
20978 (cgraph_maybe_hot_edge_p): Likewise.
20979
314f64eb
RB		 209802013-04-04 Richard Biener <rguenther@suse.de>
20981
20982 PR tree-optimization/56826
20983 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20984 more accurately.
20985
51a905b2
RB		 209862013-04-04 Richard Biener <rguenther@suse.de>
20987
20988 PR tree-optimization/56213
20989 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 20990 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 20991
f1bf4f3a
RB		 209922013-04-04 Richard Biener <rguenther@suse.de>
20993
20994 PR tree-optimization/56837
20995 * tree-loop-distribution.c (classify_partition): For non-zero
20996 values require that the value has the same precision as its
20997 mode to be useful as memset value.
20998
0bca7ded
NC		 209992013-04-03 Nick Clifton <nickc@redhat.com>
21000
33159866 21001 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC		 21002 (fmssf4): Use fmsf.s on E3V5 architectures.
21003 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21004 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21005
b4019227
JL		 210062013-04-03 Jeff Law <law@redhat.com>
21007
21008 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21009 (lra-eliminations.o): Likewise.
21010
f57ddb5b
TJ		 210112013-04-03 Teresa Johnson <tejohnson@google.com>
21012
21013 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 21014 compute_working_sets here from profile.c.
f57ddb5b 21015 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB		 21016 (gcov_working_set_t): Moved typedef here from basic-block.h
21017 (compute_working_set): Declare.
f57ddb5b
TJ		 21018 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21019 (get_working_sets): Renamed from compute_working_set,
33159866 21020 replace most of body with call to new compute_working_sets.
f57ddb5b 21021 (get_exec_counts): Replace call to compute_working_sets
33159866
UB		 21022 to get_working_sets.
21023 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 21024 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 21025 to get_working_sets.
f57ddb5b
TJ		 21026 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21027 * gcov-dump.c (dump_working_sets): New function.
21028
12211b99 210292013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ		 21030
21031 * hwint.c (sext_hwi, zext_hwi): New functions.
21032 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21033 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21034 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21035 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21036 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21037 (sext_hwi, zext_hwi): New functions.
0bca7ded 21038
be672e08
JL		 210392013-04-03 Jeff Law <law@redhat.com>
21040
21041 PR tree-optimization/56799
33159866
UB		 21042 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21043 back test for widening conversion erroneously dropped in prior change.
be672e08 21044
9d821fa5
KT		 210452013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21046
21047 PR target/56809
21048 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21049 instead of next_real_insn.
21050
71c581e7
MP		 210512013-04-03 Marek Polacek <polacek@redhat.com>
21052
21053 PR sanitizer/55702
33159866 21054 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 21055
4a32ef80
KT		 210562013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21057
21058 PR target/56809
21059 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21060 next_real_insn.
21061 (thumb1_output_casesi): Likewise.
21062 (thumb2_output_casesi): Likewise.
21063
1b2253d4
RB		 210642013-04-03 Richard Biener <rguenther@suse.de>
21065
21066 PR tree-optimization/56817
21067 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21068 Split out ...
21069 (tree_unroll_loops_completely_1): ... new function to manually
21070 walk the loop tree, properly defering outer loops of unrolled
21071 loops to later iterations.
21072
38000232
MG		 210732013-04-03 Marc Glisse <marc.glisse@inria.fr>
21074
21075 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21076 (vectorizable_load): Likewise.
21077 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21078 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21079
3fa35298
MG		 210802013-04-03 Marc Glisse <marc.glisse@inria.fr>
21081
21082 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21083 BIT_FIELD_REF.
21084
b3d45ff0
UW		 210852013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21086
21087 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21088
ec9202a8
BC		 210892013-04-03 Bin Cheng <bin.cheng@arm.com>
21090
21091 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21092
6805bd36
MG		 210932013-04-03 Marc Glisse <marc.glisse@inria.fr>
21094
21095 PR tree-optimization/56790
33159866
UB		 21096 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21097 folding.
6805bd36 21098
da694a77
MG		 210992013-04-03 Marc Glisse <marc.glisse@inria.fr>
21100
21101 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21102 Handle VEC_MERGE.
21103 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21104 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21105 equal arguments.
21106
4e7d7b3d
JJ		 211072013-04-03 Jakub Jelinek <jakub@redhat.com>
21108
21109 PR c/19449
21110 * tree.h (force_folding_builtin_constant_p): New decl.
21111 * builtins.c (force_folding_builtin_constant_p): New variable.
21112 (fold_builtin_constant_p): Fold immediately also if
21113 force_folding_builtin_constant_p.
21114
e6c9d234
RB		 211152013-04-03 Richard Biener <rguenther@suse.de>
21116
21117 PR tree-optimization/56812
21118 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21119 DRs of the same interleaving chain are independent.
21120
984d07dd
JM		 211212013-04-02 Jason Merrill <jason@redhat.com>
21122
21123 * gdbinit.in (pbb): Use debug fn.
21124
622849c9
LC		 211252013-04-02 Lawrence Crowl <crowl@google.com>
21126
21127 * sese.h (struct ivtype_map_elt_s): Remove unused.
21128 (extern debug_ivtype_map): Remove unused.
21129 (extern eq_ivtype_map_elts): Remove unused.
21130 * sese.c (debug_ivtype_map): Removed unused.
21131 (debug_ivtype_map_1): Removed unused.
21132 (debug_ivtype_elt): Remove unused.
21133 (eq_ivtype_map_elts): Remove unused.
21134
21135
82c0e1a0
KT		 211362013-04-02 Kai Tietz <ktietz@redhat.com>
21137
21138 PR target/52790
21139 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21140 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB		 21141 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21142 function.
82c0e1a0
KT		 21143 (legitimize_pe_coff_symbol): Likewise.
21144 (is_imported_p): New helper-function.
21145 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21146 for Windows x64 targets.
21147 (ix86_expand_prologue): Optimize for pe-coff targets.
21148 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21149 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21150 medium/large code-model.
21151 (legitimize_pic_address): Likewise.
21152 (legitimize_tls_address): Likewise.
21153 (ix86_expand_call): Likewise.
21154 (x86_output_mi_thunk): Likewise.
21155 (get_dllimport_decl): Add new beimport argument.
21156 (construct_plt_address): Don't assert for x64 pe-coff targets.
21157 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21158 targets.
21159 (SYMBOL_FLAG_STUBVAR): New macro.
21160 (SYMBOL_REF_STUBVAR_P): Likewise.
21161 * config/i386/winnt.c (stub_list): New structure.
21162 (stub_head): New local variable.
21163 (i386_pe_record_stub): New function.
21164 (i386_pe_file_end): Emit refptr-stubs.
21165
5d751b0c
JJ		 211662013-04-02 Jakub Jelinek <jakub@redhat.com>
21167
09bb4c99
JJ		 21168 PR rtl-optimization/56745
21169 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21170 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21171
a1d8947a
JJ		 21172 PR c++/34949
21173 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21174 and both of them are MEM_REFs, just compare first argument for
21175 equality and attempt to deal even with differing offsets.
21176
5d751b0c
JJ		 21177 PR c++/34949
21178 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21179 of gimple_clobber_p to be MEM_REF.
21180 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21181 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21182 after gimplification.
21183 * asan.c (get_mem_ref_of_assignment): Don't instrument
21184 gimple_clobber_p stmts.
21185 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21186 gimple_clobber_p stmt if they have MEM_REF lhs and
21187 are dead because of another gimple_clobber_p stmt.
21188 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21189 gimple_clobber_p stmts like debug stmts.
21190 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21191 that refer to unused VAR_DECLs or uninitialized values.
21192 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21193 gimple_clobber_p stmts if they refer to removed parameters.
21194 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21195 formatting.
21196
e594716a
UB		 211972013-04-02 Uros Bizjak <ubizjak@gmail.com>
21198
21199 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21200 using SWI48 mode attribute.
21201
7a80735b
WM		 212022013-04-02 Wei Mi <wmi@google.com>
21203
21204 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21205 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21206 *<rotate_insn><mode>3_mask in i386.md.
21207
f423a9e4
AI		 212082013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21209
21210 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21211
90eb75f2
RB		 212122013-04-02 Richard Biener <rguenther@suse.de>
21213
21214 PR tree-optimization/56778
21215 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21216 Runtime alias tests are not supported for gather loads.
21217 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21218 stmts referenced from SSA operands before updating SSA form.
21219
d8c69a92
IC		 212202013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21221 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21222
21223 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB		 21224 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21225 * config/arm/cortex-a53.md: New file.
21226 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21227 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21228 * config/arm/arm.c (arm_issue_rate): Likewise.
21229 * config/arm/arm-tune.md: Regenerate
21230 * config/arm/arm-tables.opt: Regenerate.
21231 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 21232
239eb04c
ZC		 212332013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21234
21235 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21236 non-static link.
21237
c902d3c8
SN		 212382013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21239
d8c69a92
IC		 21240 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21241 scalar load/store operations using B/H registers.
c902d3c8
SN		 21242 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21243
051d0e2f
SN		 212442013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21245
21246 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21247 scalar move.
21248 * config/aarch64/aarch64.c
21249 (aarch64_simd_scalar_immediate_valid_for_move): New.
21250 * config/aarch64/aarch64-protos.h
21251 (aarch64_simd_scalar_immediate_valid_for_move): New.
21252 * config/aarch64/constraints.md (Dh, Dq): New.
21253 * config/aarch64/iterators.md (hq): New.
21254
0ee1e3d9
EB		 212552013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21256
21257 * reorg.c (get_branch_condition): Deal with conditional returns.
21258 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21259
136fb3f7
WM		 212602013-04-01 Wei Mi <wmi@google.com>
21261
21262 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21263 Truncate operand 2 using %b asm operand modifier.
21264 (*<shift_insn><mode>3_mask): Ditto.
21265 (*<rotate_insn><mode>3_mask): Ditto.
21266
6388c738
SB		 212672013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21268
21269 PR middle-end/56798
21270 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21271
f7a4d826
KK		 212722013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21273
21274 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21275 of next_real_insn.
21276 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21277
4b943a49
LC		 212782013-03-30 Lawrence Crowl <crowl@google.com>
21279
21280 * dse.c (clear_alias_sets): Remove never set.
21281 (disqualified_clear_alias_sets): Remove never set.
21282 (clear_alias_mode_pool): Remove never set.
21283 (dse_step0): Remove condition that is never true.
21284 (canon_address): Remove condition that is never true.
21285 (dse_step7): Remove condition that is never true.
21286 (rest_of_handle_dse): Remove condition that is never true.
21287 (rest_of_handle_dse::did_global): Remove never read from above.
21288 (dse_step2_spill): Remove never called from above.
21289 (dse_step5_spill): Remove never called from above.
21290
39718607
SB		 212912013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21292
da5c6bde
SB		 21293 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21294 JUMP_TABLE_DATA changes.
21295 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21296 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21297 (Insns) <jump_table_data>: New entry.
21298 * doc/tm.texi: Regenerate.
21299
39718607
SB		 21300 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21301
21302 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21303 for table jump at the end of a basic block using tablejump_p.
21304 * targhooks.c (default_invalid_within_doloop): Likewise.
21305 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21306 target hook implementation that is identical to the default hook.
21307 (rs6000_invalid_within_doloop): Remove.
21308
21309 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21310 unused variable from tablejump_p call.
21311
21312 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21313 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21314 (INSN_DELETED_P): Likewise.
21315 (emit_jump_table_data): New prototype.
21316 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21317 after 4th as unused.
21318 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21319 * sched-vis.c (print_insn): Likewise.
21320 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21321 insn for compatibility with back ends that use next_active_insn to
21322 identify jump table data.
21323 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21324 (remove_insn): Likewise.
21325 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21326 to be emitted.
21327 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21328 (emit_jump_table_data): New function.
21329
21330 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21331 basic block, a JUMP_TABLE_DATA never is.
21332 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21333 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21334 off from code handling real insns.
21335 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21336 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21337 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21338 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21339 is not a NONDEBUG_INSN_P.
21340 * ira-costs.c (scan_one_insn): Likewise.
21341 * jump.c (mark_all_labels): Likewise.
21342 (mark_jump_label_1): Likewise.
21343 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21344 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21345 a basic block.
21346 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21347 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21348 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21349 JUMP_TABLE_DATA_P insns.
21350 (calculate_elim_costs_all_insns): Likewise.
21351 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21352 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21353 (delete_output_reload): Code style fixups.
21354 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21355 insn flags on this non-insn.
21356 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21357 as scheduling barriers, for pre-change compatibility.
21358 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21359 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21360
21361 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21362 redundant JUMP_TABLE_DATA_P test.
21363 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21364 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21365 (frv_for_each_packet): Likewise.
21366 * config/i386/i386.c (min_insn_size): Likewise.
21367 (ix86_avoid_jump_mispredicts): Likewise.
21368 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21369 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21370 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21371 (mips16_insn_length): Robustify.
21372 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21373 (mips16_split_long_branches): Likewise.
21374 * config/pa/pa.c (pa_combine_instructions): Likewise.
21375 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21376 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21377 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21378 as contributing to pool range lengths.
21379 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21380 Remove redundant JUMP_TABLE_DATA_P test.
21381 (sh_loop_align): Likewise.
21382 (split_branches): Likewise.
21383 (sh_insn_length_adjustment): Likewise.
21384 * config/spu/spu.c (get_branch_target): Likewise.
21385
0208f7da
JH		 213862013-03-29 Jan Hubicka <jh@suse.cz>
21387
21388 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21389 gcov streaming; stream hot bb threshold to ltrans.
21390 * predict.c (get_hot_bb_threshold): Break out from ....
21391 (maybe_hot_count_p): ... here.
21392 (set_hot_bb_threshold): New function.
21393 * lto-section-in.c (lto_section_name): Add profile.
21394 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21395 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21396 and data-streamer.h
21397 (histogram_entry): New structure.
21398 (histogram, histogram_pool): New global vars.
21399 (histogram_hash): New structure.
21400 (histogram_hash::hash): New method.
21401 (histogram_hash::equal): Likewise.
21402 (account_time_size): New function.
21403 (cmp_counts): New function.
21404 (dump_histogram): New function.
21405 (ipa_profile_generate_summary): New function.
21406 (ipa_profile_write_summary): New function.
21407 (ipa_profile_read_summary): New function.
21408 (ipa_profile): Decide on threshold.
e594716a
UB		 21409 (pass_ipa_profile): Add ipa_profile_write_summary and
21410 ipa_profile_read_summary.
0208f7da
JH		 21411 * Makefile.in (ipa.o): Update dependencies.
21412 * lto-streamer.h (LTO_section_ipa_profile): New section.
21413
5a6ccc94
GDR		 214142013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21415
21416 * tree.h (VAR_P): New.
21417
39385fa6
PC		 214182013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21419
21420 PR lto/56777
21421 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21422
34f0d87a
SB		 214232013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21424
21425 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21426 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21427 (control_flow_insn_p): Likewise.
21428 * cfgrtl.c (duplicate_insn_chain): Likewise.
21429 * final.c (get_attr_length_1): Likewise.
21430 (shorten_branches): Likewise.
21431 (final_scan_insn): Likewise.
21432 * function.c (instantiate_virtual_regs): Likewise.
21433 * gcse.c (insert_insn_end_basic_block): Likewise.
21434 * ira-costs.c (scan_one_insn): Likewise.
21435 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21436 * lra.c (check_rtl): Likewise.
21437 * reload1.c (elimination_costs_in_insn): Likewise.
21438 * reorg.c (follow_jumps): Likewise.
21439
21440 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21441 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21442 (thumb_far_jump_used_p): Likewise.
21443 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21444 (workaround_speculation): Likewise.
21445 (add_sched_insns_for_speculation): Likewise.
21446 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21447 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21448 (frv_for_each_packet): Likewise.
21449 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21450 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21451 (final_emit_insn_group_barriers): Likewise.
21452 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21453 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21454 (mips16_insn_length): Likewise.
21455 * config/pa/pa.c (pa_reorg): Likewise.
21456 (pa_combine_instructions): Likewise.
21457 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21458 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21459 (sh_reorg): Likewise.
21460 (split_branches): Likewise.
21461 * config/spu/spu.c (get_branch_target): Likewise.
21462
21463 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21464 JUMP_TABLE_DATA_P.
21465
4ac761b0
KY		 214662013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21467
39385fa6 21468 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY		 21469 Fix declaration name.
21470
58a49189
LC		 214712013-03-28 Lawrence Crowl <crowl@google.com>
21472
21473 * graphds.h (struct graph.indicies): Remove unused.
21474 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21475 (SCOP_ORIGINAL_PDDRS): Remove unused.
21476 * sese.h (extern insert_loop_close_phis): Removed unused.
21477 (extern insert_guard_phis): Removed unused.
21478 (extern ivtype_map_elt_info): Removed unused.
21479 (new_ivtype_map_elt): Removed unused.
21480 * sese.c (ivtype_map_elt_info): Removed unused.
21481
7b3b6ae4
LC		 214822013-03-28 Lawrence Crowl <crowl@google.com>
21483
21484 * Makefile.in: Add several missing include dependences.
21485 (DUMPFILE_H): New.
21486 (test-dump.o): New. This object is not added to any executable,
21487 but is present for ad-hoc testing.
21488 * bitmap.c
21489 (debug (const bitmap_head_def &)): New.
21490 (debug (const bitmap_head_def *)): New.
21491 * bitmap.h
21492 (extern debug (const bitmap_head_def &)): New.
21493 (extern debug (const bitmap_head_def *)): New.
21494 * cfg.c
21495 (debug (edge_def &)): New.
21496 (debug (edge_def *)): New.
21497 * cfghooks.c
21498 (debug (basic_block_def &)): New.
21499 (debug (basic_block_def *)): New.
21500 * dumpfile.h
21501 (dump_node (const_tree, int, FILE *)): Correct source file.
21502 * dwarf2out.c
21503 (debug (die_struct &)): New.
21504 (debug (die_struct *)): New.
21505 * dwarf2out.h
21506 (extern debug (die_struct &)): New.
21507 (extern debug (die_struct *)): New.
21508 * gimple-pretty-print.c
21509 (debug (gimple_statement_d &)): New.
21510 (debug (gimple_statement_d *)): New.
21511 * gimple-pretty-print.h
21512 (extern debug (gimple_statement_d &)): New.
21513 (extern debug (gimple_statement_d *)): New.
21514 * ira-build.c
21515 (debug (ira_allocno_copy &)): New.
21516 (debug (ira_allocno_copy *)): New.
21517 (debug (ira_allocno &)): New.
21518 (debug (ira_allocno *)): New.
21519 * ira-int.h
21520 (extern debug (ira_allocno_copy &)): New.
21521 (extern debug (ira_allocno_copy *)): New.
21522 (extern debug (ira_allocno &)): New.
21523 (extern debug (ira_allocno *)): New.
21524 * ira-lives.c
21525 (debug (live_range &)): New.
21526 (debug (live_range *)): New.
21527 * lra-int.h
21528 (debug (lra_live_range &)): New.
21529 (debug (lra_live_range *)): New.
21530 * lra-lives.c
21531 (debug (lra_live_range &)): New.
21532 (debug (lra_live_range *)): New.
21533 * omega.c
21534 (debug (omega_pb_d &)): New.
21535 (debug (omega_pb_d *)): New.
21536 * omega.h
21537 (extern debug (omega_pb_d &)): New.
21538 (extern debug (omega_pb_d *)): New.
21539 * print-rtl.c
21540 (debug (const rtx_def &)): New.
21541 (debug (const rtx_def *)): New.
21542 * print-tree.c
21543 (debug_tree (tree): Move within file.
21544 (debug_raw (const tree_node &)): New.
21545 (debug_raw (const tree_node *)): New.
21546 (dump_tree_via_hooks (const tree_node *, int)): New.
21547 (debug (const tree_node &)): New.
21548 (debug (const tree_node *)): New.
21549 (debug_verbose (const tree_node &)): New.
21550 (debug_verbose (const tree_node *)): New.
21551 (debug_head (const tree_node &)): New.
21552 (debug_head (const tree_node *)): New.
21553 (debug_body (const tree_node &)): New.
21554 (debug_body (const tree_node *)): New.
21555 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21556 (debug (vec<tree, va_gc> &)): New.
21557 (debug (vec<tree, va_gc> *)): New.
21558 * rtl.h
21559 (extern debug (const rtx_def &)): New.
21560 (extern debug (const rtx_def *)): New.
21561 * sbitmap.c
21562 (debug_raw (simple_bitmap_def &)): New.
21563 (debug_raw (simple_bitmap_def *)): New.
21564 (debug (simple_bitmap_def &)): New.
21565 (debug (simple_bitmap_def *)): New.
21566 * sbitmap.h
21567 (extern debug (simple_bitmap_def &)): New.
21568 (extern debug (simple_bitmap_def *)): New.
21569 (extern debug_raw (simple_bitmap_def &)): New.
21570 (extern debug_raw (simple_bitmap_def *)): New.
21571 * sel-sched-dump.c
21572 (debug (vinsn_def &)): New.
21573 (debug (vinsn_def *)): New.
21574 (debug_verbose (vinsn_def &)): New.
21575 (debug_verbose (vinsn_def *)): New.
21576 (debug (expr_def &)): New.
21577 (debug (expr_def *)): New.
21578 (debug_verbose (expr_def &)): New.
21579 (debug_verbose (expr_def *)): New.
21580 (debug (vec<rtx> &)): New.
21581 (debug (vec<rtx> *)): New.
21582 * sel-sched-dump.h
21583 (extern debug (vinsn_def &)): New.
21584 (extern debug (vinsn_def *)): New.
21585 (extern debug_verbose (vinsn_def &)): New.
21586 (extern debug_verbose (vinsn_def *)): New.
21587 (extern debug (expr_def &)): New.
21588 (extern debug (expr_def *)): New.
21589 (extern debug_verbose (expr_def &)): New.
21590 (extern debug_verbose (expr_def *)): New.
21591 (extern debug (vec<rtx> &)): New.
21592 (extern debug (vec<rtx> *)): New.
21593 * sel-sched-ir.h
21594 (_list_iter_cond_expr): Make inline instead of static.
21595 * sreal.c
21596 (debug (sreal &)): New.
21597 (debug (sreal *)): New.
21598 * sreal.h
21599 (extern debug (sreal &)): New.
21600 (extern debug (sreal *)): New.
21601 * tree.h
21602 (extern debug_raw (const tree_node &)): New.
21603 (extern debug_raw (const tree_node *)): New.
21604 (extern debug (const tree_node &)): New.
21605 (extern debug (const tree_node *)): New.
21606 (extern debug_verbose (const tree_node &)): New.
21607 (extern debug_verbose (const tree_node *)): New.
21608 (extern debug_head (const tree_node &)): New.
21609 (extern debug_head (const tree_node *)): New.
21610 (extern debug_body (const tree_node &)): New.
21611 (extern debug_body (const tree_node *)): New.
21612 (extern debug (vec<tree, va_gc> &)): New.
21613 (extern debug (vec<tree, va_gc> *)): New.
21614 * tree-cfg.c
21615 (debug (struct loop &)): New.
21616 (debug (struct loop *)): New.
21617 (debug_verbose (struct loop &)): New.
21618 (debug_verbose (struct loop *)): New.
21619 * tree-dump.c: Add header dependence.
21620 * tree-flow.h
21621 (extern debug (struct loop &)): New.
21622 (extern debug (struct loop *)): New.
21623 (extern debug_verbose (struct loop &)): New.
21624 (extern debug_verbose (struct loop *)): New.
21625 * tree-data-ref.c
21626 (debug (data_reference &)): New.
21627 (debug (data_reference *)): New.
21628 (debug (vec<data_reference_p> &)): New.
21629 (debug (vec<data_reference_p> *)): New.
21630 (debug (vec<ddr_p> &)): New.
21631 (debug (vec<ddr_p> *)): New.
21632 * tree-data-ref.h
21633 (extern debug (data_reference &)): New.
21634 (extern debug (data_reference *)): New.
21635 (extern debug (vec<data_reference_p> &)): New.
21636 (extern debug (vec<data_reference_p> *)): New.
21637 (extern debug (vec<ddr_p> &)): New.
21638 (extern debug (vec<ddr_p> *)): New.
21639 * tree-ssa-alias.c
21640 (debug (pt_solution &)): New.
21641 (debug (pt_solution *)): New.
21642 * tree-ssa-alias.h
21643 (extern debug (pt_solution &)): New.
21644 (extern debug (pt_solution *)): New.
21645 * tree-ssa-alias.c
21646 (debug (_var_map &)): New.
21647 (debug (_var_map *)): New.
21648 (debug (tree_live_info_d &)): New.
21649 (debug (tree_live_info_d *)): New.
21650 * tree-ssa-alias.h
21651 (extern debug (_var_map &)): New.
21652 (extern debug (_var_map *)): New.
21653 (extern debug (tree_live_info_d &)): New.
21654 (extern debug (tree_live_info_d *)): New.
21655
be77e1e5
JH		 216562013-03-28 Jan Hubicka <jh@suse.cz>
21657
21658 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21659
777e6976
IB		 216602013-03-28 Ian Bolton <ian.bolton@arm.com>
21661
21662 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21663 record only when desired or required.
21664
9cd347ae
UB		 216652013-03-28 Uros Bizjak <ubizjak@gmail.com>
21666
21667 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21668 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21669
33159866 216702013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N		 21671
21672 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21673 (*andsi3_compare0_uxtw): New pattern.
21674 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21675 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21676
89ab31c1
JH		 216772013-03-28 Jan Hubicka <jh@suse.cz>
21678
21679 * data-streamer-in.c (streamer_read_gcov_count): New function.
21680 * gimple-streamer-out.c: Include value-prof.h.
21681 (output_gimple_stmt): Output histogram.
21682 (output_bb): Use streamer_write_gcov_count.
21683 * value-prof.c: Include data-streamer.h
21684 (dump_histogram_value): Add HIST_TYPE_MAX.
21685 (stream_out_histogram_value): New function.
21686 (stream_in_histogram_value): New function.
21687 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21688 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21689 * data-streamer-out.c (streamer_write_gcov_count): New function.
21690 (streamer_write_gcov_count_stream): New function.
21691 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21692 (lto_output_node): Likewise.
21693 (input_node, input_edge): Likewise.
21694 * lto-streamer-out.c (output_cfg): Update streaming.
21695 * lto-streamer-in.c (input_cfg): Likewise.
21696 * data-streamer.h (streamer_write_gcov_count,
21697 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21698 * gimple-streamer-in.c: Include value-prof.h
21699 (input_gimple_stmt): Input histograms.
21700 (input_bb): Update profile streaming.
21701
e594716a 217022013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 21703
33159866 21704 * genmodes.c (emit_max_int): New function.
8697be17 21705 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC		 21706 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21707 Added doc.
8697be17 21708 * machmode.def: Fixed comment.
89ab31c1 21709
e594716a 217102013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ		 21711
21712 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ		 21713 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21714
74946978
MP		 217152013-03-28 Marek Polacek <polacek@redhat.com>
21716 Richard Biener <rguenther@suse.de>
21717
21718 PR tree-optimization/56695
21719 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21720 build signed result of a vector comparison.
21721 * tree-cfg.c (verify_gimple_comparison): Check that a result
21722 of a vector comparison has signed type.
21723
a64b9c26
RB		 217242013-03-28 Richard Biener <rguenther@suse.de>
21725
21726 PR tree-optimization/37021
21727 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21728 do not restrict gaps between groups.
21729 * tree-vect-stmts.c (vectorizable_load): Properly account for
21730 a gap between groups.
21731
a9dc2a2f
EB		 217322013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21733
21734 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21735 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21736 is not enabled.
21737
53cb97f9
GP		 217382013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21739
21740 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21741 * doc/extend.texi (Named Address Spaces): Ditto.
21742 (Variable Attributes): Ditto.
21743
b802ae5c
KT		 217442013-03-27 Kai Tietz <ktietz@redhat.com>
21745
eddae10a
KT		 21746 * config.build: Add support for cygwin x64 target.
21747 * config.gcc: Likewise.
21748 * config.host: Likewise.
21749 * configure.ac: Likewise
21750 * configure: Regenerated.
b802ae5c 21751
371e77e3 217522013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT		 21753
21754 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21755 * config/i386/t-cygwin-w64: New file.
21756 * config/i386/cygwin-w64.h: New file.
21757 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21758 and add support for x64-cygwin target.
21759 (CPP_SPEC): Likewise.
21760 (CXX_WRAP_SPEC_LIST): Undefine before define.
21761 (LIBGCJ_SONAME): Use 15 as version.
21762
f49b33cb
RB		 217632013-03-27 Richard Biener <rguenther@suse.de>
21764
21765 PR tree-optimization/56716
21766 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21767 dumping for ref nodes.
21768
b37a6ce5
MJ		 217692013-03-27 Martin Jambor <mjambor@suse.cz>
21770
21771 PR tree-optimization/55334
21772 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21773 restricted pointers to arrays.
21774
9469b9b2
GDR		 217752013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21776
21777 * Makefile.in (.SUFFIXES): Add .cc.
21778 (.c.o): Apply same recipe for implicit rule .cc.o.
21779
7d24f650
RB		 217802013-03-27 Richard Biener <rguenther@suse.de>
21781
21782 PR tree-optimization/37021
21783 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21784 REALPART/IMAGPART_EXPRs around the supported refs.
21785 * tree-ssa-structalias.c (find_func_aliases): Assume that
21786 floating-point values are not used to transfer pointers.
21787
2f251a05
AI		 217882013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21789
f43245d1
UB		 21790 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21791 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21792 * doc/tm.texi: Regenerate.
21793 * targhooks.h (default_has_ifunc_p): New.
21794 * targhooks.c (default_has_ifunc_p): Ditto.
21795 * config/linux-protos.h: New file.
21796 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21797 hook for linux which disables support of indirect functions in android.
21798 * config/linux-android.c: New file.
21799 * config/t-linux-android.c: Ditto.
21800 * config.gcc: Added new object file linux-android.o.
21801 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 21802 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 21803 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI		 21804 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21805 doesn't support indirect functions.
21806 * configure: Regenerate.
21807
78b4e425
BC		 218082013-03-27 Bin Cheng <bin.cheng@arm.com>
21809
21810 PR target/56102
21811 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21812 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21813 mult-word mode.
21814
a5ba7b92
AK		 218152013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21816
21817 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21818
be624986
TG		 218192013-03-27 Terry Guo <terry.guo@arm.com>
21820
21821 * config/arm/arm-cores.def: Added core cortex-r7.
21822 * config/arm/arm-tune.md: Regenerated.
21823 * config/arm/arm-tables.opt: Regenerated.
21824 * doc/invoke.texi: Added entry for core cortex-r7.
21825
0a514f47
WL		 218262013-03-27 Walter Lee <walt@tilera.com>
21827
21828 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21829 double-decrement of next_scratch_regno.
21830
9332b0d2
WL		 218312013-03-27 Walter Lee <walt@tilera.com>
21832
801d9b2a 21833 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL		 21834 input operands.
21835 (insn_v1mulus): Ditto.
21836 (insn_v2muls): Ditto.
21837
f54ea5dd
WL		 218382013-03-27 Walter Lee <walt@tilera.com>
21839
e594716a 21840 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL		 21841 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21842
bf60f4ca
WL		 218432013-03-27 Walter Lee <walt@tilera.com>
21844
e594716a 21845 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL		 21846 (*sibcall_value): Ditto.
21847
e3b51eeb
WL		 218482013-03-27 Walter Lee <walt@tilera.com>
21849
21850 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21851 (insn_mnz_v8qi): ... this ...
21852 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21853 vector equivalent.
21854 (insn_v<n>mnz): Replaced by ...
21855 (insn_v1mnz): ... this ...
21856 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21857 equivalent.
21858 (insn_mz_<mode>): Replaced by ...
21859 (insn_mz_v8qi): ... this ...
21860 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21861 vector equivalent.
21862 (insn_v<n>mz): Replaced by ...
21863 (insn_v1mz): ... this ...
21864 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21865 equivalent.
21866
4fc7b145
EB		 218672013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21868
21869 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21870
6f33817e
RM		 218712013-03-26 Roland McGrath <mcgrathr@google.com>
21872
21873 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21874 than fprintf with a non-constant, non-format string.
21875
93a4145b
UB		 218762013-03-26 Uros Bizjak <ubizjak@gmail.com>
21877
21878 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21879 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21880 operand 0 predicate.
21881 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21882 attribute. Use general_x64nomem_operand as operand 1 predicate.
21883 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21884 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21885 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21886 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21887 (mov<mode>_insv_1): Remove expander. Merge insn with
21888 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21889 Use general_x64nomem_operand as operand 1 predicate.
21890 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21891 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21892 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21893 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21894 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21895 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21896 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21897 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21898 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21899 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21900 (general_x64nomem_operand): Ditto.
21901
c6a9ed5a 219022013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH		 21903
21904 * config/rtems.opt: Add -pthread option.
21905
31b0a960
RB		 219062013-03-26 Richard Biener <rguenther@suse.de>
21907
93a4145b
UB		 21908 * alias.c (find_base_term): Avoid redundant and not used recursion.
21909 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB		 21910 (true_dependence_1): Compute and pass base terms to base_alias_check.
21911 (write_dependence_p): Likewise.
21912 (may_alias_p): Likewise.
21913
79517551
SN		 219142013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21915
21916 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21917 PC-relative load in SI modes and above only.
21918
a76213b9
XQ		 219192013-03-26 Xinyu Qi <xyqi@marvell.com>
21920
21921 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB		 21922 * config/arm/iwmmxt.md (WCGR0): Update.
21923 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 21924
37ff9355
UB		 219252013-03-26 Uros Bizjak <ubizjak@gmail.com>
21926
21927 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21928 Use x64 and nox64 isa attributes.
21929
30b0317c
RB		 219302013-03-26 Richard Biener <rguenther@suse.de>
21931
21932 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21933 alignment computations and rely on get_object_alignment_1
21934 for the !TYPE_P case.
21935 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21936
cbcdb140
WL		 219372013-03-26 Walter Lee <walt@tilera.com>
21938
21939 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21940 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21941
dd884525 219422013-03-25 Jeff Law <law@redhat.com>
be672e08 21943
33159866
UB		 21944 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21945 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL		 21946 wrong version of prior patch.
21947
0051d3ec
WL		 219482013-03-25 Walter Lee <walt@tilera.com>
21949
21950 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21951 TILEGX_INSN_SHUFFLEBYTES1.
21952 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21953 shufflebytes1.
21954 (tilegx_builtins): Ditto.
21955 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21956
065a3605
WL		 219572013-03-25 Walter Lee <walt@tilera.com>
21958
21959 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL		 21960 (floatunssisf2): New pattern.
21961 (floatsidf2): New pattern.
21962 (floatunssidf2): New pattern.
065a3605 21963
5b2a3c82
WL		 219642013-03-25 Walter Lee <walt@tilera.com>
21965
21966 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21967 tests for constraint J, K, N, P.
21968
192ea533
WL		 219692013-03-25 Walter Lee <walt@tilera.com>
21970
21971 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21972 Use indirect/pcrel encoding.
21973 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21974 Ditto.
21975
855e0d0b
SE		 219762013-03-25 Steve Ellcey <sellcey@mips.com>
21977
21978 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21979 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21980 * config/mips/mips.c (mips_option_override): Set IMADD default.
21981 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21982 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21983 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21984 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21985 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21986
39e45653
JL		 219872013-03-25 Jeff Law <law@redhat.com>
21988
21989 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21990 slightly to avoid creating and folding useless trees. Simplify
21991 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21992
7482c470
UB		 219932013-03-25 Uros Bizjak <ubizjak@gmail.com>
21994
21995 * config/i386/i386.md (*zero_extendsidi2): Merge with
21996 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21997 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21998 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21999 targets. Clarify comment.
22000
4a53743e
MJ		 220012013-03-25 Martin Jambor <mjambor@suse.cz>
22002
22003 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22004 pass-through jump functions differently.
22005 (ipa_read_jump_function): Likewise. Also use setter functions to set
22006 up jump functions.
22007
162712de
MJ		 220082013-03-25 Martin Jambor <mjambor@suse.cz>
22009
22010 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22011 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22012 process it.
22013 (ipa_get_indirect_edge_target): New function.
22014 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22015 ipa_get_indirect_edge_target. Update all callers.
22016 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22017 ipa_get_indirect_edge_target_1 instead of calling
22018 ipa_get_indirect_edge_target.
22019 (create_specialized_node): Pass aggvlas to
22020 ipcp_discover_new_direct_edges.
22021
a5a4c20a
KT		 220222013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22023
22024 * config/arm/arm.md (f_sels, f_seld): New types.
22025 (*cmov<mode>): New pattern.
22026 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22027 predicate.
22028
f992b9fc
KT		 220292013-03-25 Kai Tietz <ktietz@redhat.com>
22030
22031 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22032 POSIX-printf for mingw-hosted builds.
22033
c16fd676
RB		 220342013-03-25 Richard Biener <rguenther@suse.de>
22035
22036 PR middle-end/56694
22037 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22038 must-not-throw stmt location.
22039
5ad29f12
KT		 220402013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22041
22042 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22043 Emit load-acquire versions when acq is true.
22044 (arm_emit_store_exclusive): Add rel parameter.
22045 Emit store-release versions when rel is true.
22046 (arm_split_compare_and_swap): Use acquire-release instructions
22047 instead.
22048 of barriers when appropriate.
22049 (arm_split_atomic_op): Likewise.
22050 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22051 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22052 (VUNSPEC_SLX): Likewise.
22053 (VUNSPEC_LDA): Likewise.
22054 (VUNSPEC_STL): Likewise.
22055 * config/arm/sync.md (atomic_load<mode>): New pattern.
22056 (atomic_store<mode>): Likewise.
22057 (arm_load_acquire_exclusive<mode>): Likewise.
22058 (arm_load_acquire_exclusivesi): Likewise.
22059 (arm_load_acquire_exclusivedi): Likewise.
22060 (arm_store_release_exclusive<mode>): Likewise.
22061
03a7dddb
CM		 220622013-03-25 Catherine Moore <clm@codesourcery.com>
22063
22064 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22065 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22066 * config/mip/predicates.md (lwsp_swsp_operand,
22067 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22068 sb16_operand, db4_operand, db7_operand, ib3_operand,
22069 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22070 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22071 andi16_operand): New predicates.
22072 * config/mips/mips.md (compression): New attribute.
22073 (enabled): New attribute.
22074 (length): Consider compression in computing length.
22075 (shift_compression): New code attribute.
22076 (*add<mode>3): New operands. Record compression.
22077 (sub<mode>3): Likewise.
22078 (one_cmpl<mode>2): Likewise.
22079 (*and<mode>3): Likewise.
22080 (*ior<mode>3): Likewise.
22081 (unnamed pattern for xor): Likewise.
22082 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22083 (*<optab><mode>3): Likewise.
22084 (*mov<mode>_internal: Likewise.
22085 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22086 (mips_unsigned_immediate_p): New.
22087 (umips_lwsp_swsp_address_p): New.
22088 (m16_based_address_p): New.
22089 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22090 (mips_unsigned_immediate_p): New prototype.
22091 (lwsp_swsp_address_p): New prototype.
22092 (m16_based_address_p): New prototype.
22093 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22094 (mips_signed_immediate_p): New function.
22095 (m16_based_address_p): New function.
22096 (lwsp_swsp_address_p): New function.
22097 (mips_print_operand_punctuation): Recognize short delay slot insns
22098 for microMIPS.add<mode>3"
22099
f35c297f 221002013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 22101
f35c297f
KT		 22102 PR target/56720
22103 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22104 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22105
051b9446
RB		 221062013-03-25 Richard Biener <rguenther@suse.de>
22107
22108 PR tree-optimization/56689
22109 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22110 any edge.
22111
374001cb
RB		 221122013-03-25 Richard Biener <rguenther@suse.de>
22113
22114 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22115 of bitmap.
22116 (memory_references): Likewise.
22117 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22118 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22119 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22120 (gather_mem_refs_in_loops): Fold into ...
22121 (analyze_memory_references): ... this. Move initialization
22122 to tree_ssa_lim_initialize.
22123 (fill_always_executed_in): Rename to ...
22124 (fill_always_executed_in_1): ... this.
22125 (fill_always_executed_in): Move contains_call computation to
22126 this new function from ...
22127 (tree_ssa_lim_initialize): ... here.
22128 (tree_ssa_lim): Call fill_always_executed_in.
22129
57895947
EB		 221302013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22131
22132 * postreload.c (reload_combine): Fix code detecting returns.
22133
1e1b18c1
EB		 221342013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22135
22136 * function.c (emit_use_return_register_into_block): On cc0 targets,
22137 do not emit the sequence between cc0 setter and user.
22138
9216baf1
KT		 221392013-03-25 Kai Tietz <ktietz@redhat.com>
22140
1e1b18c1
EB		 22141 * config/i386/predicates.md (local_symbolic_operand): Interpret
22142 dll-imported symbols as none-local.
9216baf1 22143
76421b44
RB		 221442013-03-25 Richard Biener <rguenther@suse.de>
22145
22146 * tree-ssa-loop-im.c (struct depend): Remove.
22147 (struct lim_aux_data): Make depends a vec of gimples.
22148 (free_lim_aux_data): Adjust.
22149 (add_dependency): Likewise.
22150 (set_level): Likewise.
22151
d154bfa2
RB		 221522013-03-25 Richard Biener <rguenther@suse.de>
22153
22154 PR middle-end/56434
22155 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22156 the pointer returned by calls with ECF_MALLOC set.
22157
2cd9804e
UB		 221582013-03-24 Uros Bizjak <ubizjak@gmail.com>
22159
acef5fe0
UB		 22160 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22161
221622013-03-24 Uros Bizjak <ubizjak@gmail.com>
22163
22164 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB		 22165 using MMXMODE mode iterator.
22166 (*move<mode>_internal): Merge with *movv2sf_internal and
22167 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22168
225ccc68
SB		 221692013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22170
85c0f02d
SB		 22171 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22172 (record_last_mem_set_info): Likewise.
22173
225ccc68
SB		 22174 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22175 of XNEWVEC followed by memset.
22176 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22177
b64925dc
SB		 221782013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22179
22180 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22181 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22182 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22183 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22184 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22185 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22186 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22187 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22188 BARRIER_P instead of GET_CODE.
22189
b9aaf52e
EB		 221902013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22191
22192 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22193 inaccuracy in the probing code.
22194
22195 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22196 (ctrapdi4): Likewise.
22197
66de4d7c
EB		 221982013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22199
22200 * calls.c (expand_call): Add missing guard to code handling return
22201 of non-BLKmode structures in MSB.
22202 * function.c (expand_function_end): Likewise.
22203
4f25c8fb
EB		 222042013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22205
22206 * combine.c (try_combine): Adjust comment. Do not add the set of
22207 insn #0 if the destination indirectly is set or dies in insn #2.
22208 Tidy up code to distribute a new note.
22209
23b7850d
UB		 222102013-03-22 Uros Bizjak <ubizjak@gmail.com>
22211
22212 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22213 also for alternatives 16 and 17.
22214
bed852cf
UB		 222152013-03-22 Uros Bizjak <ubizjak@gmail.com>
22216
22217 * config/i386/sse.md (*mov<mode>_internal): Merge with
22218 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22219 Emit insn template depending on type attribute. Use
22220 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22221 movd instead of movq mnemonic for interunit moves. Rewrite mode
22222 attribute calculation. Remove unit attribute calculation.
22223 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22224 Set prefix_data16 attribute for DImode ssemov types.
22225 Use Ym instead of y for SSE-MMX conversion alternatives.
22226 Reorder operand constraints.
22227
1e8a7937
SB		 222282013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22229
22230 * df.h (df_insn_delete): Adjust prototype.
22231 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22232 and let it decide whether mark the basic block dirty.
22233 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22234 * df-scan.c (df_insn_info_delete): New helper function, split
22235 off from df_insn_delete.
22236 (df_scan_free_bb_info): Use it.
22237 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22238 Likewise.
22239 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22240 that the insn is actually an insn and it has a non-NULL basic block.
22241 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22242
e14d094c
RB		 222432013-03-22 Richard Biener <rguenther@suse.de>
22244
22245 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22246 dep_ref members.
22247 (mem_ref_alloc): Do not allocate them.
22248 (refs_independent_p): Do not query or maintain a cache.
22249
e9cf7316
RB		 222502013-03-22 Richard Biener <rguenther@suse.de>
22251
22252 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22253 (gather_mem_refs_in_loops): Do not compute it.
22254 (analyze_memory_references): Do not allocate it.
22255 (tree_ssa_lim_finalize): Do not free it.
22256 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22257
f046e81b
RB		 222582013-03-22 Richard Biener <rguenther@suse.de>
22259
22260 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 22261
a05c0ddf
IB		 222622013-03-22 Ian Bolton <ian.bolton@arm.com>
22263
22264 * config/aarch64/aarch64.c (aarch64_print_operand): New
22265 format specifier for printing a constant in hex.
22266 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22267 format specifier for printing second operand.
22268
c00217fc
RB		 222692013-03-22 Richard Biener <rguenther@suse.de>
22270
22271 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22272 bitmaps.
22273 (gather_mem_refs_in_loops): Perform store accumulation here.
22274 (create_vop_ref_mapping_loop): Remove.
22275 (create_vop_ref_mapping): Likewise.
22276 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 22277 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB		 22278 (record_indep_loop): Remove.
22279 (record_dep_loop): New function.
22280 (ref_indep_loop_p_1): Adjust to only walk over references
22281 in the loop, not its subloops.
22282 (ref_indep_loop_p): Rename to ...
22283 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22284 maintaining a more fine-grained cache.
22285 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22286 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22287
15d19bf8
RB		 222882013-03-22 Richard Biener <rguenther@suse.de>
22289
22290 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22291 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22292 aggregate mem_ref_loc.
22293 (free_mem_ref_locs): Inline into ...
22294 (memref_free): ... this and adjust.
22295 (mem_ref_alloc): Adjust.
22296 (mem_ref_locs_alloc): Remove.
22297 (record_mem_ref_loc): Adjust.
22298 (get_all_locs_in_loop): Rewrite into ...
22299 (for_all_locs_in_loop): ... this iterator.
22300 (rewrite_mem_ref_loc): New functor.
22301 (rewrite_mem_refs): Use for_all_locs_in_loop.
22302 (sm_set_flag_if_changed): New functor.
22303 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22304 (ref_always_accessed): New functor.
22305 (ref_always_accessed_p): Use for_all_locs_in_loop.
22306
6f37411d
MG		 223072013-03-21 Marc Glisse <marc.glisse@inria.fr>
22308
22309 * tree-pass.h (PROP_gimple_lvec): New.
22310 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22311 (init_optimization_passes): Move pass_lower_vector.
22312 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22313 PROP_gimple_lvec.
22314 (pass_lower_vector): Provide PROP_gimple_lvec.
22315 (pass_lower_vector_ssa): Likewise.
22316 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22317
511ef689
MW		 223182013-03-21 Mark Wielaard <mjw@redhat.com>
22319
22320 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22321
cc1df30b
UB		 223222013-03-21 Uros Bizjak <ubizjak@gmail.com>
22323
22324 * config/i386/i386.md (*movdi_internal): Disparage slightly
22325 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22326 conversion alternatives.
22327
c845cfe1
JJ		 223282013-03-21 Jakub Jelinek <jakub@redhat.com>
22329
37e99116
JJ		 22330 PR middle-end/48087
22331 * diagnostic.def (DK_WERROR): New kind.
22332 * diagnostic.h (werrorcount): Define.
22333 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22334 promoted to DK_ERROR, increment DK_WERROR counter instead of
22335 DK_ERROR counter.
22336 * toplev.c (toplev_main): Call print_ignored_options even if
22337 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22338 even if just werrorcount is non-zero.
22339
c845cfe1
JJ		 22340 PR debug/55608
22341 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22342 on failure.
22343 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22344 (string_cst_pool_decl): New function.
22345 (optimize_one_addr_into_implicit_ptr): New function.
22346 (resolve_addr_in_expr): Optimize DWARF location expression
22347 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22348 which doesn't live in memory, but has DW_AT_location or
22349 DW_AT_const_value, or refers to a string literal, into
22350 DW_OP_GNU_implicit_pointer.
22351 (optimize_location_into_implicit_ptr): New function.
22352 (resolve_addr): If removing DW_AT_location of a variable because
22353 it was DW_OP_addr of address of the variable, but the variable doesn't
22354 live in memory, try to emit const value attribute for the initializer.
22355
08e0cda6
MG		 223562013-03-21 Marc Glisse <marc.glisse@inria.fr>
22357
22358 * tree.h (VECTOR_TYPE_P): New macro.
22359 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22360 TYPE_MODE): Use it.
22361 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22362 VEC_COND_EXPR cannot be lvalues.
22363 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22364
d08633b4
MG		 223652013-03-21 Marc Glisse <marc.glisse@inria.fr>
22366
22367 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22368 Restrict the transformation to equal modes.
22369
e6647190
RB		 223702013-03-21 Richard Biener <rguenther@suse.de>
22371
22372 PR tree-optimization/39326
22373 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22374 (MEM_ANALYZABLE): Adjust.
22375 (record_mem_ref_loc): Move bitmap ops ...
22376 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22377 unanalyzable refs, do not record locations for it.
22378 (analyze_memory_references): Allocate ref zero as shared
22379 unanalyzable ref.
22380 (refs_independent_p): Do not test for unanalyzed mems here.
22381 (ref_indep_loop_p_1): Special-case disambiguation against
22382 the unanalyzed ref.
cc1df30b 22383 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 22384
65074f54
CL		 223852013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22386
22387 * config/arm/arm-protos.h (tune_params): Add
22388 prefer_neon_for_64bits field.
22389 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22390 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22391 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22392 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22393 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22394 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22395 (arm_option_override): Handle -mneon-for-64bits new option.
22396 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22397 (prefer_neon_for_64bits): Declare new variable.
22398 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 22399 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL		 22400 (arch_enabled): Handle new arch types. Remove support for onlya8
22401 and nota8.
22402 (one_cmpldi2): Use new arch names.
bc5faa5b 22403 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL		 22404 * config/arm/arm.opt (mneon-for-64bits): Add option.
22405 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22406 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22407 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22408 of onlya8.
22409 * doc/invoke.texi (-mneon-for-64bits): Document.
22410
5a2d2a79
RB		 224112013-03-21 Richard Biener <rguenther@suse.de>
22412
22413 PR tree-optimization/39326
22414 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22415 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 22416 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 22417
5abe1e05
RB		 224182013-03-21 Richard Biener <rguenther@suse.de>
22419
22420 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22421 (vect_insert_into_interleaving_chain): Likewise.
22422 (vect_drs_dependent_in_basic_block): Inline ...
22423 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22424 split out from ...
22425 (vect_analyze_data_ref_dependence): ... here. Simplify.
22426 (vect_check_interleaving): Simplify.
22427 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22428 (vect_slp_analyze_data_ref_dependences): ... this new function.
22429 (dr_group_sort_cmp): New function.
22430 (vect_analyze_data_ref_accesses): Compute data-reference groups
22431 here instead of in vect_analyze_data_ref_dependence. Use
22432 a more efficient algorithm.
22433 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22434 vect_slp_analyze_data_ref_dependences. Call
22435 vect_analyze_data_ref_accesses earlier.
22436 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22437 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22438 (vect_slp_analyze_data_ref_dependences): New prototype.
22439
cad1735b
RB		 224402013-03-21 Richard Biener <rguenther@suse.de>
22441
22442 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22443 ref is stored in the loop.
22444 (find_refs_for_sm): Walk only over all stores.
22445 (store_motion_loop): Allocate from lim_bitmap_obstack.
22446 (store_motion): Likewise.
22447
141310ef
RB		 224482013-03-21 Richard Biener <rguenther@suse.de>
22449
22450 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22451 Update virtual SSA form.
22452
5022315a
RO		 224532013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22454
22455 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22456 * configure: Regenerate.
22457 * config.in: Regenerate.
22458 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22459 if !HAVE_LD_EH_FRAME_CIEV3.
22460
50d4421c
RB		 224612013-03-21 Richard Biener <rguenther@suse.de>
22462
22463 * tree-cfg.c (verify_expr_no_block): New function.
22464 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22465 nor DECL_VALUE_EXPR have locations with associated blocks.
22466 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22467 (clear_unused_block_pointer): Remove code dealing with
22468 blocks in DECL_DEBUG_EXPR locations.
22469
839b422f
RB		 224702013-03-21 Richard Biener <rguenther@suse.de>
22471
22472 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22473 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22474 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22475 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22476 instead of DECL_DEBUG_EXPR_IS_FROM.
22477 * gimplify.c (gimplify_modify_expr): Likewise.
22478 * tree-cfg.c (verify_expr_location_1): Likewise.
22479 * tree-complex.c (create_one_component_var): Likewise.
22480 * tree-sra.c (create_access_replacement): Likewise.
22481 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22482 (clear_unused_block_pointer): Likewise.
22483 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22484 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22485 * var-tracking.c (var_debug_decl): Likewise.
22486 (track_expr_p): Likewise.
22487 * tree-inline.c (add_local_variables): Likewise. Set
22488 DECL_HAS_DEBUG_EXPR_P after copying it.
22489 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22490 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22491
fe04878d
UB		 224922013-03-21 Uros Bizjak <ubizjak@gmail.com>
22493
22494 PR bootstrap/56656
22495 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22496 * configure: Regenerate.
22497 * config.in: Regenerate.
22498 * config/i386/i386.md (*movdf_internal): Use
22499 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22500 movd instead of movq mnemonic for interunit moves.
22501 (*movdi_internal): Ditto.
22502
bd059b26 225032013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N		 22504
22505 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22506 (abd<mode>_3): New pattern.
22507 (aba<mode>_3): New pattern.
22508 (fabd<mode>_3): New pattern.
22509
bd059b26 225102013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N		 22511
22512 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22513 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22514 occurrence of REGISTER_PREFIX as its empty string.
22515
4f1f78b9
JL		 225162013-03-20 Jeff Law <law@redhat.com>
22517
22518 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22519 addititional equivalences for equality comparisons between an SSA_NAME
22520 and a constant where the SSA_NAME was set from a widening conversion.
22521
327a1118
WL		 225222013-03-20 Walter Lee <walt@tilera.com>
22523
22524 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22525
813e0036
UB		 225262013-03-20 Uros Bizjak <ubizjak@gmail.com>
22527
22528 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22529 depending on type attribute.
22530 (*movti_internal): Ditto.
22531 (*movtf_internal): Ditto.
22532 (*movxf_internal): Ditto.
22533 (*movdf_internal): Ditto.
22534 (*movsf_internal): Ditto.
22535
7cf34aae
UB		 225362013-03-20 Uros Bizjak <ubizjak@gmail.com>
22537
22538 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22539 maybe_vex for sselog1 and ssemov types.
22540 (*movdi_internal): Reorder operand constraints.
22541 (*movsi_internal): Ditto. Set prefix attribute to
22542 maybe_vex for sselog1 and ssemov types.
22543 (*movtf_internal): Set prefix attribute to maybe_vex
22544 for sselog1 and ssemov types.
22545 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22546 DImode ssemov types. Reorder operand constraints.
22547 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22548 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22549 attribute for SImode ssemov types. Reorder operand constraints.
22550
19321415
MJ		 225512013-03-20 Martin Jambor <mjambor@suse.cz>
22552
22553 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22554 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22555
12211b99 225562013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM		 22557
22558 * config/rs6000/predicates.md (indexed_address, update_address_mem
22559 update_indexed_address_mem): New predicates.
22560 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22561 attribute for load/store instructions.
22562 * config/rs6000/dfp.md (movsd_store): Likewise.
22563 (movsd_load): Likewise.
22564 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22565 (unnamed HI->DI extend define_insn): Likewise.
22566 (unnamed SI->DI extend define_insn): Likewise.
22567 (unnamed QI->SI extend define_insn): Likewise.
22568 (unnamed QI->HI extend define_insn): Likewise.
22569 (unnamed HI->SI extend define_insn): Likewise.
22570 (unnamed HI->SI extend define_insn): Likewise.
22571 (extendsfdf2_fpr): Likewise.
22572 (movsi_internal1): Likewise.
22573 (movsi_internal1_single): Likewise.
22574 (movhi_internal): Likewise.
22575 (movqi_internal): Likewise.
22576 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22577 attribute for load/store instructions.
22578 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22579 instructions.
22580 (mov<mode>_softfloat): Likewise.
22581 (mov<mode>_hardfloat32): Likewise.
22582 (mov<mode>_hardfloat64): Likewise.
22583 (mov<mode>_softfloat64): Likewise.
22584 (movdi_internal32): Likewise.
22585 (movdi_internal64): Likewise.
22586 (probe_stack_<mode>): Likewise.
22587
225882013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22589
22590 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22591 floating point, and decimal floating point to reload iterator.
22592
22593 * config/rs6000/constraints.md (wl constraint): New constraints to
22594 return FLOAT_REGS if certain options are used to reduce the number
22595 of separate patterns that exist in the file.
22596 (wx constraint): Likewise.
22597 (wz constraint): Likewise.
22598
22599 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22600 -mdebug=reg, print wg, wl, wx, and wz constraints.
22601 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22602 Initialize the reload functions for 64-bit binary/decimal floating
22603 point types.
22604 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22605 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22606 create the buffer on the stack to overcome not having a 32-bit
22607 load and store.
22608 (rs6000_emit_move): Likewise.
22609 (rs6000_secondary_memory_needed_rtx): Likewise.
22610 (rs6000_alloc_sdmode_stack_slot): Likewise.
22611 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22612 via xxlxor, just like DFmode 0.0.
22613
22614 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22615 define as 1 if we are running on a power7 or newer.
22616 (enum r6000_reg_class_enum): Add new constraints.
22617
22618 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22619 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22620 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 22621 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM		 22622 (movsd splitter): Likewise.
22623 (movsd_hardfloat): Likewise.
22624 (movsd_softfloat): Likewise.
22625
22626 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22627 binary and decimal floating point moves.
22628 (fmove_ok): New attributes to combine binary and decimal floating
22629 point moves, and to combine power6x (mfpgpr) moves along normal
22630 floating moves.
22631 (real_value_to_target): Likewise.
22632 (f32_lr): Likewise.
22633 (f32_lm): Likewise.
22634 (f32_li): Likewise.
22635 (f32_sr): Likewise.
22636 (f32_sm): Likewise.
22637 (f32_si): Likewise.
22638 (movsf): Combine binary and decimal floating point moves. Combine
22639 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 22640 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM		 22641 (mov<mode> for SFmode/SDmode); Likewise.
22642 (SFmode/SDmode splitters): Likewise.
22643 (movsf_hardfloat): Likewise.
22644 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22645 (movsf_softfloat): Likewise.
22646 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22647
22648 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22649 wx and wz constraints.
22650
22651 * config/rs6000/constraints.md (wg constraint): New constraint to
22652 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22653
22654 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22655 constraint.
22656
22657 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22658 -mdebug=reg, print wg, wl, wx, and wz constraints.
22659 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22660 Initialize the reload functions for 64-bit binary/decimal floating
22661 point types.
22662 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22663 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22664 create the buffer on the stack to overcome not having a 32-bit
22665 load and store.
22666 (rs6000_emit_move): Likewise.
22667 (rs6000_secondary_memory_needed_rtx): Likewise.
22668 (rs6000_alloc_sdmode_stack_slot): Likewise.
22669 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22670 via xxlxor, just like DFmode 0.0.
22671
c6d5ff83
MM		 22672 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22673 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22674 with other moves by using conditional constraits (wg). Use LFIWZX
22675 and STFIWX for loading SDmode on power7.
22676 (movdd splitters): Likewise.
22677 (movdd_hardfloat32): Likewise.
22678 (movdd_softfloat32): Likewise.
22679 (movdd_hardfloat64_mfpgpr): Likewise.
22680 (movdd_hardfloat64): Likewise.
22681 (movdd_softfloat64): Likewise.
22682
22683 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22684 64-bit binary and decimal floating point moves.
22685 (FMOVE64X): Likewise.
22686 (movdf): Combine 64-bit binary and decimal floating point moves.
22687 Combine power6x (mfpgpr) moves with other moves by using
22688 conditional constraits (wg).
22689 (mov<mode> for DFmode/DDmode): Likewise.
22690 (DFmode/DDmode splitters): Likewise.
22691 (movdf_hardfloat32): Likewise.
22692 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22693 (movdf_softfloat32): Likewise.
22694 (movdf_hardfloat64_mfpgpr): Likewise.
22695 (movdf_hardfloat64): Likewise.
22696 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22697 (movdf_softfloat64): Likewise.
22698 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22699 (reload_<mode>_load): Move to later in the file so they aren't in
22700 the middle of the floating point move insns.
22701 (reload_<mode>_store): Likewise.
22702
22703 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22704 constraint.
22705
22706 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22707 constraint if -mdebug=reg.
bd059b26
UB		 22708 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22709 Enable using dd reload support if needed.
c6d5ff83
MM		 22710
22711 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22712 binary and decimal floating point moves in rs6000.md.
22713 (movtd_internal): Likewise.
22714
22715 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22716 decimal floating point moves.
22717 (movtf): Likewise.
22718 (movtf_internal): Likewise.
22719 (mov<mode>_internal, TDmode/TFmode): Likewise.
22720 (movtf_softfloat): Likewise.
22721 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22722
22723 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22724 movdi_internal64, using wg constraint for move direct operations.
22725 (movdi_internal64): Likewise.
22726
22727 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22728 MODES_TIEABLE_P for selected modes. Print the numerical value of
22729 the various virtual registers. Use GPR/FPR first/last values,
22730 instead of hard coding the register numbers. Print which modes
22731 have reload functions registered.
bd059b26
UB		 22732 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22733 settings before/after setting cpu, target and subtarget settings.
22734 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22735 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM		 22736 (rs6000_secondary_reload_fail): Likewise.
22737 (rs6000_secondary_reload_inner): Likewise.
22738
22739 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22740 macros for first/last GPR and FPR registers.
22741 (LAST_GPR_REGNO): Likewise.
22742 (FIRST_FPR_REGNO): Likewise.
22743 (LAST_FPR_REGNO): Likewise.
22744
22745 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22746 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22747 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22748 (vcond<mode><mode>): Likewise.
22749 (vcondu<mode><mode>): Likewise.
22750 (vector_gtu<mode>): Likewise.
22751 (vector_gte<mode>): Likewise.
22752 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 22753 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM		 22754 (ior<mode>3): Likewise.
22755 (and<mode>3): Likewise.
22756 (one_cmpl<mode>2): Likewise.
22757 (nor<mode>3): Likewise.
22758 (andc<mode>3): Likewise.
22759
22760 * config/rs6000/constraints.md (wt constraint): New constraint
22761 that returns VSX_REGS if TImode is allowed in VSX registers.
22762
22763 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22764 constant under VSX.
22765
22766 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22767 similar to TImode, but it is restricted to being in the GPRs.
22768
22769 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22770 TImode to occupy a single VSX register.
22771
22772 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22773 -mvsx-timode for power7/power8.
22774 (power7 cpu): Likewise.
22775 (power8 cpu): Likewise.
22776
22777 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22778 sure that TFmode/TDmode take up two registers if they are ever
22779 allowed in the upper VSX registers.
22780 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22781 registers.
22782 (rs6000_init_hard_regno_mode_ok): Likewise.
22783 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22784 constraint. Print if LRA is turned on.
22785 (rs6000_option_override_internal): Give an error if -mvsx-timode
22786 and VSX is not enabled.
22787 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22788 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22789 to reg+offset addressing. Use PTImode when checking offset
22790 addresses for validity.
22791 (reg_offset_addressing_ok_p): Likewise.
22792 (rs6000_legitimate_offset_address_p): Likewise.
22793 (rs6000_legitimize_address): Likewise.
22794 (rs6000_legitimize_reload_address): Likewise.
22795 (rs6000_legitimate_address_p): Likewise.
22796 (rs6000_eliminate_indexed_memrefs): Likewise.
22797 (rs6000_emit_move): Likewise.
22798 (rs6000_secondary_reload): Likewise.
22799 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22800 reloads to fpr registers to continue to use reg+offset addressing,
22801 but 64-bit reloads to altivec registers need reg+reg addressing.
22802 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22803 it. Treat LO_SUM like a PLUS operation.
22804 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 22805 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM		 22806 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22807 registers to share a register with a smaller sized type, since VSX
22808 puts scalars in the upper 64-bits.
22809 (print_operand): Add support for PTImode.
22810 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22811 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22812 registers, but don't have arithmetic support.
22813 (rs6000_memory_move_cost): Add test for VSX.
22814 (rs6000_opt_masks): Add -mvsx-timode.
22815
22816 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22817 for TImode.
22818 (VSs): Likewise.
22819 (VSr): Use wt constraint for TImode.
22820 (VSv): Drop TImode support.
22821 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22822 (vsx_movti_64bit): Likewise.
22823 (vsx_movti_32bit): Likewise.
22824 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22825 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22826 one '?' on the appropriate output constraint. Do not allow TImode
22827 logical operations on 32-bit systems.
22828 (vsx_ior<mode>3): Likewise.
22829 (vsx_xor<mode>3): Likewise.
22830 (vsx_one_cmpl<mode>2): Likewise.
22831 (vsx_nor<mode>3): Likewise.
22832 (vsx_andc<mode>3): Likewise.
22833 (vsx_concat_<mode>): Likewise.
22834 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22835
22836 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22837 OPTION_MASK_VSX_TIMODE.
22838 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22839 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22840
22841 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22842 (TI2 iterator): New iterator for TImode, PTImode.
22843 (wd mode attribute): Add values for vector types.
bd059b26
UB		 22844 (movti_string): Replace TI move operations with operations for TImode
22845 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM		 22846 (mov<mode>_string, TImode/PTImode): Likewise.
22847 (movti_ppc64): Likewise.
22848 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22849 (TI mode splitters): Likewise.
22850
22851 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22852 constraint.
22853
1fc5eced
MG		 228542013-03-20 Marc Glisse <marc.glisse@inria.fr>
22855
22856 PR tree-optimization/56355
22857 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22858 Also handle integers with undefined overflow.
22859
22c4c869
CM		 228602013-03-20 Catherine Moore <clm@codesourcery.com>
22861 Maciej W. Rozycki <macro@codesourcery.com>
22862 Tom de Vries <tom@codesourcery.com>
12211b99 22863 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM		 22864 Iain Sandoe <iain@codesourcery.com>
22865 Nathan Froyd <froydnj@codesourcery.com>
12211b99 22866 Chao-ying Fu <fu@mips.com>
22c4c869
CM		 22867
22868 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 22869 Document new function attributes.
22c4c869
CM		 22870 * doc/invoke.texi (minterlink-compressed, mmicromips,
22871 m14k, m14ke, m14kec): Document new options.
22872 (minterlink-mips16): Update documentation.
22873 * doc/md.texi (ZC, ZD): Document new constraints.
22874 * configure.ac (gcc_cv_as_micromips): Check if linker
22875 supports the .set micromips directive.
22876 * configure: Regenerate.
22877 * config.in: Regenerate.
22878 * config/mips/mips-tables.opt: Regenerate.
22879 * config/mips/micromips.md: New file.
22880 * constraints.md (ZC, ZD): New constraints.
22881 * config/mips/predicates.md (movep_src_register): New predicate.
22882 (movep_src_operand): New predicate.
22883 (non_volatile_mem_operand): New predicate.
22884 * config/mips/mips.md (multimem): New type.
22885 (length): Differentiate between 17-bit and 18-bit branch offsets.
22886 (MOVEP1, MOVEP2): New mode iterator.
33159866 22887 (mov_<load>l): Use ZC constraint.
22c4c869
CM		 22888 (mov_<load>r): Likewise.
22889 (mov_<store>l): Likewise.
22890 (mov_<store>r): Likewise.
22891 (*branch_equality<mode>_inverted): Add microMIPS support.
22892 (*branch_equality<mode>): Likewise.
22893 (*jump_absolute): Likewise.
22894 (indirect_jump_<mode>): Likewise.
22895 (tablejump_<mode>): Likewise.
22896 (<optab>_internal): Likewise.
22897 (sibcall_internal): Likewise.
22898 (sibcall_value_internal): Likewise.
22899 (prefetch): Use constraint ZD.
22900 * config/mips/mips.opt (minterlink-compressed): New option.
22901 (minterlink-mips16): Now an alias for minterlink-compressed.
22902 (mmicromips): New option.
22903 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22904 (compare_and_swap_12): Likewise.
22905 (sync_add<mode>): Likewise.
22906 (sync_<optab>_12): Likewise.
22907 (sync_old_<optab>_12): Likewise.
22908 (sync_new_<optab>_12): Likewise.
22909 (sync_nand_12): Likewise.
22910 (sync_old_nand_12): Likewise.
22911 (sync_new_nand_12): Likewise.
22912 (sync_sub<mode>): Likewise.
22913 (sync_old_add<mode>): Likewise.
22914 (sync_old_sub<mode>): Likewise.
22915 (sync_new_add<mode>): Likewise.
22916 (sync_new_sub<mode>): Likewise.
22917 (sync_<optab><mode>): Likewise.
22918 (sync_old_<optab><mode>): Likewise.
22919 (sync_new_<optab><mode>): Likewise.
22920 (sync_nand<mode>): Likewise.
22921 (sync_old_nand<mode>): Likewise.
22922 (sync_new_nand<mode>): Likewise.
22923 (sync_lock_test_and_set<mode>): Likewise.
22924 (test_and_set_12): Likewise.
22925 (atomic_compare_and_swap<mode>): Likewise.
22926 (atomic_exchange<mode>_llsc): Likewise.
22927 (atomic_fetch_add<mode>_llsc): Likewise.
22928 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22929 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22930 (umips_save_restore_pattern_p): Likewise.
22931 (umips_load_store_pair_p): Likewise.
22932 (umips_output_load_store_pair): Likewise.
22933 (umips_movep_target_p): Likewise.
22934 (umips_12bit_offset_address_p): Likewise.
22935 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22936 (mips_base_mips16): Rename this...
22937 (mips_base_compression_flags): ...to this. Update all uses.
22938 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22939 (mips_mips16_decl_p): Delete.
22940 (mips_nomips16_decl_p): Delete.
33159866 22941 (mips_get_compress_on_flags): New function.
22c4c869
CM		 22942 (mips_get_compress_off_flags): New function.
22943 (mips_get_compress_mode): New function.
22944 (mips_get_compress_on_name): New function.
22945 (mips_get_compress_off_name): New function.
22946 (mips_insert_attributes): Support multiple compression types.
22947 (mips_merge_decl_attributes): Likewise.
22948 (umips_12bit_offset_address_p): New function.
22949 (mips_start_function_definition): Emit .set micromips directive.
22950 (mips_call_may_need_jalx_p): New function.
22951 (mips_function_ok_for_sibcall): Add microMIPS support.
22952 (mips_print_operand_punctuation): Support short delay slots and
22953 compact jumps.
22954 (umips_swm_mask, umips_swm_encoding): New.
22955 (umips_build_save_restore): New function.
22956 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22957 (was_mips16_p): Remove.
22958 (old_compression_mode): New.
22959 (mips_set_compression_mode): New function.
22960 (mips_set_current_function): Add microMIPS support.
22961 (mips_option_override): Likewise.
22962 (umips_save_restore_pattern_p): New function.
22963 (umips_output_save_restore): New function.
22964 (umips_load_store_pair_p_1): New function.
22965 (umips_load_store_pair_p): New function.
22966 (umips_output_load_store_pair_1): New function.
22967 (umips_output_load_store_pair): New function.
22968 (umips_movep_target_p) New function.
22969 (mips_prepare_pch_save): Add microMIPS support.
22970 * config/mips/mips.h (TARGET_COMPRESSION): New.
22971 (TARGET_CPU_CPP_BUILTINS): Update macro
22972 to use new compression flags and to support microMIPS.
22973 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22974 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22975 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22976 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22977 (ASM_SPEC): Support mmicromips and mno-micromips.
22978 (M16STORE_REG_P): New macro.
22979 (MIPS_CALL): Support TARGET_MICROMIPS.
22980 (MICROMIPS_J): New macro.
22981 (mips_base_mips16): Rename this...
22982 (mips_base_compression_flags): ...to this.
22983 (UMIPS_12BIT_OFFSET_P): New macro.
22984 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22985 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB		 229862013-03-20 Richard Biener <rguenther@suse.de>
22987
22988 PR tree-optimization/56661
22989 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22990 the result does not have to be distinct.
22991
54714c68
RB		 229922013-03-20 Richard Biener <rguenther@suse.de>
22993
22994 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22995 remap_gimple_op_r.
22996
cca1130d 229972013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 22998 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS		 22999
23000 PR rtl-optimization/56605
23001 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23002
2a930377
UB		 230032013-03-20 Uros Bizjak <ubizjak@gmail.com>
23004
23005 PR bootstrap/56656
23006 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23007 that require movd instead of movq.
23008
d6d305fe
RB		 230092013-03-20 Richard Biener <rguenther@suse.de>
23010
23011 * tree-ssa-structalias.c (struct variable_info): Add pointer
23012 to the first field of an aggregate with sub-vars. Make
23013 this and the pointer to the next subfield its ID.
23014 (vi_next): New function.
23015 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23016 storedanything_id, integer_id): Increment by one.
23017 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23018 get_call_clobber_vi): Adjust.
23019 (solution_set_expand): Simplify and speedup.
23020 (solution_set_add): Inline into ...
23021 (set_union_with_increment): ... this. Adjust accordingly.
23022 (do_sd_constraint): Likewise.
23023 (do_ds_constraint): Likewise.
23024 (do_complex_constraint): Simplify.
23025 (build_pred_graph): Adjust.
23026 (solve_graph): Likewise. Simplify and speedup.
23027 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23028 get_constraint_for_component_ref, get_constraint_for_1,
23029 first_vi_for_offset, first_or_preceding_vi_for_offset,
23030 create_function_info_for, create_variable_info_for_1,
23031 create_variable_info_for, intra_create_variable_infos): Adjust.
23032 (init_base_vars): Push NULL for ID zero.
23033 (compute_points_to_sets): Adjust.
23034
a271b42d
RB		 230352013-03-20 Richard Biener <rguenther@suse.de>
23036
23037 * cfgloop.c (verify_loop_structure): Streamline and avoid
23038 ICEing on corrupt loop tree.
23039 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23040 loop tree.
23041
ebd65954
RB		 230422013-03-20 Richard Biener <rguenther@suse.de>
23043
23044 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23045 check whether an SSA update is needed.
23046
4547b7ee
RS		 230472013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23048
23049 * config/mips/constraints.md (T): Rename to...
23050 (Yf): ...this.
23051 (U): Rename to...
23052 (Yd): ...this.
23053 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23054 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23055
0b8cdc58
IB		 230562013-03-19 Ian Bolton <ian.bolton@arm.com>
23057
23058 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23059 (*subsi3_carryin_uxtw): Likewise.
23060
5977a10d
IB		 230612013-03-19 Ian Bolton <ian.bolton@arm.com>
23062
23063 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23064 (*rorsi3_insn_uxtw): Likewise.
23065
bd83ff2c
IB		 230662013-03-19 Ian Bolton <ian.bolton@arm.com>
23067
23068 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23069 (*extrsi5_insn_uxtw): Likewise.
23070
ba65123c
RB		 230712013-03-19 Richard Biener <rguenther@suse.de>
23072
23073 PR tree-optimization/56273
23074 * passes.c (init_optimization_passes): Move second VRP after DOM.
23075
2eac0476
UB		 230762013-03-19 Uros Bizjak <ubizjak@gmail.com>
23077
23078 * config/i386/i386.md (*movti_internal): Merge from
23079 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23080 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23081 nox64 isa attributes.
23082
6e55eda7
RB		 230832013-03-18 Richard Biener <rguenther@suse.de>
23084
23085 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23086 (unite): Likewise.
23087 (merge_node_constraints): Likewise.
23088 (build_succ_graph): Likewise.
23089 (valid_graph_edge): Inline into single caller.
23090 (unify_nodes): Likewise. Use bitmap_set_bit return value
23091 and cache varinfo.
23092 (scc_visit): Fix formatting and variable use.
23093 (do_sd_constraint): Use gcc_checking_assert.
23094 (do_ds_constraint): Likewise.
23095 (do_complex_constraint): Likewise.
23096 (condense_visit): Likewise. Cleanup.
23097 (dump_pred_graph): New function.
23098 (perform_var_substitution): Dump the pred-graph before
23099 variable substitution.
23100 (find_equivalent_node): Use gcc_checking_assert.
23101 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23102
4bdd44c4
RB		 231032013-03-18 Richard Biener <rguenther@suse.de>
23104
23105 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23106 Remove cond_expr_stmt_list argument and do not gimplify the
23107 built expression.
23108 (vect_loop_versioning): Adjust.
23109 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23110 Cleanup to use less temporaries.
23111 (vect_create_data_ref_ptr): Cleanup.
23112
38c56a5b
JJ		 231132013-03-18 Jakub Jelinek <jakub@redhat.com>
23114
23115 PR tree-optimization/56635
23116 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23117 require types_compatible_p types.
23118
20b2e6a0
NC		 231192013-03-18 Nick Clifton <nickc@redhat.com>
23120
a6178a25
NC		 23121 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23122 spurious backslash.
23123
20b2e6a0
NC		 23124 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23125 Add missing line to comment describing function.
23126
92608d0e
RB		 231272013-03-18 Richard Biener <rguenther@suse.de>
23128
23129 PR tree-optimization/56210
23130 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23131 Handle string / character search functions.
23132 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23133
31348d52
RB		 231342013-03-18 Richard Biener <rguenther@suse.de>
23135
23136 PR middle-end/56483
23137 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23138 and implement properly.
23139 * gimple.h (gimple_cond_single_var_p): Remove.
23140
fcac74a1
RB		 231412013-03-18 Richard Biener <rguenther@suse.de>
23142
23143 * tree-data-ref.h (find_data_references_in_loop): Declare.
23144 * tree-data-ref.c (get_references_in_stmt): Use a stack
23145 vector pre-allocated in the callers.
23146 (find_data_references_in_stmt): Adjust.
23147 (graphite_find_data_references_in_stmt): Likewise.
23148 (create_rdg_vertices): Likewise.
23149 (find_data_references_in_loop): Export.
23150 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23151 Compute dependences here...
23152 (vect_analyze_data_refs): ...not here. When we encounter
23153 a non-vectorizable data reference in basic-block vectorization
23154 truncate the data reference vector. Do not bother to
23155 fixup data-dependence information for gather loads.
23156 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23157 of data references, as reported.
23158
0d5a1b56
RB		 231592013-03-18 Richard Biener <rguenther@suse.de>
23160
23161 PR tree-optimization/3713
23162 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23163 has_constants and expr.
23164 (stmt_has_constants): Properly valueize SSA names when deciding
23165 whether the stmt has constants.
23166
789c34e3
RB		 231672013-03-18 Richard Biener <rguenther@suse.de>
23168
23169 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23170 whole function when there is nothing to do.
23171 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23172 * tree-vectorizer.c (vectorize_loops): Update virtual and
23173 loop-closed SSA once.
23174 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23175
076b4605
RB		 231762013-03-18 Richard Biener <rguenther@suse.de>
23177
23178 PR middle-end/56113
23179 * domwalk.c (bb_postorder): New global static.
23180 (cmp_bb_postorder): New function.
23181 (walk_dominator_tree): Replace scheme imposing an order for
23182 visiting dominator sons by one sorting them at the time they
23183 are pushed on the stack.
23184
bdb01696
RB		 231852013-03-18 Richard Biener <rguenther@suse.de>
23186
23187 PR tree-optimization/39326
23188 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23189 (struct mem_ref): Replace mem member with ao_ref typed member.
23190 (MEM_ANALYZABLE): Adjust.
23191 (memref_eq): Likewise.
23192 (mem_ref_alloc): Likewise.
23193 (gather_mem_refs_stmt): Likewise.
23194 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23195 (execute_sm_if_changed_flag_set): Adjust.
23196 (execute_sm): Likewise.
23197 (ref_always_accessed_p): Likewise.
23198 (refs_independent_p): Likewise.
23199 (can_sm_ref_p): Likewise.
23200
12d2dc5e
JJ		 232012013-03-18 Jakub Jelinek <jakub@redhat.com>
23202
23203 PR c/56566
23204 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23205 return 1 even for !unsignedp.
23206
286fb677
UB		 232072013-03-17 Uros Bizjak <ubizjak@gmail.com>
23208
23209 * config/i386/i386.md (isa): Add x64 and nox64.
23210 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23211 (*pushtf): Enable *roF alternative for x64 isa only.
23212 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23213 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23214 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23215 (*movtf_internal): Merge from *movtf_internal_rex64 and
23216 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23217 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23218 nox64 isa attributes.
23219 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23220 nox64 isa attributes.
23221 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23222
88b97037
UB		 232232013-03-17 Uros Bizjak <ubizjak@gmail.com>
23224
23225 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23226
9f4f1735
JJ		 232272013-03-17 Jakub Jelinek <jakub@redhat.com>
23228
23229 PR target/56640
23230 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23231 class names. Remove trailing comma after "ALL_REGS".
23232
48f4a6fa
JH		 232332013-03-16 Jan Hubicka <jh@suse.cz>
23234
23235 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23236 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23237 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23238 of cgraph_get_create_node.
23239 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23240
98e81123
JM		 232412013-03-16 Jason Merrill <jason@redhat.com>
23242
23243 PR debug/49090
23244 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23245 with DW_AT_default_value.
23246
8a298c04
JJ		 232472013-03-16 Jakub Jelinek <jakub@redhat.com>
23248
23249 * BASE-VER: Set to 4.9.0.
23250
4323afa0
AK		 232512013-03-14 Andi Kleen <ak@linux.intel.com>
23252
23253 PR target/56619
23254 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23255 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23256 Document _x* TSX intrinsics.
23257
b3c0d469
JJ		 232582013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23259 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI		 23260
23261 * configure.ac: Add MicroBlaze TLS support detection.
23262 * configure: Regenerate.
88b97037
UB		 23263 * config/microblaze/microblaze-protos.h
23264 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23265 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23266 * config/microblaze/microblaze.c (microblaze_address_type): Add
23267 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI		 23268 (microblaze_address_info): Add tls_reloc.
23269 (TARGET_HAVE_TLS): Define.
23270 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23271 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB		 23272 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23273 load_tls_operand, microblaze_call_tls_get_addr,
23274 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI		 23275 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23276 (get_base_reg): Use microblaze_tls_symbol_p.
23277 (microblaze_classify_address): Handle TLS.
88b97037
UB		 23278 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23279 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI		 23280 (microblaze_legitimize_address): Handle TLS.
23281 (microblaze_address_insns): Handle ADDRESS_TLS.
23282 (pic_address_needs_scratch): Handle TLS.
23283 (print_operand_address): Handle TLS.
23284 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23285 (microblaze_expand_move): Handle TLS.
88b97037
UB		 23286 (microblaze_legitimate_constant_p): Check
23287 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI		 23288 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23289 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23290 (PIC_OFFSET_TABLE_REGNUM): Set.
23291 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23292 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23293 (addsi3, movsi_internal2, movdf_internal): Update constraints
23294 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB		 23295 (move_operand): Redefine as move_src_operand,
23296 check microblaze_tls_referenced_p.
8cc9a5a5 23297
d803a491
IB		 232982013-03-14 Ian Bolton <ian.bolton@arm.com>
23299
23300 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23301 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23302
a8504f22
IB		 233032013-03-14 Ian Bolton <ian.bolton@arm.com>
23304
23305 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23306 CC mode for AND.
23307
df2dfaea
JJ		 233082013-03-14 Jakub Jelinek <jakub@redhat.com>
23309
fbd28bc3
JJ		 23310 PR tree-optimization/53265
23311 * common.opt (Waggressive-loop-optimizations): New option.
23312 * tree-ssa-loop-niter.c: Include tree-pass.h.
23313 (do_warn_aggressive_loop_optimizations): New function.
23314 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23315 if number_of_latch_executions returned constant.
23316 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23317 early. If number_of_latch_executions returned constant, set
23318 nb_iterations_upper_bound back to it.
23319 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23320 field.
23321 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23322 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23323
df2dfaea
JJ		 23324 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23325 (MULTILIB_OSDIRNAMES): Set.
23326 * genmultilib: If defaultosdirname doesn't start with :: , set
23327 defaultosdirname2 instead, clear it and emit two . multilib_raw
23328 entries instead of just one.
23329
ee0d2b68
KK		 233302013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23331
23332 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23333 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23334 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23335 (SUBTARGET_OVERRIDE_OPTIONS): New.
23336
decc676e
OE		 233372013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23338
23339 PR target/49880
23340 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23341 (musermode): Convert to Var(TARGET_USERMODE).
23342 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23343 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23344 * config/sh/sh.c (sh_option_override): Use
23345 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23346 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23347 condition.
23348 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23349 TARGET_SH4.
23350 (udivsi3_i4_single, divsi3_i4_single): Use
23351 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23352
f4b719c7
DK		 233532013-03-13 Dave Korn <dave.korn.cygwin@....>
23354
23355 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23356 default setting.
23357
c40eced0
RB		 233582013-03-13 Richard Biener <rguenther@suse.de>
23359
23360 PR tree-optimization/56608
23361 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23362 calls when vectorizing basic-blocks.
23363
1bfa5973
JJ		 233642013-03-13 Jakub Jelinek <jakub@redhat.com>
23365
23366 PR plugins/45078
23367 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23368 tm_file.
23369
6fcf5434
JJ		 233702013-03-12 Jakub Jelinek <jakub@redhat.com>
23371
23372 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23373
4f38fa8c
JH		 233742013-03-11 Jan Hubicka <jh@suse.cz>
23375
23376 PR lto/56557
23377 * lto-streamer-out.c (output_symbol_p): Skip references from
23378 constructors of external variables.
23379
c5c5ba89
JH		 233802013-03-11 Jan Hubicka <jh@suse.cz>
23381
23382 PR middle-end/56571
23383 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23384 from pseudos.
23385 * emit-rtl.c (verify_rtx_sharing): Likewise.
23386 (copy_insn_1): Likewise.
23387 * rtl.c (copy_rtx): Likewise.
23388
c2a939b1
GJL		 233892013-03-11 Georg-Johann Lay <avr@gjlay.de>
23390
23391 PR target/56591
23392 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23393 output_operand_lossage message.
88b97037 23394
3c3279fb
RE		 233952013-03-11 Richard Earnshaw <rearnsha@arm.com>
23396
23397 PR target/56470
23398 * arm.c (shift_op): Validate RTL pattern on the fly.
23399 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23400 the RTL.
23401
aef5ef9d
JDA		 234022013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23403
23404 PR target/56347
23405 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23406 new call patterns that clobber %fr12.
23407 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23408 split and postreload patterns.
23409 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23410 registers %fr12 and %fr12R as call used.
23411
456610d3
SB		 234122013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23413
23414 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23415 (canon_address, record_store, replace_read, check_mem_read_rtx,
23416 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23417 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23418 rest_of_handle_dse): Likewise.
23419
4b1baac8
RS		 234202013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23421
23422 PR middle-end/56524
23423 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23424 Add base_optabs.
23425 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23426 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23427 (save_optabs_if_changed): Replace with...
23428 (init_tree_optimization_optabs): ...this.
23429 * optabs.c (save_optabs_if_changed): Rename to...
23430 (init_tree_optimization_optabs): ...this. Take the optimization node
23431 as argument. Do nothing if the base optabs are already correct.
23432 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23433 to recompute optabs.
23434 * function.h (function): Remove optabs field.
23435 * function.c (invoke_set_current_function_hook): Call
23436 init_tree_optimization_optabs. Use the result to initialize
23437 this_fn_optabs.
23438
b7a78683
AH		 234392013-02-27 Aldy Hernandez <aldyh@redhat.com>
23440
23441 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23442 if GTMA_HAS_NO_INSTRUMENTATION.
23443 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23444 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23445 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23446 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23447 GTMA_HAS_NO_INSTRUMENTATION.
23448
6384c29b
JJ		 234492013-03-08 Jakub Jelinek <jakub@redhat.com>
23450
23451 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23452 libasan_preinit.o.
23453
ecd4f20a
MP		 234542013-03-08 Marek Polacek <polacek@redhat.com>
23455 Jakub Jelinek <jakub@redhat.com>
23456
23457 PR tree-optimization/56478
23458 * predict.c (is_comparison_with_loop_invariant_p): Change the
23459 type of loop_step to tree.
23460 (predict_loops): Adjust.
23461 (predict_iv_comparison): Perform the computations on double_ints.
23462
64366d35
RB		 234632013-03-08 Richard Biener <rguenther@suse.de>
23464
23465 PR tree-optimization/56570
23466 * tree-cfg.c (verify_expr_location_1): Verify locations for
23467 DECL_DEBUG_EXPR.
23468 * tree-sra.c (create_access_replacement): Strip locations
23469 from DECL_DEBUG_EXPRs.
23470
a9d5a059
RB		 234712013-03-08 Richard Biener <rguenther@suse.de>
23472
23473 * tree-inline.c (expand_call_inline): Do not associate
23474 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23475 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23476
b7aa4e9a
RB		 234772013-03-08 Richard Biener <rguenther@suse.de>
23478
23479 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23480 or block changes with -Og. Fix for location / block encoding
23481 changes and PHI arguments with locations.
23482
c4c2f9fa
SB		 234832013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23484
23485 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23486 for all counters.
23487 (struct output_info): Likewise.
23488 (register_overhead): Remove bad gcc_assert.
23489 (bitmap_find_bit): If there is only a single bitmap element, do not
23490 count a miss as a search.
23491 (print_statistics): Update for counter type changes.
23492 (dump_bitmap_statistics): Likewise. Print headers such that they
23493 are properly lined up with the printed counters.
23494
5bf6606a
JJ		 234952013-03-07 Jakub Jelinek <jakub@redhat.com>
23496
23497 PR tree-optimization/56559
23498 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23499 check that it has only a single use.
23500
2c653d46
RB		 235012013-03-07 Richard Biener <rguenther@suse.de>
23502
23503 * doc/invoke.texi (fwhole-program): Discourage use in combination
23504 with -flto.
23505
a72d8780
JJ		 235062013-03-06 Jakub Jelinek <jakub@redhat.com>
23507
01a454df
JJ		 23508 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23509
c1781047
JJ		 23510 PR tree-optimization/56539
23511 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23512 instead of GSI_CONTINUE_LINKING as last argument to
23513 force_gimple_operand_gsi. Adjust function comment.
23514
9772c47a
JJ		 23515 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23516 aarch64-cores.def.
23517
a72d8780
JJ		 23518 PR middle-end/56548
23519 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23520 promoted mode, convert the result back to the original mode.
23521
fa5556de
RB		 235222013-03-06 Richard Biener <rguenther@suse.de>
23523
23524 PR middle-end/56294
23525 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23526 (insert_updated_phi_nodes_compare_uids): New function.
23527 (update_ssa): Sort symbols_to_rename after UID before
23528 traversing it to insert PHI nodes.
23529
010403d1
RB		 235302013-03-06 Richard Biener <rguenther@suse.de>
23531
23532 PR middle-end/50494
23533 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23534 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23535
23536 Revert
23537 2013-02-13 Richard Biener <rguenther@suse.de>
23538
23539 PR lto/50494
23540 * varasm.c (output_constant_def_1): Get the decl representing
23541 the constant as argument.
23542 (output_constant_def): Wrap output_constant_def_1.
23543 (make_decl_rtl): Use output_constant_def_1 with the decl
23544 representing the constant.
23545 (build_constant_desc): Optionally re-use a decl already
23546 representing the constant.
23547 (tree_output_constant_def): Adjust.
23548
3c27ce4c
JY		 235492013-03-06 Joey Ye <joey.ye@arm.com>
23550
23551 PR lto/50293
23552 * gcc.c (convert_white_space): New function.
23553 (main): Handles white space in function name.
23554
8f6d1c86
OE		 235552013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23556
23557 PR target/56529
23558 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23559 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23560 to SH_DIV_CALL_TABLE for TARGET_SH2.
23561 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23562 list.
23563 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23564 call-table options.
23565
798a209f
SA		 235662013-03-05 Sterling Augustine <saugustine@google.com>
23567 Cary Coutant <ccoutant@google.com>
23568
23569 PR debug/55364
23570 * dwarf2out.c (resolve_addr): Don't call
23571 remove_loc_list_addr_table_entries a second time for the same
23572 expression.
23573
6cfa417f 235742013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 23575
6cfa417f
JJ		 23576 PR debug/56510
23577 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23578 (avoid_complex_debug_insns): New function.
23579 (expand_debug_locations): Call it.
23580
23581 PR rtl-optimization/56484
23582 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23583 lifetimes of hard registers on small register class machines.
23584
235852013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23586
23587 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 23588 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 23589 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH		 23590 fast_interrupt.
23591 (microblaze_fast_interrupt_function_p): New function.
23592 (microblaze_is_interrupt_handler): Rename to
23593 microblaze_is_interrupt_variant and add fast_interrupt check.
23594 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23595 (save_restore_insns): Likewise.
23596 (compute_frame_size): Likewise.
23597 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23598 (microblaze_globalize_label): Likewise.
6cfa417f
JJ		 23599 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23600 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH		 23601 microblaze_is_interrupt_variant.
23602
731aefac
KT		 236032013-03-05 Kai Tietz <ktietz@redhat.com>
23604
23605 * sdbout.c (sdbout_one_type): Switch to current function's section
23606 supporting cold/hot.
23607
a72d8780 236082013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH		 23609
23610 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23611 -mxl-reorder.
23612
334e71e8
JJ		 236132013-03-05 Jakub Jelinek <jakub@redhat.com>
23614
0b50e654
JJ		 23615 PR middle-end/56461
23616 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23617 if VALGRIND_GET_VBITS is defined, temporarily make object
23618 memory all defined, and restore previous valgrind addressability
23619 and definability afterwards. Free this_object at the end.
23620
4ccf8f43
JJ		 23621 PR middle-end/56461
23622 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23623 right before calling lra_create_live_ranges, also call it
23624 when clearing live_p. Only call lra_clear_live_ranges
23625 at the end if live_p.
23626
334e71e8
JJ		 23627 PR middle-end/56461
23628 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23629
9ca966ca
RB		 236302013-03-05 Richard Biener <rguenther@suse.de>
23631
23632 PR tree-optimization/56521
23633 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23634 value-id.
23635
d273b176
SB		 236362013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23637
23638 PR c++/55135
23639 * except.h (remove_unreachable_eh_regions): New prototype.
23640 * except.c (remove_eh_handler_splicer): New function, split out
23641 of remove_eh_handler.
23642 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23643 warning about running it on many EH regions one at a time.
23644 (remove_unreachable_eh_regions_worker): New function, walk the
23645 EH tree in depth-first order and remove non-marked regions.
23646 (remove_unreachable_eh_regions): New function.
23647 * tree-eh.c (mark_reachable_handlers): New function, split out
23648 from remove_unreachable_handlers.
23649 (remove_unreachable_handlers): Use mark_reachable_handlers and
23650 remove_unreachable_eh_regions.
23651 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23652 and remove_unreachable_eh_regions.
23653
5e7f1aef
RB		 236542013-03-05 Richard Biener <rguenther@suse.de>
23655
23656 PR middle-end/56525
23657 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23658 not freeing them until the end.
23659
f276b762
AK		 236602013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23661
23662 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23663
12a43ab8
RB		 236642013-03-05 Richard Biener <rguenther@suse.de>
23665
23666 PR tree-optimization/56270
23667 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23668 of loads after scheduling an SLP instance.
23669
93675444
JJ		 236702013-03-05 Jakub Jelinek <jakub@redhat.com>
23671
db4138e3
JJ		 23672 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23673 tic6x.exp.
23674 (check_gcc_parallelize): Run guality.exp as a separate job from
23675 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23676 struct-layout-1.exp with stackalign.exp.
23677
dd3d1ec0
JJ		 23678 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23679
b8d381a3
JJ		 23680 PR middle-end/56461
23681 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23682 load_index sbitmap even if some bit in it isn't set.
23683
b4f9786b
JJ		 23684 PR middle-end/56461
23685 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23686 (discover_iteration_bound_by_body_walk): Change queues to
23687 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23688 spelling in comment. Call safe_push on queues[bound_index] directly.
23689 Release queues[queue_index] in every iteration unconditionally.
23690 Release bounds vector.
23691
93675444
JJ		 23692 PR middle-end/56461
23693 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23694 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23695 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23696 inner_phis vector.
23697
3e492e9c
RB		 236982013-03-05 Richard Biener <rguenther@suse.de>
23699
23700 PR lto/56515
23701 * tree-inline.c (remap_blocks_to_null): New function.
23702 (expand_call_inline): When expanding a call stmt without
23703 an associated block inline remap all callee blocks to NULL.
23704
a5d1569a
JJ		 237052013-03-05 Jakub Jelinek <jakub@redhat.com>
23706
86efb5cd
JJ		 23707 PR rtl-optimization/56494
23708 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23709 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23710 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23711
85f5dbea
JJ		 23712 PR middle-end/56461
23713 * sel-sched-ir.c (free_sched_pools): Release
23714 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23715 if succs_info_pool.max_top isn't -1.
23716
a5d1569a
JJ		 23717 PR bootstrap/56509
23718 * opts.c (opts_obstack, opts_concat): Moved to...
23719 * opts-common.c (opts_obstack, opts_concat): ... here.
23720
4432aa6c
JJ		 237212013-03-04 Jakub Jelinek <jakub@redhat.com>
23722
23723 PR middle-end/56461
23724 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23725
f3a1fb91
MJ		 237262013-03-04 Martin Jambor <mjambor@suse.cz>
23727
23728 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23729 all appropriate places.
23730
22110e6c
EB		 237312013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23732
23733 PR tree-optimization/56424
23734 * ipa-split.c (split_function): Do not set the RSO flag if result is
23735 not by reference and its type is a register type.
23736
a72d8780 237372013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 23738
88b97037 23739 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 23740 (microblaze_legitimate_pic_operand): Likewise
88b97037 23741 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH		 23742 new function microblaze_legitimate_pic_operand
23743 * config/microblaze/microblaze-protos.h
23744 (microblaze_legitimate_pic_operand): Declare.
23745
a72d8780 237462013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 23747
a72d8780 23748 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 23749 New predicate for supported rtx code types.
a72d8780 23750 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI		 23751 call_insn_simple_operand predicate.
23752
541d9ac8
JJ		 237532013-03-04 Jakub Jelinek <jakub@redhat.com>
23754
5eb010bc
JJ		 23755 PR middle-end/56461
23756 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23757 partitions.ordered_remove.
23758
30862efc
JJ		 23759 PR middle-end/56461
23760 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23761 vec_oprnds0.create (1) for modifier == NONE.
23762
8930f723
JJ		 23763 PR middle-end/56461
23764 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23765 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23766 vec_oprnds1 right before pushing anything to it for
23767 scalar_shift_arg.
23768
541d9ac8
JJ		 23769 PR middle-end/56461
23770 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23771 set nbbs to 0 instead of having separate code path.
23772 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23773 instead of false as last argument if returning NULL.
23774
7aa7f2e3
SL		 237752013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23776
23777 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23778 the attribute is now called "target" instead of "option".
23779 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23780 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23781 attribute/pragma name for TARGET_OPTION_VALID_P and
23782 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23783 * doc/tm.texi: Regenerated.
23784
8930f723 237852013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 23786
8930f723 23787 * config/microblaze/microblaze.c:
cb8a1637 23788 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ		 23789 * config/microblaze/microblaze.h: Add -mxl-reorder to
23790 DRIVER_SELF_SPECS.
23791 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23792 instructions emitted if TARGET_REORDER.
88b97037
UB		 23793 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23794 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 23795 separately.
cb8a1637 23796
4dc1d68c
XDL		 237972013-03-01 Xinliang David Li <davidxl@google.com>
23798
23799 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23800 walk length.
23801
689eaba3
JJ		 238022013-03-01 Jakub Jelinek <jakub@redhat.com>
23803
9d676bf6
JJ		 23804 PR middle-end/56461
23805 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23806 vector even when returning true. Fix up function comment formatting.
23807
caff7edf
JJ		 23808 PR middle-end/56461
23809 * ira-build.c (ira_loop_nodes_count): New variable.
23810 (create_loop_tree_nodes): Initialize it.
23811 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23812
b6b9227d
JJ		 23813 PR middle-end/56461
23814 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23815 method on dr_chain and result_chain.
23816 * tree-vect-stmts.c (vectorizable_store): Only call
23817 result_chain.create if j == 0.
23818
689eaba3
JJ		 23819 PR middle-end/56461
23820 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23821 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23822 before overwriting it.
23823
40bf31ed
TB		 238242013-03-01 Tobias Burnus <burnus@net-b.de>
23825
23826 * doc/extended.texi (C Extensions): Change order in @menu
23827 to match @node.
23828 (Other MIPS Built-in Functions): Move last MIPS entry before
23829 "picoChip Built-in Functions".
23830 (SH Built-in Functions): Move after RX Built-in Functions.
23831 * doc/gcc.texi (Introduction): Change order in @menu
23832 to match @node.
23833 * doc/md.texi (Constraints): Ditto.
23834 * gty.texi (Type Information): Ditto.
23835 (User-provided marking routines for template types): Make
23836 subsection.
23837 * doc/invoke.texi (AArch64 Options): Move before
23838 "Adapteva Epiphany Options".
23839
e664c61c
KS		 238402013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23841 Jakub Jelinek <jakub@redhat.com>
23842
23843 PR sanitizer/56454
23844 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23845 no_address_safety_analysis attribute.
23846 * doc/extend.texi (no_address_safety_attribute): Rename to
23847 no_sanitize_address attribute, mention no_address_safety_analysis
23848 attribute as deprecated alias.
23849
37b5ec8f
JJ		 238502013-02-28 Jakub Jelinek <jakub@redhat.com>
23851
23852 PR middle-end/56461
23853 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23854 type to vec<vec<tree> > *.
23855 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23856 to be vec<tree> instead of vec<tree> *, set vec_defs
23857 to vNULL and call vec_defs.create (number_of_vects), adjust other
23858 uses of vec_defs.
23859 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23860 vectorizable_condition): Adjust vect_get_slp_defs callers.
23861
ba96cdfb
JG		 238622013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23863
23864 * config/aarch64/aarch64.c
23865 (aarch64_float_const_representable): Remove unused variable.
23866
6f549691
JG		 238672013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23868
23869 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23870
af55e82d
JG		 238712013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23872
23873 * config/aarch64/aarch64-builtins.c
23874 (aarch64_init_simd_builtins): Make static.
23875
1df3f464
JG		 238762013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23877
23878 * config/aarch64/aarch64.c
23879 (aarch64_simd_make_constant): Make static.
23880
f8f42513
MJ		 238812013-02-28 Martin Jambor <mjambor@suse.cz>
23882
23883 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23884 with no initialization to the RHS of debug statements.
23885
b48b3fc4
MJ		 238862013-02-28 Martin Jambor <mjambor@suse.cz>
23887
23888 PR tree-optimization/56294
23889 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23890 Adjust dumping.
23891 (get_access_replacement): Do not call create_access_replacement.
23892 Assert a replacement exists.
23893 (get_repl_default_def_ssa_name): Create the replacement declaration
23894 itself.
23895
c3ae224c
RR		 238962013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23897
23898 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23899 final_end_function.
23900
45fa0eef
MP		 239012013-02-28 Marek Polacek <polacek@redhat.com>
23902
23903 PR rtl-optimization/56466
23904 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23905 if we're changing a loop.
23906 (peel_loops_completely): Likewise.
23907
502c067d
PC		 239082013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23909
23910 PR c++/55813
23911 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23912
ff24afc8
GJL		 239132013-02-28 Georg-Johann Lay <avr@gjlay.de>
23914
23915 PR target/56445
23916 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23917 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23918 INTX_FTYPE_FX, FX_FTYPE_INTX.
23919 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23920
c7b1fc1b
GJL		 239212013-02-28 Georg-Johann Lay <avr@gjlay.de>
23922
23923 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23924 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23925 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23926 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23927 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23928 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23929 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23930 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23931 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23932 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23933 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23934 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23935 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23936 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23937 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23938 (avrxmega6): Increase max flash segments from 5 to 6.
23939 * config/avr/t-multilib: Regenerate.
23940 * config/avr/avr-tables.opt: Regenerate.
23941 * doc/avr-mmcu.texi: Regenerate.
23942
4a0e3cfe
GJL		 239432013-02-28 Georg-Johann Lay <avr@gjlay.de>
23944
23945 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23946 (avr_device_to_arch): Rename to avr_device_to_ld.
23947 (avr_device_to_as): New prototype.
23948 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23949 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23950 * config/avr/driver-avr.c (avr_device_to_as): New.
23951 (avr_device_to_arch): Rename to avr_device_to_ld.
23952
97785e52
JJ		 239532013-02-27 Jakub Jelinek <jakub@redhat.com>
23954
3f292312
JJ		 23955 PR middle-end/56461
23956 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23957 method on dr_chain and result_chain.
23958
a344216b
JJ		 23959 PR middle-end/56461
23960 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23961 pointer_set_destroy on not_executed_last_iteration.
23962
f121ad02 23963 PR middle-end/56461
88b97037 23964 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 23965
307f83a3
JJ		 23966 PR middle-end/56461
23967 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23968 FOR_EACH_DEFINED_FUNCTION when freeing state.
23969
e19624ee
JJ		 23970 PR middle-end/56461
23971 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23972 pool_free.
23973 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23974 overwriting it.
23975
90e709fd
JJ		 23976 PR middle-end/56461
23977 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23978 known_aggs[i].items and release known_aggs vector.
23979
97785e52
JJ		 23980 PR middle-end/56461
23981 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23982
239832013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 23984
97785e52
JJ		 23985 * config/microblaze/microblaze.c (microblaze_emit_compare):
23986 Use xor for EQ/NE comparisions.
23987 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23988 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 23989
6fa5e0ed
JJ		 239902013-02-27 Jakub Jelinek <jakub@redhat.com>
23991
23992 PR middle-end/56461
23993 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23994 vector.
23995 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23996 vec_safe_push, always update *slot.
23997 (redirect_edge_var_map_clear): Use vec_free.
23998 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23999 (free_var_map_entry): Use vec_free.
24000 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24001 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24002
436a956a
AB		 240032013-02-27 Andrey Belevantsev <abel@ispras.ru>
24004
24005 PR middle-end/45472
436a956a
AB		 24006 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24007 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB		 24008 Reorder tests for speculativeness in the logical and operator.
24009
f5c2caca
JJ		 240102013-02-27 Jakub Jelinek <jakub@redhat.com>
24011
0fcb564b
JJ		 24012 * incpath.c (add_standard_paths): Use reconcat instead of concat
24013 where appropriate and avoid leaking memory.
24014
dc357798
JJ		 24015 * opts.h: Include obstack.h.
24016 (opts_concat): New prototype.
24017 (opts_obstack): New declaration.
24018 * opts.c (opts_concat): New function.
24019 (opts_obstack): New variable.
24020 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24021 (finish_options): Use opts_concat instead of concat
24022 and XOBNEWVEC instead of XNEWVEC.
24023 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24024 generate_option): Likewise.
24025 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24026 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24027
f5c2caca
JJ		 24028 PR target/56455
24029 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24030 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24031
d0163673
JJ		 240322013-02-26 Jakub Jelinek <jakub@redhat.com>
24033
24034 PR middle-end/56461
24035 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24036
b5ad2b8e
JR		 240372013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24038
24039 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24040 (arm_block_move_unaligned_straight): Likewise.
24041 (arm_adjust_block_mem): Likewise.
24042
9b639e2c
JR		 240432013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24044
24045 PR target/48901
24046 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24047 temp, cond and label.
24048 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24049
e13a0ccb
JR		 24050 PR target/52500
24051 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24052 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24053
fbe4f171
JR		 24054 PR target/52501
24055 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24056 of prologue/epilogue functions.
24057
ae006543
JR		 24058 PR target/52550
24059 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24060 Remove unused variable cfa_offset.
24061 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24062
c81369fa
JR		 24063 PR target/54639
24064 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24065 type promotion to unsigned.
24066
f8a8fea7
JR		 24067 PR target/54640
24068 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24069 for HOST_WIDE_INT of 32 bit / same size as int.
24070 (arm_block_move_unaligned_straight): Likewise.
24071 (arm_adjust_block_mem): Likewise.
24072
f8be5169
JR		 24073 PR target/54662
24074 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24075 ALL_CFLAGS.
24076
f1ad3354
MP		 240772013-02-26 Marek Polacek <polacek@redhat.com>
24078
24079 PR tree-optimization/56426
88b97037 24080 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 24081
a6af0f42
RB		 240822013-02-26 Richard Biener <rguenther@suse.de>
24083
24084 PR target/56444
24085 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24086 unused variable loops.
24087
aca43c6c
JJ		 240882013-02-26 Jakub Jelinek <jakub@redhat.com>
24089
cecbe5d9
JJ		 24090 PR tree-optimization/56448
24091 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24092 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24093 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24094 later operands of the references, or even first operand for
24095 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24096
aca43c6c
JJ		 24097 PR tree-optimization/56443
24098 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24099 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24100 to type_for_mode langhook.
24101
53e2e141
MT		 241022013-02-25 Matt Turner <mattst88@gmail.com>
24103
24104 * doc/invoke.texi: Document r4700.
24105
259ee451
RB		 241062013-02-25 Richard Biener <rguenther@suse.de>
24107
24108 PR tree-optimization/56175
24109 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24110 split out from ...
24111 (simplify_bitwise_binary): ... here. Also guard the conversion
24112 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24113
6d65e8f1
CM		 241142013-02-25 Catherine Moore <clm@codesourcery.com>
24115
24116 Revert:
24117 2013-02-24 Catherine Moore <clm@codesourcery.com>
24118 Maciej W. Rozycki <macro@codesourcery.com>
24119 Tom de Vries <tom@codesourcery.com>
a72d8780 24120 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM		 24121 Iain Sandoe <iain@codesourcery.com>
24122 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 24123 Chao-ying Fu <fu@mips.com>
88b97037 24124
6d65e8f1 24125 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 24126 Document new function attributes.
6d65e8f1
CM		 24127 * doc/invoke.texi (minterlink-compressed, mmicromips,
24128 m14k, m14ke, m14kec): Document new options.
24129 (minterlink-mips16): Update documentation.
24130 * doc/md.texi (ZC, ZD): Document new constraints.
24131 * configure.ac (gcc_cv_as_micromips): Check if linker
24132 supports the .set micromips directive.
24133 * configure: Regenerate.
24134 * config.in: Regenerate.
24135 * config/mips/mips-tables.opt: Regenerate.
24136 * config/mips/micromips.md: New file.
24137 * constraints.md (ZC, AD): New constraints.
24138 * config/mips/predicates.md (movep_src_register): New predicate.
24139 (movep_src_operand): New predicate.
24140 (non_volatile_mem_operand): New predicate.
24141 * config/mips/mips.md (multimem): New type.
24142 (length): Differentiate between 17-bit and 18-bit branch offsets.
24143 (MOVEP1, MOVEP2): New mode iterator.
24144 (mov_<load>l): Use ZC constraint.
24145 (mov_<load>r): Likewise.
24146 (mov_<store>l): Likewise.
24147 (mov_<store>r): Likewise.
24148 (*branch_equality<mode>_inverted): Add microMIPS support.
24149 (*branch_equality<mode>): Likewise.
24150 (*jump_absolute): Likewise.
24151 (indirect_jump_<mode>): Likewise.
24152 (tablejump_<mode>): Likewise.
24153 (<optab>_internal): Likewise.
24154 (sibcall_internal): Likewise.
24155 (sibcall_value_internal): Likewise.
24156 (prefetch): Use constraint ZD.
24157 * config/mips/mips.opt (minterlink-compressed): New option.
24158 (minterlink-mips16): Now an alias for minterlink-compressed.
24159 (mmicromips): New option.
24160 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24161 (compare_and_swap_12): Likewise.
24162 (sync_add<mode>): Likewise.
24163 (sync_<optab>_12): Likewise.
24164 (sync_old_<optab>_12): Likewise.
24165 (sync_new_<optab>_12): Likewise.
24166 (sync_nand_12): Likewise.
24167 (sync_old_nand_12): Likewise.
24168 (sync_new_nand_12): Likewise.
24169 (sync_sub<mode>): Likewise.
24170 (sync_old_add<mode>): Likewise.
24171 (sync_old_sub<mode>): Likewise.
24172 (sync_new_add<mode>): Likewise.
24173 (sync_new_sub<mode>): Likewise.
24174 (sync_<optab><mode>): Likewise.
24175 (sync_old_<optab><mode>): Likewise.
24176 (sync_new_<optab><mode>): Likewise.
24177 (sync_nand<mode>): Likewise.
24178 (sync_old_nand<mode>): Likewise.
24179 (sync_new_nand<mode>): Likewise.
24180 (sync_lock_test_and_set<mode>): Likewise.
24181 (test_and_set_12): Likewise.
24182 (atomic_compare_and_swap<mode>): Likewise.
24183 (atomic_exchange<mode>_llsc): Likewise.
24184 (atomic_fetch_add<mode>_llsc): Likewise.
24185 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24186 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24187 (umips_save_restore_pattern_p): Likewise.
24188 (umips_load_store_pair_p): Likewise.
24189 (umips_output_load_store_pair): Likewise.
24190 (umips_movep_target_p): Likewise.
24191 (umips_12bit_offset_address_p): Likewise.
24192 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24193 (mips_base_mips16): Rename this...
24194 (mips_base_compression_flags): ...to this. Update all uses.
24195 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24196 (mips_mips16_decl_p): Delete.
24197 (mips_nomips16_decl_p): Delete.
24198 (mips_get_compress_on_flags): New function.
24199 (mips_get_compress_off_flags): New function.
24200 (mips_get_compress_mode): New function.
24201 (mips_get_compress_on_name): New function.
24202 (mips_get_compress_off_name): New function.
24203 (mips_insert_attributes): Support multiple compression types.
24204 (mips_merge_decl_attributes): Likewise.
24205 (umips_12bit_offset_address_p): New function.
24206 (mips_start_function_definition): Emit .set micromips directive.
24207 (mips_call_may_need_jalx_p): New function.
24208 (mips_function_ok_for_sibcall): Add microMIPS support.
24209 (mips_print_operand_punctuation): Support short delay slots and
24210 compact jumps.
24211 (umips_swm_mask, umips_swm_encoding): New.
24212 (umips_build_save_restore): New function.
24213 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24214 (was_mips16_p): Remove.
24215 (old_compression_mode): New.
24216 (mips_set_compression_mode): New function.
24217 (mips_set_current_function): Add microMIPS support.
24218 (mips_option_override): Likewise.
24219 (umips_save_restore_pattern_p): New function.
24220 (umips_output_save_restore): New function.
24221 (umips_load_store_pair_p_1): New function.
24222 (umips_load_store_pair_p): New function.
24223 (umips_output_load_store_pair_1): New function.
24224 (umips_output_load_store_pair): New function.
24225 (umips_movep_target_p) New function.
24226 (mips_prepare_pch_save): Add microMIPS support.
24227 * config/mips/mips.h (TARGET_COMPRESSION): New.
24228 (TARGET_CPU_CPP_BUILTINS): Update macro
24229 to use new compression flags and to support microMIPS.
24230 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24231 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24232 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24233 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24234 (ASM_SPEC): Support mmicromips and mno-micromips.
24235 (M16STORE_REG_P): New macro.
24236 (MIPS_CALL): Support TARGET_MICROMIPS.
24237 (MICROMIPS_J): New macro.
24238 (mips_base_mips16): Rename this...
24239 (mips_base_compression_flags): ...to this.
24240 (UMIPS_12BIT_OFFSET_P): New macro.
24241 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24242 (MULTILIB_DIRNAMES): Likewise.
24243
5e5df392
TV		 242442013-02-25 Tom de Vries <tom@codesourcery.com>
24245
24246 PR rtl-optimization/56131
24247 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24248 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24249 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24250
3fdb53c1
TB		 242512013-02-25 Tobias Burnus <burnus@net-b.de>
24252
24253 * doc/invoke.texi (-fsanitize=): Move from optimization
24254 to debugging options.
24255
ed358aea
AB		 242562013-02-25 Andrey Belevantsev <abel@ispras.ru>
24257
24258 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24259
f45e9053
AB		 242602013-02-25 Andrey Belevantsev <abel@ispras.ru>
24261 Alexander Monakov <amonakov@ispras.ru>
24262
24263 PR middle-end/56077
24264 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24265 flush pending lists also on non-jumps. Adjust comment.
24266
6941b508
CM		 242672013-02-24 Catherine Moore <clm@codesourcery.com>
24268 Maciej W. Rozycki <macro@codesourcery.com>
24269 Tom de Vries <tom@codesourcery.com>
a72d8780 24270 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM		 24271 Iain Sandoe <iain@codesourcery.com>
24272 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 24273 Chao-ying Fu <fu@mips.com>
6941b508
CM		 24274
24275 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 24276 Document new function attributes.
6941b508
CM		 24277 * doc/invoke.texi (minterlink-compressed, mmicromips,
24278 m14k, m14ke, m14kec): Document new options.
24279 (minterlink-mips16): Update documentation.
24280 * doc/md.texi (ZC, ZD): Document new constraints.
24281 * configure.ac (gcc_cv_as_micromips): Check if linker
24282 supports the .set micromips directive.
24283 * configure: Regenerate.
24284 * config.in: Regenerate.
24285 * config/mips/mips-tables.opt: Regenerate.
24286 * config/mips/micromips.md: New file.
24287 * constraints.md (ZC, AD): New constraints.
24288 * config/mips/predicates.md (movep_src_register): New predicate.
24289 (movep_src_operand): New predicate.
24290 (non_volatile_mem_operand): New predicate.
24291 * config/mips/mips.md (multimem): New type.
24292 (length): Differentiate between 17-bit and 18-bit branch offsets.
24293 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 24294 (mov_<load>l): Use ZC constraint.
6941b508
CM		 24295 (mov_<load>r): Likewise.
24296 (mov_<store>l): Likewise.
24297 (mov_<store>r): Likewise.
24298 (*branch_equality<mode>_inverted): Add microMIPS support.
24299 (*branch_equality<mode>): Likewise.
24300 (*jump_absolute): Likewise.
24301 (indirect_jump_<mode>): Likewise.
24302 (tablejump_<mode>): Likewise.
24303 (<optab>_internal): Likewise.
24304 (sibcall_internal): Likewise.
24305 (sibcall_value_internal): Likewise.
24306 (prefetch): Use constraint ZD.
24307 * config/mips/mips.opt (minterlink-compressed): New option.
24308 (minterlink-mips16): Now an alias for minterlink-compressed.
24309 (mmicromips): New option.
24310 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24311 (compare_and_swap_12): Likewise.
24312 (sync_add<mode>): Likewise.
24313 (sync_<optab>_12): Likewise.
24314 (sync_old_<optab>_12): Likewise.
24315 (sync_new_<optab>_12): Likewise.
24316 (sync_nand_12): Likewise.
24317 (sync_old_nand_12): Likewise.
24318 (sync_new_nand_12): Likewise.
24319 (sync_sub<mode>): Likewise.
24320 (sync_old_add<mode>): Likewise.
24321 (sync_old_sub<mode>): Likewise.
24322 (sync_new_add<mode>): Likewise.
24323 (sync_new_sub<mode>): Likewise.
24324 (sync_<optab><mode>): Likewise.
24325 (sync_old_<optab><mode>): Likewise.
24326 (sync_new_<optab><mode>): Likewise.
24327 (sync_nand<mode>): Likewise.
24328 (sync_old_nand<mode>): Likewise.
24329 (sync_new_nand<mode>): Likewise.
24330 (sync_lock_test_and_set<mode>): Likewise.
24331 (test_and_set_12): Likewise.
24332 (atomic_compare_and_swap<mode>): Likewise.
24333 (atomic_exchange<mode>_llsc): Likewise.
24334 (atomic_fetch_add<mode>_llsc): Likewise.
24335 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24336 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24337 (umips_save_restore_pattern_p): Likewise.
24338 (umips_load_store_pair_p): Likewise.
24339 (umips_output_load_store_pair): Likewise.
24340 (umips_movep_target_p): Likewise.
24341 (umips_12bit_offset_address_p): Likewise.
24342 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24343 (mips_base_mips16): Rename this...
24344 (mips_base_compression_flags): ...to this. Update all uses.
24345 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24346 (mips_mips16_decl_p): Delete.
24347 (mips_nomips16_decl_p): Delete.
3fdb53c1 24348 (mips_get_compress_on_flags): New function.
6941b508
CM		 24349 (mips_get_compress_off_flags): New function.
24350 (mips_get_compress_mode): New function.
24351 (mips_get_compress_on_name): New function.
24352 (mips_get_compress_off_name): New function.
24353 (mips_insert_attributes): Support multiple compression types.
24354 (mips_merge_decl_attributes): Likewise.
24355 (umips_12bit_offset_address_p): New function.
24356 (mips_start_function_definition): Emit .set micromips directive.
24357 (mips_call_may_need_jalx_p): New function.
24358 (mips_function_ok_for_sibcall): Add microMIPS support.
24359 (mips_print_operand_punctuation): Support short delay slots and
24360 compact jumps.
24361 (umips_swm_mask, umips_swm_encoding): New.
24362 (umips_build_save_restore): New function.
24363 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24364 (was_mips16_p): Remove.
24365 (old_compression_mode): New.
24366 (mips_set_compression_mode): New function.
24367 (mips_set_current_function): Add microMIPS support.
24368 (mips_option_override): Likewise.
24369 (umips_save_restore_pattern_p): New function.
24370 (umips_output_save_restore): New function.
24371 (umips_load_store_pair_p_1): New function.
24372 (umips_load_store_pair_p): New function.
24373 (umips_output_load_store_pair_1): New function.
24374 (umips_output_load_store_pair): New function.
24375 (umips_movep_target_p) New function.
24376 (mips_prepare_pch_save): Add microMIPS support.
24377 * config/mips/mips.h (TARGET_COMPRESSION): New.
24378 (TARGET_CPU_CPP_BUILTINS): Update macro
24379 to use new compression flags and to support microMIPS.
24380 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24381 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24382 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24383 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24384 (ASM_SPEC): Support mmicromips and mno-micromips.
24385 (M16STORE_REG_P): New macro.
24386 (MIPS_CALL): Support TARGET_MICROMIPS.
24387 (MICROMIPS_J): New macro.
24388 (mips_base_mips16): Rename this...
24389 (mips_base_compression_flags): ...to this.
24390 (UMIPS_12BIT_OFFSET_P): New macro.
24391 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24392 (MULTILIB_DIRNAMES): Likewise.
24393
cdebbc6d
JJ		 243942013-02-24 Jakub Jelinek <jakub@redhat.com>
24395
24396 PR target/52555
24397 * target-globals.c (save_target_globals): For init_reg_sets and
24398 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24399
18c63565
JG		 244002013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24401
24402 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24403 * config/aarch64/t-aarch64
24404 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24405
2194f7a2
VM		 244062013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24407
24408 PR inline-asm/56148
24409 * lra-constraints.c (process_alt_operands): Reload operand
24410 conflicting with earlier clobber only if no more other conflicting
24411 operands.
24412
7d613735
JJ		 244132013-02-22 Jakub Jelinek <jakub@redhat.com>
24414
24415 PR sanitizer/56393
24416 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24417 if not linking a shared library.
24418
ac8d93a7
SL		 244192013-02-22 Seth LaForge <sethml@google.com>
24420
24421 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24422
e0237780
GY		 244232013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24424
3fdb53c1
TB		 24425 * config/arm/arm.md (split for extendsidi): Update condition.
24426 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24427 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24428 (qhs_zextenddi_cstr): Likewise.
e0237780 24429
d7fde18c
JJ		 244302013-02-21 Jakub Jelinek <jakub@redhat.com>
24431
be63b77d
JJ		 24432 PR middle-end/56420
24433 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24434 avoid signed wrapping.
24435 (expand_mult): Handle properly multiplication by
24436 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24437 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24438 in the compiler if coeff is HOST_WIDE_INT_MIN.
24439 (expand_divmod): Don't make ext_op1 static, change it's type to
24440 uhwi. Avoid undefined behavior in -INTVAL (op1).
24441
d7fde18c
JJ		 24442 PR rtl-optimization/50339
24443 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24444 field.
24445 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24446 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24447 into splitting_ashiftrt field.
24448 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24449 ASHIFTRT.
24450 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24451 choices.
24452
6aad4455
AH		 244532013-02-20 Aldy Hernandez <aldyh@redhat.com>
24454
24455 PR middle-end/56108
24456 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24457 are sure to go irrevocable.
24458
38fe784d
HPN		 244592013-02-21 Hans-Peter Nilsson <hp@axis.com>
24460
24461 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24462 scalars are valid operands.
24463
0fd44da3
MJ		 244642013-02-21 Martin Jambor <mjambor@suse.cz>
24465
24466 PR tree-optimization/56310
24467 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24468 only matching indices and non-negative final offsets.
24469 (intersect_aggregates_with_edge): Pass src_idx to
24470 agg_replacements_to_vector. Pass src_idx insstead of index to
24471 intersect_with_agg_replacements.
24472
7a92038b
MJ		 244732013-02-21 Martin Jambor <mjambor@suse.cz>
24474
24475 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24476 instead of hard-wired defaults.
24477
c0da9c37
MR		 244782013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24479
24480 * doc/invoke.texi (MIPS Options): Update documentation of the
24481 floating-point multiply-accumulate instruction restrictions.
24482
d247ea0c 244832013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS		 24484
24485 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24486 asan_shadow_offset on x86_64 linux.
24487
22deefcb
RB		 244882013-02-21 Richard Biener <rguenther@suse.de>
24489
24490 PR tree-optimization/56415
24491 Revert
24492 2013-02-11 Richard Biener <rguenther@suse.de>
24493
24494 PR tree-optimization/56273
24495 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24496 first VRP run.
24497
7bcc6e75
JJ		 244982013-02-21 Jakub Jelinek <jakub@redhat.com>
24499
a0ad148f
JJ		 24500 PR bootstrap/56258
24501 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24502 instead of @itemx.
24503
7bcc6e75
JJ		 24504 PR inline-asm/56405
24505 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24506 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24507
a0a7b611
JH		 245082013-02-20 Jan Hubicka <jh@suse.cz>
24509
24510 PR tree-optimization/56265
88b97037
UB		 24511 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24512 when target is referenced for first time.
a0a7b611 24513
c0e50f72
RB		 245142013-02-20 Richard Biener <rguenther@suse.de>
24515
24516 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24517 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24518 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24519 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24520 not return anything.
24521 (rename_ssa_copies): Do not remove unused locals.
24522 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 24523 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB		 24524 * passes.c (execute_function_todo): Do not schedule unused locals
24525 removal if cleanup_tree_cfg did something.
24526 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24527 about the number of removed locals.
24528
a52ca739
RB		 245292013-02-20 Richard Biener <rguenther@suse.de>
24530
24531 PR tree-optimization/56398
88b97037 24532 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 24533
ec9c9d1b
MJ		 245342013-02-20 Martin Jambor <mjambor@suse.cz>
24535
24536 PR tree-optimization/55334
24537 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24538 restricted pointers to arrays.
24539
e91c8ed6 245402013-02-20 Richard Biener <rguenther@suse.de>
88b97037 24541 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB		 24542
24543 PR tree-optimization/56396
24544 * tree-ssa-ccp.c (n_const_val): New static variable.
24545 (get_value): Return NULL for SSA names we don't have a lattice
24546 entry for.
24547 (ccp_initialize): Initialize n_const_val.
24548 * tree-ssa-copy.c (n_copy_of): New static variable.
24549 (init_copy_prop): Initialize n_copy_of.
24550 (get_value): Return NULL_TREE for SSA names we don't have a
24551 lattice entry for.
24552
3d916479
MJ		 245532013-02-20 Martin Jambor <mjambor@suse.cz>
24554
24555 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24556
71a86758
RB		 245572013-02-20 Richard Biener <rguenther@suse.de>
24558
24559 * genpreds.c (write_lookup_constraint): Do not compare first
24560 letter of the constraint again.
24561
79836a12
RB		 245622013-02-20 Richard Biener <rguenther@suse.de>
24563
24564 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24565 and ceil_log2.
24566 (get_use_iv_cost): Terminate hashtable walk when coming across
24567 an empty entry.
24568
bbe4fb2c
IZ		 245692013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24570
24571 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24572 reassociation for avx2 targets.
24573
a72d8780 245742012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 24575
c77f83d5
EI		 24576 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24577 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 24578 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI		 24579 version and TARGET_PATTERN_COMPARE check
24580 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 24581
a72d8780 245822012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 24583
a72d8780 24584 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI		 24585 function before branching.
24586
242387fa
AB		 245872012-02-19 Andrey Belevantsev <abel@ispras.ru>
24588
24589 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24590 DUMP_INSN_RTX_UID.
24591 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24592
b4979ab9
AB		 245932012-02-19 Andrey Belevantsev <abel@ispras.ru>
24594
24595 PR middle-end/55889
b4979ab9
AB		 24596 * sel-sched.c: Include ira.h.
24597 (implicit_clobber_conflict_p): New function.
24598 (moveup_expr): Use it.
88b97037 24599 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 24600
24d63016
RB		 246012013-02-19 Richard Biener <rguenther@suse.de>
24602
24603 PR tree-optimization/56384
24604 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24605 (vn_hash_type): Split out from ...
24606 (vn_hash_constant_with_type): ... here.
24607 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24608 (vn_phi_eq): Compare types from vn_phi_s structure.
24609 (vn_phi_lookup): Populate vn_phi_s type.
24610 (vn_phi_insert): Likewise.
24611
a475fd3d
JJ		 246122013-02-19 Jakub Jelinek <jakub@redhat.com>
24613
47cc28f5
JJ		 24614 PR tree-optimization/56350
24615 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24616 if haven't found reduction or nested cycle operand, rather than
24617 asserting we must find it.
24618
a475fd3d
JJ		 24619 PR tree-optimization/56381
24620 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24621 to fold_build3.
24622
246232013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH		 24624 Jakub Jelinek <jakub@redhat.com>
24625
24626 PR target/52555
24627 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24628 (swap_optab_enable): Same.
24629 (init_all_optabs): Use argument instead of global.
88b97037 24630 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH		 24631 * expr.h (init_all_optabs): Add argument to prototype.
24632 (TREE_OPTIMIZATION_OPTABS): New.
24633 (save_optabs_if_changed): Protoize.
24634 * optabs.h: Declare this_fn_optabs.
24635 * optabs.c (save_optabs_if_changed): New.
24636 Declare this_fn_optabs.
24637 (init_optabs): Add argument to init_all_optabs() call.
24638 * function.c (invoke_set_current_function_hook): Handle per
24639 function optabs.
24640 * function.h (struct function): New field optabs.
24641 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24642 optimization_current_node has changed.
24643 * target-globals.h (save_target_globals_default_opts): Protoize.
24644 * target-globals.c (save_target_globals_default_opts): New.
24645
3f587ca3
JDA		 246462013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24647
24648 PR target/56347
24649 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24650 registers %fr12 and %fr12R as call used.
ceaca33e
JDA		 24651
24652 PR target/56214
24653 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB		 24654 and HImode, require all displacements to be an integer multiple of
24655 their mode size.
ceaca33e
JDA		 24656 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24657 only allow QImode and HImode when reload is in progress and strict is
24658 true. Likewise for symbolic addresses. Use base14_operand to check
24659 displacements in REG+BASE addresses.
24660
fe0b4796
RB		 246612013-02-18 Richard Biener <rguenther@suse.de>
24662
24663 PR tree-optimization/56366
24664 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24665 handle sign-conversion of outer-loop initial induction value.
24666
6aaf596b
RB		 246672013-02-18 Richard Biener <rguenther@suse.de>
24668
73db8ff1 24669 PR middle-end/56349
6aaf596b
RB		 24670 * cfghooks.c (merge_blocks): If we merge a latch into another
24671 block adjust references to it.
24672 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24673 (verify_loop_structure): Verify that a recorded latch is in fact
24674 a latch.
24675
5e97dfb6
RB		 246762013-02-18 Richard Biener <rguenther@suse.de>
24677
24678 PR tree-optimization/56321
24679 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24680 order SSA name release and virtual operand unlinking.
24681
825527e8
EI		 246822013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24683
24684 * config/microblaze/microblaze.md (save_stack_block): Define.
24685 (restore_stack_block): Likewise.
24686
debd11d9
EI		 246872013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24688
24689 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24690 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24691 * config/microblaze/microblaze.c (microblaze_option_override):
24692 Bail out early for PIC modes when target does not support PIC.
24693
8ec77be0
EI		 246942013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24695
88b97037 24696 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI		 24697 Replace with a microblaze version.
24698 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 24699 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI		 24700 microblaze.
24701
b41288b3
JJ		 247022013-02-16 Jakub Jelinek <jakub@redhat.com>
24703 Dodji Seketeli <dodji@redhat.com>
24704
24705 PR asan/56330
88b97037 24706 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ		 24707 (instrument_mem_region_access): Do not forget to always put
24708 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24709 0) statement, even for cases where either 'base' or 'base + len'
24710 are not instrumented -- because they have been previously
24711 instrumented. Simplify the logic by putting all the statements
24712 instrument 'base + len' inside a sequence, and then insert that
24713 sequence right before the current insertion point. Then, to
24714 instrument 'base + len', just get an iterator on that statement.
24715 And do not forget to update the pointer to iterator the function
24716 received as argument.
24717
47918951
VM		 247182013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24719
24720 PR rtl-optimization/56348
24721 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24722
3fb7c699
SB		 247232013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24724
24725 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24726 (clean_graph_dump_file): Pass base to start_graph_dump.
24727
576fe41a
RH		 247282013-02-14 Richard Henderson <rth@redhat.com>
24729
24730 PR target/55941
24731 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24732
bb3accfa
SB		 247332013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24734
24735 * collect2-aix.h: Define F_LOADONLY.
24736
03143140
RB		 247372013-02-14 Richard Biener <rguenther@suse.de>
24738
24739 PR lto/50494
24740 * varasm.c (output_constant_def_1): Get the decl representing
24741 the constant as argument.
24742 (output_constant_def): Wrap output_constant_def_1.
24743 (make_decl_rtl): Use output_constant_def_1 with the decl
24744 representing the constant.
24745 (build_constant_desc): Optionally re-use a decl already
24746 representing the constant.
24747 (tree_output_constant_def): Adjust.
24748
8fb06726
DS		 247492013-02-14 Dodji Seketeli <dodji@redhat.com>
24750
24751 Fix an asan crash
24752 * asan.c (instrument_builtin_call): Really put the length of the
24753 second source argument into src1_len.
24754
c4bfe8bf
JJ		 247552013-02-13 Jakub Jelinek <jakub@redhat.com>
24756
24757 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24758 argument. If it is false, don't create edge from then_bb to
24759 fallthru_bb.
24760 (insert_if_then_before_iter): Pass true to it.
24761 (build_check_stmt): Pass false to it.
24762 (transform_statements): Flush hash table only on extended basic
24763 block boundaries, rather than at the beginning of every bb.
24764 Don't flush hash table on nonfreeing_call_p calls.
24765 * tree-flow.h (nonfreeing_call_p): New prototype.
24766 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24767
7afe2801
DM		 247682013-02-13 David S. Miller <davem@davemloft.net>
24769
24770 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24771
70cc3288
VM		 247722013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24773
24774 PR target/56184
24775 * ira.c (max_regno_before_ira): Move from ...
24776 (ira): ... here.
24777 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24778 vec_safe_length.
24779
6422242b
JJ		 247802013-02-13 Jakub Jelinek <jakub@redhat.com>
24781
24782 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24783
976a81ee
RB		 247842013-02-13 Richard Biener <rguenther@suse.de>
24785
24786 PR lto/56295
24787 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24788 globals in MEM_REFs.
24789
c1874a87
RB		 247902013-02-13 Richard Biener <rguenther@suse.de>
24791
24792 * loop-init.c (loop_optimizer_init): Clear loop state when
24793 re-initializing preserved loops.
24794 * loop-unswitch.c (unswitch_single_loop): Return whether
24795 we unswitched the loop. Do not verify loop state here.
88b97037 24796 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 24797
b4ab7d34
KS		 247982013-02-13 Kostya Serebryany <kcc@google.com>
24799
88b97037
UB		 24800 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24801 on x86_64 linux.
b4ab7d34
KS		 24802 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24803
bdcbe80c
DS		 248042013-02-12 Dodji Seketeli <dodji@redhat.com>
24805
24806 Avoid instrumenting duplicated memory access in the same basic block
24807 * Makefile.in (asan.o): Add new dependency on hash-table.h
24808 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24809 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24810 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24811 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24812 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24813 (get_mem_ref_of_assignment): New functions.
24814 (get_mem_refs_of_builtin_call): Extract from
24815 instrument_builtin_call and tweak a little bit to make it fit with
24816 the new signature.
24817 (instrument_builtin_call): Use the new
24818 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24819 of is_gimple_builtin_call.
24820 (instrument_derefs, instrument_mem_region_access): Insert the
24821 instrumented memory reference into the hash table.
24822 (maybe_instrument_assignment): Renamed instrument_assignment into
24823 this, and change it to advance the iterator when instrumentation
24824 actually happened and return true in that case. This makes it
24825 homogeneous with maybe_instrument_assignment, and thus give a
24826 chance to callers to be more 'regular'.
24827 (transform_statements): Clear the memory reference hash table
24828 whenever we enter a new BB, when we cross a function call, or when
24829 we are done transforming statements. Use
24830 maybe_instrument_assignment instead of instrumentation. No more
24831 need to special case maybe_instrument_assignment and advance the
24832 iterator after calling it; it's now handled just like
24833 maybe_instrument_call. Update comment.
24834
4861a1f7
RB		 248352013-02-13 Richard Biener <rguenther@suse.de>
24836
24837 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24838 Fix loop discovery code.
24839
1c86bd80
VM		 248402013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24841
24842 PR inline-asm/56148
24843 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 24844 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM		 24845 the operand is not reloaded. Prefer to reload conflicting operand
24846 if earlyclobber and matching operands are the same.
24847
7cbda518
RB		 248482013-02-12 Richard Biener <rguenther@suse.de>
24849
24850 PR lto/56297
24851 * lto-streamer-out.c (write_symbol): Do not output symbols
24852 for hard register variables.
24853
e68a4ef6
GJL		 248542013-02-12 Georg-Johann Lay <avr@gjlay.de>
24855
24856 PR target/54222
24857 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24858 (umulsidi3_insn, mulsidi3_insn): New insns.
24859
a72d8780 248602013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL		 24861
24862 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24863 (struct tune_params): Add vec_costs field.
24864 * config/arm/arm.c (arm_builtin_vectorization_cost)
24865 (arm_add_stmt_cost): New functions.
24866 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24867 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24868 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24869 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24870 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24871 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24872 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24873
43320568
RB		 248742013-02-12 Richard Biener <rguenther@suse.de>
24875
24876 PR lto/56295
24877 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24878 decls again if possible.
24879
a011aa39
RB		 248802013-02-12 Richard Biener <rguenther@suse.de>
24881
24882 PR middle-end/56288
24883 * tree-ssa.c (verify_ssa_name): Fix check, move
24884 SSA_NAME_IN_FREE_LIST check up.
24885
6da26889
JJ		 248862013-02-12 Jakub Jelinek <jakub@redhat.com>
24887 Steven Bosscher <steven@gcc.gnu.org>
24888
24889 PR rtl-optimization/56151
24890 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24891 equal to op0 or op1, and last_insn pattern is CODE operation
24892 with MEM dest and one of the operands matches that MEM.
24893
f80e0faf
ST		 248942013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24895
24896 * doc/extend.texi: Document Function Multiversioning and "default"
24897 parameter string to target attribute.
24898 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24899 target attribute parameter is "default".
24900 (ix86_compare_version_priority): Remove checks for target attribute.
24901 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24902 Remove check for target attribute equal to NULL. Add assert.
24903 (ix86_generate_version_dispatcher_body): Change error to sorry.
24904
6c59ffd1
IS		 249052013-02-11 Iain Sandoe <iain@codesourcery.com>
24906 Jack Howarth <howarth@bromo.med.uc.edu>
24907 Patrick Marlier <patrick.marlier@gmail.com>
24908
24909 PR libitm/55693
24910 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24911 define ENDFILE_SPEC as TM_DESTRUCTOR.
24912 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24913
a72d8780 249142013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP		 24915 Jack Howarth <howarth@bromo.med.uc.edu>
24916 Jakub Jelinek <jakub@redhat.com>
24917
24918 PR sanitizer/55617
24919 * config/darwin.c (cdtor_record): Rename ctor_record.
24920 (sort_cdtor_records): Rename sort_ctor_records.
24921 (finalize_dtors): New routine to sort destructors by
24922 priority before use in assemble_integer.
24923 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24924
b63fe007
UB		 249252013-02-11 Uros Bizjak <ubizjak@gmail.com>
24926
24927 PR rtl-optimization/56275
24928 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24929 offset is non-negative and less than cmode size before
24930 calling simplify_subreg.
24931
8e89b5b5
RB		 249322013-02-11 Richard Biener <rguenther@suse.de>
24933
24934 PR tree-optimization/56264
24935 * cfgloop.h (fix_loop_structure): Adjust prototype.
24936 * loop-init.c (fix_loop_structure): Return the number of
24937 newly discovered loops.
24938 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24939 are discovered, do a full loop-closed SSA rewrite.
24940
b4a4b56d
RB		 249412013-02-11 Richard Biener <rguenther@suse.de>
24942
24943 PR tree-optimization/56273
24944 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24945 first VRP run.
24946 (check_array_ref): Fix missing newline in dumps.
24947 (search_for_addr_array): Likewise.
24948
0c885229
DE		 249492013-02-09 David Edelsohn <dje.gcc@gmail.com>
24950
24951 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24952
59ac9a55
JJ		 249532013-02-09 Jakub Jelinek <jakub@redhat.com>
24954
24955 PR target/56256
24956 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24957
25bb0bb5
VM		 249582013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24959
24960 PR rtl-optimization/56246
0c885229 24961 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM		 24962 reload pseudo.
24963 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24964 constraints are satisfied.
24965
a698cc03
JL		 249662013-02-08 Jeff Law <law@redhat.com>
24967
24968 PR debug/53948
24969 * emit-rtl.c (reg_is_parm_p): New function.
24970 * regs.h (reg_is_parm_p): New prototype.
24971 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24972 callee-clobbered registers.
24973
e1122ddd
MM		 249742013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24975
24976 PR target/56043
24977 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24978 If there is no implicit builtin declaration, just return NULL.
24979
19c5f6e6
UB		 249802013-02-08 Uros Bizjak <ubizjak@gmail.com>
24981
24982 * config/i386/sse.md (FMAMODEM): New mode iterator.
24983 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24984 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24985
2480f2ca 249862013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 24987
2480f2ca
UB		 24988 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24989 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24990 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24991
249922013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24993
24994 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24995 (microblaze*-*-elf): Likewise.
24996 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24997 LINK_SPEC.
24998 * config/microblaze/microblaze-c.c: Add builtin defines for
24999 _LITTLE_ENDIAN and _BIG_ENDIAN.
25000 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25001 add to TARGET_DEFAULT flags.
76ef61fb 25002 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB		 25003 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25004 * config/microblaze/microblaze.md: Update extendsidi2 and
25005 movdi_internal instructions to use low-order / high-order reg
25006 print_operands.
25007 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25008 options and inversemask / mask of LITTLE_ENDIAN.
25009 * config/microblaze/t-microblaze: Expand multilib options to
25010 include mlittle-endian (le) and update exceptions patterns.
25011
600a5961
JJ		 250122013-02-08 Jakub Jelinek <jakub@redhat.com>
25013
5df81313
JJ		 25014 PR rtl-optimization/56195
25015 * lra-constraints.c (get_reload_reg): Don't reuse regs
25016 if they have smaller mode than requested, if they have
25017 wider mode than requested, try to return a SUBREG.
25018
600a5961
JJ		 25019 PR tree-optimization/56250
25020 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25021 if type is unsigned and code isn't MULT_EXPR.
25022
ff544649
GJL		 250232013-02-08 Georg-Johann Lay <avr@gjlay.de>
25024
25025 PR tree-optimization/56064
25026 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25027 bits according to mode.
25028 * fixed-value.h (fixed_from_double_int)
25029 (const_fixed_from_double_int): Adjust comments.
25030
e45cde98
RB		 250312013-02-08 Richard Biener <rguenther@suse.de>
25032
25033 PR lto/56231
25034 * lto-streamer.h (struct data_in): Remove current_file, current_line
25035 and current_col members.
25036 * lto-streamer-out.c (lto_output_location): Stream changed bits
25037 en-block for efficiency.
25038 * lto-streamer-in.c (clear_line_info): Remove.
25039 (lto_input_location): Cache current file, line and column
25040 globally via local statics. Read changed bits en-block.
25041 (input_function): Do not call clear_line_info.
25042 (lto_read_body): Likewise.
25043 (lto_input_toplevel_asms): Likewise.
25044
c1ca73d8
MM		 250452013-02-08 Michael Matz <matz@suse.de>
25046
25047 PR tree-optimization/52448
25048 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25049 (nt_call_phase): New static.
25050 (add_or_mark_expr): Only mark accesses with newer phase than any
25051 call seen.
25052 (nonfreeing_call_p): New.
25053 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25054 (nt_fini_block): Keep blocks marked as visited.
25055 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25056
57548aa2
RB		 250572013-02-08 Richard Biener <rguenther@suse.de>
25058
25059 * ira.c (ira): Free broken dominator information.
25060
8e10366f
UB		 250612013-02-08 Uros Bizjak <ubizjak@gmail.com>
25062
25063 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25064
f64fb0fa
MP		 250652013-02-08 Marek Polacek <polacek@redhat.com>
25066
8e10366f 25067 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 25068
0375167b
RB		 250692013-02-08 Richard Biener <rguenther@suse.de>
25070
25071 PR middle-end/56181
25072 * cfgloop.h (flow_loops_find): Adjust.
25073 (bb_loop_header_p): Declare.
25074 * cfgloop.c (bb_loop_header_p): New function split out from ...
25075 (flow_loops_find): ... here. Adjust function signature,
25076 support incremental loop structure update.
25077 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25078 * cfgloopmanip.c (fix_loop_structure): Move ...
25079 * loop-init.c (fix_loop_structure): ... here.
25080 (apply_loop_flags): Split out from ...
25081 (loop_optimizer_init): ... here.
25082 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25083 in incremental mode, only remove dead loops here.
25084
85d768f3
GJL		 250852013-02-08 Georg-Johann Lay <avr@gjlay.de>
25086
25087 PR target/54222
25088 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25089 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25090 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25091 (*round<mode>3.libgcc): New insns for fixed-modes.
25092 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25093 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25094 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25095 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25096 implementations. Define to __builtin_avr_absFX,
25097 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25098 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25099 __builtin_avr_countlsFX, respectively.
25100 * config/avr/avr-c.c (target.h): Include it.
25101 (enum avr_builtin_id): New enum.
25102 (avr_resolve_overloaded_builtin): New static function.
25103 (avr_register_target_pragmas): Use it to set
25104 targetm.resolve_overloaded_builtin.
25105 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25106 tree nodes used by DEF_BUILTIN.
25107 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25108 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25109 <AVR_BUILTIN_xxBITS>: Same.
25110
661bc682
RB		 251112013-02-08 Richard Biener <rguenther@suse.de>
25112
25113 * cfgloop.c (verify_loop_structure): Properly handle
25114 a loop exiting to another loop header.
25115 * ira-int.h (ira_loops): Remove.
25116 * ira.c (ira_loops): Remove.
25117 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25118 (do_reload): Use loop_optimizer_finalize.
25119 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25120 number_of_loops to access the loop tree.
25121 (more_one_region_p): Likewise.
25122 (finish_loop_tree_nodes): Likewise.
25123 (rebuild_regno_allocno_maps): Likewise.
25124 (mark_loops_for_removal): Likewise.
25125 (mark_all_loops_for_removal): Likewise.
25126 (remove_unnecessary_regions): Likewise.
25127 (ira_build): Likewise.
25128 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25129
0d5049b2
RB		 251302013-02-08 Richard Biener <rguenther@suse.de>
25131
25132 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25133 * ipa-pure-const.c (analyze_function): Avoid calling
25134 mark_irreducible_loops twice.
8e10366f 25135 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 25136
499f32e8
DM		 251372013-02-07 David S. Miller <davem@davemloft.net>
25138
25139 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25140 on 'reg'.
25141 * var-tracking.c (vt_add_function_parameter): Test the presence of
25142 HAVE_window_save properly and do not remap argument registers when
25143 we have a leaf function.
25144
6edc3e32
UB		 251452013-02-07 Uros Bizjak <ubizjak@gmail.com>
25146
25147 PR bootstrap/56227
25148 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25149 instead of "ll".
25150 * config/i386/i386.c (ix86_print_operand): Ditto.
25151
5306401f
VM		 251522013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25153
6edc3e32 25154 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 25155
027ece11
VM		 251562013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25157
25158 PR rtl-optimization/56225
25159 * lra-constraints.c (process_alt_operands): Check that reload hard
25160 reg can hold value for strict_low_part.
25161
f980dfdb
JJ		 251622013-02-07 Jakub Jelinek <jakub@redhat.com>
25163
25164 PR debug/56154
25165 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25166 dwarf2out_end_function.
25167 (in_first_function_p, maybe_at_text_label_p,
25168 first_loclabel_num_not_at_text_label): New variables.
25169 (dwarf2out_var_location): In the first function find out
6edc3e32 25170 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ		 25171 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25172 functions.
25173
73dd3123
EB		 251742013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25175
25176 PR rtl-optimization/56178
25177 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25178 SUBREG of a register. Tidy up related block of code.
25179 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25180 note if the source is a register or a SUBREG of a register.
25181
e3936f47
JJ		 251822013-02-07 Jakub Jelinek <jakub@redhat.com>
25183
25184 PR target/56228
25185 * config/rs6000/rs6000.md (ptrm): New mode attr.
25186 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25187 call_value_indirect_aix<pttrsize>,
25188 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25189 m in constraints.
25190
d96d674b
MH		 251912013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25192
25193 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25194 if -bnortl. Convert to strcmp and strncmp.
25195
921f2dee
AM		 251962013-02-07 Alan Modra <amodra@gmail.com>
25197
25198 PR target/54009
25199 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25200 addresses won't wrap when offsetting.
25201 (rs6000_secondary_reload): Provide secondary reloads needed for
25202 wrapping LO_SUM addresses.
25203
d09c7dba
TS		 252042013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25205
25206 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25207 MACH, just __MACH__.
25208
a44bbd48
RB		 252092013-02-06 Richard Biener <rguenther@suse.de>
25210
25211 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25212 instead of calling fix_loop_structure.
25213
6d840d99
JJ		 252142013-02-06 Jakub Jelinek <jakub@redhat.com>
25215
25216 PR middle-end/56217
25217 * omp-low.c (use_pointer_for_field): Return false if
25218 lower_send_shared_vars doesn't generate any copy-out code.
25219
0f33baa9
TV		 252202013-02-06 Tom de Vries <tom@codesourcery.com>
25221
25222 PR rtl-optimization/56131
25223 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25224 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25225 of the label is NULL. Add comment.
25226
11f1e3ab
JJ		 252272013-02-05 Jakub Jelinek <jakub@redhat.com>
25228
31502f9f
JJ		 25229 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25230
a74db9bd
JJ		 25231 PR sanitizer/55374
25232 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25233 (STATIC_LIBTSAN_LIBS): Likewise.
25234 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25235 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25236 is defined, don't add anything else beyond that.
25237 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25238 (LINK_COMMAND_SPEC): Use them.
25239
11f1e3ab
JJ		 25240 PR tree-optimization/56205
25241 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25242 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25243 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25244
14ac6aa2
RB		 252452013-02-05 Richard Biener <rguenther@suse.de>
25246
25247 PR tree-optimization/53342
25248 PR tree-optimization/53185
25249 * tree-vectorizer.h (vect_check_strided_load): Remove.
25250 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25251 not disallow peeling for vectorized strided loads.
25252 (vect_check_strided_load): Make static and simplify.
25253 (vect_analyze_data_refs): Adjust.
25254 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25255 correctly when vectorizing strided loads.
25256
6f22445a
RB		 252572013-02-05 Richard Biener <rguenther@suse.de>
25258
25259 * doc/install.texi: Refer to ISL, not PPL.
25260
39f9719e
JH		 252612013-02-05 Jan Hubicka <jh@suse.cz>
25262
ec4224ac
JH		 25263 PR tree-optimization/55789
25264 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25265
252662013-02-05 Jan Hubicka <jh@suse.cz>
25267
25268 PR tree-optimization/55789
39f9719e
JH		 25269 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25270 the dead call anyway.
25271
956d3b33
EB		 252722013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25273
25274 PR sanitizer/55374
25275 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25276
252772013-02-04 Alexander Potapenko <glider@google.com>
25278 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP		 25279 Jakub Jelinek <jakub@redhat.com>
25280
25281 PR sanitizer/55617
25282 * config/darwin.c (sort_ctor_records): Stabilized qsort
25283 on constructor priority by using original position.
25284 (finalize_ctors): New routine to sort constructors by
25285 priority before use in assemble_integer.
25286 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25287
7ac3af38
JJ		 252882013-02-04 Jakub Jelinek <jakub@redhat.com>
25289
25290 PR libstdc++/54314
25291 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25292 about visibility on artificial decls.
25293 * config/sol2.c (solaris_assemble_visibility): Likewise.
25294
152689dc
KT		 252952013-02-04 Kai Tietz <ktietz@redhat.com>
25296
25297 PR target/56186
25298 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25299 argument and improve checking of return-argument types for 16-byte
25300 modes.
25301 (ix86_function_value_1): Add additional valtype argument on call
25302 of function_value_64.
25303 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25304 handling infunction_value_64 function.
25305
2b5987b5
MGD		 253062013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25307
7ac3af38 25308 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 25309
14597080
RB		 253102013-02-04 Richard Biener <rguenther@suse.de>
25311
25312 PR tree-optimization/56188
25313 * tree-ssa-structalias.c (label_visit): Consider case with
25314 initially non-empty points-to set.
25315 (perform_var_substitution): Dump node mapping and clean up.
25316
ed73881e
RG		 253172013-02-04 Richard Guenther <rguenther@suse.de>
25318
25319 PR lto/56168
25320 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25321 node prevail as last resort.
25322 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 25323 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 25324
139a0707
RB		 253252013-02-04 Richard Biener <rguenther@suse.de>
25326
25327 PR tree-optimization/56113
25328 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25329 Merge into ...
25330 (equiv_class_lookup_or_add): ... this.
25331 (label_visit): Adjust and fix error in previous patch.
25332 (perform_var_substitution): Adjust.
25333
50fe8924
OE		 253342013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25335
25336 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25337 * config/sh/sh4-300.md: Likewise.
25338 * config/sh/sh4a.md: Likewise.
25339 * config/sh/constraints.md: Likewise.
25340 * config/sh/sh.md: Likewise.
25341 * config/sh/netbsd-elf.h: Likewise.
25342 * config/sh/predicates.md: Likewise.
25343 * config/sh/sh-protos.h: Likewise.
25344 * config/sh/ushmedia.h: Likewise.
25345 * config/sh/linux.h: Likewise.
25346 * config/sh/sh.c: Likewise.
25347 * config/sh/superh.h: Likewise.
25348 * config/sh/elf.h: Likewise.
25349 * config/sh/sh4.md: Likewise.
25350 * config/sh/sh.h: Likewise.
25351
1a04ac2b
JDA		 253522013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25353
25354 * config/pa/constraints.md: Adjust unused letters. Change "T"
25355 constraint to match_test floating_point_store_memory_operand().
25356 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25357 (base14_operand): New.
25358 (floating_point_store_memory_operand): New.
25359 (integer_store_memory_operand): Revise to use base14_operand and
25360 reg_plus_base_memory_operand.
25361 (move_dest_operand): Allow symbolic_memory_operands.
25362 (symbolic_memory_operand): Check for LO_SOM.
25363 (symbolic_operand): Change default case to break.
25364 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25365 CONST_DOUBLE values to be reloaded by putting them into memory when
25366 the destination is a floating point register.
25367 (movdf): Remove code to handle CONST_DOUBLE.
25368 (movsf): Likewise.
25369 (reload_indf_r1): New.
25370 (reload_insf_r1): New.
25371 Consistently use "Q" and "T" constraints with integer and floating
25372 point move instructions, respectively.
25373 (movdi): Remove FAIL.
25374 Change predicate for source operand unamed DImode move from
25375 general_operand to move_src_operand.
25376 (umulsidi3): Change predicate for destination operand to
25377 register_operand.
25378 Likewise for similar unamed patterns.
25379 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25380 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25381 (hppa_legitimize_address): Simplify mask calculation.
25382 (pa_emit_move_sequence): Revised handling of secondary reloads from
25383 REG+D addresses for floating point loads and stores. Directly handle
25384 loading CONST0_RTX (mode) to a floating point register.
25385 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25386 to floating point registers. Don't restrict secondary reloads to
25387 floating point registers to integer modes. Revise some comments and
25388 cleanup some code.
25389 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25390 (pa_legitimate_address_p): New.
25391 (pa_legitimize_reload_address): New.
25392 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25393 (STRICT_REG_OK_FOR_BASE_P): New.
25394 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25395 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25396
3bf65591
DE		 253972013-02-03 David Edelsohn <dje.gcc@gmail.com>
25398 Andrew Dixie <andrewd@gentrack.com>
25399
25400 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25401 flag set.
25402
c0a8a3e6
RS		 254032013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25404
25405 * expmed.c (extract_bit_field_1): Pass the full width of the
25406 structure to get_best_reg_extraction_insn.
25407
99113dff
DE		 254082013-02-01 David Edelsohn <dje.gcc@gmail.com>
25409
25410 PR target/54601
25411 * configure.ac (use_cxa_atexit): Add AIX.
25412 * configure: Regenerate.
25413
25414 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25415
ddd84654
JJ		 254162013-02-01 Jakub Jelinek <jakub@redhat.com>
25417
25418 PR debug/54793
25419 * final.c (need_profile_function): New variable.
25420 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25421 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25422 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25423 notes, targetm.asm_out.function_prologue doesn't emit anything,
25424 HAVE_prologue and profiler should be emitted before prologue,
25425 set need_profile_function instead of emitting it.
25426 (final_scan_insn): If need_profile_function, emit
25427 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25428 NOTE_INSN_FUNCTION_BEG note.
25429
254302013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH		 25431
25432 * config/rs6000/rs6000.md (smulditi3): New.
25433 (umulditi3): New.
25434
ff2a9d88
RH		 25435 * config/alpha/alpha.md (umulditi3): New.
25436
14d52b90
DE		 254372013-02-01 David Edelsohn <dje.gcc@gmail.com>
25438
25439 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25440 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25441
8c7ca45c
RB		 254422013-02-01 Richard Biener <rguenther@suse.de>
25443
25444 PR tree-optimization/56113
25445 * tree-ssa-structalias.c (label_visit): Reduce work for
25446 single-predecessor nodes.
25447
9f419393
EB		 254482013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25449
25450 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25451 range isn't testing for zero.
25452
c5f4be84
SB		 254532013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25454
25455 PR middle-end/56113
25456 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25457
dbdbd982
NC		 254582013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25459 Nick Clifton <nickc@redhat.com>
25460
25461 * config/v850/constraints.md (Q): Define as a memory constraint.
25462 * config/v850/predicates.md (label_ref_operand): New predicate.
25463 (e3v5_shift_operand): New predicate.
25464 (ior_operator): New predicate.
25465 * config/v850/t-v850: Add e3v5 multilib.
25466 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25467 (v850_gen_movdi): Prototype.
25468 * config/v850/v850.c: Add support for e3v5 architecture.
25469 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25470 TARGET_V850E_UP.
25471 (construct_save_jarl): Add e3v5 long JARL support.
25472 (v850_adjust_insn_length): New function. Adjust length of call
25473 insns when using e3v5 instructions.
25474 (v850_gen_movdi): New function: Generate instructions to move a
25475 DImode value.
25476 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25477 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25478 (TARGET_USE_FPU): Enable for e3v5.
25479 (CONST_OK_FOR_W): New macro.
25480 (ADJUST_INSN_LENGTH): Define.
25481 * config/v850/v850.md (UNSPEC_LOOP): Define.
25482 (attr cpu): Add v850e3v5.
25483 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25484 (movdi): New pattern.
25485 (movdi_internal): New pattern.
25486 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25487 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25488 (cstoresf4): Likewise.
25489 (cstoredf4): Likewise.
25490 (insv): New pattern.
25491 (rotlso3_a): New pattern.
25492 (rotlsi3_b): New pattern
25493 (rotlsi3_v850e3v5): New pattern.
25494 (doloop_begin): New pattern.
25495 (fix_loop_counter): New pattern.
25496 (doloop_end): New pattern.
25497 (branch_normal): Add e3v5 long branch support.
25498 (branch_invert): Likewise.
25499 (branch_z_normal): Likewise.
25500 (branch_z_invert): Likewise.
25501 (branch_nz_normal): Likewise.
25502 (branch_nz_invert): Likewise.
25503 (call_internal_short): Add e3v5 register-indirect JARL support.
25504 (call_internal_long): Likewise.
25505 (call_value_internal_short): Likewise.
25506 (call_value_internal_long): Likewise.
25507 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25508 (mloop): New option.
25509 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 25510 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 25511
73861a41
PK		 255122013-01-31 Paul Koning <ni1d@arrl.net>
25513
25514 PR debug/55059
25515 PR debug/54508
25516 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25517 children if parent is a class.
25518 (prune_unused_types_prune): Don't add DW_AT_declaration.
25519
e44978dc
RB		 255202013-01-31 Richard Biener <rguenther@suse.de>
25521
25522 PR tree-optimization/56157
25523 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25524 match up operand with SLP child.
25525
90b10dec
JM		 255262013-01-31 Jason Merrill <jason@redhat.com>
25527
04d2dadd 25528 PR debug/54410
90b10dec
JM		 25529 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25530 parameters the first time.
25531 (gen_scheduled_generic_parms_dies): Check completeness here.
25532
6e616110
RB		 255332013-01-31 Richard Biener <rguenther@suse.de>
25534
25535 PR middle-end/53073
25536 * common.opt (faggressive-loop-optimizations): New flag,
25537 enabled by default.
25538 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25539 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25540 infer_loop_bounds_from_undefined by it.
25541
636f59cf
RB		 255422013-01-31 Richard Biener <rguenther@suse.de>
25543
25544 PR tree-optimization/56150
25545 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25546 visit virtual operands.
25547 (find_uses_to_rename_bb): Likewise.
25548
255492013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB		 25550
25551 PR tree-optimization/56150
25552 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25553 mixed store non-store stmts.
25554
32887460
JJ		 255552013-01-30 Jakub Jelinek <jakub@redhat.com>
25556
e60e09a0
JJ		 25557 PR sanitizer/55374
25558 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25559 LIBASAN_EARLY_SPEC is defined.
25560 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25561 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25562 before %o.
25563 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25564
32887460
JJ		 25565 PR c++/55742
25566 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25567 invalid args instead of ICEing on it.
25568 (ix86_valid_target_attribute_tree): Return error_mark_node if
25569 ix86_valid_target_attribute_inner_p failed.
25570 (ix86_valid_target_attribute_p): Return false only if
25571 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25572 target("default") attribute.
25573 (sorted_attr_string): Change argument from const char * to tree,
25574 merge in all target attribute arguments rather than just one.
25575 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25576 instead of free. Avoid using strcat.
25577 (ix86_mangle_function_version_assembler_name): Mangle
25578 target("default") as if no target attribute is present. Adjust
25579 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25580 instead of xmalloc and XDELETEVEC instead of free.
25581 (ix86_function_versions): Don't return true if one of the decls
25582 doesn't have target attribute. If they don't and one of the decls
25583 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25584 sorted_attr_string caller. Use XDELETEVEC instead of free.
25585 (ix86_supports_function_versions): Remove.
25586 (make_name): Fix up formatting.
25587 (make_dispatcher_decl): Remove resolver_name and its initialization.
25588 Avoid leaking memory.
25589 (is_function_default_version): Return true if there is
25590 target("default") attribute rather than no target attribute at all.
25591 (make_resolver_func): Avoid leaking memory.
25592 (ix86_generate_version_dispatcher_body): Likewise.
25593 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25594 * target.def (supports_function_versions): Remove.
25595 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25596 * doc/tm.texi: Regenerated.
25597
73cca0cc
VM		 255982013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25599
25600 PR rtl-optimization/56144
25601 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25602 for values with side effects.
25603
11452e7b
RB		 256042013-01-30 Richard Biener <rguenther@suse.de>
25605
25606 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25607 (sparseset_pop): Likewise.
25608 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25609 to be able to use quick_push in the worker loop.
25610
01cb1ef5
MP		 256112013-01-30 Marek Polacek <polacek@redhat.com>
25612
25613 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25614
e892936e
RB		 256152013-01-30 Richard Biener <rguenther@suse.de>
25616
25617 PR lto/56147
6edc3e32 25618 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 25619
cc06c01d
GJL		 256202013-01-30 Georg-Johann Lay <avr@gjlay.de>
25621
25622 PR tree-optimization/56064
25623 * fixed-value.c (fixed_from_double_int): New function.
25624 * fixed-value.h (fixed_from_double_int): New prototype.
25625 (const_fixed_from_double_int): New static inline function.
25626 * fold-const.c (native_interpret_fixed): New static function.
25627 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25628 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25629 (native_encode_fixed): New static function.
25630 (native_encode_expr) <FIXED_CST>: Use it.
25631 (native_interpret_int): Move double_int worker code to...
25632 * double-int.c (double_int::from_buffer): ...this new static method.
25633 * double-int.h (double_int::from_buffer): Prototype it.
25634
d394a308
RB		 256352013-01-30 Richard Biener <rguenther@suse.de>
25636
25637 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25638 New pointer-map and obstack.
25639 (init_alias_vars): Allocate pointer-map and obstack.
25640 (delete_points_to_sets): Free them.
25641 (find_what_var_points_to): Cache result.
25642 (find_what_p_points_to): Adjust for changed interface of
25643 find_what_var_points_to.
25644 (compute_points_to_sets): Likewise.
25645 (ipa_pta_execute): Likewise.
25646
20804d96
RO		 256472013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25648
25649 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25650 * configure: Regenerate.
25651 * config.in: Regenerate.
25652 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25653 #nobits/#progbits if supported.
25654
a7ad88a2
OE		 256552013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25656
25657 PR target/56121
25658 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25659 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25660 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25661
7af79f92
GY		 256622013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25663
91bfca59
OE		 25664 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25665 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 25666
753bcf7b
GY		 256672013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25668
91bfca59
OE		 25669 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25670 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 25671
8cbc2ea8
GY		 256722013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25673
91bfca59
OE		 25674 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25675 declaration.
25676 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25677 * config/arm/cortex-a7.md: New bypasses using
25678 arm_mac_accumulator_is_result.
8cbc2ea8 25679
697a3325
GY		 256802013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25681
91bfca59 25682 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 25683 (cortex_a7_neon_mla): Likewise.
91bfca59
OE		 25684 (cortex_a7_fpfmad): New reservation.
25685 (cortex_a7_fpmacs): Use ffmas and update required units.
25686 (cortex_a7_fpmuld): Update required units and latency.
25687 (cortex_a7_fpmacd): Likewise.
25688 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25689 (cortex_a7_neon). Likewise.
25690 (bypass) Update participating units.
697a3325 25691
29637783
GY		 256922013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25693
91bfca59
OE		 25694 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25695 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25696 from fmac to ffma.
25697 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25698 (vfp_fmul): Use ffmad.
25699 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25700 (cortex_r4_fmacd): Use ffmad.
25701 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25702 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25703 (cortex_a9_fmacd): Use ffmad.
25704 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25705 (cortex_a8_vfp_macd): Use ffmad.
25706 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25707 (cortex_a5_fpmacd): Use ffmad.
25708 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25709 (cortex_a15_vfp_macd): Use ffmad.
25710 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 25711
2ee8a2d5
JM		 257122013-01-29 Jason Merrill <jason@redhat.com>
25713
25714 PR libstdc++/54314
25715 * varasm.c (default_assemble_visibility): Don't warn about
25716 visibility on artificial decls.
25717
78d087bc
RB		 257182013-01-29 Richard Biener <rguenther@suse.de>
25719
25720 PR tree-optimization/56113
25721 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25722 the bitmap leader.
25723 (label_visit): Free duplicate bitmaps and record the leader instead.
25724 (perform_var_substitution): Adjust.
25725
83ba4d6f
RB		 257262013-01-29 Richard Biener <rguenther@suse.de>
25727
25728 PR tree-optimization/55270
25729 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25730 the CFG, schedule loops for fixup.
25731
66dfe13f
NC		 257322013-01-29 Nick Clifton <nickc@redhat.com>
25733
25734 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25735 SP_REG.
25736
5a579c3b
LE		 257372013-01-28 Leif Ekblad <leif@rdos.net>
25738
25739 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25740 * config/i386/i386.h (TARGET_RDOS): New macro.
25741 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25742 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25743 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25744 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25745 DEFAULT_LARGE_SECTION_THRESHOLD.
25746 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25747 * config/i386/rdos.h: New file.
25748 * config/i386/rdos64.h: New file.
25749
51e44392
BS		 257502013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25751
25752 PR other/54814
25753 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25754 TEST_HARD_REG_BIT.
25755
db1fb332
JJ		 257562013-01-28 Jakub Jelinek <jakub@redhat.com>
25757
25758 PR rtl-optimization/56117
25759 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25760 call cselib_lookup_from_insn on the MEM before calling
25761 add_insn_mem_dependence.
25762
16917761
RB		 257632013-01-28 Richard Biener <rguenther@suse.de>
25764
25765 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25766 to a stmt that didn't have one.
25767 (copy_phis_for_bb): Likewise for PHI arguments.
25768 (copy_debug_stmt): Likewise for debug stmts.
25769
b9fc0497
RB		 257702013-01-28 Richard Biener <rguenther@suse.de>
25771
25772 PR tree-optimization/56034
6edc3e32 25773 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB		 25774 (partition_builtin_p): Adjust.
25775 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25776 it is the last partition.
25777 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25778 up the vertex for the definition.
25779 (classify_partition): Classify whether a partition is a
25780 PKIND_REDUCTION, thus has uses outside of the loop.
25781 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25782 Merge all PKIND_REDUCTION partitions into the last partition.
25783 (tree_loop_distribution): Seed partitions from reductions as well.
25784
aa710d25
JJ		 257852013-01-28 Jakub Jelinek <jakub@redhat.com>
25786
0bfbca58
JJ		 25787 PR tree-optimization/56125
25788 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25789 pow(x,c) into sqrt(x) * powi(x, n/2) or
25790 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25791 optimizing for size.
25792 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25793 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25794 integer.
25795
aa710d25
JJ		 25796 PR tree-optimization/56094
25797 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25798 to UNKNOWN_LOCATION while gimplifying expr.
25799
77dc5297
UB		 258002013-01-27 Uros Bizjak <ubizjak@gmail.com>
25801
25802 PR target/56114
25803 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25804 operand 0 in movabs insn template for -masm=intel asm alternative.
25805 (*movabs<mode>_2): Ditto for operand 1.
25806
0bfbca58 258072013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH		 25808
25809 PR target/54663
25810 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25811 of microblaze-c.o
25812
0bfbca58 258132013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI		 25814
25815 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25816 tm_file.
25817
0bfbca58 258182013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N		 25819
25820 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25821 Undef to avoid warning.
25822
478f60f9
MH		 258232013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25824
25825 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25826 * configure: Regenerate.
25827
d7fa6ee2
JJ		 258282013-01-25 Jakub Jelinek <jakub@redhat.com>
25829
25830 PR tree-optimization/56098
25831 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25832 for stmts with volatile ops.
25833 (cond_store_replacement): Don't optimize if assign has volatile ops.
25834 (cond_if_else_store_replacement_1): Don't optimize if either
25835 then_assign or else_assign have volatile ops.
25836 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25837 volatile ops.
25838
f8fe87bd
GJL		 258392013-01-25 Georg-Johann Lay <avr@gjlay.de>
25840
25841 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25842
28398d0d
GJL		 258432013-01-25 Georg-Johann Lay <avr@gjlay.de>
25844
25845 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25846 missing ':' in asm example.
25847
b7d7d917
TB		 258482013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25849
25850 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25851 entries into lane and laneq entries.
77dc5297
UB		 25852 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25853 Remove AdvSIMD scalar modes.
b7d7d917
TB		 25854 (aarch64_sq<r>dmulh_laneq<mode>): New.
25855 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25856 modes.
25857 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25858 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25859 * config/aarch64/iterators.md (VCOND): New.
25860 (VCONQ): New.
25861
556f9906
GJL		 258622013-01-25 Georg-Johann Lay <avr@gjlay.de>
25863
25864 PR target/54222
25865 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25866 Add NULL LIBNAME argument to existing definitions.
25867 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25868 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25869 * config/avr/avr.c (DEF_BUILTIN): Same.
25870 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25871 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25872 implementation is available (DECL_ASSEMBLER_NAME is set).
25873 (avr_fold_absfx): New static function.
25874 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25875 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25876 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25877 AVR_BUILTIN_ABSLLK.
25878 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25879 (abshk, absk, abslk, absllk): Provide as static inline functions.
25880
1f546bbb
MP		 258812013-01-25 Marek Polacek <polacek@redhat.com>
25882
25883 PR tree-optimization/56035
25884 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25885
7e184bd7
UB		 258862012-01-24 Uros Bizjak <ubizjak@gmail.com>
25887
25888 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25889 (*movtf_internal_rex64): Add (!o,C) alternative
25890 (*movxf_internal_rex64): Ditto.
25891 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25892
f5ba49ac
SM		 258932013-01-24 Shenghou Ma <minux.ma@gmail.com>
25894
25895 * doc/invoke.texi: fix typo.
25896 * doc/objc.texi: fix typo.
25897
a6343728
RS		 258982013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25899
25900 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25901 for the first two alternatives.
25902
cd030c07
DN		 259032013-01-24 Diego Novillo <dnovillo@google.com>
25904
77dc5297 25905 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN		 25906 (ggc-zone.o): Remove.
25907 * configure.ac: Remove option --with-gc.
25908 * configure: Re-generate.
25909 * doc/install.texi: Remove documentation for --with-gc.
25910 * gengtype.c (write_enum_defn): Remove. Update all users.
25911 (write_Types_process_field): Remove generation of gt_e_* argument.
25912 (output_type_enum): Remove. Update all users.
25913 (write_enum_defn): Remove. Update all users.
25914 (enum alloc_zone): Remove. Update all users.
77dc5297 25915 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN		 25916 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25917 Update all callers.
25918 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 25919 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN		 25920 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25921 Update all users.
aaf1e810 25922 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN		 25923 (struct alloc_zone): Remove.
25924 (ggc_internal_alloc_zone_stat): Remove.
25925 (ggc_internal_cleared_alloc_zone_stat): Remove.
25926 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB		 25927 (ggc_pch_count_object): Remove last argument. Update all users.
25928 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN		 25929 (struct alloc_zone): Remove.
25930 * ggc-zone.c: Remove.
77dc5297 25931 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN		 25932 (struct alloc_zone): Remove.
25933 (ggc_alloc_typed_stat): Remove.
25934 (ggc_alloc_typed): Remove.
25935 (ggc_splay_alloc): Remove first argument.
25936 (rtl_zone): Remove. Update all users.
25937 (tree_zone): Remove. Update all users.
25938 (tree_id_zone): Remove. Update all users.
25939 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 25940 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 25941 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 25942 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 25943
a861ffa4
GJL		 259442013-01-24 Georg-Johann Lay <avr@gjlay.de>
25945
25946 * config/avr/avr.c (avr_out_fract): Make register numbers that
25947 might be outside of source operand signed.
25948
593c0ddd
UB		 259492013-01-24 Uros Bizjak <ubizjak@gmail.com>
25950
25951 * config/i386/constraints.md (Yf): New constraint.
25952 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25953 of f constraint to conditionaly disable x87 register preferences.
25954 (*movdf_internal): Ditto.
25955 (*movsf_internal): Ditto.
25956
e86c0101
SB		 259572013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25958
25959 PR inline-asm/55934
25960 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25961 that have operands with impossible constraints.
25962 Add a FIXME for a speed-up opportunity.
25963 * lra-constraints.c (process_alt_operands): Verify that a class
25964 selected from constraints on asms is valid for the operand mode.
25965 (curr_insn_transform): Remove incorrect comment.
25966
f6fee35f
DE		 259672013-01-23 David Edelsohn <dje.gcc@gmail.com>
25968
25969 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25970 TOC operand is a valid symbol ref in the constant pool.
25971
259722013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 25973
aaf1e810 25974 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 25975
dc62d7d1
GJL		 259762013-01-23 Georg-Johann Lay <avr@gjlay.de>
25977
25978 PR target/54222
25979 * config/avr/stdfix.h: New file.
25980 * t-avr (stdfix-gcc.h): New rule to build it.
25981 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25982
2660d12d
KS		 259832013-01-23 Kostya Serebryany <kcc@google.com>
25984
77dc5297
UB		 25985 * config/darwin.h: remove dependency on
25986 CoreFoundation (asan on Mac OS).
2660d12d 25987
a70418fc
JJ		 259882013-01-23 Jakub Jelinek <jakub@redhat.com>
25989
25990 PR target/49069
25991 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25992 instead of cmpdi_operand for first comparison operand.
25993 Don't assert that comparison operands aren't both constants.
25994
47876a2a
JW		 259952013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25996
25997 * doc/install.texi (Downloading the Source): Update references to
25998 downloading separate components.
25999
260002013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW		 26001
26002 * doc/extend.texi (__int128): Improve grammar.
26003
47876a2a 260042013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB		 26005
26006 PR target/56028
26007 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26008 alternative to (o,r).
26009 (*movdi_internal_rex64): Remove (!o,n) alternative.
26010 (DImode immediate->memory splitter): Remove.
26011 (DImode immediate->memory peephole2): Remove.
26012 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26013 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26014 alternative to (!o,*r).
26015 (*movtf_internal_sse): New pattern.
26016 (*movxf_internal_rex64): New pattern.
26017 (*movxf_internal): Disable for TARGET_64BIT.
26018 (*movdf_internal_rex64): Remove (!o,F) alternative.
26019
3a984f10
JJ		 260202013-01-22 Jakub Jelinek <jakub@redhat.com>
26021
502498d5
JJ		 26022 PR middle-end/56074
26023 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26024 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26025 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26026 stmt locations where LOCATION_LOCUS of the stmt location is
26027 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26028
3a984f10
JJ		 26029 PR target/55686
26030 * config/i386/i386.md (UNSPEC_STOS): New.
26031 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26032 *strsetqi_1): Add UNSPEC_STOS.
26033
fa817f7f
PC		 260342013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26035
26036 PR c++/56067
26037 * doc/invoke.texi: Remove left over -Wsynth example.
26038
8f498c1b
JJ		 260392013-01-21 Jakub Jelinek <jakub@redhat.com>
26040
26041 PR tree-optimization/56051
26042 * fold-const.c (fold_binary_loc): Don't fold
26043 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26044 a narrowing conversion, or widening conversion from signed
26045 to unsigned.
26046
47876a2a 260472013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB		 26048
26049 PR rtl-optimization/56023
26050 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26051 dependent on debug instruction.
26052
5b9db1bc
MJ		 260532013-01-21 Martin Jambor <mjambor@suse.cz>
26054
26055 PR middle-end/56022
26056 * function.c (allocate_struct_function): Call
26057 invoke_set_current_function_hook earlier.
26058
e8bb7d68
JJ		 260592013-01-21 Jakub Jelinek <jakub@redhat.com>
26060
26061 * reload1.c (init_reload): Only initialize reload_obstack
26062 during the first call.
26063
616a4e32
MP		 260642013-01-21 Marek Polacek <polacek@redhat.com>
26065
26066 * cfgloop.c (verify_loop_structure): Fix up grammar.
26067
4401981b
YHH		 260682013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26069
26070 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26071 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26072
8e87740b
RR		 260732013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26074
26075 PR target/56058
26076 * config/arm/marvell-pj4.md: Update copyright year.
26077 Fix up use of alu to alu_reg and simple_alu_imm.
26078
47876a2a 260792013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB		 26080
26081 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26082
89d56d79
VM		 260832013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26084
26085 PR target/55433
26086 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 26087 insn for secondary memory move when memory mode should be different.
89d56d79 26088
fe603553
JDA		 260892013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26090
26091 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26092 atomic_storedi_1): New patterns.
26093
01284895
VK		 260942013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26095
26096 btver2 pipeline descriptions.
26097 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26098 descriptions.
26099 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 26100 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK		 26101 type attributes.
26102 * config/i386/btver2.md: New file describing btver2 pipelines.
26103
5630e3e1
JL		 261042013-01-19 Andrew Pinski <apinski@cavium.com>
26105
26106 PR tree-optimization/52631
26107 * tree-ssa-sccvn (visit_use): Before looking up the original
26108 statement, try looking up the simplified expression.
26109
650ae806
AG		 261102013-01-19 Anthony Green <green@moxielogic.com>
26111
26112 * config/moxie/moxie.c (moxie_expand_prologue): Set
26113 current_function_static_stack_size.
26114
e300ec2d
JJ		 261152013-01-18 Jakub Jelinek <jakub@redhat.com>
26116
26117 PR tree-optimization/56029
26118 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26119 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26120
a3d7af04
SS		 261212013-01-18 Sharad Singhai <singhai@google.com>
26122
26123 PR tree-optimization/55995
26124 * dumpfile.c (dump_loc): Print location only if available.
26125 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26126
66aa7879
VM		 261272013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26128
26129 PR target/55433
26130 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26131 secondary memory move.
26132 (inherit_reload_reg): Use rclass instead of cl for
26133 check_secondary_memory_needed_p.
26134
3f0fee7b
JJ		 261352013-01-18 Jakub Jelinek <jakub@redhat.com>
26136
26137 PR middle-end/56015
26138 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 26139 the case where writing real complex part of target modifies op1.
3f0fee7b 26140
70c67693
JG		 261412013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26142
26143 * config/aarch64/aarch64-simd.md
26144 (aarch64_vcond_internal<mode>): Handle unordered cases.
26145 * config/aarch64/iterators.md (v_cmp_result): New.
26146
df8de9b3
YHH		 261472013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26148 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26149
26150 * config/arm/marvell-pj4.md: New file.
26151 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26152 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26153 (generic_vfp): Likewise.
26154 * config/arm/arm-cores.def: Add marvell-pj4.
26155 * config/arm/arm-tune.md: Regenerate.
26156 * config/arm/arm-tables.opt: Regenerate.
26157 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26158 * doc/invoke.texi: Document marvell-pj4.
26159
be30c356
TB		 261602013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26161
26162 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26163
0bfbca58 261642013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO		 26165
26166 PR debug/54114
26167 PR debug/54402
26168 PR debug/49888
26169 * var-tracking.c (negative_power_of_two_p): New.
26170 (global_get_addr_cache, local_get_addr_cache): New.
26171 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26172 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26173 heading comment.
26174 (vt_stack_offset_p): Remove.
26175 (vt_canon_true_dep): Always canonicalize loc's address.
26176 (clobber_overlapping_mems): Make sure we have a MEM.
26177 (local_get_addr_clear_given_value): New.
26178 (val_reset): Clear local cached entries.
26179 (compute_bb_dataflow): Create and release the local cache.
26180 Disable duplicate MEMs clobbering.
26181 (emit_notes_in_bb): Clobber MEMs likewise.
26182 (vt_emit_notes): Create and release the local cache.
26183 (vt_initialize, vt_finalize): Create and release the global
26184 cache, respectively.
1f6bc337 26185 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 26186
0bfbca58 261872013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO		 26188
26189 PR libmudflap/53359
26190 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26191 not found in the symtab.
26192
0bfbca58 261932013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 26194
c350ba53 26195 PR debug/56006
3aa03517
AO		 26196 PR rtl-optimization/55547
26197 PR rtl-optimization/53827
26198 PR debug/53671
26199 PR debug/49888
26200 * alias.c (offset_overlap_p): New, factored out of...
26201 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26202 the conservative special case for symbolic constants. Don't
26203 adjust zero sizes on alignment.
26204
c664546f
JL		 262052013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26206
26207 PR rtl-optimization/52573
26208 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26209 REG_UNUSED for the same register.
26210
1bd3f750
MP		 262112013-01-17 Richard Biener <rguenther@suse.de>
26212 Marek Polacek <polacek@redhat.com>
26213
26214 PR rtl-optimization/55833
26215 * loop-unswitch.c (unswitch_loops): Move loop verification...
26216 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26217 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26218 Set it to true when we're removing a loop from hierarchy tree in
26219 an irreducible region.
26220 (fix_bb_placements): Adjust caller.
26221 (fix_loop_placements): Likewise.
26222
e52a8b71
GJL		 262232013-01-17 Georg-Johann Lay <avr@gjlay.de>
26224
26225 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26226 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26227 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26228 Remove ID. Adjust comments.
26229 * config/avr/avr-c.c (avr_builtin_name): Remove.
26230 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26231 * config/avr/avr.c (avr_tolower): New static function.
26232 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26233 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26234 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26235 default expansion.
26236
8386a7ea
JH		 262372013-01-17 Jan Hubicka <jh@suse.cz>
26238
610fb637 26239 PR tree-optimization/55273
8386a7ea
JH		 26240 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26241
47876a2a 262422013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB		 26243
26244 PR target/55981
26245 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26246 store through atomic_store<mode>_1.
26247 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26248
8222c37e
MJ		 262492013-01-17 Martin Jambor <mjambor@suse.cz>
26250
26251 PR tree-optimizations/55264
26252 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26253 for virtual methods.
26254 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26255 virtual methods before inlining is over.
26256 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26257 virtual functions.
26258 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26259 non-virtual.
26260
79f01c76
VM		 262612013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26262
26263 PR rtl-optimization/56005
26264 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26265 pending reads for prefetch.
26266
d0b6bb1b
IB		 262672013-01-16 Ian Bolton <ian.bolton@arm.com>
26268
aaf1e810 26269 * config/aarch64/aarch64.md
d0b6bb1b
IB		 26270 (*cstoresi_neg_uxtw): New pattern.
26271 (*cmovsi_insn_uxtw): New pattern.
26272 (*<optab>si3_uxtw): New pattern.
26273 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26274 (*<optab>si3_insn_uxtw): New pattern.
26275 (*bswapsi2_uxtw): New pattern.
26276
cb9cf03b
RB		 262772013-01-16 Richard Biener <rguenther@suse.de>
26278
26279 * tree-inline.c (tree_function_versioning): Remove set but
26280 never used variable.
26281
2cfc56b9
RB		 262822013-01-16 Richard Biener <rguenther@suse.de>
26283
26284 PR tree-optimization/55964
26285 * tree-flow.h (rename_variables_in_loop): Remove.
26286 (rename_variables_in_bb): Likewise.
26287 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26288 (copy_loop_before): Adjust and delete update-ssa status.
26289 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 26290 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB		 26291 (rename_variables_in_loop): Remove.
26292 (slpeel_update_phis_for_duplicate_loop): Likewise.
26293 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26294 use available cfg machinery instead of duplicating it.
26295 Update PHI nodes and perform poor-mans SSA update here.
26296 (slpeel_tree_peel_loop_to_edge): Adjust.
26297
c25a0c60
RB		 262982013-01-16 Richard Biener <rguenther@suse.de>
26299
26300 PR tree-optimization/54767
26301 PR tree-optimization/53465
26302 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26303 (vrp_visit_phi_node): For PHI arguments coming via backedges
26304 drop all symbolical range information.
26305 (execute_vrp): Compute backedges.
26306
04b535af
RB		 263072013-01-16 Richard Biener <rguenther@suse.de>
26308
26309 * doc/install.texi: Update CLooG and ISL requirements to
26310 0.18.0 and 0.11.1.
26311
8b0a1e0b
CB		 263122013-01-16 Christian Bruel <christian.bruel@st.com>
26313
26314 PR target/55301
26315 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26316 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26317 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26318
263192013-01-16 DJ Delorie <dj@redhat.com>
26320
26321 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26322 (UNSPECV_SP_SWITCH_E): New.
26323 (sp_switch_1): Change to an unspec.
26324 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26325 replace $r15.
26326
263272013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB		 26328
26329 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26330 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26331 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26332 (expand_mem_thread_fence): Ditto.
26333 (expand_mem_signal_fence): Ditto.
26334 (expand_atomic_load): Ditto.
26335 (expand_atomic_store): Ditto.
26336
0bfbca58 263372013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO		 26338
26339 PR rtl-optimization/55547
26340 PR rtl-optimization/53827
26341 PR debug/53671
26342 PR debug/49888
26343 * alias.c (memrefs_conflict_p): Set sizes to negative after
26344 AND adjustments.
26345
305e3ac1
JJ		 263462013-01-15 Jakub Jelinek <jakub@redhat.com>
26347
26348 PR target/55940
26349 * function.c (thread_prologue_and_epilogue_insns): Always
26350 add crtl->drap_reg to set_up_by_prologue.set, even if
26351 stack_realign_drap is false.
26352
f78ac4f2
JBG		 263532013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26354
26355 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26356 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26357 *call): Fix indention.
26358
a78a8cc4
TV		 263592013-01-15 Tom de Vries <tom@codesourcery.com>
26360
26361 PR target/55876
26362 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26363 Update comment.
26364
0e80383f
VM		 263652013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26366
305e3ac1 26367 PR rtl-optimization/55153
0e80383f
VM		 26368 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26369
207b5956
MJ		 263702013-01-15 Martin Jambor <mjambor@suse.cz>
26371
26372 PR tree-optimization/55920
26373 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26374 accesses as grp_to_be_debug_replaced.
26375
a7818b54
JJ		 263762013-01-15 Jakub Jelinek <jakub@redhat.com>
26377
26378 PR tree-optimization/55920
26379 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26380 there is non-useless type conversion needed from debug rhs to lhs,
26381 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26382
b0fe107e
JM		 263832013-01-15 Joseph Myers <joseph@codesourcery.com>
26384 Mikael Pettersson <mikpe@it.uu.se>
26385
26386 PR target/43961
26387 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26388 Thumb.
26389 (ASM_OUTPUT_CASE_LABEL): Remove.
26390 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26391 * final.c (shorten_branches): Update alignment of labels before
26392 jump tables if CASE_VECTOR_SHORTEN_MODE.
26393
34ab62ee
RB		 263942013-01-15 Richard Biener <rguenther@suse.de>
26395
26396 PR bootstrap/55961
26397 * system.h: Do not include gmp.h for building host tools.
26398
783a3a05
RB		 263992013-01-15 Richard Biener <rguenther@suse.de>
26400
26401 PR middle-end/55882
26402 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26403 account for bitpos when computing alignment.
26404
3a579e09
VY		 264052013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26406
26407 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26408 (ix86_target_macros_internal): Likewise.
26409
26410 * config/i386/i386.c (m_CORE2I7): Removed.
26411 (m_CORE_HASWELL): New macro.
26412 (m_CORE_ALL): Likewise.
26413 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26414 (initial_ix86_arch_features): Likewise.
26415 (processor_target_table): Initializations for Core avx2.
26416 (cpu_names): New names "core-avx2".
26417 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26418 PROCESSOR_CORE_HASWELL.
26419 (ix86_issue_rate): New case.
26420 (ia32_multipass_dfa_lookahead): Likewise.
26421 (ix86_sched_init_global): Likewise.
26422
26423 * config/i386/i386.h (TARGET_HASWELL): New macro.
26424 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26425 (processor_type): New PROCESSOR_HASWELL.
26426
955f5a07
JJ		 264272013-01-15 Jakub Jelinek <jakub@redhat.com>
26428
ff784829
JJ		 26429 PR tree-optimization/55955
26430 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26431 *SHIFT_EXPR and *ROTATE_EXPR codes.
26432
955f5a07
JJ		 26433 PR tree-optimization/48766
26434 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26435 -ftrapv disable -fwrapv.
26436
aeb8b4e9
GJL		 264372013-01-14 Georg-Johann Lay <avr@gjlay.de>
26438
26439 PR target/55974
26440 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26441 etc. to 1 and not to __flash.
26442 Use LL suffix for __INT24_MAX__ with -mint8.
26443 Use ULL suffix for __UINT24_MAX__ with -mint8.
26444
1c494c6a
GJL		 264452013-01-14 Georg-Johann Lay <avr@gjlay.de>
26446
26447 * config/avr/avr-arch.h
26448 (struct base_arch_s): Use typedef avr_arch_t instead.
26449 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26450 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26451 * config/avr/avr.c: Same.
26452 * config/avr/avr-devices.c: Same.
26453 * config/avr/driver-avr.c: Same.
26454 * config/avr/gen-avr-mmcu-texi.c: Same.
26455 * config/avr/avr-mcus.def: Adjust comment.
26456
a50344cb
TB		 264572013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26458
88e784e6
UB		 26459 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26460 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 26461
47876a2a 264622013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB		 26463 Andi Kleen <ak@linux.intel.com>
26464
26465 PR target/55948
26466 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26467 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26468 memmodel flag.
26469
00892272
GJL		 264702013-01-14 Georg-Johann Lay <avr@gjlay.de>
26471
26472 * config/avr/avr-stdint.h: Remove trailing blanks.
26473 * config/avr/avr-log.h: Same.
26474 * config/avr/avr-arch.h: Same.
26475 * config/avr/avr-devices.c: Same.
26476 * config/avr/avr-dimode.md: Same.
26477 * config/avr/predicates.md: Same.
26478 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 26479
00892272
GJL		 26480 * config/avr/avr-protos.h: Same. And:
26481 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26482 (init_cumulative_args): Rename to avr_init_cumulative_args.
26483 (expand_prologue): Rename to avr_expand_prologue.
26484 (expand_epilogue): Rename to avr_expand_epilogue.
26485 (adjust_insn_length): Rename to avr_adjust_insn_length.
26486 (notice_update_cc): Rename to avr_notice_update_cc.
26487 (final_prescan_insn): Rename to avr_final_prescan_insn.
26488 * config/avr/avr.c: Same.
26489 * config/avr/avr.h: Same.
26490 * config/avr/avr.md: Remove trailing blanks.
26491 (prologue): Use avr_expand_prologue.
26492 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26493
7bb01996
RB		 264942013-01-14 Richard Biener <rguenther@suse.de>
26495
26496 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26497 verify_location, collect_subblocks): New functions.
26498 (verify_gimple_in_cfg): Verify that locations only reference
26499 BLOCKs in the functions BLOCK tree.
26500
2724573f
RB		 265012013-01-14 Richard Biener <rguenther@suse.de>
26502
26503 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26504 PHI argument.
26505 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26506 unshare reference.
26507 (insert_out_of_ssa_copy_on_edge): Likewise.
26508 (rewrite_close_phi_out_of_ssa): Likewise.
26509 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26510 debug expressions.
26511 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26512 propagated constants.
26513 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26514 can not be shared.
26515
9a0bbab6
GJL		 265162013-01-14 Georg-Johann Lay <avr@gjlay.de>
26517
26518 * config/avr/avr-modes.def: Add GPL copyright notice.
26519
45805f17
UB		 265202013-01-13 Uros Bizjak <ubizjak@gmail.com>
26521
26522 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26523 MEMMODEL_MASK to determine memory model.
26524 (atomic_store<mode>): Ditto from operands[2].
26525 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26526
9d60be38
JJ		 265272013-01-13 Jakub Jelinek <jakub@redhat.com>
26528
26529 PR fortran/55935
45805f17 26530 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ		 26531 (fold_gimple_assign): Don't call unshare_expr here.
26532 (fold_ctor_reference): Call unshare_expr.
26533
e7f49d92
TG		 265342013-01-13 Terry Guo <terry.guo@arm.com>
26535
aaf1e810
EB		 26536 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26537 * doc/fragments.texi: Document MULTILIB_REUSE.
26538 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 26539 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 26540 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG		 26541 (tmpmultilib4): Ditto.
26542 (multilib_reuse): New multilib argument.
26543
fbd03a27
RS		 265442013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26545
26546 * Makefile.in: Update copyright.
26547
b3681f13
TV		 265482013-01-12 Tom de Vries <tom@codesourcery.com>
26549
26550 PR middle-end/55890
26551 * calls.c (expand_call): Check if arg_nr is valid.
26552
3f287e4b
MM		 265532013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26554
26555 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26556 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM		 26557 documentation. Add missing '__' in front of
26558 __builtin_ia32_packssdw256.
3f287e4b 26559
1abcd5eb
AK		 265602013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26561
26562 PR target/55719
26563 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26564 NO_REGS for larl operands.
26565 (s390_reload_larl_operand): Use s390_load_address instead of
26566 emit_move_insn.
26567
980d0812
RB		 265682013-01-11 Richard Biener <rguenther@suse.de>
26569
26570 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26571 (verify_node_sharing): ... here.
26572 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26573
7cb9fd07
EB		 265742013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26575
26576 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26577 Substitute TREECHECKING.
26578 * configure: Regenerate.
26579 * Makefile.in (TREECHECKING): New.
26580
47876a2a 265812013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG		 26582
26583 PR tree-optimization/44061
26584 * tree-vrp.c (extract_range_basic): Compute zero as
26585 value-range for __builtin_constant_p of function parameters.
26586
d1e082c2
RS		 265872013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26588
45805f17 26589 Update copyright years.
d1e082c2 26590
f85021b0
VM		 265912013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26592
9d60be38 26593 PR rtl-optimization/55672
0160442c
VM		 26594 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26595 const to be eliminable.
f85021b0 26596
7a8b1ec4
DE		 265972013-01-10 David Edelsohn <dje.gcc@gmail.com>
26598
26599 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26600 * configure: Regenerate.
26601
ad2e5b71
RB		 266022013-01-10 Richard Biener <rguenther@suse.de>
26603
ddf9322d 26604 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 26605
ffc5b2bb
RB		 266062013-01-10 Richard Biener <rguenther@suse.de>
26607
26608 PR bootstrap/55792
26609 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26610 locations for virtual PHI arguments.
26611 (rewrite_update_phi_arguments): Likewise.
26612
e1f674e4
JS		 266132013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26614
26615 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26616 on to assembler.
26617
a57fc743
JJ		 266182013-01-10 Jakub Jelinek <jakub@redhat.com>
26619
26620 PR tree-optimization/55921
26621 * tree-complex.c (expand_complex_asm): New function.
26622 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26623
0ff4390d
AK		 266242013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26625
26626 PR target/55718
26627 * config/s390/s390.c (s390_symref_operand_p)
26628 (s390_loadrelative_operand_p): Merge the two functions.
26629 (s390_check_qrst_address, print_operand_address): Add parameters
26630 to s390_loadrelative_operand_p invokation.
26631 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26632 (s390_reload_larl_operand, s390_secondary_reload): Use
26633 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26634 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26635
6f557e0e
MS		 266362013-01-09 Mike Stump <mikestump@comcast.net>
26637
26638 * dse.c (record_store): Remove unnecessary assert.
26639
7770c9e9
JH		 266402013-01-09 Jan Hubicka <jh@suse.cz>
26641
26642 PR tree-optimization/55569
26643 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26644 * cfgloop.h (scale_loop_profile): Likewise.
26645
a19b1432
JH		 266462013-01-09 Jan Hubicka <jh@suse.cz>
26647
26648 PR lto/45375
ddf9322d
UB		 26649 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26650 functions.
a19b1432
JH		 26651 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26652
9a002da8
RS		 266532013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26654
26655 PR middle-end/55114
26656 * expr.h (maybe_emit_group_store): Declare.
26657 * expr.c (maybe_emit_group_store): New function.
26658 * builtins.c (expand_builtin_int_roundingfn): Call it.
26659 (expand_builtin_int_roundingfn_2): Likewise.
26660
511dcace
VM		 266612013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26662
e1f2b729 26663 PR rtl-optimization/55829
511dcace
VM		 26664 * lra-constraints.c (match_reload): Add code for absent output.
26665 (curr_insn_transform): Add code for reloads of matched inputs
26666 without output.
26667
7b0fe4f4
UB		 266682013-01-09 Uros Bizjak <ubizjak@gmail.com>
26669
26670 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26671 attribute of movddup insn to DF.
26672 (*vec_interleave_lowv2df): Ditto.
26673 (vec_dupv2df): Ditto.
26674
870ca331
JH		 266752013-01-09 Jan Hubicka <jh@suse.cz>
26676
26677 PR tree-optimiation/55875
26678 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26679 EVERY_ITERATION parameter.
7b0fe4f4 26680 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH		 26681 (idx_infer_loop_bounds): Similarly here.
26682 (n_of_executions_at_most): Simplify
26683 to only test for cases where statement is dominated by the
7b0fe4f4 26684 particular bound; handle correctly the "postdominance" test.
870ca331
JH		 26685 (scev_probably_wraps_p): Use max loop iterations info
26686 as a global bound first.
26687
6f575fe4 266882013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD		 26689 Nick Clifton <nickc@redhat.com>
26690
26691 * config/v850/v850.md (cbranchsf4): New pattern.
26692 (cstoresf4): New pattern.
26693 (cbranchdf4): New pattern.
26694 (cstoredf4): New pattern.
26695 (movsicc): Disallow floating point comparisons.
26696 (cmpsf_le_insn): Fix order of operators.
26697 (cmpsf_lt_insn): Likewise.
26698 (cmpsf_eq_insn): Likewise.
26699 (cmpdf_le_insn): Likewise.
26700 (cmpdf_lt_insn): Likewise.
26701 (cmpdf_eq_insn): Likewise.
26702 (cmpsf_ge_insn): Use LE comparison.
26703 (cmpdf_ge_insn): Likewise.
26704 (cmpsf_gt_insn): Use LT comparison.
26705 (cmpdf_gt_insn): Likewise.
26706 (cmpsf_ne_insn): Delete pattern.
26707 (cmpdf_ne_insn): Delete pattern.
26708 * config/v850/v850.c (v850_gen_float_compare): Use
26709 gen_cmpdf_eq_insn for NE comparison.
26710 (v850_float_z_comparison_operator)
26711 (v850_float_nz_comparison_operator): Move from here ...
26712 * config/v850/predicates.md: ... to here. Move GT and GE
26713 comparisons into v850_float_z_comparison_operator.
26714 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26715 Delete prototype.
26716 (v850_float_nz_comparison_operator): Likewise.
26717
f0d54148
JDA		 267182013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26719
26720 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26721 with calls to gen_insvsi/gen_insvdi.
26722
8f01beca
VK		 267232013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26724
26725 * config/i386/i386.c (initial_ix86_tune_features): Set up
26726 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26727
2aa579ad
JJ		 267282013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26729 Jakub Jelinek <jakub@redhat.com>
26730
26731 PR tree-optimization/48189
26732 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26733 If nitercst is 0, don't predict the exit edge.
26734
6edc3e32 267352013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N		 26736
26737 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26738 in asm_fprintf with reg_names.
26739 (aarch64_print_operand_address): Likewise.
26740 (aarch64_return_addr): Likewise.
26741 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26742
f6f94d94
JDA		 267432013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26744
26745 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26746 (INT_U6_BITS): Likewise.
26747 * config/pa/predicates.md (uint6_operand): New predicate.
26748 (shift5_operand, shift6_operand): Likewise.
26749 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26750 arith32_operand.
26751 (lshrdi3): Use shift6_operand.
26752 (shrpsi4, shrpdi4): New insn patterns.
26753 (extzv): Delete expander.
26754 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26755 predicates in unamed zero extract patterns. Tighten common constraint.
26756 (extv): Delete expander.
26757 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26758 predicates in unamed sign extract patterns. Tighten common constraint.
26759 (insv): Delete expander.
26760 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26761 predicates in unamed insert patterns. Tighten common constraint.
26762 Change uint32_operand predicate to uint6_operand predicate in unamed
26763 DImode pattern to insert constant values of type 1...1xxxx.
26764
36b72910
JH		 267652013-01-04 Jan Hubicka <jh@suse.cz>
26766
26767 PR tree-optimization/55823
7b0fe4f4
UB		 26768 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26769 issue.
36b72910 26770
47876a2a 267712013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ		 26772 Uros Bizjak <ubizjak@gmail.com>
26773
26774 PR rtl-optimization/55845
26775 * df-problems.c (can_move_insns_across): Stop scanning at
26776 volatile_insn_p source instruction or give up if
26777 across_from .. across_to range contains any volatile_insn_p
26778 instructions.
26779
4369c11e
TB		 267802013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26781
7b0fe4f4
UB		 26782 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26783 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26784 Declare.
4369c11e 26785 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 26786 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 26787
aeb7e7c1
JJ		 267882013-01-08 Jakub Jelinek <jakub@redhat.com>
26789
26790 PR fortran/55341
26791 * asan.c (asan_clear_shadow): New function.
26792 (asan_emit_stack_protection): Use it.
26793
a02ad1aa
TB		 267942013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26795
26796 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26797 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26798 with tab instead of space.
26799
f90d87f5
NC		 268002013-01-08 Nick Clifton <nickc@redhat.com>
26801
26802 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26803 register bank 0 at the start of an interrupt handler.
83ffd964
NC		 26804 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26805 MDBH registers.
f90d87f5 26806
385eb93d
JG		 268072013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26808
26809 * config/aarch64/aarch64-simd.md
26810 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26811 (aarch64_simd_bsl): Likewise.
26812 (aarch64_vcond_internal<mode>): Likewise.
26813 (vcond<mode><mode>): Likewise.
26814 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26815 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26816
4dcd1054
JG		 268172013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26818
26819 * config/aarch64/aarch64-builtins.c
26820 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26821
4267a4a6
MJ		 268222013-01-08 Martin Jambor <mjambor@suse.cz>
26823
26824 PR debug/55579
26825 * tree-sra.c (analyze_access_subtree): Return true also after
26826 potentially creating a debug-only replacement.
26827
5f4e6de3
JJ		 268282013-01-08 Jakub Jelinek <jakub@redhat.com>
26829
3138f224
JJ		 26830 PR middle-end/55890
26831 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26832
5f4e6de3
JJ		 26833 PR tree-optimization/54120
26834 * tree-vrp.c (range_fits_type_p): Don't allow
26835 src_precision < precision from signed vr to unsigned_p
26836 if vr->min or vr->max is negative.
26837 (simplify_float_conversion_using_ranges): Test can_float_p
26838 against CODE_FOR_nothing.
26839
f3ef18ff
JJ		 268402013-01-08 Jakub Jelinek <jakub@redhat.com>
26841 Richard Biener <rguenther@suse.de>
26842
26843 PR middle-end/55851
26844 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26845 types instead of just INTEGER_TYPE types.
26846
25c210f9
MK		 268472013-01-07 Mark Kettenis <kettenis@openbsd.org>
26848
26849 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26850 TF_SIZE): Define.
f90d87f5 26851
81826a7b
SE		 268522013-01-07 Steve Ellcey <sellcey@mips.com>
26853
26854 PR target/42661
26855 * config/mips/mips.opt: Change mad to mmad to match documentation.
26856
35678514
GJL		 268572013-01-07 Georg-Johann Lay <avr@gjlay.de>
26858
26859 PR target/55897
26860 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26861 .progmemx.data now.
26862
e5669488
GJL		 268632013-01-07 Georg-Johann Lay <avr@gjlay.de>
26864
26865 PR target/55897
26866 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26867 (avr_addrspace_t): Add .section_name field.
26868 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26869 array size.
26870 (avr_addrspace): Same. Initialize .section_name. Remove last
26871 NULL entry. Put __memx into .progmemx.data.
26872 (progmem_section_prefix): Remove.
26873 (avr_asm_init_sections): No need to initialize progmem_section.
26874 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26875 section name prefix.
26876 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26877 retrieve the progmem section.
26878 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26879 boundary to run over avr_addrspace[].
26880 (avr_register_target_pragmas): Ditto.
26881
656e6f37
JJ		 268822013-01-06 Jakub Jelinek <jakub@redhat.com>
26883
26884 * varasm.c (output_constant_def_contents): For asan_protect_global
26885 protected strings, adjust DECL_ALIGN if needed, before testing for
26886 anchored symbols.
26887 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26888 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26889 normal decls.
26890 (output_object_block): For asan protected decls, emit asan padding
26891 after their contents.
26892 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26893 (asan_finish_file): Test it here instead.
26894
6edc3e32
UB		 268952013-01-07 Nick Clifton <nickc@redhat.com>
26896 Matthias Klose <doko@debian.org>
26897 Doug Kwan <dougkwan@google.com>
26898 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC		 26899
26900 PR driver/55470
26901 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26902
26903 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26904
26905 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26906
7b0fe4f4 26907 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC		 26908
26909 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26910
2a095093
GJL		 269112013-01-07 Georg-Johann Lay <avr@gjlay.de>
26912
f3b503f4 26913 PR target/54461
2a095093
GJL		 26914 * doc/install.texi (Cross-Compiler-Specific Options): Document
26915 --with-avrlibc.
26916
383f9b34
TB		 269172013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26918
26919 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26920 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26921 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26922 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26923 vqmovun_high_s64): Fix source operand number and update copyright.
26924
3626621a
RB		 269252013-01-07 Richard Biener <rguenther@suse.de>
26926
26927 PR middle-end/55890
26928 * gimple.h (gimple_call_builtin_p): New overload.
26929 * gimple.c (validate_call): New function.
26930 (gimple_call_builtin_p): Likewise.
26931 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26932 Use gimple_call_builtin_p.
26933 (find_func_clobbers): Likewise.
26934 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26935 (strlen_optimize_stmt): Likewise.
26936
8b2ea410
JG		 269372013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26938
26939 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26940 (vld1q_dup_*): Likewise.
26941 (vld1_*): Likewise.
26942 (vld1q_*): Likewise.
26943 (vld1_lane_*): Likewise.
26944 (vld1q_lane_*): Likewise.
26945
9713d329
RB		 269462013-01-07 Richard Biener <rguenther@suse.de>
26947
26948 * lto-streamer.h (LTO_minor_version): Bump to 2.
26949
3520f7cc
JG		 269502013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26951
26952 * config/aarch64/aarch64-protos.h
26953 (aarch64_const_double_zero_rtx_p): Rename to...
26954 (aarch64_float_const_zero_rtx_p): ...this.
26955 (aarch64_float_const_representable_p): New.
26956 (aarch64_output_simd_mov_immediate): Likewise.
26957 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26958 move immediate case.
26959 * config/aarch64/aarch64.c
26960 (aarch64_const_double_zero_rtx_p): Rename to...
26961 (aarch64_float_const_zero_rtx_p): ...this.
26962 (aarch64_print_operand): Allow printing of new constants.
26963 (aarch64_valid_floating_const): New.
26964 (aarch64_legitimate_constant_p): Check for valid floating-point
26965 constants.
26966 (aarch64_simd_valid_immediate): Likewise.
26967 (aarch64_vect_float_const_representable_p): New.
26968 (aarch64_float_const_representable_p): Likewise.
26969 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26970 (aarch64_output_simd_mov_immediate): New.
26971 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26972 (*movdf_aarch64): Likewise.
26973 * config/aarch64/constraints.md (Ufc): New.
26974 (Y): call aarch64_float_const_zero_rtx.
26975 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26976
e167c04d
RB		 269772013-01-07 Richard Biener <rguenther@suse.de>
26978
26979 PR tree-optimization/55888
26980 PR tree-optimization/55862
26981 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26982 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26983 not if it is contained therein.
26984
0139aaab
GJL		 269852013-01-07 Georg-Johann Lay <avr@gjlay.de>
26986
26987 * config/avr/t-avr: Typo.
26988
4a176b23
GJL		 269892013-01-07 Georg-Johann Lay <avr@gjlay.de>
26990
26991 PR55243
26992 * config/avr/t-avr: Don't automatically rebuild
26993 $(srcdir)/config/avr/t-multilib
26994 $(srcdir)/config/avr/avr-tables.opt
26995 $(srcdir)/doc/avr-mmcu.texi
26996 (avr-mcus): New phony target to build them on request.
26997 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26998 * avr/avr-mcus.def: Adjust comments.
26999
c7afdc98
UB		 270002013-01-07 Uros Bizjak <ubizjak@gmail.com>
27001
27002 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27003
1ab05c31
RS		 270042013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27005
27006 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27007
488e3acc
RS		 270082013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27009
27010 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27011
a89599d2
DE		 270122013-01-05 David Edelsohn <dje.gcc@gmail.com>
27013
27014 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27015 to generate profiling.
27016 * config/rs6000/aix64.h (LIB_SPEC): Same.
27017
70f09188
AP		 270182013-01-04 Andrew Pinski <apinski@cavium.com>
27019
27020 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27021 New function.
27022 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27023
918d445f
UB		 270242013-01-04 Uros Bizjak <ubizjak@gmail.com>
27025
27026 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27027 unconditionally.
27028 (ix86_expand_move): Ditto.
27029 (ix86_zero_extend_to_Pmode): Ditto.
27030 (ix86_expand_call): Ditto.
27031 (ix86_expand_special_args_builtin): Ditto.
27032 (ix86_expand_builtin): Ditto.
27033
361618ec
RB		 270342013-01-04 Richard Biener <rguenther@suse.de>
27035
27036 PR tree-optimization/55862
27037 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27038 translating them through PHI nodes.
27039
87eab554
MJ		 270402013-01-04 Martin Jambor <mjambor@suse.cz>
27041
27042 PR tree-optimization/55755
27043 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27044 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27045
33879b9f
MP		 270462013-01-04 Marek Polacek <polacek@redhat.com>
27047
27048 PR middle-end/55859
27049 * opts.c (default_options_optimization): Clarify error message.
27050
3068819a
RB		 270512013-01-04 Richard Biener <rguenther@suse.de>
27052
27053 PR middle-end/55863
27054 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27055 reassociation.
27056
0e4ae794
JDA		 270572013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27058
27059 PR target/53789
27060 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA		 27061 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27062 references.
0e4ae794 27063
65c88cef
RH		 270642013-01-03 Richard Henderson <rth@redhat.com>
27065
27066 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27067 after eliminating TLS symbols.
27068
8ac16127
MG		 270692013-01-03 Marc Glisse <marc.glisse@inria.fr>
27070
27071 PR bootstrap/50167
27072 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27073 * graphite-poly.c (debug_gmp_value): Likewise.
27074
bb664f09
UB		 270752013-01-03 Uros Bizjak <ubizjak@gmail.com>
27076
27077 PR target/55712
27078 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27079 selected code model, define __code_mode_small__, __code_model_medium__,
27080 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27081 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27082 xchg temporary register with %k. Declare temporary register as
27083 early clobbered.
27084 [__x86_64__]: For medium and large code models, preserve %rbx register.
27085
bcf1ef00
RB		 270862013-01-03 Richard Biener <rguenther@suse.de>
27087
0506634a 27088 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB		 27089 (dump_subscript): Adjust.
27090 (finalize_ddr_dependent): Do not dump redundant info.
27091 (analyze_siv_subscript): Adjust.
27092 (subscript_dependence_tester): Likewise.
27093 (compute_affine_dependence): Likewise.
27094
59fd17e3
RB		 270952013-01-03 Richard Biener <rguenther@suse.de>
27096
27097 Revert
27098 2013-01-03 Richard Biener <rguenther@suse.de>
27099
27100 PR tree-optimization/55857
27101 * tree-vect-stmts.c (vectorizable_load): Do not setup
27102 re-alignment for invariant loads.
27103
27104 2013-01-02 Richard Biener <rguenther@suse.de>
27105
27106 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 27107 invariant load do not generate a vector load from the scalar location.
59fd17e3 27108
595c2679
RB		 271092013-01-03 Richard Biener <rguenther@suse.de>
27110
27111 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27112 for not vectorizing.
27113 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27114 not build INDIRECT_REFs, call get_name once only.
27115 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27116 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27117
90bb7d7a
RB		 271182013-01-03 Richard Biener <rguenther@suse.de>
27119
27120 PR tree-optimization/55857
27121 * tree-vect-stmts.c (vectorizable_load): Do not setup
27122 re-alignment for invariant loads.
27123
f09b77ca
RB		 271242013-01-03 Richard Biener <rguenther@suse.de>
27125
27126 PR lto/55848
27127 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27128 prefer a built-in decl.
27129
3461a16e
JJ		 271302013-01-03 Jakub Jelinek <jakub@redhat.com>
27131
df8e2b4f
JJ		 27132 * gcc.c (process_command): Update copyright notice dates.
27133 * gcov.c (print_version): Likewise.
27134 * gcov-dump.c (print_version): Likewise.
27135
3461a16e
JJ		 27136 PR rtl-optimization/55838
27137 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27138 iv0.step, iv1.step and step.
27139
8b5546d6
JJ		 271402013-01-03 Jakub Jelinek <jakub@redhat.com>
27141 Marc Glisse <marc.glisse@inria.fr>
27142
27143 PR tree-optimization/55832
27144 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27145 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27146 integer_{one,zero}_node.
27147
8ab1d2e9
JJ		 271482013-01-03 Jakub Jelinek <jakub@redhat.com>
27149
27150 PR debug/54402
27151 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27152 * var-tracking.c (reverse_op): Don't add reverse ops to
27153 VALUEs that have already
0506634a 27154 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 27155
5b9ad1d4
GP		 271562013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27157
27158 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27159
e25a6711
TJ		 271602013-01-02 Teresa Johnson <tejohnson@google.com>
27161
27162 * dumpfile.c (dump_loc): Print filename with location.
27163 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27164 new location_t parameter to emit complete unroll message with
27165 new dump framework.
27166 (canonicalize_loop_induction_variables): Compute loops location
27167 and pass to try_unroll_loop_completely.
27168 * loop-unroll.c (report_unroll_peel): New function.
27169 (peel_loops_completely): Use new dump format with location
27170 for main dumpfile message, and invoke report_unroll_peel on success.
27171 (decide_unrolling_and_peeling): Ditto.
27172 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27173 by report_unroll_peel.
27174 (decide_peel_completely): Ditto.
27175 (decide_unroll_constant_iterations): Ditto.
27176 (decide_unroll_runtime_iterations): Ditto.
27177 (decide_peel_simple): Ditto.
27178 (decide_unroll_stupid): Ditto.
27179 * cfgloop.c (get_loop_location): New function.
27180 * cfgloop.h (get_loop_location): Declare.
27181
77878621
ST		 271822013-01-02 Sriraman Tallam <tmsriram@google.com>
27183
27184 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27185 NULL.
27186
9e65d03e
JDA		 271872013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27188
27189 PR middle-end/55198
27190 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27191 BLKmode objects when EXPAND_MEMORY is specified.
27192
6a7da30f
ST		 271932013-01-02 Sriraman Tallam <tmsriram@google.com>
27194
27195 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27196 in loop predicate.
27197 (fold_builtin_cpu): Do not share cpu model decls across statements.
27198
e78167f2
JM		 271992013-01-02 Jason Merrill <jason@redhat.com>
27200
27201 PR c++/55804
27202 * tree.c (build_array_type_1): Revert earlier change.
27203
8c075fb4
YZ		 272042013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27205
27206 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27207 "cortex-a57".
27208 * config/aarch64/aarch64-tune.md: Re-generate.
27209
0682ed3e
RB		 272102013-01-02 Richard Biener <rguenther@suse.de>
27211
27212 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 27213 invariant load do not generate a vector load from the scalar location.
0682ed3e 27214
b8f6a302
RB		 272152013-01-02 Richard Biener <rguenther@suse.de>
27216
27217 PR bootstrap/55784
27218 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27219 * configure: Regenerate.
27220
04b80dbb
RS		 272212013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27222
27223 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27224 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27225 (expand_builtin_int_roundingfn_2): Keep the original target around
27226 for the fallback case.
27227
635b0b0c
RS		 272282013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27229
27230 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27231 to be clear for sign changes.
27232
42d57399
JH		 272332013-01-01 Jan Hubicka <jh@suse.cz>
27234
27235 * ipa-inline-analysis.c: Fix formatting.
27236
5bb6669d
JJ		 272372013-01-01 Jakub Jelinek <jakub@redhat.com>
27238
27239 PR tree-optimization/55831
27240 * tree-vect-loop.c (get_initial_def_for_induction): Use
27241 gsi_after_labels instead of gsi_start_bb.
ad41bd84 27242\f
86a2db33 27243Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM		 27244
27245Copying and distribution of this file, with or without modification,
27246are permitted in any medium without royalty provided the copyright
27247notice and this notice are preserved.
Mirror of https://gcc.gnu.org/git/gcc.git