]> git.ipfire.org Git - thirdparty/gcc.git/blob - gcc/ChangeLog
re PR tree-optimization/58794 (ICE in set_lattice_value, at tree-ssa-ccp.c:455 on...
[thirdparty/gcc.git] / gcc / ChangeLog
1 2013-10-21 Richard Biener <rguenther@suse.de>
2
3 PR tree-optimization/58794
4 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
5 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
6
7 2013-10-21 Richard Biener <rguenther@suse.de>
8
9 PR middle-end/58742
10 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
11 to (T) X for sign-changing conversions (or no conversion).
12
13 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
14
15 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
16
17 2013-10-20 Jan Hubicka <jh@suse.cz>
18
19 * config/i386/i386-tune.def: Add comment; organize into categories
20
21 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
22
23 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
24 argument. Update function comment.
25 (expand_set_or_movmem_via_rep): New function combining
26 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
27 (expand_movmem_via_rep_mov): Remove.
28 expand_setmem_via_rep_stos): Remove.
29 (expand_movmem_epilogue): Update calls correspondingly.
30 (expand_setmem_epilogue_via_loop): Likewise.
31 (emit_memset): New.
32 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
33 (expand_set_or_movmem_prologue): New function combining
34 expand_movmem_prologue and expand_setmem_prologue.
35 (expand_movmem_prologue): Remove.
36 (expand_setmem_prologue): Remove.
37 (expand_set_or_movmem_constant_prologue): New function combining
38 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
39 (expand_constant_movmem_prologue): Remove.
40 (expand_constant_setmem_prologue): Remove.
41 (promote_duplicated_reg): Allow vector-const0 value.
42 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
43 and ix86_expand_setmem.
44 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
45 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
46
47 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
48
49 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
50 little endian.
51 (vec_unpacku_hi_v8hi): Likewise.
52 (vec_unpacku_lo_v16qi): Likewise.
53 (vec_unpacku_lo_v8hi): Likewise.
54
55 2013-10-20 Jan Hubicka <jh@suse.cz>
56
57 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
58 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
59 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
60
61 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
62
63 * config/mips/mips.h (ISA_HAS_WSBH): Define.
64 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
65 constants.
66 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
67
68 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
69
70 PR target/58603
71 * system.h: Undef m_slot.
72
73 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
74
75 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
76 all elements for both endian flavors.
77
78 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
79
80 PR target/58792
81 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
82 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
83 and SI_REG for 64bit SYSV ABI targets.
84
85 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
86
87 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
88 to multi_reg_return. Clarify that we are skipping USEs of multiple
89 return registers. Use bool type where appropriate.
90
91 2013-10-18 Jan Hubicka <jh@suse.cz>
92
93 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
94 for cold functions.
95 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
96 (X86_TUNE_PUSH_MEMORY): Likewise.
97 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
98 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
99 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
100 New.
101 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
102 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
103 Remove.
104 (ix86_option_override_internal): Update to use tune features instead
105 of variables.
106
107 2013-10-18 Cong Hou <congh@google.com>
108
109 PR tree-optimization/58508
110 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
111 statement that contains data refs with zero-step.
112
113 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
114
115 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
116 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
117 * gimple-low.c (gimple_check_call_arg,
118 gimple_check_call_matching_types): Move to cgraph.c.
119 * gimple-low.h: Remove prototype.
120 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
121 Relocate from gimple-low.c.
122 * cgraph.h: Add prototype. Don't include basic-block.h.
123 * gimplify.c: Add gimple-low to include list.
124 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
125 * tree-eh.c: Add gimple-low to include list.
126 * tree-nested.c: Likewise.
127 * cfgexpand.c: Add tree-ssa-address.h to include list.
128 * expr.c: Likewise.
129 * gimple-fold.c: Likewise.
130 * gimple-ssa-strength-reduction.c: Likewise.
131 * trans-mem.c: Likewise.
132 * tree-mudflap.c: Likewise.
133 * tree-ssa-loop-ivopts.c: Likewise.
134 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
135 (degenerate_phi_result): Move to tree-phinodes.c.
136 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
137 * tree-ssa-threadedge.c: Likewise.
138 * tree-vrp.c: Likewise.
139 * tree-phinodes.c (degenerate_phi_result): Relocate here.
140 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
141 * tree-phinodes.h (degenerate_phi_result): Add prototype.
142 * tree-ssa-copy.c: Include tree-ssa-dom.h.
143 * tree-ssa-forwprop.c: Likewise.
144 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
145 pass_data_cleanup_cfg_post_optimizing,
146 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
147 * tree-optimize.c: Delete File.
148 * graphite.c: Include tree-cfgcleanup.h.
149 * passes.c: Likewise.
150 * tree-cfg.c: Likewise.
151 * tree-profile.c: Likewise.
152 * tree-ssa-dse.c: Likewise.
153 * tree-ssa-loop-ivcanon.c: Likewise.
154 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
155 * tree-outof-ssa.c: Include sbitmap.h.
156 * tree-ssa-live.c: Likewise.
157 * tree-ssa-propagate.c: Likewise.
158 * tree-ssa-structalias.c: Likewise.
159 * tree-stdarg.c: Likewise.
160 * Makefile.in (OBJS): Delete tree-optimize.o.
161 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
162 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
163 * varasm.c: Include basic-block.h.
164 * cfgloop.h: Include function.h instead of basic-block.h
165 (bb_loop_depth): Move to cfgloop.c.
166 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
167
168 2013-10-18 Teresa Johnson <tejohnson@google.com>
169
170 * predict.c (probably_never_executed): Compare frequency-based
171 count to number of training runs.
172 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
173
174 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
175
176 * config/arm/arm.c (cortexa9_extra_costs): New table.
177 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
178
179 2013-10-18 Jeff Law <law@redhat.com>
180
181 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
182
183 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
184 (register_jump_thread): Add "registered_jump_thread" debug
185 counter support.
186 * dbgcnt.def (registered_jump_thread): New debug counter.
187
188 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
189
190 * config/rs6000/rs6000.c: Include cgraph.h.
191
192 2013-10-18 Teresa Johnson <tejohnson@google.com>
193
194 * tree-ssa-tail-merge.c (replace_block_by): Update edge
195 weights during merging.
196
197 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
198
199 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
200 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
201 * tree-ssa-operands.h: Remove prototype.
202 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
203 * gimple.c (virtual_operand_p): Relocate from gimple.c.
204 * gimple.h: Add prototype.
205 * gimple-ssa.h: Include tree-ssa-operands.h.
206 * tree-dump.c: Add tree-cfg.h to include list.
207 * tree-ssa-alias.c: Add ipa-reference.h to include list.
208 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
209 * config/i386/i386.c: Don't include tree-flow.h.
210 * config/rs6000/rs6000.c: Likewise.
211
212 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
213
214 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
215
216 2013-10-18 Richard Biener <rguenther@suse.de>
217
218 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
219 or TYPE_UNSIGNED of integral types.
220 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
221 NULL_TREE for zero-precision integral types.
222
223 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
224
225 * config/aarch64/arm_neon.h
226 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
227
228 2013-10-17 Sriraman Tallam <tmsriram@google.com>
229
230 * opth-gen.awk: Define target_flags_explicit.
231
232 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
233
234 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
235 fields to the reg_addr array that describes the valid addressing
236 mode for any register, general purpose registers, floating point
237 registers, and Altivec registers.
238 (FIRST_RELOAD_REG_CLASS): Likewise.
239 (LAST_RELOAD_REG_CLASS): Likewise.
240 (struct reload_reg_map_type): Likewise.
241 (reload_reg_map_type): Likewise.
242 (RELOAD_REG_VALID): Likewise.
243 (RELOAD_REG_MULTIPLE): Likewise.
244 (RELOAD_REG_INDEXED): Likewise.
245 (RELOAD_REG_OFFSET): Likewise.
246 (RELOAD_REG_PRE_INCDEC): Likewise.
247 (RELOAD_REG_PRE_MODIFY): Likewise.
248 (reg_addr): Likewise.
249 (mode_supports_pre_incdec_p): New helper functions to say whether
250 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
251 (mode_supports_pre_modify_p): Likewise.
252 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
253 print the valid address mode bits for each mode.
254 (rs6000_debug_print_mode): Likewise.
255 (rs6000_debug_reg_global): Likewise.
256 (rs6000_setup_reg_addr_masks): New function to set up the address
257 mask bits for each type.
258 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
259 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
260 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
261 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
262 PRE_MODIFY are supported.
263 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
264 registers, instead of {src,dest}_av_p.
265 (rs6000_print_options_internal): Tweak the debug output slightly.
266
267 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
268
269 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
270 isa attribute.
271
272 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
273
274 * tree-flow.h (struct omp_region): Move to omp-low.c.
275 Remove omp_ prototypes and variables.
276 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
277 (copy_var_decl): Relocate prototype from tree-flow.h.
278 * gimple.c (copy_var_decl): Relocate from omp-low.c.
279 * tree.h: Move prototype to omp-low.h.
280 * omp-low.h: New File. Relocate prototypes here.
281 * omp-low.c (struct omp_region): Make local here.
282 (root_omp_region): Make static.
283 (copy_var_decl) Move to gimple.c.
284 (new_omp_region): Make static.
285 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
286 * tree-cfg.c: Include omp-low.h.
287 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
288 * gimplify.c: Include omp-low.h.
289 * tree-parloops.c: Likewise.
290
291 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
292
293 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
294 operands are in memory, prefer to force non-matched operand 1 to
295 the register.
296
297 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
298
299 PR target/58673
300 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
301 restrict TImode addresses to single indirect registers if both
302 -mquad-memory and -mvsx-timode are used.
303 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
304 we should emit load/store quad. Remove using %y for quad memory
305 addresses.
306
307 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
308 constraints to allow load/store quad on machines where TImode is
309 not allowed in VSX registers. Use 'n' instead of 'F' constraint
310 for TImode to load integer constants.
311
312 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
313
314 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
315
316 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
317
318 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
319 handling of STACK_REG.
320
321 2013-10-17 Richard Biener <rguenther@suse.de>
322
323 PR tree-optimization/58143
324 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
325 New function.
326 (rewrite_to_defined_overflow): Likewise.
327 (move_computations_dom_walker::before_dom): Rewrite stmts
328 with undefined signed overflow that are not always executed
329 into unsigned arithmetic.
330
331 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
332
333 PR target/57756
334 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
335 explicit isa flag to be an options variable, instead of using
336 global_options_set. Remove define from rs6000.h.
337 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
338
339 * config/rs6000/rs6000.c (rs6000_option_override_internal):
340 Initialize rs6000_isa_flags_explicit.
341 (rs6000_function_specific_save): Add gcc_options* parameter, so
342 that the powerpc builds after the 2013-10-15 changes.
343 (rs6000_function_specific_restore): Likewise.
344
345 2013-10-16 DJ Delorie <dj@redhat.com>
346
347 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
348 op is a REG before checking REGNO.
349 (rl78_alloc_physical_registers): Verify pattern is a SET before
350 checking SET_SRC.
351
352 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
353
354 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
355 endianness.
356 (vec_unpacks_lo_v4sf): Likewise.
357 (vec_unpacks_float_hi_v4si): Likewise.
358 (vec_unpacks_float_lo_v4si): Likewise.
359 (vec_unpacku_float_hi_v4si): Likewise.
360 (vec_unpacku_float_lo_v4si): Likewise.
361
362 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
363
364 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
365 (vsx_concat_v2sf): Likewise.
366
367 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
368
369 * config/aarch64/aarch64.md
370 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
371
372 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
373
374 PR tree-optimization/58697
375 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
376 estimated_loop_iterations_int.
377 (max_stmt_executions_int): Call get_max_loop_iterations_int.
378 (get_max_loop_iterations_int): New. HWINT version of
379 get_max_loop_iterations.
380 * cfgloop.h: Add prototypes.
381 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
382 * loop-unroll.c (decide_peel_once_rolling): Call
383 get_estimated_loop_iterations_int.
384 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
385 * tree-ssa-loop-niter.h: Tweak prototypes.
386
387 2013-10-16 David Malcolm <dmalcolm@redhat.com>
388
389 * gengtype-parse.c (struct_field_seq): Ignore access-control
390 keywords ("public:" etc).
391
392 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
393
394 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
395 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
396
397 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
398
399 * config/arm/arm.opt (mlra): New option.
400 * config/arm/arm.c (arm_lra_p): New function.
401 (TARGET_LRA_P): Define.
402
403 2013-10-16 Paulo Matos <pmatos@broadcom.com>
404
405 * tree-core.h (tree_code_name): Remove.
406 * tree.h (get_tree_code_name): New prototype.
407 * tree.c (tree_code_name): Make static.
408 (get_tree_code_name): New function.
409 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
410 tree_class_check_failed, tree_range_check_failed,
411 tree_not_class_check_failed, omp_clause_check_failed,
412 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
413 wrapper get_tree_code_name instead of calling tree_code_name directly.
414 * tree-vrp.c (dump_asserts_for): Likewise.
415 * tree-dump.c (dequeue_and_dump): Likewise.
416 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
417 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
418 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
419 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
420 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
421 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
422 dump_gimple_omp_for): Likewise.
423 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
424 * tree-ssa-pre.c (print_pre_expr): Likewise.
425 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
426 * print-tree.c (print_node_brief, print_node): Likewise.
427 * gimple.c (gimple_check_failed): Likewise.
428 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
429 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
430 * config/mep/mep.c (mep_validate_vliw): Likewise.
431 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
432 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
433
434 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
435
436 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
437 for AMD bdver3.
438
439 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
440
441 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
442 (MULTILIB_MATCHES): Add multilib for -march=v8.
443
444 2013-10-15 Sriraman Tallam <tmsriram@google.com>
445
446 PR target/57756
447 * optc-save-gen.awk: Add extra parameter to the save and restore
448 target calls.
449 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
450 * tree.c (build_optimization_node): New parameter. Add extra parameter
451 to call to cl_optimization_save.
452 (build_target_option_node): New parameter. Add extra parameter
453 to call to cl_target_option_save.
454 * tree.h (build_optimization_node): New parameter.
455 (build_target_option_node): New parameter.
456 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
457 build_optimization_node and build_target_option_node.
458 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
459 (handle_pragma_push_options): Ditto.
460 * toplev.c (process_options): Ditto.
461 * opts.c (init_options_struct): Check for opts_set non-null.
462 * target.def (target_option.save): New parameter.
463 (target_option.restore): New parameter.
464 * tm.texi: Generate.
465 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
466 (ix86_pragma_target_parse): Ditto.
467 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
468 parameters.
469 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
470 to build_optimization_node and build_target_option_node.
471 (rs6000_valid_attribute_p): Ditto.
472 (rs6000_pragma_target_parse): Ditto.
473 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
474 data.
475 * config/i386/i386.h:
476 TARGET_64BIT_P: New Macro
477 TARGET_MMX_P: New Macro.
478 TARGET_3DNOW_P: New Macro.
479 TARGET_3DNOW_A_P: New Macro.
480 TARGET_SSE_P: New Macro.
481 TARGET_SSE2_P: New Macro.
482 TARGET_SSE3_P: New Macro.
483 TARGET_SSSE3_P: New Macro.
484 TARGET_SSE4_1_P: New Macro.
485 TARGET_SSE4_2_P: New Macro.
486 TARGET_AVX_P: New Macro.
487 TARGET_AVX2_P: New Macro.
488 TARGET_AVX512F_P: New Macro.
489 TARGET_AVX512PF_P: New Macro.
490 TARGET_AVX512ER_P: New Macro.
491 TARGET_AVX512CD_P: New Macro.
492 TARGET_FMA_P: New Macro.
493 TARGET_SSE4A_P: New Macro.
494 TARGET_FMA4_P: New Macro.
495 TARGET_XOP_P: New Macro.
496 TARGET_LWP_P: New Macro.
497 TARGET_ABM_P: New Macro.
498 TARGET_BMI_P: New Macro.
499 TARGET_BMI2_P: New Macro.
500 TARGET_LZCNT_P: New Macro.
501 TARGET_TBM_P: New Macro.
502 TARGET_POPCNT_P: New Macro.
503 TARGET_SAHF_P: New Macro.
504 TARGET_MOVBE_P: New Macro.
505 TARGET_CRC32_P: New Macro.
506 TARGET_AES_P: New Macro.
507 TARGET_PCLMUL_P: New Macro.
508 TARGET_CMPXCHG16B_P: New Macro.
509 TARGET_FSGSBASE_P: New Macro.
510 TARGET_RDRND_P: New Macro.
511 TARGET_F16C_P: New Macro.
512 TARGET_RTM_P: New Macro.
513 TARGET_HLE_P: New Macro.
514 TARGET_RDSEED_P: New Macro.
515 TARGET_PRFCHW_P: New Macro.
516 TARGET_ADX_P: New Macro.
517 TARGET_FXSR_P: New Macro.
518 TARGET_XSAVE_P: New Macro.
519 TARGET_XSAVEOPT_P: New Macro.
520 TARGET_LP64_P: New Macro.
521 TARGET_X32_P: New Macro.
522 TARGET_FPMATH_DEFAULT_P: New Macro.
523 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
524 * config/i386/i386.c (ix86_option_override_internal): New parameters.
525 opts and opts_set.
526 Change ix86_tune_string to access opts->x_ix86_tune_string.
527 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
528 Change ix86_arch_string to access opts->x_ix86_arch_string.
529 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
530 Change ix86_pmode to access opts->x_ix86_pmode.
531 Change ix86_abi to access opts->x_ix86_abi.
532 Change ix86_cmodel to access opts->x_ix86_cmodel.
533 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
534 Change ix86_isa_flags_explicit to access
535 opts->x_ix86_isa_flags_explicit.
536 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
537 Change ix86_regparm to access opts->x_ix86_regparm.
538 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
539 Change ix86_preferred_stack_boundary_arg to access
540 opts->x_ix86_preferred_stack_boundary_arg.
541 Change ix86_force_align_arg_pointer to access
542 opts->x_ix86_force_align_arg_pointer.
543 Change ix86_incoming_stack_boundar_arg to access
544 opts->x_ix86_incoming_stack_boundar_arg.
545 Change ix86_fpmath to access opts->x_ix86_fpmath.
546 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
547 Change ix86_recip_name to access opts->x_ix86_recip_name.
548 Change ix86_stack_protector_guard to access
549 opts->x_ix86_stack_protector_guard.
550 Change ix86_tune_memcpy_strategy to access
551 opts->x_ix86_tune_memcpy_strategy.
552 Change ix86_tune_memset_strategy to access
553 opts->x_ix86_tune_memset_strategy.
554 Change global_options to access opts.
555 Change global_options_set to access opts_set.
556 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
557 Change TARGET_MMX to TARGET_MMX_P (opts->...).
558 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
559 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
560 Change TARGET_SSE to TARGET_SSE_P (opts->...).
561 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
562 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
563 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
564 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
565 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
566 Change TARGET_AVX to TARGET_AVX_P (opts->...).
567 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
568 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
569 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
570 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
571 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
572 Change TARGET_FMA to TARGET_FMA_P (opts->...).
573 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
574 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
575 Change TARGET_XOP to TARGET_XOP_P (opts->...).
576 Change TARGET_LWP to TARGET_LWP_P (opts->...).
577 Change TARGET_ABM to TARGET_ABM_P (opts->...).
578 Change TARGET_BMI to TARGET_BMI_P (opts->...).
579 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
580 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
581 Change TARGET_TBM to TARGET_TBM_P (opts->...).
582 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
583 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
584 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
585 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
586 Change TARGET_AES to TARGET_AES_P (opts->...).
587 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
588 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
589 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
590 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
591 Change TARGET_F16C to TARGET_F16C_P (opts->...).
592 Change TARGET_RTM to TARGET_RTM_P (opts->...).
593 Change TARGET_HLE to TARGET_HLE_P (opts->...).
594 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
595 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
596 Change TARGET_ADX to TARGET_ADX_P (opts->...).
597 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
598 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
599 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
600 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
601 Change TARGET_X32 to TARGET_X32_P (opts->...).
602 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
603 Change TARGET_FLOAT_RETURNS_IN_80387 to
604 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
605 (ix86_function_specific_save): New parameter. Use opts-> fields
606 to replace global fields.
607 (ix86_function_specific_restore): Ditto.
608 (ix86_valid_target_attribute_inner_p): New parameters.
609 Fix recursive call.
610 Fix call to ix86_handle_option and set_option.
611 (ix86_valid_target_attribute_tree): New parameters.
612 Change global_options to access opts.
613 Change global_options_set to access opts_set.
614 Fix call to ix86_valid_target_attribute_inner_p.
615 Change ix86_tune_string to access opts->x_ix86_tune_string.
616 Change ix86_arch_string to access opts->x_ix86_arch_string.
617 Change ix86_fpmath to access opts->x_ix86_fpmath
618 Fix call to ix86_option_override_internal.
619 Fix call to ix86_add_new_builtins.
620 Fix calls to build_optimization_node and build_target_option_node.
621 (ix86_valid_target_attribute_p): Remove access to global_options.
622 Use new gcc_options structure func_options.
623 Fix call to ix86_valid_target_attribute_tree.
624 Fix call to build_optimization_node.
625 (get_builtin_code_for_version): Fix call to
626 ix86_valid_target_attribute_tree.
627
628 2013-10-15 David Malcolm <dmalcolm@redhat.com>
629
630 * Makefile.in (PICFLAG): New.
631 (enable_host_shared): New.
632 (INTERNAL_CFLAGS): Use PICFLAG.
633 (LIBIBERTY): Use pic build of libiberty.a if configured with
634 --enable-host-shared.
635 * configure.ac: Add --enable-host-shared, setting up new
636 PICFLAG variable.
637 * configure: Regenerate.
638 * doc/install.texi (--enable-shared): Add note contrasting it with ...
639 (--enable-host-shared): New option.
640
641 2013-10-15 Richard Biener <rguenther@suse.de>
642
643 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
644 for built-in functions.
645
646 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
647
648 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
649 (optimize_range_tests_1): New function,
650 extracted from optimize_range_tests.
651 (optimize_range_tests_xor): Similarly.
652 (optimize_range_tests_diff): New function.
653 (optimize_range_tests): Use optimize_range_tests_1.
654
655 2013-10-15 Cong Hou <congh@google.com>
656
657 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
658 requirement of the reduction pattern so that one operand of the
659 reduction operation can come from outside of the loop.
660
661 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
662
663 * config/arm/neon-schedgen.ml: Remove.
664 * config/arm/cortex-a9-neon.md: Remove comment regarding
665 neon-schedgen.ml.
666
667 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
668
669 * config/arm/types: Remove old neon types.
670
671 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
672
673 * config/arm/cortex-a7.md
674 (cortex_a7_neon_type): New.
675 (cortex_a7_neon_mul): Update for new types.
676 (cortex_a7_neon_mla): Likewise.
677 (cortex_a7_neon): Likewise.
678
679 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
680
681 * config/arm/cortex-a15-neon.md
682 (cortex_a15_neon_type): New,
683
684 (cortex_a15_neon_int_1): Remove.
685 (cortex_a15_neon_int_2): Likewise.
686 (cortex_a15_neon_int_3): Likewise.
687 (cortex_a15_neon_int_4): Likewise.
688 (cortex_a15_neon_int_5): Likewise.
689 (cortex_a15_neon_vqneg_vqabs): Likewise.
690 (cortex_a15_neon_vmov): Likewise.
691 (cortex_a15_neon_vaba): Likewise.
692 (cortex_a15_neon_vaba_qqq): Likewise.
693 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
694 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
695 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
696 Likewise.
697 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
698 (cortex_a15_neon_mla_qqq_8_16): Likewise.
699 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
700 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
701 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
702 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
703 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
704 (cortex_a15_neon_shift_1): Likewise.
705 (cortex_a15_neon_shift_2): Likewise.
706 (cortex_a15_neon_shift_3): Likewise.
707 (cortex_a15_neon_vshl_ddd): Likewise.
708 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
709 (cortex_a15_neon_vsra_vrsra): Likewise.
710 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
711 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
712 (cortex_a15_neon_bp_3cycle): Likewise.
713 (cortex_a15_neon_ldm_2): Likewise.
714 (cortex_a15_neon_stm_2): Likewise.
715 (cortex_a15_neon_mcr): Likewise.
716 (cortex_a15_neon_mrc): Likewise.
717 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
718 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
719 (cortex_a15_neon_fp_vmul_ddd): Likewise.
720 (cortex_a15_neon_fp_vmul_qqd): Likewise.
721 (cortex_a15_neon_fp_vmla_ddd): Likewise.
722 (cortex_a15_neon_fp_vmla_qqq): Likewise.
723 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
724 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
725 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
726 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
727 (cortex_a15_neon_bp_simple): Likewise.
728 (cortex_a15_neon_bp_2cycle): Likewise.
729 (cortex_a15_neon_bp_3cycle): Likewise.
730 (cortex_a15_neon_vld1_1_2_regs): Likewise.
731 (cortex_a15_neon_vld1_3_4_regs): Likewise.
732 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
733 (cortex_a15_neon_vld2_4_regs): Likewise.
734 (cortex_a15_neon_vld3_vld4): Likewise.
735 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
736 (cortex_a15_neon_vst1_3_4_regs): Likewise.
737 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
738 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
739 (cortex_a15_neon_vst3_vst4): Rename to...
740 (cortex_a15_neon_vst4): This, update for new attributes.
741 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
742 (cortex_a15_neon_vld3_vld4_lane): Likewise.
743 (cortex_a15_neon_vst1_vst2_lane): Likewise.
744 (cortex_a15_neon_vst3_vst4_lane): Likewise.
745 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
746 (cortex_a15_neon_ldm_2): Likewise.
747 (cortex_a15_neon_stm_2): Likewise.
748 (cortex_a15_neon_mcr): Likewise.
749 (cortex_a15_neon_mcr_2_mcrr): Likewise.
750 (cortex_a15_neon_mrc): Likewise.
751 (cortex_a15_neon_mrrc): Likewise.
752
753 (cortex_a15_neon_abd): New.
754 (cortex_a15_neon_abd_q): Likewise.
755 (cortex_a15_neon_aba): Likewise.
756 (cortex_a15_neon_aba_q): Likewise.
757 (cortex_a15_neon_acc): Likewise.
758 (cortex_a15_neon_acc_q): Likewise.
759 (cortex_a15_neon_arith_basic): Likewise.
760 (cortex_a15_neon_arith_complex): Likewise.
761 (cortex_a15_neon_multiply): Likewise.
762 (cortex_a15_neon_multiply_q): Likewise.
763 (cortex_a15_neon_mla): Likewise.
764 (cortex_a15_neon_mla_q): Likewise.
765 (cortex_a15_neon_sat_mla_long): Likewise.
766 (cortex_a15_neon_shift_acc): Likewise.
767 (cortex_a15_neon_shift_imm_basic): Likewise.
768 (cortex_a15_neon_shift_imm_complex): Likewise.
769 (cortex_a15_neon_shift_reg_basic): Likewise.
770 (cortex_a15_neon_shift_reg_basic_q): Likewise.
771 (cortex_a15_neon_shift_reg_complex): Likewise.
772 (cortex_a15_neon_shift_reg_complex_q): Likewise.
773 (cortex_a15_neon_fp_negabs): Likewise
774 (cortex_a15_neon_fp_arith): Likewise
775 (cortex_a15_neon_fp_arith_q): Likewise
776 (cortex_a15_neon_fp_cvt_int): Likewise
777 (cortex_a15_neon_fp_cvt_int_q): Likewise
778 (cortex_a15_neon_fp_cvt_16): Likewise
779 (cortex_a15_neon_fp_mul): Likewise
780 (cortex_a15_neon_fp_mul_q): Likewise
781 (cortex_a15_neon_fp_mla): Likewise
782 (cortex_a15_neon_fp_mla_q): Likewise
783 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
784 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
785 (cortex_a15_neon_bitops): Likewise.
786 (cortex_a15_neon_bitops_q): Likewise.
787 (cortex_a15_neon_from_gp): Likewise.
788 (cortex_a15_neon_from_gp_q): Likewise.
789 (cortex_a15_neon_tbl3_tbl4): Likewise.
790 (cortex_a15_neon_zip_q): Likewise.
791 (cortex_a15_neon_to_gp): Likewise.
792 (cortex_a15_neon_load_a): Likewise.
793 (cortex_a15_neon_load_b): Likewise.
794 (cortex_a15_neon_load_c): Likewise.
795 (cortex_a15_neon_load_d): Likewise.
796 (cortex_a15_neon_load_e): Likewise.
797 (cortex_a15_neon_load_f): Likewise.
798 (cortex_a15_neon_store_a): Likewise.
799 (cortex_a15_neon_store_b): Likewise.
800 (cortex_a15_neon_store_c): Likewise.
801 (cortex_a15_neon_store_d): Likewise.
802 (cortex_a15_neon_store_e): Likewise.
803 (cortex_a15_neon_store_f): Likewise.
804 (cortex_a15_neon_store_g): Likewise.
805 (cortex_a15_neon_store_h): Likewise.
806 (cortex_a15_vfp_to_from_gp): Likewise.
807
808 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
809
810 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
811
812 (cortex_a9_neon_vshl_ddd): Remove.
813 (cortex_a9_neon_vst3_vst4): Likewise.
814 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
815
816 (cortex_a9_neon_bit_ops_q): New.
817
818 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
819 (cortex_a9_neon_int_2): Likewise.
820 (cortex_a9_neon_int_3): Likewise.
821 (cortex_a9_neon_int_4): Likewise.
822 (cortex_a9_neon_int_5): Likewise.
823 (cortex_a9_neon_vqneg_vqabs): Likewise.
824 (cortex_a9_neon_vmov): Likewise.
825 (cortex_a9_neon_vaba): Likewise.
826 (cortex_a9_neon_vaba_qqq): Likewise.
827 (cortex_a9_neon_shift_1): Likewise.
828 (cortex_a9_neon_shift_2): Likewise.
829 (cortex_a9_neon_shift_3): Likewise.
830 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
831 (cortex_a9_neon_vsra_vrsra): Likewise.
832 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
833 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
834 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
835 Likewise.
836 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
837 (cortex_a9_neon_mla_qqq_8_16): Likewise.
838 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
839 Likewise.
840 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
841 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
842 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
843 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
844 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
845 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
846 (cortex_a9_neon_fp_vsum): Likewise.
847 (cortex_a9_neon_fp_vmul_ddd): Likewise.
848 (cortex_a9_neon_fp_vmul_qqd): Likewise.
849 (cortex_a9_neon_fp_vmla_ddd): Likewise.
850 (cortex_a9_neon_fp_vmla_qqq): Likewise.
851 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
852 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
853 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
854 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
855 (cortex_a9_neon_bp_simple): Likewise.
856 (cortex_a9_neon_bp_2cycle): Likewise.
857 (cortex_a9_neon_bp_3cycle): Likewise.
858 (cortex_a9_neon_ldr): Likewise.
859 (cortex_a9_neon_str): Likewise.
860 (cortex_a9_neon_vld1_1_2_regs): Likewise.
861 (cortex_a9_neon_vld1_3_4_regs): Likewise.
862 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
863 (cortex_a9_neon_vld2_4_regs): Likewise.
864 (cortex_a9_neon_vld3_vld4): Likewise.
865 (cortex_a9_neon_vld1_vld2_lane): Likewise.
866 (cortex_a9_neon_vld3_vld4_lane): Likewise.
867 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
868 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
869 (cortex_a9_neon_vst1_3_4_regs): Likewise.
870 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
871 (cortex_a9_neon_vst1_vst2_lane): Likewise.
872 (cortex_a9_neon_vst3_vst4_lane): Likewise.
873 (cortex_a9_neon_mcr): Likewise.
874 (cortex_a9_neon_mcr_2_mcrr): Likewise.
875 (cortex_a9_neon_mrc): Likewise.
876 (cortex_a9_neon_mrrc): Likewise.
877
878 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
879
880 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
881
882 (cortex_a8_neon_vshl_ddd): Remove.
883 (cortex_a8_neon_vst3_vst4): Likewise.
884 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
885
886 (cortex_a8_neon_bit_ops_q): New.
887
888 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
889 (cortex_a8_neon_int_2): Likewise..
890 (cortex_a8_neon_int_3): Likewise.
891 (cortex_a8_neon_int_5): Likewise.
892 (cortex_a8_neon_vqneg_vqabs): Likewise.
893 (cortex_a8_neon_int_4): Likewise.
894 (cortex_a8_neon_vaba): Likewise.
895 (cortex_a8_neon_vaba_qqq): Likewise.
896 (cortex_a8_neon_shift_1): Likewise.
897 (cortex_a8_neon_shift_2): Likewise.
898 (cortex_a8_neon_shift_3): Likewise.
899 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
900 (cortex_a8_neon_vsra_vrsra): Likewise.
901 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
902 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
903 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
904 Likewise.
905 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
906 (cortex_a8_neon_mla_qqq_8_16): Likewise.
907 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
908 Likewise.
909 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
910 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
911 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
912 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
913 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
914 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
915 (cortex_a8_neon_fp_vsum): Likewise.
916 (cortex_a8_neon_fp_vmul_ddd): Likewise.
917 (cortex_a8_neon_fp_vmul_qqd): Likewise.
918 (cortex_a8_neon_fp_vmla_ddd): Likewise.
919 (cortex_a8_neon_fp_vmla_qqq): Likewise.
920 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
921 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
922 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
923 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
924 (cortex_a8_neon_bp_simple): Likewise.
925 (cortex_a8_neon_bp_2cycle): Likewise.
926 (cortex_a8_neon_bp_3cycle): Likewise.
927 (cortex_a8_neon_ldr): Likewise.
928 (cortex_a8_neon_str): Likewise.
929 (cortex_a8_neon_vld1_1_2_regs): Likewise.
930 (cortex_a8_neon_vld1_3_4_regs): Likewise.
931 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
932 (cortex_a8_neon_vld2_4_regs): Likewise.
933 (cortex_a8_neon_vld3_vld4): Likewise.
934 (cortex_a8_neon_vld1_vld2_lane): Likewise.
935 (cortex_a8_neon_vld3_vld4_lane): Likewise.
936 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
937 (cortex_a8_neon_vst1_3_4_regs): Likewise.
938 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
939 (cortex_a8_neon_vst1_vst2_lane): Likewise.
940 (cortex_a8_neon_vst3_vst4_lane): Likewise.
941 (cortex_a8_neon_mcr): Likewise.
942 (cortex_a8_neon_mcr_2_mcrr): Likewise.
943 (cortex_a8_neon_mrc): Likewise.
944 (cortex_a8_neon_mrrc): Likewise.
945
946 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
947
948 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
949 (fp): New.
950 * config/aarch64/aarch64-simd.md (neon_type): Remove.
951 (aarch64_simd_dup<mode>): Add "type" attribute.
952 (aarch64_dup_lane<mode>): Likewise.
953 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
954 (*aarch64_simd_mov<mode>): Likewise.
955 (aarch64_simd_mov_from_<mode>low): Likewise.
956 (aarch64_simd_mov_from_<mode>high): Likewise.
957 (orn<mode>3): Likewise.
958 (bic<mode>3): Likewise.
959 (add<mode>3): Likewise.
960 (sub<mode>3): Likewise.
961 (mul<mode>3): Likewise.
962 (*aarch64_mul3_elt<mode>): Likewise.
963 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
964 (*aarch64_mul3_elt_to_128df): Likewise.
965 (*aarch64_mul3_elt_to_64v2df): Likewise.
966 (neg<mode>2): Likewise.
967 (abs<mode>2): Likewise.
968 (abd<mode>_3): Likewise.
969 (aba<mode>_3): Likewise.
970 (fabd<mode>_3): Likewise.
971 (*fabd_scalar<mode>3): Likewise.
972 (and<mode>3): Likewise.
973 (ior<mode>3): Likewise.
974 (xor<mode>3): Likewise.
975 (one_cmpl<mode>2): Likewise.
976 (aarch64_simd_vec_set<mode>): Likewise.
977 (aarch64_simd_lshr<mode>): Likewise.
978 (aarch64_simd_ashr<mode>): Likewise.
979 (aarch64_simd_imm_shl<mode>): Likewise.
980 (aarch64_simd_reg_sshl<mode): Likewise.
981 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
982 (aarch64_simd_reg_shl<mode>_signed): Likewise.
983 (aarch64_simd_vec_setv2di): Likewise.
984 (aarch64_simd_vec_set<mode>): Likewise.
985 (aarch64_mla<mode>): Likewise.
986 (*aarch64_mla_elt<mode>): Likewise.
987 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
988 (aarch64_mls<mode>): Likewise.
989 (*aarch64_mls_elt<mode>): Likewise.
990 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
991 (<su><maxmin><mode>3): Likewise.
992 (move_lo_quad_<mode>): Likewise.
993 (aarch64_simd_move_hi_quad_<mode>): Likewise.
994 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
995 (vec_pack_trunc_<mode>): Likewise.
996 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
997 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
998 (*aarch64_<su>mlal_lo<mode>): Likewise.
999 (*aarch64_<su>mlal_hi<mode>): Likewise.
1000 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1001 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1002 (*aarch64_<su>mlal<mode>): Likewise.
1003 (*aarch64_<su>mlsl<mode>): Likewise.
1004 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1005 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1006 (add<mode>3): Likewise.
1007 (sub<mode>3): Likewise.
1008 (mul<mode>3): Likewise.
1009 (div<mode>3): Likewise.
1010 (neg<mode>2): Likewise.
1011 (abs<mode>2): Likewise.
1012 (fma<mode>4): Likewise.
1013 (*aarch64_fma4_elt<mode>): Likewise.
1014 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1015 (*aarch64_fma4_elt_to_128df): Likewise.
1016 (*aarch64_fma4_elt_to_64v2df): Likewise.
1017 (fnma<mode>4): Likewise.
1018 (*aarch64_fnma4_elt<mode>): Likewise.
1019 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1020 (*aarch64_fnma4_elt_to_128df): Likewise.
1021 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1022 (<frint_pattern><mode>2): Likewise.
1023 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1024 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1025 (vec_unpacks_lo_v4sf): Likewise.
1026 (aarch64_float_extend_lo_v2df): Likewise.
1027 (vec_unpacks_hi_v4sf): Likewise.
1028 (aarch64_float_truncate_lo_v2sf): Likewise.
1029 (aarch64_float_truncate_hi_v4sf): Likewise.
1030 (aarch64_vmls<mode>): Likewise.
1031 (<su><maxmin><mode>3): Likewise.
1032 (<maxmin_uns><mode>3): Likewise.
1033 (reduc_<sur>plus_<mode>): Likewise.
1034 (reduc_<sur>plus_v2di): Likewise.
1035 (reduc_<sur>plus_v2si): Likewise.
1036 (reduc_<sur>plus_<mode>): Likewise.
1037 (aarch64_addpv4sf): Likewise.
1038 (clz<mode>2): Likewise.
1039 (reduc_<maxmin_uns>_<mode>): Likewise.
1040 (reduc_<maxmin_uns>_v2di): Likewise.
1041 (reduc_<maxmin_uns>_v2si): Likewise.
1042 (reduc_<maxmin_uns>_<mode>): Likewise.
1043 (reduc_<maxmin_uns>_v4sf): Likewise.
1044 (aarch64_simd_bsl<mode>_internal): Likewise.
1045 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1046 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1047 (aarch64_get_lane<mode>): Likewise.
1048 (*aarch64_combinez<mode>): Likewise.
1049 (aarch64_combine<mode>): Likewise.
1050 (aarch64_simd_combine<mode>): Likewise.
1051 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1052 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1053 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1054 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1055 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1056 (aarch64_<sur>h<addsub><mode>): Likewise.
1057 (aarch64_<sur><addsub>hn<mode>): Likewise.
1058 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1059 (aarch64_pmul<mode>): Likewise.
1060 (aarch64_<su_optab><optab><mode>): Likewise.
1061 (aarch64_<sur>qadd<mode>): Likewise.
1062 (aarch64_sqmovun<mode>): Likewise.
1063 (aarch64_<sur>qmovn<mode>): Likewise.
1064 (aarch64_s<optab><mode>): Likewise.
1065 (aarch64_sq<r>dmulh<mode>): Likewise.
1066 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1067 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1068 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1069 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1070 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1071 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1072 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1073 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1074 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1075 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1076 (aarch64_sqdmull<mode>): Likewise.
1077 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1078 (aarch64_sqdmull_n<mode>): Likewise.
1079 (aarch64_sqdmull2<mode>_internal): Likewise.
1080 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1081 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1082 (aarch64_<sur>shl<mode>): Likewise.
1083 (aarch64_<sur>q<r>shl<mode>
1084 (aarch64_<sur>shll_n<mode>): Likewise.
1085 (aarch64_<sur>shll2_n<mode>): Likewise.
1086 (aarch64_<sur>shr_n<mode>): Likewise.
1087 (aarch64_<sur>sra_n<mode>): Likewise.
1088 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1089 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1090 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1091 (aarch64_cm<optab><mode>): Likewise.
1092 (aarch64_cm<optab>di): Likewise.
1093 (aarch64_cm<optab><mode>): Likewise.
1094 (aarch64_cm<optab>di): Likewise.
1095 (aarch64_cmtst<mode>): Likewise.
1096 (aarch64_cmtstdi): Likewise.
1097 (aarch64_cm<optab><mode>): Likewise.
1098 (*aarch64_fac<optab><mode>): Likewise.
1099 (aarch64_addp<mode>): Likewise.
1100 (aarch64_addpdi): Likewise.
1101 (sqrt<mode>2): Likewise.
1102 (vec_load_lanesoi<mode>): Likewise.
1103 (vec_store_lanesoi<mode>): Likewise.
1104 (vec_load_lanesci<mode>): Likewise.
1105 (vec_store_lanesci<mode>): Likewise.
1106 (vec_load_lanesxi<mode>): Likewise.
1107 (vec_store_lanesxi<mode>): Likewise.
1108 (*aarch64_mov<mode>): Likewise.
1109 (aarch64_ld2<mode>_dreg): Likewise.
1110 (aarch64_ld2<mode>_dreg): Likewise.
1111 (aarch64_ld3<mode>_dreg): Likewise.
1112 (aarch64_ld3<mode>_dreg): Likewise.
1113 (aarch64_ld4<mode>_dreg): Likewise.
1114 (aarch64_ld4<mode>_dreg): Likewise.
1115 (aarch64_tbl1<mode>): Likewise.
1116 (aarch64_tbl2v16qi): Likewise.
1117 (aarch64_combinev16qi): Likewise.
1118 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1119 (aarch64_st2<mode>_dreg): Likewise.
1120 (aarch64_st2<mode>_dreg): Likewise.
1121 (aarch64_st3<mode>_dreg): Likewise.
1122 (aarch64_st3<mode>_dreg): Likewise.
1123 (aarch64_st4<mode>_dreg): Likewise.
1124 (aarch64_st4<mode>_dreg): Likewise.
1125 (*aarch64_simd_ld1r<mode>): Likewise.
1126 (aarch64_frecpe<mode>): Likewise.
1127 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1128 (aarch64_frecps<mode>): Likewise.
1129
1130 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1131
1132 * config/arm/iterators.md (V_elem_ch): New.
1133 (q): Likewise.
1134 (VQH_type): Likewise.
1135 * config/arm/arm.md (is_neon_type): New.
1136 (conds): Use is_neon_type.
1137 (anddi3_insn): Update type attribute.
1138 (xordi3_insn): Likewise.
1139 (one_cmpldi2): Likewise.
1140 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1141 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1142 (*movmisalign<mode>_neon_store): Likewise.
1143 (*movmisalign<mode>_neon_load): Likewise.
1144 (vec_set<mode>_internal): Likewise.
1145 (vec_set<mode>_internal): Likewise.
1146 (vec_setv2di_internal): Likewise.
1147 (vec_extract<mode>): Likewise.
1148 (vec_extract<mode>): Likewise.
1149 (vec_extractv2di): Likewise.
1150 (*add<mode>3_neon): Likewise.
1151 (adddi3_neon): Likewise.
1152 (*sub<mode>3_neon): Likewise.
1153 (subdi3_neon): Likewise.
1154 (fma<VCVTF:mode>4): Likewise.
1155 (fma<VCVTF:mode>4_intrinsic): Likewise.
1156 (*fmsub<VCVTF:mode>4): Likewise.
1157 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1158 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1159 (ior<mode>3): Likewise.
1160 (and<mode>3): Likewise.
1161 (orn<mode>3_neon): Likewise.
1162 (orndi3_neon): Likewise.
1163 (bic<mode>3_neon): Likewise.
1164 (bicdi3_neon): Likewise.
1165 (xor<mode>3): Likewise.
1166 (one_cmpl<mode>2): Likewise.
1167 (abs<mode>2): Likewise.
1168 (neg<mode>2): Likewise.
1169 (negdi2_neon): Likewise.
1170 (*umin<mode>3_neon): Likewise.
1171 (*umax<mode>3_neon): Likewise.
1172 (*smin<mode>3_neon): Likewise.
1173 (*smax<mode>3_neon): Likewise.
1174 (vashl<mode>3): Likewise.
1175 (vashr<mode>3_imm): Likewise.
1176 (vlshr<mode>3_imm): Likewise.
1177 (ashl<mode>3_signed): Likewise.
1178 (ashl<mode>3_unsigned): Likewise.
1179 (neon_load_count): Likewise.
1180 (ashldi3_neon_noclobber): Likewise.
1181 (ashldi3_neon): Likewise.
1182 (signed_shift_di3_neon): Likewise.
1183 (unsigned_shift_di3_neon): Likewise.
1184 (ashrdi3_neon_imm_noclobber): Likewise.
1185 (lshrdi3_neon_imm_noclobber): Likewise.
1186 (<shift>di3_neon): Likewise.
1187 (widen_ssum<mode>3): Likewise.
1188 (widen_usum<mode>3): Likewise.
1189 (quad_halves_<code>v4si): Likewise.
1190 (quad_halves_<code>v4sf): Likewise.
1191 (quad_halves_<code>v8hi): Likewise.
1192 (quad_halves_<code>v16qi): Likewise.
1193 (reduc_splus_v2di): Likewise.
1194 (neon_vpadd_internal<mode>): Likewise.
1195 (neon_vpsmin<mode>): Likewise.
1196 (neon_vpsmax<mode>): Likewise.
1197 (neon_vpumin<mode>): Likewise.
1198 (neon_vpumax<mode>): Likewise.
1199 (*ss_add<mode>_neon): Likewise.
1200 (*us_add<mode>_neon): Likewise.
1201 (*ss_sub<mode>_neon): Likewise.
1202 (*us_sub<mode>_neon): Likewise.
1203 (neon_vadd<mode>_unspec): Likewise.
1204 (neon_vaddl<mode>): Likewise.
1205 (neon_vaddw<mode>): Likewise.
1206 (neon_vhadd<mode>): Likewise.
1207 (neon_vqadd<mode>): Likewise.
1208 (neon_vaddhn<mode>): Likewise.
1209 (neon_vmul<mode>): Likewise.
1210 (neon_vfms<VCVTF:mode>): Likewise.
1211 (neon_vmlal<mode>): Likewise.
1212 (neon_vmls<mode>): Likewise.
1213 (neon_vmlsl<mode>): Likewise.
1214 (neon_vqdmulh<mode>): Likewise.
1215 (neon_vqdmlal<mode>): Likewise.
1216 (neon_vqdmlsl<mode>): Likewise.
1217 (neon_vmull<mode>): Likewise.
1218 (neon_vqdmull<mode>): Likewise.
1219 (neon_vsub<mode>_unspec): Likewise.
1220 (neon_vsubl<mode>): Likewise.
1221 (neon_vsubw<mode>): Likewise.
1222 (neon_vqsub<mode>): Likewise.
1223 (neon_vhsub<mode>): Likewise.
1224 (neon_vsubhn<mode>): Likewise.
1225 (neon_vceq<mode>): Likewise.
1226 (neon_vcge<mode>): Likewise.
1227 (neon_vcgeu<mode>): Likewise.
1228 (neon_vcgt<mode>): Likewise.
1229 (neon_vcgtu<mode>): Likewise.
1230 (neon_vcle<mode>): Likewise.
1231 (neon_vclt<mode>): Likewise.
1232 (neon_vcage<mode>): Likewise.
1233 (neon_vcagt<mode>): Likewise.
1234 (neon_vtst<mode>): Likewise.
1235 (neon_vabd<mode>): Likewise.
1236 (neon_vabdl<mode>): Likewise.
1237 (neon_vaba<mode>): Likewise.
1238 (neon_vabal<mode>): Likewise.
1239 (neon_vmax<mode>): Likewise.
1240 (neon_vmin<mode>): Likewise.
1241 (neon_vpaddl<mode>): Likewise.
1242 (neon_vpadal<mode>): Likewise.
1243 (neon_vpmax<mode>): Likewise.
1244 (neon_vpmin<mode>): Likewise.
1245 (neon_vrecps<mode>): Likewise.
1246 (neon_vrsqrts<mode>): Likewise.
1247 (neon_vqabs<mode>): Likewise.
1248 (neon_vqneg<mode>): Likewise.
1249 (neon_vcls<mode>): Likewise.
1250 (clz<mode>2): Likewise.
1251 (popcount<mode>2): Likewise.
1252 (neon_vrecpe<mode>): Likewise.
1253 (neon_vrsqrte<mode>): Likewise.
1254 (neon_vget_lane<mode>_sext_internal): Likewise.
1255 (neon_vget_lane<mode>_zext_internal): Likewise.
1256 (neon_vdup_n<mode>): Likewise.
1257 (neon_vdup_n<mode>): Likewise.
1258 (neon_vdup_nv2di): Likewise.
1259 (neon_vdup_lane<mode>_interal): Likewise.
1260 (*neon_vswp<mode>): Likewise.
1261 (neon_vcombine<mode>): Likewise.
1262 (float<mode><V_cvtto>2): Likewise.
1263 (floatuns<mode><V_cvtto>2): Likewise.
1264 (fix_trunc<mode><V_cvtto>2): Likewise.
1265 (fixuns_trunc<mode><V_cvtto>2
1266 (neon_vcvt<mode>): Likewise.
1267 (neon_vcvt<mode>): Likewise.
1268 (neon_vcvtv4sfv4hf): Likewise.
1269 (neon_vcvtv4hfv4sf): Likewise.
1270 (neon_vcvt_n<mode>): Likewise.
1271 (neon_vcvt_n<mode>): Likewise.
1272 (neon_vmovn<mode>): Likewise.
1273 (neon_vqmovn<mode>): Likewise.
1274 (neon_vqmovun<mode>): Likewise.
1275 (neon_vmovl<mode>): Likewise.
1276 (neon_vmul_lane<mode>): Likewise.
1277 (neon_vmul_lane<mode>): Likewise.
1278 (neon_vmull_lane<mode>): Likewise.
1279 (neon_vqdmull_lane<mode>): Likewise.
1280 (neon_vqdmulh_lane<mode>): Likewise.
1281 (neon_vqdmulh_lane<mode>): Likewise.
1282 (neon_vmla_lane<mode>): Likewise.
1283 (neon_vmla_lane<mode>): Likewise.
1284 (neon_vmlal_lane<mode>): Likewise.
1285 (neon_vqdmlal_lane<mode>): Likewise.
1286 (neon_vmls_lane<mode>): Likewise.
1287 (neon_vmls_lane<mode>): Likewise.
1288 (neon_vmlsl_lane<mode>): Likewise.
1289 (neon_vqdmlsl_lane<mode>): Likewise.
1290 (neon_vext<mode>): Likewise.
1291 (neon_vrev64<mode>): Likewise.
1292 (neon_vrev32<mode>): Likewise.
1293 (neon_vrev16<mode>): Likewise.
1294 (neon_vbsl<mode>_internal): Likewise.
1295 (neon_vshl<mode>): Likewise.
1296 (neon_vqshl<mode>): Likewise.
1297 (neon_vshr_n<mode>): Likewise.
1298 (neon_vshrn_n<mode>): Likewise.
1299 (neon_vqshrn_n<mode>): Likewise.
1300 (neon_vqshrun_n<mode>): Likewise.
1301 (neon_vshl_n<mode>): Likewise.
1302 (neon_vqshl_n<mode>): Likewise.
1303 (neon_vqshlu_n<mode>): Likewise.
1304 (neon_vshll_n<mode>): Likewise.
1305 (neon_vsra_n<mode>): Likewise.
1306 (neon_vsri_n<mode>): Likewise.
1307 (neon_vsli_n<mode>): Likewise.
1308 (neon_vtbl1v8qi): Likewise.
1309 (neon_vtbl2v8qi): Likewise.
1310 (neon_vtbl3v8qi): Likewise.
1311 (neon_vtbl4v8qi): Likewise.
1312 (neon_vtbl1v16qi): Likewise.
1313 (neon_vtbl2v16qi): Likewise.
1314 (neon_vcombinev16qi): Likewise.
1315 (neon_vtbx1v8qi): Likewise.
1316 (neon_vtbx2v8qi): Likewise.
1317 (neon_vtbx3v8qi): Likewise.
1318 (neon_vtbx4v8qi): Likewise.
1319 (*neon_vtrn<mode>_insn): Likewise.
1320 (*neon_vzip<mode>_insn): Likewise.
1321 (*neon_vuzp<mode>_insn): Likewise.
1322 (neon_vld1<mode>): Likewise.
1323 (neon_vld1_lane<mode>): Likewise.
1324 (neon_vld1_lane<mode>): Likewise.
1325 (neon_vld1_dup<mode>): Likewise.
1326 (neon_vld1_dup<mode>): Likewise.
1327 (neon_vld1_dupv2di): Likewise.
1328 (neon_vst1<mode>): Likewise.
1329 (neon_vst1_lane<mode>): Likewise.
1330 (neon_vst1_lane<mode>): Likewise.
1331 (neon_vld2<mode>): Likewise.
1332 (neon_vld2<mode>): Likewise.
1333 (neon_vld2_lane<mode>): Likewise.
1334 (neon_vld2_lane<mode>): Likewise.
1335 (neon_vld2_dup<mode>): Likewise.
1336 (neon_vst2<mode>): Likewise.
1337 (neon_vst2<mode>): Likewise.
1338 (neon_vst2_lane<mode>): Likewise.
1339 (neon_vst2_lane<mode>): Likewise.
1340 (neon_vld3<mode>): Likewise.
1341 (neon_vld3qa<mode>): Likewise.
1342 (neon_vld3qb<mode>): Likewise.
1343 (neon_vld3_lane<mode>): Likewise.
1344 (neon_vld3_lane<mode>): Likewise.
1345 (neon_vld3_dup<mode>): Likewise.
1346 (neon_vst3<mode>): Likewise.
1347 (neon_vst3qa<mode>): Likewise.
1348 (neon_vst3qb<mode>): Likewise.
1349 (neon_vst3_lane<mode>): Likewise.
1350 (neon_vst3_lane<mode>): Likewise.
1351 (neon_vld4<mode>): Likewise.
1352 (neon_vld4qa<mode>): Likewise.
1353 (neon_vld4qb<mode>): Likewise.
1354 (neon_vld4_lane<mode>): Likewise.
1355 (neon_vld4_lane<mode>): Likewise.
1356 (neon_vld4_dup<mode>): Likewise.
1357 (neon_vst4<mode>): Likewise.
1358 (neon_vst4qa<mode>): Likewise.
1359 (neon_vst4qb<mode>): Likewise.
1360 (neon_vst4_lane<mode>): Likewise.
1361 (neon_vst4_lane<mode>): Likewise.
1362 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1363 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1364 (neon_vec_<US>mult_lo_<mode>): Likewise.
1365 (neon_vec_<US>mult_hi_<mode>): Likewise.
1366 (neon_vec_<US>shiftl_<mode>): Likewise.
1367 (neon_unpack<US>_<mode>): Likewise.
1368 (neon_vec_<US>mult_<mode>): Likewise.
1369 (vec_pack_trunc_<mode>): Likewise.
1370 (neon_vec_pack_trunc_<mode>): Likewise.
1371 (neon_vabd<mode>_2): Likewise.
1372 (neon_vabd<mode>_3): Likewise.
1373
1374 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1375
1376 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1377 (load_pair): Update type attribute.
1378 (store_pair): Update type attribute.
1379 * config/aarch64/iterators.md (q): New.
1380
1381 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1382
1383 * config/arm/types.md: Add new types for Neon insns.
1384
1385 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1386 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1387 Sergey Lega <sergey.s.lega@intel.com>
1388 Anna Tikhonova <anna.tikhonova@intel.com>
1389 Ilya Tocar <ilya.tocar@intel.com>
1390 Andrey Turetskiy <andrey.turetskiy@intel.com>
1391 Ilya Verbin <ilya.verbin@intel.com>
1392 Kirill Yukhin <kirill.yukhin@intel.com>
1393 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1394
1395 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1396 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1397 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1398 (rcp14<mode>): New.
1399 (srcp14<mode>): Ditto.
1400 (rsqrt14<mode>): Ditto.
1401 (rsqrt14<mode>): Ditto.
1402 (avx512f_vmscalef<mode>): Ditto.
1403 (avx512f_scalef<mode>): Ditto.
1404 (avx512f_getexp<mode>): Ditto.
1405 (avx512f_sgetexp<mode>): Ditto.
1406 (avx512f_fixupimm<mode>): Ditto.
1407 (avx512f_sfixupimm<mode>): Ditto.
1408 (avx512f_rndscale<mode>): Ditto.
1409 (*avx512er_exp2<mode>): Ditto.
1410 (*avx512er_rcp28<mode>): Ditto.
1411 (avx512er_rsqrt28<mode>): Ditto.
1412 (avx512f_getmant<mode>): Ditto.
1413 (avx512f_getmant<mode>): Ditto.
1414 (avx512f_rndscale<mode>): Fix formatting.
1415
1416 2013-10-15 Martin Jambor <mjambor@suse.cz>
1417
1418 * ipa-utils.h (ipa_edge_within_scc): Declare.
1419 * ipa-cp.c (edge_within_scc): Moved...
1420 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1421
1422 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1423 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1424 Sergey Lega <sergey.s.lega@intel.com>
1425 Anna Tikhonova <anna.tikhonova@intel.com>
1426 Ilya Tocar <ilya.tocar@intel.com>
1427 Andrey Turetskiy <andrey.turetskiy@intel.com>
1428 Ilya Verbin <ilya.verbin@intel.com>
1429 Kirill Yukhin <kirill.yukhin@intel.com>
1430 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1431
1432 * config/i386/predicates.md (const_8_to_15_operand): New.
1433 (const_16_to_31_operand): Ditto.
1434 * config/i386/sse.md (V8FI): New.
1435 (V16FI): Ditto.
1436 (reduc_splus_v8df): Ditto.
1437 (reduc_splus_v16sf): Ditto.
1438 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1439 (vec_extract_hi_<mode>): Ditto.
1440 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1441 (vec_set_lo_<mode>): Ditto.
1442 (vec_set_hi_<mode>): Ditto.
1443 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1444 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1445 (avx512f_pshufd_1): Ditto.
1446 (avx512f_broadcast<mode>): Ditto.
1447 (avx512f_broadcast<mode>): Ditto.
1448 (define_split): Split vec_extract_lo into move.
1449 (ssequartermode): Ditto.
1450 (ssedoublemode): Extened with wider modes.
1451 (vec_extract_lo_<mode>): Ditto.
1452
1453 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1454 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1455 Sergey Lega <sergey.s.lega@intel.com>
1456 Anna Tikhonova <anna.tikhonova@intel.com>
1457 Ilya Tocar <ilya.tocar@intel.com>
1458 Andrey Turetskiy <andrey.turetskiy@intel.com>
1459 Ilya Verbin <ilya.verbin@intel.com>
1460 Kirill Yukhin <kirill.yukhin@intel.com>
1461 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1462
1463 * config/i386/predicates.md (register_or_constm1_operand): New.
1464 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1465 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1466 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1467 UNSPEC_SCATTER_PREFETCH
1468 (VI48_512): New.
1469 (avx512f_ucmp<mode>3): Ditto.
1470 (avx512f_vternlog<mode>): Ditto.
1471 (avx512f_align<mode>): Ditto.
1472 (<shift_insn><mode>3): Ditto.
1473 (avx512f_<rotate>v<mode>): Ditto.
1474 (avx512f_<rotate><mode>): Ditto.
1475 (avx512f_eq<mode>3): Ditto.
1476 (avx512f_eq<mode>3_1): Ditto.
1477 (avx512f_gt<mode>3): Ditto.
1478 (avx512f_testm<mode>3): Ditto.
1479 (avx512f_testnm<mode>3): Ditto.
1480 (avx512pf_gatherpf<mode>): Ditto.
1481 (*avx512pf_gatherpf<mode>_mask): Ditto.
1482 (*avx512pf_gatherpf<mode>): Ditto.
1483 (avx512pf_scatterpf<mode>): Ditto.
1484 (*avx512pf_scatterpf<mode>_mask): Ditto.
1485 (*avx512pf_scatterpf<mode>): Ditto.
1486 (avx512f_vec_dup_gpr<mode>): Ditto.
1487 (clz<mode>2): Ditto.
1488 (conflict<mode>): Ditto.
1489 (REDUC_SMINMAX_MODE): Extened with wider modes.
1490 (reduc_<code>_<mode>): Ditto.
1491 (vlshr<mode>3): Ditto.
1492 (vashl<mode>3): Ditto.
1493
1494 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1495 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1496 Sergey Lega <sergey.s.lega@intel.com>
1497 Anna Tikhonova <anna.tikhonova@intel.com>
1498 Ilya Tocar <ilya.tocar@intel.com>
1499 Andrey Turetskiy <andrey.turetskiy@intel.com>
1500 Ilya Verbin <ilya.verbin@intel.com>
1501 Kirill Yukhin <kirill.yukhin@intel.com>
1502 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1503
1504 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1505 UNSPEC_SCATTER.
1506 (VI48F_512): New.
1507 (avx512fmaskmode): Ditto.
1508 (bcstscalarsuff): Ditto.
1509 (avx512f_blendm<mode>): Ditto.
1510 (cmp_imm_predicate): Ditto.
1511 (avx512f_cmp<mode>3): Ditto.
1512 (avx512f_vec_dup<mode>): Ditto.
1513 (avx512f_vec_dup_mem<mode>): Ditto.
1514 (avx512f_vpermi2var<mode>3): Ditto.
1515 (avx512f_vpermt2var<mode>3): Ditto.
1516 (vec_init<mode>): Ditto.
1517 (avx512f_gathersi<mode>): Ditto.
1518 (*avx512f_gathersi<mode>): Ditto.
1519 (*avx512f_gathersi<mode>_2): Ditto.
1520 (avx512f_gatherdi<mode>): Ditto.
1521 (*avx512f_gatherdi<mode>): Ditto.
1522 (*avx512f_gatherdi<mode>_2): Ditto.
1523 (avx512f_scattersi<mode>): Ditto.
1524 (*avx512f_scattersi<mode>): Ditto.
1525 (avx512f_scatterdi<mode>): Ditto.
1526 (*avx512f_scatterdi<mode>): Ditto.
1527 (sseintprefix): Extened with wider modes.
1528 (VEC_GATHER_IDXSI): Ditto.
1529 (VEC_GATHER_IDXDI): Ditto.
1530 (VEC_GATHER_SRCDI): Ditto.
1531
1532 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1533 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1534
1535 * config/arm/t-aprofile: New file.
1536 * config.gcc: Handle --with-multilib-list option.
1537
1538 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1539
1540 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1541 the input for r2.
1542
1543 2013-10-15 Richard Biener <rguenther@suse.de>
1544
1545 * tree-loop-distribution.c (build_empty_rdg): Inline into
1546 single user.
1547 (rdg_flag_vertex): Inline into single user.
1548 (rdg_flag_vertex_and_dependent): Likewise.
1549 (build_rdg_partition_for_vertex): Remove processed bitmap.
1550 (rdg_build_partitions): Simplify.
1551
1552 2013-10-15 Richard Biener <rguenther@suse.de>
1553
1554 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1555 Restructure forwarding through conversions and copies to
1556 avoid performing copy-propagation the wrong way. Adjust
1557 recursion invocations.
1558 (forward_propagate_addr_expr): Add argument stating if we
1559 are recursing from a single-use.
1560 (ssa_forward_propagate_and_combine): Adjust.
1561
1562 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1563
1564 * dumpfile.h (gcc::dump_manager): New class, to hold state
1565 relating to dumpfile management.
1566 (get_dump_file_name): Remove in favor of method of dump_manager.
1567 (dump_initialized_p): Likewise.
1568 (dump_start): Likewise.
1569 (dump_finish): Likewise.
1570 (dump_switch_p): Likewise.
1571 (dump_register): Likewise.
1572 (get_dump_file_info): Likewise.
1573 * context.c (gcc::context::context): Construct the dump_manager
1574 instance.
1575 * context.h (gcc::context::get_dumps): New.
1576 (gcc::context::m_dumps): New.
1577 * coverage.c (coverage_init): Port to dump_manager API.
1578 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1579 (extra_dump_files_in_use): Likewise.
1580 (extra_dump_files_alloced): Likewise.
1581 (gcc::dump_manager::dump_manager): New.
1582 (dump_register): Convert to...
1583 (gcc::dump_manager::dump_register): ...method, replacing
1584 function-static next_dump with m_next_dump field.
1585 (get_dump_file_info): Convert to...
1586 (gcc::dump_manager::get_dump_file_info): ...method.
1587 (get_dump_file_name): Convert to...
1588 (gcc::dump_manager::get_dump_file_name): ...method.
1589 (dump_start): Convert to...
1590 (gcc::dump_manager::dump_start): ...method.
1591 (dump_finish): Convert to...
1592 (gcc::dump_manager::dump_finish): ...method.
1593 (dump_begin): Replace body with...
1594 (gcc::dump_manager::dump_begin): ...new method.
1595 (dump_phase_enabled_p): Convert to...
1596 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1597 (dump_phase_enabled_p): Convert to...
1598 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1599 (dump_initialized_p): Convert to...
1600 (gcc::dump_manager::dump_initialized_p): ...method.
1601 (dump_flag_name): Replace body with...
1602 (gcc::dump_manager::dump_flag_name): ...new method.
1603 (dump_enable_all): Convert to...
1604 (gcc::dump_manager::dump_enable_all): ...new method.
1605 (opt_info_enable_passes): Convert to...
1606 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1607 (dump_switch_p_1): Convert to...
1608 (gcc::dump_manager::dump_switch_p_1): ...new method.
1609 (dump_switch_p): Convert to...
1610 (gcc::dump_manager::dump_switch_p): ...new method.
1611 (opt_info_switch_p): Port to dump_manager API.
1612 (enable_rtl_dump_file): Likewise.
1613 * opts-global.c (handle_common_deferred_options): Port to new
1614 dump_manager API.
1615 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1616 (pass_manager::register_one_dump_file): Likewise.
1617 (pass_manager::register_pass): Likewise.
1618 (pass_init_dump_file): Likewise.
1619 (pass_fini_dump_file): Likewise.
1620 * statistics.c (statistics_early_init): Likewise.
1621
1622 2013-10-14 Richard Biener <rguenther@suse.de>
1623
1624 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1625 iterative_hash_canonical_type, gimple_canonical_type_hash,
1626 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1627 gimple_register_canonical_type, print_gimple_types_stats,
1628 free_gimple_type_tables): Move to lto/lto.c
1629 (gt-gimple.h): Do not include.
1630 * gimple.h (gimple_register_canonical_type,
1631 print_gimple_types_stats, free_gimple_type_tables): Remove.
1632 * Makefile.in (GTFILES): Remove gimple.c.
1633
1634 2013-10-14 Travis Snoozy <quandary@remstate.com>
1635
1636 PR target/58716
1637 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1638 scanning for msp430x targets.
1639
1640 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1641
1642 PR bootstrap/58509
1643 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1644 (registers_ok_for_ldd_peep): Move around.
1645 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1646 * config/sparc/sparc.md (widening peepholes): Use it.
1647
1648 2013-10-14 Richard Biener <rguenther@suse.de>
1649
1650 PR middle-end/58712
1651 PR middle-end/55358
1652 * gimple.c (iterative_hash_canonical_type): Make sure to
1653 record the hash into the correct hashtable slot.
1654
1655 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1656
1657 PR rtl-optimization/58662
1658 * combine.c (try_combine): Take into account death nodes on I2 when
1659 splitting a PARALLEL of two independent SETs. Fix dump message.
1660
1661 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1662
1663 PR target/51244
1664 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1665 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1666 * config/sh/t-sh (sh_treg_combine.o): New entry.
1667 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1668 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1669 (register_sh_passes): New function. Register sh_treg_combine pass.
1670 (sh_option_override): Invoke it.
1671 (sh_canonicalize_comparison): Handle op0_preserve_value.
1672 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1673 opportunities. Canonicalize branch condition.
1674 (nott): Allow only if pseudos can be created for non-SH2A.
1675
1676 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1677
1678 PR target/58690
1679 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1680 (ix86_expand_movmem): Replace copy_addr_to_reg with
1681 ix86_copy_addr_to_reg.
1682 (ix86_expand_setmem): Likewise.
1683
1684 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1685
1686 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1687 provided by ix86_fp_compare_mode instead of CCFPUmode.
1688
1689 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1690
1691 * config/aarch64/arm_neon.h
1692 (vtbx<1,3>_<psu>8): Fix register constriants.
1693
1694 2013-10-11 Jeff Law <law@redhat.com>
1695
1696 PR tree-optimization/58640
1697 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1698 threading paths that cross over two loop entry points.
1699
1700 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1701
1702 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1703 handle vector float as well.
1704 (*vsx_le_perm_load_v4si): Likewise.
1705 (*vsx_le_perm_store_v2di): Likewise.
1706 (*vsx_le_perm_store_v4si): Likewise.
1707
1708 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1709
1710 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1711 directly to circumvent subtract from splat{31} workaround.
1712 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1713 prototype.
1714 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1715 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1716 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1717 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1718 separate big and little endian logic.
1719 (*altivec_vperm_<mode>_internal): New define_insn.
1720 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1721 separate big and little endian logic.
1722 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1723 (vec_permv16qi): Add little endian logic.
1724
1725 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1726
1727 * doc/extend.texi (returns_nonnull): Remove arguments.
1728
1729 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1730 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1731 Sergey Lega <sergey.s.lega@intel.com>
1732 Anna Tikhonova <anna.tikhonova@intel.com>
1733 Ilya Tocar <ilya.tocar@intel.com>
1734 Andrey Turetskiy <andrey.turetskiy@intel.com>
1735 Ilya Verbin <ilya.verbin@intel.com>
1736 Kirill Yukhin <kirill.yukhin@intel.com>
1737 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1738
1739 * config/i386/sse.md (VI48F_256_512): New.
1740 (avx2_permvar<mode>): Change to ...
1741 (<avx2_avx512f>_permvar<mode>): This.
1742
1743 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1744 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1745 Sergey Lega <sergey.s.lega@intel.com>
1746 Anna Tikhonova <anna.tikhonova@intel.com>
1747 Ilya Tocar <ilya.tocar@intel.com>
1748 Andrey Turetskiy <andrey.turetskiy@intel.com>
1749 Ilya Verbin <ilya.verbin@intel.com>
1750 Kirill Yukhin <kirill.yukhin@intel.com>
1751 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1752
1753 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1754 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1755 * config/i386/sse.md (VI4_AVX): New.
1756 (sf2simodelower): Ditto.
1757 (sse2_cvtps2dq): Change to ...
1758 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1759
1760 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1761 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1762 Sergey Lega <sergey.s.lega@intel.com>
1763 Anna Tikhonova <anna.tikhonova@intel.com>
1764 Ilya Tocar <ilya.tocar@intel.com>
1765 Andrey Turetskiy <andrey.turetskiy@intel.com>
1766 Ilya Verbin <ilya.verbin@intel.com>
1767 Kirill Yukhin <kirill.yukhin@intel.com>
1768 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1769
1770 * config/i386/sse.md (V_512): New.
1771 (VI_512): Ditto.
1772 (vcond<V_512:mode><VF_512:mode>): Ditto.
1773 (vcond<V_512:mode><VI_512:mode>): Ditto.
1774 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1775
1776 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1777 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1778 Sergey Lega <sergey.s.lega@intel.com>
1779 Anna Tikhonova <anna.tikhonova@intel.com>
1780 Ilya Tocar <ilya.tocar@intel.com>
1781 Andrey Turetskiy <andrey.turetskiy@intel.com>
1782 Ilya Verbin <ilya.verbin@intel.com>
1783 Kirill Yukhin <kirill.yukhin@intel.com>
1784 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1785
1786 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1787 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1788 (FMAMODE): Ditto.
1789 (fma<mode>4): Removed condition.
1790 (fms<mode>4): Ditto.
1791 (fnma<mode>4): Ditto.
1792 (fnms<mode>4): Ditto.
1793 (fma4i_fmadd_<mode>): Ditto.
1794 (*fma_fmadd_<mode>): Ditto.
1795 (*fma_fmsub_<mode>): Ditto.
1796 (*fma_fnmadd_<mode>): Ditto.
1797 (*fma_fnmsub_<mode>): Ditto.
1798 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1799 (*fma_fmaddsub_<mode>): Ditto.
1800 (*fma_fmsubadd_<mode>): Ditto.
1801 (*fmai_fmadd_<mode>): Ditto.
1802 (*fmai_fmsub_<mode>): Ditto.
1803 (*fmai_fnmadd_<mode>): Ditto.
1804 (*fmai_fnmsub_<mode>): Ditto.
1805
1806 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1807 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1808 Sergey Lega <sergey.s.lega@intel.com>
1809 Anna Tikhonova <anna.tikhonova@intel.com>
1810 Ilya Tocar <ilya.tocar@intel.com>
1811 Andrey Turetskiy <andrey.turetskiy@intel.com>
1812 Ilya Verbin <ilya.verbin@intel.com>
1813 Kirill Yukhin <kirill.yukhin@intel.com>
1814 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1815
1816 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1817 (VI124_256): Changed to ...
1818 (VI124_256_48_512): This.
1819 (ssepackmode): Extended with wider modes.
1820 (<code><mode>3): Changed iterator.
1821 (*avx2_<code><mode>3): Ditto.
1822 (vec_pack_trunc_<mode>): Ditto.
1823
1824 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1825 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1826 Sergey Lega <sergey.s.lega@intel.com>
1827 Anna Tikhonova <anna.tikhonova@intel.com>
1828 Ilya Tocar <ilya.tocar@intel.com>
1829 Andrey Turetskiy <andrey.turetskiy@intel.com>
1830 Ilya Verbin <ilya.verbin@intel.com>
1831 Kirill Yukhin <kirill.yukhin@intel.com>
1832 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1833
1834 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1835 (VI8F_256_512): Ditto.
1836 (abs<mode>2): Changed iterator.
1837 (avx2_perm<mode>): Changed to ...
1838 (<avx2_avx512f>_perm<mode>): This.
1839 (avx2_perm<mode>_1): Changed to ...
1840 (<avx2_avx512f>_perm<mode>_1): This.
1841
1842 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1843 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1844 Sergey Lega <sergey.s.lega@intel.com>
1845 Anna Tikhonova <anna.tikhonova@intel.com>
1846 Ilya Tocar <ilya.tocar@intel.com>
1847 Andrey Turetskiy <andrey.turetskiy@intel.com>
1848 Ilya Verbin <ilya.verbin@intel.com>
1849 Kirill Yukhin <kirill.yukhin@intel.com>
1850 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1851
1852 * config/i386/sse.md (VI48_AVX512F): New.
1853 (VI48_AVX2): Changed to ...
1854 (VI48_AVX2_48_AVX512F): This.
1855 (avx2_ashrv<mode>): Changed to ...
1856 (<avx2_avx512f>_ashrv<mode>): This.
1857 (avx2_<shift_insn>v<mode>): Changed to ...
1858 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1859
1860 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1861 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1862 Sergey Lega <sergey.s.lega@intel.com>
1863 Anna Tikhonova <anna.tikhonova@intel.com>
1864 Ilya Tocar <ilya.tocar@intel.com>
1865 Andrey Turetskiy <andrey.turetskiy@intel.com>
1866 Ilya Verbin <ilya.verbin@intel.com>
1867 Kirill Yukhin <kirill.yukhin@intel.com>
1868 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1869
1870 * config/i386/sse.md (VI4_AVX512F): New.
1871 (VI8_AVX2_AVX512F): Ditto.
1872 (mul<mode>3): Extended with wider modes.
1873 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1874 (mul<mode>3): Ditto.
1875 (vec_widen_<s>mult_odd_<mode>): Ditto.
1876
1877 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1878 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1879 Sergey Lega <sergey.s.lega@intel.com>
1880 Anna Tikhonova <anna.tikhonova@intel.com>
1881 Ilya Tocar <ilya.tocar@intel.com>
1882 Andrey Turetskiy <andrey.turetskiy@intel.com>
1883 Ilya Verbin <ilya.verbin@intel.com>
1884 Kirill Yukhin <kirill.yukhin@intel.com>
1885 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1886
1887 * config/i386/sse.md (VI2_AVX512F): New.
1888 (VI124_AVX512F): Ditto.
1889 (sseunpackmode): Extended with wider modes.
1890 (sseunpackfltmode): Ditto.
1891 (vec_unpacks_float_hi_<mode>): Ditto.
1892 (vec_unpacks_float_lo_<mode>): Ditto.
1893 (vec_unpacku_float_hi_<mode>): Ditto.
1894 (vec_unpacku_float_lo_<mode>): Ditto.
1895 (vec_unpacks_lo_<mode>): Ditto.
1896 (vec_unpacks_hi_<mode>): Ditto.
1897 (vec_unpacku_lo_<mode>): Ditto.
1898 (vec_unpacku_hi_<mode>): Ditto.
1899
1900 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1901 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1902 Sergey Lega <sergey.s.lega@intel.com>
1903 Anna Tikhonova <anna.tikhonova@intel.com>
1904 Ilya Tocar <ilya.tocar@intel.com>
1905 Andrey Turetskiy <andrey.turetskiy@intel.com>
1906 Ilya Verbin <ilya.verbin@intel.com>
1907 Kirill Yukhin <kirill.yukhin@intel.com>
1908 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1909
1910 * config/i386/i386.md (multdiv): New.
1911 (multdiv_mnemonic): Ditto.
1912 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1913 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1914 (<sse>_vmdiv<mode>3): Removed.
1915
1916 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1917 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1918 Sergey Lega <sergey.s.lega@intel.com>
1919 Anna Tikhonova <anna.tikhonova@intel.com>
1920 Ilya Tocar <ilya.tocar@intel.com>
1921 Andrey Turetskiy <andrey.turetskiy@intel.com>
1922 Ilya Verbin <ilya.verbin@intel.com>
1923 Kirill Yukhin <kirill.yukhin@intel.com>
1924 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1925
1926 * config/i386/sse.md (V): Extended with wider modes.
1927 (VF2): Ditto.
1928 (ssehalfvecmode): Ditto.
1929 (i128): Ditto.
1930 (ssepackfltmode): Ditto.
1931 (avx_vec_concat<mode>): Ditto.
1932 (V_256_512): New iterator.
1933 (VF2_512_256): Ditto.
1934 (si2dfmode): New attribute.
1935 (si2dfmodelower): Ditto.
1936 (sf2dfmode): Ditto.
1937 (concat_tg_mode): Ditto.
1938 (floatv4siv4df2): Changed to ...
1939 (float<si2dfmodelower><mode>2): This.
1940 (avx_cvtps2pd256): Changed to ...
1941 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1942 (vec_pack_trunc_v4df): Changed to ...
1943 (vec_pack_trunc_<mode>): This.
1944 (avx_vpermil<mode>): Changed to ...
1945 (<sse2_avx_avx512f>_vpermil<mode>): This.
1946 (<fixsuffix>fix_truncv8dfv8si2): New.
1947 (vec_pack_sfix_trunc_v8df): Ditto.
1948 (avx512f_rndscale<mode>): Ditto.
1949 (avx512f_roundpd512): Ditto.
1950 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1951
1952 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1953 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1954 Sergey Lega <sergey.s.lega@intel.com>
1955 Anna Tikhonova <anna.tikhonova@intel.com>
1956 Ilya Tocar <ilya.tocar@intel.com>
1957 Andrey Turetskiy <andrey.turetskiy@intel.com>
1958 Ilya Verbin <ilya.verbin@intel.com>
1959 Kirill Yukhin <kirill.yukhin@intel.com>
1960 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1961
1962 * config/i386/i386.md (any_fix): New iterator.
1963 (fixsuffix): New attribute.
1964 * config/i386/sse.md (VF1): Extened with wider modes.
1965 (VI): Ditto.
1966 (VI_AVX2): Ditto.
1967 (VI8): Ditto.
1968 (sseintvecmodelower): Ditto.
1969 (ssescalarmode): Ditto.
1970 (ssescalarnum): Ditto.
1971 (VF1_128_256): New.
1972 (ssexmmmode): Ditto.
1973 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1974 (<sse>_rcp<mode>2): Change iterator.
1975 (rsqrt<mode>2): Ditto.
1976 (<sse>_rsqrt<mode>2): Ditto.
1977 (avx2_vec_dup<mode>): Ditto.
1978 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1979 (round<mode>2_sfix): Ditto.
1980 (avx2_pbroadcast<mode>): Ditto.
1981 (*andnot<mode>3): Handle XI mode.
1982 (*<code><mode>3): Ditto.
1983 (AVXTOSSEMODE): Removed.
1984 (avx_vpermil<mode>): Changed to ...
1985 (<sse2_avx_avx512f>_vpermil<mode>): This.
1986
1987 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1988 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1989 Sergey Lega <sergey.s.lega@intel.com>
1990 Anna Tikhonova <anna.tikhonova@intel.com>
1991 Ilya Tocar <ilya.tocar@intel.com>
1992 Andrey Turetskiy <andrey.turetskiy@intel.com>
1993 Ilya Verbin <ilya.verbin@intel.com>
1994 Kirill Yukhin <kirill.yukhin@intel.com>
1995 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1996
1997 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1998 (<sse>_comi): Ditto.
1999 (<sse>_ucomi): Ditto.
2000 (sse_cvtss2siq_2): Ditto.
2001 (sse2_cvtsd2si): Ditto.
2002 (sse2_cvtsd2siq): Ditto.
2003 (sse2_cvttsd2si): Ditto.
2004 (sse2_cvttsd2siq): Ditto.
2005 (<shift_insn><mode>3): Ditto.
2006 (sse2_cvtsi2sdq): Update constraint and prefix.
2007 (sse_cvtsi2ss): Update prefix.
2008 (sse_cvtsi2ssq): Ditto.
2009
2010 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2011
2012 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2013 ignore internal calls.
2014
2015 2013-10-11 Richard Biener <rguenther@suse.de>
2016
2017 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2018 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2019 TREE_OVERFLOW is set.
2020
2021 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2022
2023 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2024
2025 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2026 * gimple.h: Likewise.
2027
2028 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2029
2030 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2031 macros.
2032
2033 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2034
2035 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2036 * configure: Regenerate.
2037
2038 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2039
2040 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2041 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2042 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2043 similarly pp_right_paren.
2044 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2045 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2046 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2047 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2048 clauses.
2049 (omp_declare_simd_clauses_equal,
2050 omp_remove_redundant_declare_simd_attrs): New functions.
2051 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2052 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2053 * tree.h (OMP_LOOP_CHECK): Define.
2054 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2055 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2056 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2057 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2058 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2059 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2060 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2061 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2062 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2063 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2064 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2065 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2066 (omp_remove_redundant_declare_simd_attrs): New prototype.
2067 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2068 GIMPLE_OMP_TEAMS): New codes.
2069 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2070 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2071 fields.
2072 (target_nesting_level): New variable.
2073 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2074 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2075 for collapse > 1 static schedule unless ordered.
2076 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2077 (determine_parallel_type): Adjust get_ws_args_for caller.
2078 (install_var_field): Handle mask & 4 for double indirection.
2079 (scan_sharing_clauses): Ignore shared clause on teams construct.
2080 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2081 (create_omp_child_function): If inside target or declare target
2082 constructs, set "omp declare target" attribute on the child function.
2083 (find_combined_for): New function.
2084 (scan_omp_parallel): Handle combined loops.
2085 (scan_omp_target, scan_omp_teams): New functions.
2086 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2087 restrictions and set ctx->cancellable for cancellable constructs.
2088 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2089 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2090 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2091 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2092 (omp_clause_aligned_alignment): New function.
2093 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2094 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2095 on teams constructs. Handle user defined reductions and new
2096 OpenMP 4.0 clauses.
2097 (lower_reduction_clauses): Don't set placeholder to address of ref
2098 if it has already the right type.
2099 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2100 (expand_parallel_call): Use the new non-_start suffixed builtins,
2101 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2102 and GOMP_parallel_end after the call.
2103 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2104 (expand_omp_for_init_counts): Handle combined loops.
2105 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2106 loops.
2107 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2108 end of cancellable loops.
2109 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2110 Likewise. Handle collapse > 1 loops.
2111 (expand_omp_simd): Handle combined loops.
2112 (expand_omp_for): Add inner_stmt argument, adjust callers of
2113 expand_omp_for* functions, use expand_omp_for_static*chunk even
2114 for collapse > 1 unless ordered.
2115 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2116 of cancellable sections.
2117 (expand_omp_single): Remove need_barrier variable, just rely on
2118 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2119 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2120 (expand_omp_atomic_load, expand_omp_atomic_store,
2121 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2122 (expand_omp_target): New function.
2123 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2124 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2125 (build_omp_regions_1): Immediately close region for
2126 GF_OMP_TARGET_KIND_UPDATE.
2127 (maybe_add_implicit_barrier_cancel): New function.
2128 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2129 cancellation.
2130 (lower_omp_single): Likewise. Add clobber after the barrier.
2131 (lower_omp_taskgroup): New function.
2132 (lower_omp_for): Handle combined loops. Adjust
2133 lower_rec_input_clauses caller. Handle cancellation.
2134 (lower_depend_clauses): New function.
2135 (lower_omp_taskreg): Lower depend clauses. Adjust
2136 lower_rec_input_clauses caller. Add clobber after the call. Handle
2137 cancellation.
2138 (lower_omp_target, lower_omp_teams): New functions.
2139 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2140 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2141 and GOMP_cancellation_point calls.
2142 (lower_omp): Fold stmts inside of target region.
2143 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2144 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2145 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2146 (BT_FN_VOID_OMPFN_PTR_UINT,
2147 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2148 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2149 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2150 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2151 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2152 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2153 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2154 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2155 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2156 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2157 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2158 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2159 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2160 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2161 BUILT_IN_GOMP_PARALLEL_END.
2162 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2163 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2164 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2165 GF_OMP_FOR_KIND_DISTRIBUTE.
2166 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2167 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2168 (dump_gimple_omp_return): Print lhs if it has any.
2169 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2170 gimple_omp_atomic_seq_cst_p.
2171 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2172 and GIMPLE_OMP_TEAMS.
2173 * langhooks.c (lhd_omp_mappable_type): New function.
2174 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2175 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2176 hook.
2177 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2178 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2179 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2180 (struct gimplify_omp_ctx): Add combined_loop field.
2181 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2182 on stmts inside of target region.
2183 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2184 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2185 ORT_TARGET and ORT_TARGET_DATA.
2186 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2187 Handle GOVD_MAP.
2188 (omp_notice_threadprivate_variable): Complain about threadprivate
2189 variables in target region.
2190 (omp_notice_variable): Complain about vars with non-mappable type
2191 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2192 (omp_check_private): Ignore ORT_TARGET* regions.
2193 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2194 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2195 (find_combined_omp_for): New function.
2196 (gimplify_omp_for): Handle gimplification of combined loops.
2197 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2198 OMP_TEAMS.
2199 (gimplify_omp_target_update): New function.
2200 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2201 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2202 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2203 (gimplify_body): If fndecl has "omp declare target" attribute, add
2204 implicit ORT_TARGET context around it.
2205 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2206 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2207 * tree-nested.c (convert_nonlocal_reference_stmt,
2208 convert_local_reference_stmt, convert_gimple_call): Handle
2209 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2210 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2211 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2212 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2213 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2214 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2215 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2216 BUILT_IN_GOMP_LOOP_END_CANCEL,
2217 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2218 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2219 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2220 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2221 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2222 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2223 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2224 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2225 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2226 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2227 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2228 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2229 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2230 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2231 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2232 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2233 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2234 gimple_build_omp_teams): New functions.
2235 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2236 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2237 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2238 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2239 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2240 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2241 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2242 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2243 GF_OMP_ATOMIC_SEQ_CST): New.
2244 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2245 gimple_build_omp_teams): New prototypes.
2246 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2247 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2248 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2249 GIMPLE_OMP_SINGLE as end of range.
2250 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2251 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2252 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2253 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2254 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2255 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2256 gimple_omp_target_kind, gimple_omp_target_set_kind,
2257 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2258 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2259 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2260 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2261 gimple_omp_teams_set_clauses): New inlines.
2262 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2263 and GIMPLE_OMP_TASKGROUP.
2264 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2265 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2266 enum omp_clause_proc_bind_kind): New.
2267 (union omp_clause_subcode): Add depend_kind, map_kind and
2268 proc_bind_kind fields.
2269 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2270 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2271 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2272 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2273 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2274
2275 2013-10-10 Teresa Johnson <tejohnson@google.com>
2276
2277 * predict.c (tree_estimate_probability): Add new parameter
2278 for estimate_bb_frequencies.
2279 (estimate_bb_frequencies): Add new parameter to force estimation.
2280 (rebuild_frequencies): When max frequency in function is small,
2281 recompute counts from frequencies.
2282 * predict.h (estimate_bb_frequencies): New parameter.
2283
2284 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2285
2286 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2287 optimizations are disabled.
2288
2289 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2290
2291 * coverage.c (coverage_finish): Fix leak of da_file_name.
2292
2293 2013-10-10 Jan Hubicka <jh@suse.cz>
2294
2295 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2296 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2297 Bobcat and generic.
2298
2299 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2300
2301 PR middle-end/58670
2302 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2303 if any labels are in FALLTHRU_BB, use a special label emitted
2304 immediately after the asm goto insn rather than label_rtx
2305 of the LABEL_DECL.
2306 (expand_asm_stmt): Adjust caller.
2307 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2308 edge if the last insn in predecessor is a jump with single successor,
2309 but it isn't simplejump_p.
2310
2311 2013-10-10 Richard Biener <rguenther@suse.de>
2312
2313 PR tree-optimization/58656
2314 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2315
2316 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2317
2318 * gimplify.c: Include expr.h and tm_p.h for targets with special
2319 va-arg padding requirements.
2320
2321 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2322
2323 * tree-flow.h: Move some prototypes to gimple.h.
2324 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2325 * gimple.h: Relocate some prototypes from tree-flow.h
2326 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2327 Move to gimplify.c.
2328 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2329 (build_va_arg_indirect_ref): Relocate and make static.
2330 (std_gimplify_va_arg_expr): Relocate here.
2331 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2332 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2333
2334 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2335
2336 * doc/md.texi: Document the mnemonic attribute.
2337
2338 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2339
2340 PR target/57377
2341 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2342 (set_attr_alternative x ...) when searching for user defined
2343 mnemonic attribute.
2344
2345 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2346
2347 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2348
2349 2013-10-09 Easwaran Raman <eraman@google.com>
2350
2351 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2352 * cfgexpand.c (defer_stack_allocation): ...use here
2353 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2354
2355 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2356
2357 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2358 (operand_equal_for_value_replacement): New function, extracted from
2359 value_replacement and enhanced to catch more cases.
2360 (value_replacement): Use operand_equal_for_value_replacement.
2361
2362 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2363
2364 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2365 get_max_loop_iterations.
2366
2367 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2368
2369 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2370 Place comment above function.
2371
2372 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2373
2374 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2375 are not related to tree-cfg.c.
2376 * tree-ssa-address.h: New file. Relocate prototypes.
2377 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2378 (addr_for_mem_ref): New. Combine call to get_address_description and
2379 return addr_for_mem_ref.
2380 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2381 * tree-ssa-live.h: Adjust prototypes.
2382 * passes.c: Include tree-ssa-live.h.
2383 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2384 * graphite.c (graphite_transform_loops): Make static.
2385 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2386 make_pass_graphite, pass_data_graphite_transforms,
2387 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2388 * ipa-pure-const.c (warn_function_noreturn): Make static.
2389 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2390 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2391 Relocate from tree-cfg.c
2392 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2393 static.
2394 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2395 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2396 Move to ipa-pure-const.c.
2397 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2398 Relocate from tree-optimize.c.
2399 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2400 make_pass_fixup_cfg): Move to tree-cfg.c.
2401 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2402 Relocate some prototypes.
2403 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2404 * tree-dump.c (dump_function_to_file): Remove prototype.
2405 Add tree-flow.h to the include file.
2406 * tree-dump.h: Remove prototype.
2407 * tree-parloops.h: New File. Add prototypes.
2408 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2409 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2410 from tree-ssa-loop.c.
2411 * tree-predcom.c (run_tree_predictive_commoning,
2412 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2413 Relocate here from tree-ssa-loop.c.
2414 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2415 ssa_name_values.release ().
2416 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2417 (ssa_name_values): Relocate from tree-flow.h.
2418 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2419 * tree-ssa-loop.c (run_tree_predictive_commoning,
2420 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2421 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2422 make_pass_graphite, pass_data_graphite_transforms,
2423 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2424 tree_parallelize_loops, pass_data_parallelize_loops,
2425 make_pass_parallelize_loops): Move to other files.
2426 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2427 moved here.
2428 * tree.h: Remove prototypes from tree-address.c.
2429
2430 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2431
2432 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2433 (struct int_tree_map): Move to tree-hasher.h
2434 (SCALE, LABEL, PERCENT): Move to gimple.h
2435 * tree-flow-inline.h: Delete. Move functions to other files.
2436 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2437 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2438 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2439 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2440 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2441 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2442 * tree-hasher.h: Don't include tree-flow.h.
2443 (struct int_tree_map): Relocate from tree-flow.h.
2444 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2445 tree-flow-inline.h and make static.
2446 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2447 tree-flow-inline.h.
2448 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2449 tree-flow-inline.h and make static.
2450 * tree.h (is_global_var, may_be_aliased): Relocate from
2451 tree-flow-inline.h.
2452 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2453 * value-prof.c: No longer include tree-flow-inline.h.
2454 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2455
2456 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2457
2458 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2459 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2460 (enum move_pos): Move to tree-ssa-loop-im.h
2461 * cfgloop.h: Move some prototypes.
2462 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2463 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2464 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2465 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2466 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2467 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2468 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2469 (enum move_pos): Relocate here.
2470 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2471 tree-ssa-loop.c.
2472 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2473 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2474 make_pass_lim): Relocate here from tree-ssa-loop.c.
2475 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2476 tree-ssa-loop.c.
2477 (loop_edge_to_cancel, unloop_loops): Make static.
2478 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2479 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2480 (tree_complete_unroll, gate_tree_complete_unroll,
2481 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2482 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2483 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2484 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2485 (stmt_invariant_in_loop_p): Remove unused function.
2486 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2487 * tree-ssa-loop-manip.h: New file. Add prototypes.
2488 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2489 (gcov_type_to_double_int): Move to cfgloop.h.
2490 (double_int_cmp, bound_index,
2491 estimate_numbers_of_iterations_loop): Make static.
2492 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2493 (max_loop_iterations): Factor out get_max_loop_iterations.
2494 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2495 cfgloop.c.
2496 * tree-ssa-loop-niter.h: New file. Add prototypes.
2497 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2498 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2499 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2500 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2501 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2502 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2503 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2504 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2505 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2506 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2507 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2508 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2509 pass_data_complete_unroll, make_pass_complete_unroll,
2510 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2511 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2512 tree-ssa-loop-ivcanon.c.
2513 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2514 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2515 tree-ssa-loop-prefetch.c.
2516 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2517 tree-ssa-loop-im.c.
2518 (get_lsm_tmp_name): Relocate and add suffix parameter.
2519 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2520 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2521 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2522 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2523 (get_estimated_loop_iterations): Factor out accessor from
2524 estimated_loop_iterations in tree-ssa-loop-niter.c.
2525 (get_max_loop_iterations): Factor out accessor from
2526 _max_loop_iterations in tree-ssa-niter.c.
2527 * loop-unroll.c (decide_unroll_constant_iterations,
2528 decide_unroll_runtime_iterations, decide_peel_simple,
2529 decide_unroll_stupid): Use new get_* accessors.
2530
2531 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2532
2533 PR tree-optimization/20318
2534 * doc/extend.texi (returns_nonnull): New function attribute.
2535 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2536 attribute.
2537 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2538 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2539
2540 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2541
2542 PR middle-end/58570
2543 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2544 false if both components are bitfields.
2545
2546 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2547
2548 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2549 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2550 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2551 * config/aarch64/aarch64.h
2552 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2553 * config/aarch64/aarch64-simd-builtins.def
2554 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2555
2556 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2557
2558 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2559
2560 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2561
2562 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2563
2564 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2565
2566 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2567 (vneg_f64): New intrinsic.
2568 (vneg_s8): Asm replaced with C.
2569 (vneg_s16): Likewise.
2570 (vneg_s32): Likewise.
2571 (vneg_s64): New intrinsic.
2572 (vnegq_f32): Asm replaced with C.
2573 (vnegq_f64): Likewise.
2574 (vnegq_s8): Likewise.
2575 (vnegq_s16): Likewise.
2576 (vnegq_s32): Likewise.
2577 (vnegq_s64): Likewise.
2578
2579 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2580
2581 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2582
2583 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2584
2585 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2586 packed stack special handling.
2587 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2588 back to fixed stack slots for FPRs saved due to stdarg.
2589
2590 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2591
2592 * config/s390/s390.c (s390_frame_info): Restructure function.
2593
2594 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2595
2596 * config/s390/s390.c (struct s390_frame_layout): New field
2597 gpr_save_slots.
2598 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2599 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2600 regs_ever_clobbered to char*.
2601 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2602 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2603 (s390_register_info): Enable FPR save slots. Move/Copy some
2604 functionality into ...
2605 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2606 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2607 function.
2608 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2609 not imply a stack frame.
2610 (s390_init_frame_layout): Remove variable clobbered_regs.
2611 (s390_update_register_info): Remove function.
2612 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2613 cannot be used for register renaming.
2614 (s390_hard_regno_scratch_ok): New function.
2615 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2616 (s390_initial_elimination_offset): Change offset calculation of
2617 the return address pointer.
2618 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2619 (restore_gprs): Set frame related flag.
2620 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2621 (s390_emit_prologue): Call s390_register_info instead of
2622 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2623 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2624 (s390_optimize_prologue): Call s390_optimize_register_info.
2625 Try to remove also FPR slot save/restore INSNs. Remove frame
2626 related flags from restore INSNs.
2627
2628 2013-10-08 DJ Delorie <dj@redhat.com>
2629
2630 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2631 (movhi): Likewise.
2632
2633 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2634 avoid conflict with the MI use of %c.
2635 * config/rl78/rl78-real.md: change %c to %C throughout.
2636 * config/rl78/rl78-virt.md: Likewise.
2637
2638 2013-10-08 Jan Hubicka <jh@suse.cz>
2639
2640 * config/i386/i386.c (ix86_option_override_internal): Switch
2641 to SSE math for -ffast-math when target ISA supports SSE2.
2642
2643 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2644
2645 * tree-flow.h: Remove some prototypes.
2646 * tree.h: Remove some protypes, add a couple.
2647 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2648 using_eh_for_cleanups_p): Add interface routines for front ends.
2649 * tree-eh.h: New file. Add protoptyes.
2650 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2651 (add_stmt_to_eh_lp_fn): Make static.
2652 (lower_try_finally): Use new using_eh_for_cleanups_p.
2653 * emit-rtl.c: Include tree-eh.h.
2654 * gimple.h: Include tree-eh.h.
2655
2656 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2657
2658 PR tree-optimization/58480
2659 * tree-vrp.c (infer_nonnull_range): New function.
2660 (infer_value_range): Call infer_nonnull_range.
2661
2662 2013-10-08 Dehao Chen <dehao@google.com>
2663
2664 PR tree-optimization/58619
2665 * tree-inline.c (copy_phis_for_bb): Combine location data
2666 only if non-null.
2667
2668 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2669
2670 PR target/58423
2671 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2672 RTX_FRAME_RELATED_P on INSN.
2673
2674 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2675
2676 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2677 (altivec_expand_vec_perm_const): Call it.
2678
2679 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2680
2681 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2682 sequences for LE VSX loads and stores at expand time.
2683 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2684 prototype.
2685 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2686 (rs6000_gen_le_vsx_permute): New.
2687 (rs6000_gen_le_vsx_load): New.
2688 (rs6000_gen_le_vsx_store): New.
2689 (rs6000_gen_le_vsx_move): New.
2690 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2691 (*vsx_le_perm_load_v4si): New.
2692 (*vsx_le_perm_load_v8hi): New.
2693 (*vsx_le_perm_load_v16qi): New.
2694 (*vsx_le_perm_store_v2di): New.
2695 (*vsx_le_perm_store_v4si): New.
2696 (*vsx_le_perm_store_v8hi): New.
2697 (*vsx_le_perm_store_v16qi): New.
2698 (*vsx_xxpermdi2_le_<mode>): New.
2699 (*vsx_xxpermdi4_le_<mode>): New.
2700 (*vsx_xxpermdi8_le_V8HI): New.
2701 (*vsx_xxpermdi16_le_V16QI): New.
2702 (*vsx_lxvd2x2_le_<mode>): New.
2703 (*vsx_lxvd2x4_le_<mode>): New.
2704 (*vsx_lxvd2x8_le_V8HI): New.
2705 (*vsx_lxvd2x16_le_V16QI): New.
2706 (*vsx_stxvd2x2_le_<mode>): New.
2707 (*vsx_stxvd2x4_le_<mode>): New.
2708 (*vsx_stxvd2x8_le_V8HI): New.
2709 (*vsx_stxvd2x16_le_V16QI): New.
2710
2711 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2712
2713 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2714
2715 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2716
2717 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2718 loop to work also for 31bit ABI.
2719 Save the stack pointer for frame_size > 0.
2720
2721 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2722
2723 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2724 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2725 constraint letters from expanders.
2726 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2727 retry count to general_operand.
2728 ("tabort"): Give operand 0 a mode.
2729 ("tabort_1"): Add mode and constraint letter for operand 0.
2730 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2731
2732 2013-10-04 Jeff Law <law@redhat.com>
2733
2734 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2735
2736 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2737 out of ...
2738 (thread_across_edge): Here. Call it.
2739
2740 2013-10-04 Cary Coutant <ccoutant@google.com>
2741
2742 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2743 discarding a location list expression (or a piece of one).
2744
2745 2013-10-03 Jan Hubicka <jh@suse.cz>
2746
2747 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2748 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2749 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2750
2751 2013-10-03 Jan Hubicka <jh@suse.cz>
2752
2753 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2754 accumulate-outgoing-args when producing unwind info.
2755
2756 2013-10-03 Wei Mi <wmi@google.com>
2757
2758 * lra-constraints.c (insert_move_for_subreg): New function
2759 extracted from simplify_operand_subreg.
2760 (simplify_operand_subreg): Add reload for paradoxical subreg.
2761
2762 2013-10-03 Rong Xu <xur@google.com>
2763
2764 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2765 the candidate of builtin_expect such that we should fix the
2766 size/time estimation.
2767 (estimate_function_body_sizes): Do the acutally size/time fix-up
2768 for builtin_expect.
2769
2770 2013-10-03 Rong Xu <xur@google.com>
2771
2772 * predict.c (tree_predict_by_opcode): Get the probability
2773 for builtin_expect from param builtin_expect_probability.
2774 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2775 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2776 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2777
2778 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2779
2780 PR c++/19476
2781 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2782 * calls.c (alloca_call_p): Use get_callee_fndecl.
2783 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2784 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2785 Likewise.
2786 (vrp_visit_stmt): Remove duplicated code.
2787
2788 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2789
2790 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2791 ceildf2, btruncdf2, instead of vsx_* name.
2792
2793 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2794 iterators to only do V2DF and V4SF here. Move the DF code to
2795 rs6000.md where it is combined with SF mode. Replace <VSv> with
2796 just 'v' since only vector operations are handled with these insns
2797 after moving the DF support to rs6000.md.
2798 (vsx_sub<mode>3): Likewise.
2799 (vsx_mul<mode>3): Likewise.
2800 (vsx_div<mode>3): Likewise.
2801 (vsx_fre<mode>2): Likewise.
2802 (vsx_neg<mode>2): Likewise.
2803 (vsx_abs<mode>2): Likewise.
2804 (vsx_nabs<mode>2): Likewise.
2805 (vsx_smax<mode>3): Likewise.
2806 (vsx_smin<mode>3): Likewise.
2807 (vsx_sqrt<mode>2): Likewise.
2808 (vsx_rsqrte<mode>2): Likewise.
2809 (vsx_fms<mode>4): Likewise.
2810 (vsx_nfma<mode>4): Likewise.
2811 (vsx_copysign<mode>3): Likewise.
2812 (vsx_btrunc<mode>2): Likewise.
2813 (vsx_floor<mode>2): Likewise.
2814 (vsx_ceil<mode>2): Likewise.
2815 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2816 (vsx_sminsf3): Likewise.
2817 (vsx_fmadf4): Likewise.
2818 (vsx_fmsdf4): Likewise.
2819 (vsx_nfmadf4): Likewise.
2820 (vsx_nfmsdf4): Likewise.
2821 (vsx_cmpdf_internal1): Likewise.
2822
2823 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2824 simpler to select whether a target has SPE or traditional floating
2825 point support in iterators.
2826 (TARGET_DF_SPE): Likewise.
2827 (TARGET_SF_FPR): Likewise.
2828 (TARGET_DF_FPR): Likewise.
2829 (TARGET_SF_INSN): Macros to say whether floating point support
2830 exists for a given operation for expanders.
2831 (TARGET_DF_INSN): Likewise.
2832
2833 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2834 combining of SF/DF mode operations, using both traditional and VSX
2835 registers.
2836 (Fvsx): Likewise.
2837 (Ff): Likewise.
2838 (Fv): Likewise.
2839 (Fs): Likewise.
2840 (Ffre): Likewise.
2841 (FFRE): Likewise.
2842 (abs<mode>2): Combine SF/DF modes using traditional floating point
2843 instructions. Add support for using the upper DF registers with
2844 VSX support, and SF registers with power8-vector support. Update
2845 expanders for operations supported by both the SPE and traditional
2846 floating point units.
2847 (abs<mode>2_fpr): Likewise.
2848 (nabs<mode>2): Likewise.
2849 (nabs<mode>2_fpr): Likewise.
2850 (neg<mode>2): Likewise.
2851 (neg<mode>2_fpr): Likewise.
2852 (add<mode>3): Likewise.
2853 (add<mode>3_fpr): Likewise.
2854 (sub<mode>3): Likewise.
2855 (sub<mode>3_fpr): Likewise.
2856 (mul<mode>3): Likewise.
2857 (mul<mode>3_fpr): Likewise.
2858 (div<mode>3): Likewise.
2859 (div<mode>3_fpr): Likewise.
2860 (sqrt<mode>3): Likewise.
2861 (sqrt<mode>3_fpr): Likewise.
2862 (fre<Fs>): Likewise.
2863 (rsqrt<mode>2): Likewise.
2864 (cmp<mode>_fpr): Likewise.
2865 (smax<mode>3): Likewise.
2866 (smin<mode>3): Likewise.
2867 (smax<mode>3_vsx): Likewise.
2868 (smin<mode>3_vsx): Likewise.
2869 (negsf2): Delete SF operations that are merged with DF.
2870 (abssf2): Likewise.
2871 (addsf3): Likewise.
2872 (subsf3): Likewise.
2873 (mulsf3): Likewise.
2874 (divsf3): Likewise.
2875 (fres): Likewise.
2876 (fmasf4_fpr): Likewise.
2877 (fmssf4_fpr): Likewise.
2878 (nfmasf4_fpr): Likewise.
2879 (nfmssf4_fpr): Likewise.
2880 (sqrtsf2): Likewise.
2881 (rsqrtsf_internal1): Likewise.
2882 (smaxsf3): Likewise.
2883 (sminsf3): Likewise.
2884 (cmpsf_internal1): Likewise.
2885 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2886 (negdf2): Delete DF operations that are merged with SF.
2887 (absdf2): Likewise.
2888 (nabsdf2): Likewise.
2889 (adddf3): Likewise.
2890 (subdf3): Likewise.
2891 (muldf3): Likewise.
2892 (divdf3): Likewise.
2893 (fred): Likewise.
2894 (rsqrtdf_internal1): Likewise.
2895 (fmadf4_fpr): Likewise.
2896 (fmsdf4_fpr): Likewise.
2897 (nfmadf4_fpr): Likewise.
2898 (nfmsdf4_fpr): Likewise.
2899 (sqrtdf2): Likewise.
2900 (smaxdf3): Likewise.
2901 (smindf3): Likewise.
2902 (cmpdf_internal1): Likewise.
2903 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2904 (btrunc<mode>2): Delete separate expander, and combine with the
2905 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2906 (btrunc<mode>2_fpr): Likewise.
2907 (ceil<mode>2): Likewise.
2908 (ceil<mode>2_fpr): Likewise.
2909 (floor<mode>2): Likewise.
2910 (floor<mode>2_fpr): Likewise.
2911 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2912 Add support for using the upper registers with VSX and
2913 power8-vector. Move insns to be closer to the define_expands. On
2914 VSX systems, prefer the traditional form of FMA over the VSX
2915 version, since the traditional form allows the target not to
2916 overlap with the inputs.
2917 (fms<mode>4_fpr): Likewise.
2918 (nfma<mode>4_fpr): Likewise.
2919 (nfms<mode>4_fpr): Likewise.
2920
2921 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2922 Richard Earnshaw <richard.earnshaw@arm.com>
2923
2924 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2925 (struct mult_cost_table): Likewise.
2926 (struct mem_cost_table): Likewise.
2927 (struct fp_cost_table): Likewise.
2928 (struct vector_cost_table): Likewise.
2929 (cpu_cost_table): Likewise.
2930 * config/arm/arm.opt (mold-rts-costs): New option.
2931 (mnew-generic-costs): Likewise.
2932 * config/arm/arm.c (generic_extra_costs): New table.
2933 (cortexa15_extra_costs): Likewise.
2934 (arm_slowmul_tune): Use NULL as new costs.
2935 (arm_fastmul_tune): Likewise.
2936 (arm_strongarm_tune): Likewise.
2937 (arm_xscale_tune): Likewise.
2938 (arm_9e_tune): Likewise.
2939 (arm_v6t2_tune): Likewise.
2940 (arm_cortex_a5_tune): Likewise.
2941 (arm_cortex_a9_tune): Likewise.
2942 (arm_v6m_tune): Likewise.
2943 (arm_fa726te_tune): Likewise.
2944 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2945 (arm_cortex_tune): Use generict_extra_costs.
2946 (shifter_op_p): New function.
2947 (arm_unspec_cost): Likewise.
2948 (LIBCALL_COST): Define.
2949 (arm_new_rtx_costs): New function.
2950 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2951 table is available. Use old costs otherwise unless mnew-generic-costs
2952 is specified.
2953 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2954 (cpu_cost_table): Declare.
2955
2956 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2957
2958 PR target/58460
2959 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2960 (*subs_mul_imm_<mode>)
2961 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2962 (*sub_<shift>_<mode>)
2963 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2964 Remove k constraint.
2965
2966 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2967
2968 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2969 code.
2970 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2971
2972 2013-10-02 Teresa Johnson <tejohnson@google.com>
2973
2974 * predict.c (probably_never_executed): New function.
2975 (probably_never_executed_bb_p): Invoke probably_never_executed.
2976 (probably_never_executed_edge_p): Ditto.
2977 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2978 Treat profile insanities conservatively.
2979
2980 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2981
2982 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2983
2984 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2985
2986 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2987 use it. Use smaller increase for scratch. Don't increase reject
2988 for early clobber scratch.
2989 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2990 setting eliminated regs except setting fp from hfp.
2991 (lra_eliminate): Check lra_insn_recog_data on NULL.
2992
2993 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2994
2995 PR target/58587
2996 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2997 setting -mvsx-timode by default until the underlying problem is fixed.
2998 (RS6000_CPU, power7 defaults): Likewise.
2999
3000 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3001
3002 * config/x-linux (host-linux.o): Remove header dependencies.
3003 Use $(COMPILE) and $(POSTCOMPILE).
3004 * config/t-linux-android (linux-android.o): Ditto.
3005
3006 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3007
3008 * Makefile.in (expmed.o-warn): Remove.
3009
3010 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3011
3012 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3013 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3014
3015 2013-10-02 Teresa Johnson <tejohnson@google.com>
3016
3017 * dojump.c (do_jump_1): Divide probability between
3018 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3019
3020 2013-10-02 Tom Tromey <tromey@redhat.com>
3021
3022 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3023
3024 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3025
3026 * tree-flow.h: Remove some prototypes.
3027 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3028 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3029 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3030 mark_virtual_phi_result_for_renaming): Relocate here.
3031 * tree-into-ssa.h: Add prototypes.
3032 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3033 single_pred_before_succ_order.
3034 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3035 (nonfreeing_call_p) Move to gimple.c.
3036 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3037 tree-ssa-phiopt.c.
3038 * basic-block.h (single_pred_before_succ_order): Add prototype.
3039 * gimple.c (nonfreeing_call_p): Relocate here.
3040 * gimple.h: Add prototype.
3041 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3042 * tree-ssa-dom.h: New file. Relocate prototypes here.
3043 * tree-ssa.h: Include tree-ssa-dom.h.
3044
3045 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3046
3047 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3048 Use $(COMPILE) and $(POSTCOMPILE).
3049
3050 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3051
3052 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3053
3054 * tree-flow.h: Remove some prototypes.
3055 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3056 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3057 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3058 propagate_tree_value*): Move from here to...
3059 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3060 propagate_tree_value*): Relocate here.
3061 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3062 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3063 * gimple-fold.c: Remove gimple-fold.h from include list.
3064 * tree-vrp.c: Remove gimple-fold.h from include list.
3065 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3066 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3067 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3068 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3069 * sese.c: Add tree-ssa-propagate.h to include list.
3070
3071 2013-10-02 Richard Biener <rguenther@suse.de>
3072
3073 * tree-loop-distribution.c: Include tree-vectorizer.h for
3074 find_loop_location.
3075 (enum partition_kind): Remove PKIND_REDUCTION.
3076 (struct partition_s): Remove has_writes member, add reduction_p member.
3077 (partition_alloc): Adjust.
3078 (partition_builtin_p): Likewise.
3079 (partition_has_writes): Remove.
3080 (partition_reduction_p): New function.
3081 (partition_merge_into): Likewise.
3082 (generate_code_for_partition): Commonize builtin partition
3083 handling tail.
3084 (rdg_cannot_recompute_vertex_p): Remove.
3085 (already_processed_vertex_p): Likewise.
3086 (rdg_flag_vertex): Do not set has_writes.
3087 (classify_partition): Adjust.
3088 (rdg_build_partitions): Do not set has_writes, treat all
3089 partitions as useful.
3090 (distribute_loop): Record number of library calls generated. Adjust.
3091 (tree_loop_distribution): Report number of loops and library
3092 calls generated as opt-info.
3093
3094 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3095
3096 * tree-flow.h: Include new .h files. Move prototypes.
3097 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3098 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3099 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3100 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3101 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3102 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3103 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3104 * tree.h (get_ref_base_and_extent): Move prototype out.
3105 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3106 tree-dfa.h.
3107 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3108 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3109 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3110 * tree-scalar-evolution.c: Include tree.h.
3111 * sese.c: Include tree.h.
3112 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3113 * dwarf2out.c: Include tree-dfa.h.
3114 * tree-chrec.c: Include tree.h.
3115 * tree-data-ref.c: Include tree.h.
3116
3117 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3118
3119 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3120 Fix whitespace.
3121
3122 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3123
3124 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3125 Use $(COMPILE) and $(POSTCOMPILE).
3126 (sol2-cxx.o): Likewise.
3127 (sol2-stubs.o): Likewise.
3128 (sol2.o): Likewise.
3129 * config/x-solaris (host-solaris.o): Likewise.
3130
3131 * config/sparc/t-sparc (sparc.o): Remove.
3132 (sparc-c.o): Remove header dependencies.
3133 Use $(COMPILE) and $(POSTCOMPILE).
3134 * config/sparc/x-sparc: Likewise.
3135
3136 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3137
3138 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3139 * config/arc/arc700.md: Likewise.
3140 * config/arc/arc-modes.def: Likewise.
3141 * config/arc/arc-simd.h: Likewise.
3142 * config/arc/t-arc-uClibc: Likewise.
3143 * config/arc/t-arc-newlib: Likewise.
3144
3145 2013-10-02 Renlin Li <renlin.li@arm.com>
3146
3147 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3148 plus_constant.
3149 (aarch64_expand_epilogue): Likewise.
3150
3151 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3152 Yufeng Zhang <yufeng.zhang@arm.com>
3153
3154 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3155 declaration.
3156 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3157 'base_in' represent a conversion and legal_cast_p_1 holds; set
3158 'base_in' with the returned value from get_unwidened.
3159
3160 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3161
3162 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3163 plus_constant is not used.
3164
3165 2013-10-01 Wei Mi <wmi@google.com>
3166
3167 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3168 * config/i386/i386.md: Add define_peephole2 to
3169 break partial reg stall for cvtss2sd/cvtsd2ss.
3170
3171 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3172
3173 * config/arc/arc.c (pass_arc_ifcvt::clone):
3174 Update for ctxt_ -> m_ctxt change.
3175
3176 2013-10-01 Jeff Law <law@redhat.com>
3177
3178 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3179 outgoing_edge and intermediate_edge fields. Instead store the path.
3180 (redirection_data::hash): Hash on the last edge's destination index.
3181 (redirection_data::equal): Check the entire thread path.
3182 (lookup_redirectio_data): Corresponding changes.
3183 (create_edge_and_update_destination_phis): Likewise.
3184 (thread_single_edge): Likewise.
3185
3186 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3187 Diego Novillo <dnovillo@google.com>
3188
3189 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3190 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3191 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3192 versions of these patterns.
3193
3194 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3195 (__builtin_arc_aligned): Likewise.
3196
3197 * config/arc/arc.md: Expand adc_0 comment stating the intended
3198 purpose and why it isn't ready.
3199 Replace commented out call_value_via_label_mixed with a
3200 plain comment about bl_s.
3201
3202 * config/arc/arc.c (stdio.h): Don't include directly.
3203 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3204 Remove if (1) condition.
3205 (arc_encode_section_info): Fix comment.
3206
3207 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3208
3209 * config/arc/arc.c (arc_conditional_register_usage):
3210 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3211 Also set reg_alloc_order for DMA config regs.
3212
3213 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3214 Jeremy Bennett <jeremy.bennett@embecosm.com>
3215
3216 * doc/install.texi (--with-cpu): Mention ARC.
3217 (arc-*-elf32): New paragraph.
3218 (arc-linux-uclibc): Likewise.
3219 * doc/md.texi (Machine Constraints): Add ARC part.
3220 * doc/invoke.texi: (menu): Add ARC Options.
3221 (Machine Dependent Options) <ARC Options>: Add synopsis.
3222 (node ARC Options): Add.
3223 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3224 (ARC Built-in Functions): New section defining
3225 generic ARC built-in functions.
3226 (ARC SIMD Built-in Functions): New section defining SIMD specific
3227 built-in functions.
3228 (Declaring Attributes of Functions): Extended
3229 description of short_call and long_call attributes for ARC and
3230 added index entries.
3231
3232 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3233 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3234 Joern Rennecke <joern.rennecke@embecosm.com>
3235 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3236 Brendan Kehoe <brendan@zen.org>
3237 Michael Eager <eager@eagercon.com>
3238 Simon Cook <simon.cook@embecosm.com>
3239 Jeremy Bennett <jeremy.bennett@embecosm.com>
3240
3241 * config/arc, common/config/arc: New directories.
3242
3243 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3244 Brendan Kehoe <brendan@zen.org>
3245 Simon Cook <simon.cook@embecosm.com>
3246
3247 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3248
3249 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3250
3251 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3252 * tree-ssa-coalesce.h: New. Move prototype to here.
3253 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3254 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3255 (gimple_can_coalesce_p): Move to...
3256 * gimple.c (gimple_can_coalesce_p): Here.
3257
3258 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3259
3260 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3261 (dump_decl_set): Move to gimple.c.
3262 * gimple.h: Don't include tree-ssa-operands.h.
3263 (dump_decl_set): Add prototype.
3264 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3265 Move to gimple-ssa.h.
3266 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3267 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3268 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3269 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3270 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3271 than PHI_ARG_DEF.
3272 (dump_decl_set): Relocate here.
3273 * gimple-ssa.h: New file.
3274 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3275 Relocate from gimple.h.
3276 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3277 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3278 swap_tree_operands and remove non-ssa path.
3279 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3280 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3281 swap_ssa_operands.
3282 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3283 vect_is_simple_reduction_1): Use swap_ssa_operands.
3284 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3285 (enum need_phi_state): Move to tree-into-ssa.c.
3286 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3287 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3288 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3289 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3290 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3291 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3292 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3293 num_imm_uses): Move to ssa-iterators.h.
3294 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3295 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3296 tree-phinodes.h.
3297 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3298 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3299 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3300 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3301 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3302 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3303 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3304 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3305 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3306 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3307 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3308 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3309 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3310 (set_phi_nodes): Move to tree-phinodes.h.
3311 * tree-ssa-operands.h (enum ssa_op_iter_type,
3312 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3313 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3314 (dump_decl_set): Remove prototype.
3315 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3316 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3317 (set_phi_nodes): Relocate from tree-flow-inline.h.
3318 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3319 tree-flow-inline.h
3320 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3321 include list. Temporarily add gimple.h to include list.
3322 * ssa-iterators.h: New file.
3323 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3324 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3325 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3326 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3327 Relocate from tree-ssa-operands.h.
3328 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3329 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3330 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3331 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3332 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3333 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3334 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3335 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3336 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3337 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3338 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3339 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3340 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3341 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3342 Relocate from tree-flow-inline.h.
3343 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3344
3345 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3346
3347 * aarch64-simd.md
3348 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3349 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3350 Insert '\t' to output template.
3351 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3352 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3353 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3354 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3355
3356 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3357
3358 * doc/install.texi (Host/target specific installation notes for GCC):
3359 Put @anchor before @heading.
3360 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3361 Use @email for email addresses.
3362
3363 2013-10-01 Jeff Law <law@redhat.com>
3364
3365 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3366 a vec. Only delete the path if we create one without successfully
3367 registering a jump thread.
3368 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3369 as a pointer.
3370 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3371 (paths): New vector of jump threading paths.
3372 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3373 (THREAD_PATH): New accessor macro for the entire thread path.
3374 (lookup_redirection_data): Get intermediate and final outgoing edge
3375 from the thread path.
3376 (create_edge_and_update_destination_phis): Copy the threading path.
3377 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3378 jump threading path.
3379 (ssa_redirect_edges): Get edges and block types from the jump threading
3380 path. Free the path vector.
3381 (thread_block): Get edges from the jump threading path. Look at the
3382 entire path to see if we thread to a loop exit. If we cancel a jump
3383 thread request, then free the path vector.
3384 (thread_single_edge): Get edges and block types from the jump threading
3385 path. Free the path vector.
3386 (thread_through_loop_header): Get edges and block types from the jump
3387 threading path. Free the path vector.
3388 (mark_threaded_blocks): Iterate over the vector of paths and store
3389 the path on the appropriate edge. Get edges and block types from the
3390 jump threading path.
3391 (mark_threaded_blocks): Get edges and block types from the jump
3392 threading path. Free the path vector.
3393 (thread_through_all_blocks): Use the vector of paths rather than
3394 a vector of 3-edge sets.
3395 (register_jump_thread): Accept pointer to a path vector rather
3396 than the path vector itself. Store the path vector for later use.
3397 Simplify.
3398
3399 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3400 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3401
3402 PR target/58574
3403 * config/s390/s390.c (s390_split_branches): Modify check for table
3404 jump insns.
3405 (s390_chunkify_start): Rearrange table jump insn check in order to
3406 deal with compare and branch insns correctly.
3407
3408 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3409
3410 PR target/58578
3411 Revert
3412 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3413 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3414 define_insn_and_split.
3415 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3416 (shiftsi3_compare): New pattern.
3417 (rrx): New pattern.
3418 * config/arm/unspecs.md (UNSPEC_RRX): New.
3419
3420 2013-10-01 Alan Modra <amodra@gmail.com>
3421
3422 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3423 casing inout operands.
3424
3425 2013-10-01 Richard Biener <rguenther@suse.de>
3426
3427 PR tree-optimization/58553
3428 * tree-loop-distribution.c (struct partition_s): Add niter member.
3429 (classify_partition): Populate niter member for the partition
3430 and properly identify whether the relevant store happens before
3431 or after the loop exit.
3432 (generate_memset_builtin): Use niter member from the partition.
3433 (generate_memcpy_builtin): Likewise.
3434
3435 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3436
3437 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3438 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3439
3440 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3441
3442 * basic-block.h (edge_list): Prefix member names with "m_".
3443 * context.h (context): Likewise.
3444 * domwalk.h (dom_walker): Likewise.
3445 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3446 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3447 * hash-table.h (hash_table): Likewise.
3448 * machmode.h (bit_field_mode_iterator): Likewise.
3449 * pass_manager.h (pass_list): Likewise.
3450 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3451 * tree-pass.h (pass_data): Likewise.
3452 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3453 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3454 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3455 * asan.c (pass_data_asan): Update accordingly.
3456 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3457 (control_dependences::control_dependences): Likewise.
3458 (control_dependences::~control_dependences): Likewise.
3459 (control_dependences::~control_dependences): Likewise.
3460 (control_dependences::get_edges_dependent_on): Likewise.
3461 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3462 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3463 * context.c (gcc::context::context): Likewise.
3464 * cprop.c (pass_rtl_cprop::clone): Likewise.
3465 * domwalk.c (dom_walker::walk): Likewise.
3466 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3467 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3468 * mode-switching.c (pass_mode_switching::clone): Likewise.
3469 * passes.c (opt_pass::opt_pass): Likewise.
3470 (pass_manager::pass_manager): Likewise.
3471 * predict.c (pass_strip_predict_hints::clone): Likewise.
3472 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3473 (pass_split_all_insns::clone): Likewise.
3474 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3475 Likewise.
3476 (bit_field_mode_iterator::next_mode): Likewise.
3477 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3478 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3479 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3480 * tree-complex.c (pass_lower_complex::clone): Likewise.
3481 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3482 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3483 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3484 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3485 (pass_fold_builtins::clone): Likewise.
3486 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3487 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3488 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3489 (pass_cd_dce::clone): Likewise.
3490 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3491 (pass_phi_only_cprop::clone): Likewise.
3492 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3493 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3494 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3495 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3496 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3497 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3498 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3499 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3500 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3501 * tree-vrp.c (pass_vrp::clone): Likewise.
3502 * tsan.c (pass_tsan::clone): Likewise.
3503
3504 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3505
3506 PR middle-end/58564
3507 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3508 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3509
3510 PR middle-end/58564
3511 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3512 optimization, punt if sign_bit_p looked through any zero extension.
3513
3514 2013-09-30 Teresa Johnson <tejohnson@google.com>
3515
3516 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3517 Update redirected out edge count in joiner case.
3518 (ssa_redirect_edges): Common the joiner and non-joiner cases
3519 so that joiner case gets profile updates.
3520
3521 2013-09-30 Richard Biener <rguenther@suse.de>
3522
3523 PR tree-optimization/58554
3524 * tree-loop-distribution.c (classify_partition): Require
3525 unconditionally executed stores for memcpy and memset recognition.
3526 (tree_loop_distribution): Calculate dominance info.
3527
3528 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3529
3530 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3531 (NO_PROFILE_COUNTERS): Likewise.
3532 (PROFILE_HOOK): Likewise.
3533 (FUNCTION_PROFILER): Likewise.
3534 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3535
3536 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3537
3538 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3539 calls and defines in TARGET_MACHO conditional.
3540 (load_macho_picbase_di): Likewise.
3541 (reload_macho_picbase): Likewise.
3542 (reload_macho_picbase_si): Likewise.
3543 (reload_macho_picbase_di): Likewise.
3544 (nonlocal_goto_receiver): Likewise.
3545
3546 2013-09-30 Nick Clifton <nickc@redhat.com>
3547
3548 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3549 that use the MSP430X ISA.
3550 (msp430_option_override): Scan -mmcu command line option for any
3551 MCU name that supports the MSP430X ISA.
3552 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3553 -mmcu options which enable the MSP430X ISA.
3554
3555 2013-09-30 Richard Biener <rguenther@suse.de>
3556
3557 PR middle-end/58532
3558 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3559 before looking for setjmp-like calls.
3560
3561 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3562
3563 PR target/10901
3564 * config/darwin-protos.h (machopic_get_function_picbase): New.
3565 * config/darwin.c (machopic_get_function_picbase): New.
3566 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3567 label for a new func. (load_macho_picbase_di): Likewise.
3568 (reload_macho_picbase): New expand.
3569 (reload_macho_picbase_si): New insn.
3570 (reload_macho_picbase_di): New insn.
3571 (nonlocal_goto_receiver): New define and split.
3572 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3573 (unspecv enum): Add UNSPECV_NLGR.
3574
3575 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3576
3577 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3578 that altivec registers are correctly sized on Darwin.
3579
3580 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3581
3582 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3583 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3584 * config/x-darwin (host-darwin.o): Likewise.
3585 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3586 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3587 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3588
3589 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3590
3591 * doc/invoke.texi: Fix usage of @tie{} command.
3592
3593 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3594
3595 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3596
3597 2013-09-28 Jan Hubicka <jh@suse.cz>
3598
3599 * config/i386/x86-tune.def: Add documentation for each of the options;
3600 add whitespace.
3601
3602 2013-09-28 Jan Hubicka <jh@suse.cz>
3603
3604 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3605 generic.
3606 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3607 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3608 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3609 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3610 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3611
3612 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3613
3614 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3615 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3616 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3617 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3618 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3619 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3620 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3621 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3622 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3623 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3624 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3625 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3626 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3627 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3628 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3629 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3630 gimple.h, godump.c, graphite-clast-to-gimple.c,
3631 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3632 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3633 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3634 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3635 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3636 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3637 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3638 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3639 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3640 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3641 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3642 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3643 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3644 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3645 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3646 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3647 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3648 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3649 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3650 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3651 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3652 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3653 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3654 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3655 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3656 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3657 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3658 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3659 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3660 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3661 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3662 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3663 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3664 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3665 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3666 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3667 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3668 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3669 whitespace before "(".
3670
3671 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3672
3673 * expr.h (extract_bit_field): Remove packedp parameter.
3674 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3675 from forward declaration.
3676 (store_split_bit_field): Remove packedp arg from calls to
3677 extract_fixed_bit_field.
3678 (extract_bit_field_1): Remove packedp parameter and packedp
3679 argument from recursive calls and calls to extract_fixed_bit_field.
3680 (extract_bit_field): Remove packedp parameter and corresponding
3681 arg to extract_bit_field_1.
3682 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3683 to issue warnings.
3684 (extract_split_bit_field): Remove packedp arg from call to
3685 extract_fixed_bit_field.
3686 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3687 (copy_blkmode_from_reg): Likewise.
3688 (copy_blkmode_to_reg): Likewise.
3689 (read_complex_part): Likewise.
3690 (store_field): Likewise.
3691 (expand_expr_real_1): Likewise.
3692 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3693 to extract_bit_field.
3694 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3695 call to extract_bit_field.
3696 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3697 call to extract_bit_field.
3698 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3699 and special packedp behavior from -fstrict-volatile-bitfields
3700 documentation.
3701
3702 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3703
3704 * lra-eliminations.c (init_elim_table): Guard value_p.
3705
3706 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3707
3708 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3709 DFmode, DImode, and SFmode in the upper VSX registers based on the
3710 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3711 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3712 of the VSX register handling.
3713
3714 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3715 (f32_sv): Likewise.
3716 (zero_extendsidi2_lfiwzx): Add support for loading into the
3717 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3718 only do VSX memory options on Altivec registers.
3719 (extendsidi2_lfiwax): Likewise.
3720 (extendsfdf2_fpr): Likewise.
3721 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3722 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3723 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3724 (movdi_internal64): Likewise.
3725
3726 2013-09-27 Xinliang David Li <davidxl@google.com>
3727
3728 * opts.c (finish_options): Adjust parameters
3729 according to vect cost model.
3730 (common_handle_option): Set dynamic vect cost
3731 model for FDO.
3732 targhooks.c (default_add_stmt_cost): Compute stmt cost
3733 unconditionally.
3734 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3735 Use helper function.
3736 * tree-vectorizer.h (unlimited_cost_model): New function.
3737 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3738 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3739 function.
3740 (vect_enhance_data_refs_alignment): Ditto.
3741 * flag-types.h: New enum.
3742 * common/config/i386/i386-common.c (ix86_option_init_struct):
3743 No need to initialize vect_cost_model flag.
3744 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3745 unconditionally.
3746
3747 2013-09-27 Diego Novillo <dnovillo@google.com>
3748
3749 * gimple.h (enum ssa_mode): Remove.
3750
3751 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3752
3753 * cfgloop.h (number_of_loops): Fix typo in check for null.
3754
3755 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3756
3757 PR middle-end/58551
3758 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3759 are children of outermost saved_cfun's loop, and set it up to
3760 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3761 if loop != loop0 and SESE region contains bbs that belong to loop0.
3762
3763 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3764
3765 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3766 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3767 (set_address_segment, set_address_base, set_address_index)
3768 (set_address_disp): Accept the argument unconditionally.
3769 (baseness): Remove must_be_base_p and must_be_index_p checks.
3770 (decompose_normal_address): Classify as much as possible in the
3771 main loop.
3772
3773 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3774
3775 * cse.c (count_reg_usage): Handle INT_LIST.
3776 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3777 * reginfo.c (reg_scan_mark_refs): Likewise.
3778 * reload1.c (eliminate_regs_1): Likewise.
3779
3780 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3781
3782 PR middle-end/58547
3783 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3784 signed.
3785
3786 2013-09-27 Richard Biener <rguenther@suse.de>
3787
3788 PR tree-optimization/58459
3789 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3790 restriction not propagating into loops.
3791
3792 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3793
3794 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3795 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3796 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3797
3798 2013-09-26 Richard Biener <rguenther@suse.de>
3799
3800 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3801
3802 2013-09-26 Richard Biener <rguenther@suse.de>
3803
3804 * alias.h (component_uses_parent_alias_set): Rename to ...
3805 (component_uses_parent_alias_set_from): ... this.
3806 * alias.c (component_uses_parent_alias_set): Rename to ...
3807 (component_uses_parent_alias_set_from): ... this and return
3808 the desired parent.
3809 (reference_alias_ptr_type_1): Use the result from
3810 component_uses_parent_alias_set_from instead of stripping
3811 components one at a time.
3812 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3813
3814 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3815
3816 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3817 Move prototypes to...
3818 * tree-ssa-ter.h: New File. Move prototypes here.
3819 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3820 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3821 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3822 from is_replaceable_p.
3823 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3824 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3825 (process_replaceable): Use ter_is_replaceable_p.
3826 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3827 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3828 newly refactored ssa_is_replaceable_p.
3829 * cfgexpand.c: Include tree-outof-ssa.h.
3830 * ssaexpand.h: Delete.
3831
3832 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3833
3834 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3835 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3836 tree-ssa.c
3837 (create_gimple_tmp): Delete.
3838 (get_expr_type, build_assign, build_type_cast): Move to...
3839 * gimple-builder.c: New File.
3840 (get_expr_type): Relocate from gimple.c.
3841 (build_assign, build_type_cast): Change to only create ssanames.
3842 * gimple.h: Move prototypes to...
3843 * gimple-builder.h: New File. Here.
3844 * tree-ssa.h: And here.
3845 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3846 count_uses_and_derefs): Relocate from gimple.c.
3847 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3848 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3849 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3850 gimple_replace_ssa_lhs.
3851 * asan.c: Include gimple-builder.h.
3852 * Makefile.in: Add gimple-builder.o.
3853
3854 2013-09-26 Richard Biener <rguenther@suse.de>
3855
3856 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3857 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3858 (loe_visit_block): Use gcc_checking_assert.
3859 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3860 gimple_assign_ssa_name_copy_p.
3861 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3862 change.
3863
3864 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3865
3866 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3867 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3868
3869 2013-09-26 Richard Biener <rguenther@suse.de>
3870
3871 PR tree-optimization/58539
3872 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3873 the fact that debug statements are not taking part in loop-closed
3874 SSA construction.
3875
3876 2013-09-26 Nick Clifton <nickc@redhat.com>
3877
3878 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3879 time warning message.
3880 (msp430_print_operand_raw): Delete unused letter parameter.
3881 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3882 (msp430_print_operand_address): New function.
3883 (msp430_print_operand): Move address printing code from here to
3884 new function.
3885 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3886 assembler.
3887 (zero_extendpsisi2): Likewise.
3888 (extendpsisi2): New pattern.
3889 (andneghi3): New pattern.
3890
3891 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3892
3893 * config/aarch64/aarch64.opt (mlra): New option.
3894 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3895 (TARGET_LRA_P): Define.
3896
3897 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3898
3899 * expr.c (expand_assignment): Remove obsolete comment.
3900
3901 2013-09-25 Jeff Law <law@redhat.com>
3902
3903 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3904 tree-ssa-threadupdate.h.
3905 (register_jump_thread): Similarly.
3906 * tree-ssa-threadupdate.h: New header file.
3907 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3908 * tree-vrp.c: Likewise.
3909 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3910 (thread_around_empty_blocks): Change type of path vector argument to
3911 an edge,type pair from just an edge. Initialize both elements when
3912 appending to a jump threading path. Tweak references to elements
3913 appropriately.
3914 (thread_across_edge): Similarly. Release memory for the elements
3915 as needed.
3916 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3917 (dump_jump_thread_path): New function broken out from
3918 register_jump_thread.
3919 (register_jump_thread): Use dump_jump_thread_path. Change type of
3920 path vector entries. Search the path for NULL edges and dump
3921 the path if one is found. Tweak the conversion of path to 3-edge
3922 form to use the block copy type information embedded in the path.
3923
3924 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3925
3926 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3927
3928 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3929 Vladimir Makarov <vmakarov@redhat.com>
3930
3931 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3932 from the least significant bit.
3933 (strip_address_mutations): Add bitfield operations handling.
3934 (must_be_index_p): Add shifting and rotate operations handling.
3935 (set_address_base): Use must_be_base_p predicate.
3936 (set_address_index): Use must_be_index_p predicate.
3937
3938 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3939 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3940 Sergey Lega <sergey.s.lega@intel.com>
3941 Anna Tikhonova <anna.tikhonova@intel.com>
3942 Ilya Tocar <ilya.tocar@intel.com>
3943 Andrey Turetskiy <andrey.turetskiy@intel.com>
3944 Ilya Verbin <ilya.verbin@intel.com>
3945 Kirill Yukhin <kirill.yukhin@intel.com>
3946 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3947
3948 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3949 Use new names.
3950 (ix86_expand_vector_move_misalign): Support new unaligned load and
3951 stores and use new names.
3952 (CODE_FOR_sse2_storedqu): Rename to ...
3953 (CODE_FOR_sse2_storedquv16qi): ... this.
3954 (CODE_FOR_sse2_loaddqu): Rename to ...
3955 (CODE_FOR_sse2_loaddquv16qi): ... this.
3956 (CODE_FOR_avx_loaddqu256): Rename to ...
3957 (CODE_FOR_avx_loaddquv32qi): ... this.
3958 (CODE_FOR_avx_storedqu256): Rename to ...
3959 (CODE_FOR_avx_storedquv32qi): ... this.
3960 * config/i386/i386.md (fpint_logic): New.
3961 * config/i386/sse.md (VMOVE): Extend for AVX512.
3962 (VF): Ditto.
3963 (VF_128_256): New.
3964 (VF_512): Ditto.
3965 (VI_UNALIGNED_LOADSTORE): Ditto.
3966 (sse2_avx_avx512f): Ditto.
3967 (sse2_avx2): Extend for AVX512.
3968 (sse4_1_avx2): Ditto.
3969 (avx2_avx512f): New.
3970 (sse): Extend for AVX512.
3971 (sse2): Ditto.
3972 (sse4_1): Ditto.
3973 (avxsizesuffix): Ditto.
3974 (sseintvecmode): Ditto.
3975 (ssePSmode): Ditto.
3976 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3977 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3978 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3979 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3980 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3981 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3982 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3983 (STORENT_MODE): Extend for AVX512.
3984 (*absneg<mode>2): Replace constraint "x" with "v".
3985 (*mul<mode>3): Ditto.
3986 (*ieee_smin<mode>3): Ditto.
3987 (*ieee_smax<mode>3): Ditto.
3988 (avx_cmp<mode>3): Replace VF with VF_128_256.
3989 (*<sse>_maskcmp<mode>3_comm): Ditto.
3990 (<sse>_maskcmp<mode>3): Ditto.
3991 (<sse>_andnot<mode>3): Extend for AVX512.
3992 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3993 (<code><mode>3, fpint_logic): New.
3994 (*<code><mode>3): Extend for AVX512.
3995 (avx512flogicsuff): New.
3996 (avx512f_<logic><mode>): Ditto.
3997 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3998 VF_128_256.
3999 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4000 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4001 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4002 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4003 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4004 (xop_vpermil2<mode>3): Ditto.
4005 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4006 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4007 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4008 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4009
4010 2013-09-25 Tom Tromey <tromey@redhat.com>
4011
4012 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4013 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4014 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4015 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4016 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4017 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4018 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4019 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4020 (GRAPHITE_HTAB_H): Remove.
4021
4022 2013-09-25 Tom Tromey <tromey@redhat.com>
4023
4024 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4025
4026 2013-09-25 Tom Tromey <tromey@redhat.com>
4027
4028 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4029
4030 2013-09-25 Tom Tromey <tromey@redhat.com>
4031
4032 * config/i386/t-i386 (i386.o): Remove.
4033 (i386-c.o): Use COMPILE and POSTCOMPILE.
4034
4035 2013-09-25 Tom Tromey <tromey@redhat.com>
4036
4037 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4038
4039 2013-09-25 Tom Tromey <tromey@redhat.com>
4040
4041 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4042 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4043 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4044 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4045 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4046 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4047 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4048 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4049 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4050 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4051 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4052 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4053 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4054 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4055 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4056 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4057 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4058 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4059 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4060 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4061 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4062 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4063 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4064 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4065 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4066 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4067 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4068 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4069 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4070 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4071 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4072 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4073 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4074 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4075 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4076 (graphite-clast-to-gimple.o, graphite-dependences.o)
4077 (graphite-interchange.o, graphite-poly.o)
4078 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4079 (graphite-optimize-isl.o, tree-vect-loop.o)
4080 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4081 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4082 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4083 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4084 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4085 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4086 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4087 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4088 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4089 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4090 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4091 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4092 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4093 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4094 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4095 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4096 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4097 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4098 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4099 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4100 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4101 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4102 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4103 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4104 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4105 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4106 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4107 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4108 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4109 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4110 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4111 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4112 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4113 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4114 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4115 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4116 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4117 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4118 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4119 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4120 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4121 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4122 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4123 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4124 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4125 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4126 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4127 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4128 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4129 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4130 (gcov-dump.o): Remove.
4131 (default-c.o): Use COMPILE and POSTCOMPILE.
4132 (CFLAGS-gcc.o): New variable.
4133 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4134
4135 2013-09-25 Tom Tromey <tromey@redhat.com>
4136
4137 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4138 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4139 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4140 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4141 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4142 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4143 (c-family/array-notation-common.o, c-family/stub-objc.o)
4144 (c-family/c-ubsan.o): Remove.
4145
4146 2013-09-25 Tom Tromey <tromey@redhat.com>
4147
4148 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4149
4150 2013-09-25 Tom Tromey <tromey@redhat.com>
4151
4152 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4153 to add -DENABLE_SHARED_LIBGCC.
4154 (gcc.o): Don't use subshell.
4155
4156 2013-09-25 Tom Tromey <tromey@redhat.com>
4157
4158 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4159 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4160 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4161 * configure, config.in: Rebuild.
4162
4163 2013-09-25 Tom Tromey <tromey@redhat.com>
4164
4165 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4166 (COMPILE, POSTCOMPILE): New variables.
4167 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4168 (DEPFILES): New variable.
4169 Include ".Po" files.
4170 * configure.ac: Add checks for dependency checking.
4171 * configure, aclocal.m4: Regenerate.
4172
4173 2013-09-25 Tom Tromey <tromey@redhat.com>
4174
4175 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4176 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4177
4178 2013-09-25 Tom Tromey <tromey@redhat.com>
4179
4180 * Makefile.in (generated_files): Add options.h,
4181 target-hooks-def.h, insn-opinit.h,
4182 common/common-target-hooks-def.h, pass-instances.def,
4183 c-family/c-target-hooks-def.h.
4184
4185 2013-09-25 Jeff Law <law@redhat.com>
4186
4187 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4188 than foo[foo.length () - 1] to access last member in a vec.
4189 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4190
4191 2013-09-25 Richard Biener <rguenther@suse.de>
4192
4193 PR middle-end/58521
4194 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4195
4196 2013-09-25 Jan Hubicka <jh@suse.cz>
4197
4198 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4199 test.
4200
4201 2013-09-25 Marek Polacek <polacek@redhat.com>
4202
4203 PR sanitizer/58420
4204 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4205 when determining the type name.
4206
4207 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4208
4209 * config/sh/sh.md: Fix formatting.
4210
4211 2013-09-24 Xinliang David Li <davidxl@google.com>
4212
4213 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4214 max peel iterations parameter.
4215 * param.def: New parameter.
4216 * doc/invoke.texi: Document New parameter.
4217
4218 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4219
4220 * gimple-pretty-print.c: Various whitespace tweaks.
4221 * tree-core.h: Likewise.
4222 * tree-pretty-print.c: Likewise.
4223 * tree-ssa-alias.c: Likewise.
4224 * tree-ssa-copy.c: Likewise.
4225 * tree-ssanames.c: Likewise.
4226 * tree-ssanames.h: Likewise.
4227 * tree-vrp.c: Likewise.
4228
4229 2013-09-24 Alan Modra <amodra@gmail.com>
4230
4231 PR middle-end/57134
4232 PR middle-end/57586
4233 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4234 for output operands that disallow regs. Don't use EXPAND_WRITE on
4235 inout operands.
4236
4237 2013-09-24 Richard Biener <rguenther@suse.de>
4238
4239 PR middle-end/58513
4240 * tree.c (reference_alias_ptr_type): Move ...
4241 * alias.c (reference_alias_ptr_type): ... here and implement
4242 in terms of the new reference_alias_ptr_type_1.
4243 (ref_all_alias_ptr_type_p): New helper.
4244 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4245 use ref_all_alias_ptr_type_p.
4246 (get_deref_alias_set): Add flag_strict_aliasing check here.
4247 (reference_alias_ptr_type_1): New function, split out from ...
4248 (get_alias_set): ... here.
4249 (alias_ptr_types_compatible_p): New function.
4250 * alias.h (reference_alias_ptr_type): Declare.
4251 (alias_ptr_types_compatible_p): Likewise.
4252 * tree.h (reference_alias_ptr_type): Remove.
4253 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4254 to compare MEM_REF alias types.
4255
4256 2013-09-24 Richard Biener <rguenther@suse.de>
4257
4258 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4259
4260 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4261
4262 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4263 reload helper function arrays into a single array reg_addr.
4264 (reload_fpr_gpr): Likewise.
4265 (reload_gpr_vsx): Likewise.
4266 (reload_vsx_gpr): Likewise.
4267 (struct rs6000_reg_addr): Likewise.
4268 (reg_addr): Likewise.
4269 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4270 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4271 (rs6000_init_hard_regno_mode_ok): Likewise.
4272 (rs6000_secondary_reload_direct_move): Likewise.
4273 (rs6000_secondary_reload): Likewise.
4274
4275 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4276 constraints: wu, ww, and wy. Repurpose wv constraint added during
4277 power8 changes. Put wg constraint in alphabetical order.
4278
4279 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4280 for future work to add ISA 2.07 VSX single precision support.
4281 (-mvsx-scalar-double): Change default from -1 to 1, update
4282 documentation comment.
4283 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4284 (-mupper-regs-df): New debug switch to control whether DF values
4285 can go in the traditional Altivec registers.
4286 (-mupper-regs-sf): New debug switch to control whether SF values
4287 can go in the traditional Altivec registers.
4288
4289 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4290 and wy constraints.
4291 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4292 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4293 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4294 (rs6000_debug_legitimate_address_p): Print if we are running
4295 before, during, or after reload.
4296 (rs6000_secondary_reload): Add a comment.
4297 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4298
4299 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4300 constraints. Update documentation string.
4301 (wd constraint): Likewise.
4302 (wf constraint): Likewise.
4303 (wg constraint): Likewise.
4304 (wn constraint): Likewise.
4305 (ws constraint): Likewise.
4306 (wt constraint): Likewise.
4307 (wx constraint): Likewise.
4308 (wz constraint): Likewise.
4309 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4310 instructions.
4311 (ww constraint): Likewise.
4312 (wy constraint): Likewise.
4313 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4314 in the previous submissions.
4315 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4316
4317 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4318
4319 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4320 (REG_BR_PROB): Say that the probability is stored in an int_list.
4321 * reg-notes.def: Update commentary to mention INT_LIST.
4322 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4323 (INT_LIST): New rtx.
4324 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4325 * rtlanal.c (int_reg_note_p): New function.
4326 (alloc_reg_note): Assert that the note does not have an int argument.
4327 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4328 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4329 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4330 rather than an INSN_LIST. Handle INT_LIST.
4331 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4332 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4333 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4334 Manipulate them as ints rather than rtxes.
4335 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4336 * regmove.c (copy_src_to_dest): Likewise.
4337 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4338
4339 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4340 into the cases that need it.
4341 * config/arm/arm.c (arm_unwind_emit): Likewise.
4342
4343 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4344 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4345 * loop-doloop.c (add_test, doloop_modify): Likewise.
4346 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4347 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4348 * predict.c (combine_predictions_for_insn): Likewise.
4349 * print-rtl.c (print_rtx): Handle INT_LIST.
4350 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4351 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4352 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4353 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4354 (ix86_split_fp_branch, predict_jump): Likewise.
4355 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4356 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4357 * config/spu/spu.c (ea_load_store_inline): Likewise.
4358
4359 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4360 value of a REG_BR_PROB note.
4361 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4362 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4363 * emit-rtl.c (try_split): Likewise.
4364 * predict.c (br_prob_note_reliable_p): Likewise.
4365 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4366 * reorg.c (mostly_true_jump): Likewise.
4367 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4368 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4369 * config/i386/i386.c (ix86_print_operand): Likewise.
4370 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4371 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4372 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4373 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4374 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4375 * config/sparc/sparc.c (output_cbranch): Likewise.
4376 * config/spu/spu.c (get_branch_target): Likewise.
4377 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4378 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4379
4380 2013-09-23 Jan Hubicka <jh@suse.cz>
4381
4382 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4383 for ipa-devirt.
4384 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4385 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4386 of external calls
4387 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4388 (gimple_fold_call): Dump inconsistent devirtualizations; add
4389 sanity check for type based devirtualizations.
4390 * ipa-prop.c: Include ipa-utils.h
4391 (ipa_intraprocedural_devirtualization): Add sanity check.
4392 (try_make_edge_direct_virtual_call): Likewise.
4393
4394 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4395
4396 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4397 assignment statements.
4398
4399 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4400
4401 * gimple-pretty-print.c (dump_ssaname_info): New function.
4402 (dump_gimple_phi): Call it.
4403 (pp_gimple_stmt_1): Likewise.
4404 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4405 (range_info_def): Declare.
4406 * tree-pretty-print.c (pp_double_int): New function.
4407 (dump_generic_node): Call it.
4408 * tree-pretty-print.h (pp_double_int): Declare.
4409 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4410 * tree-ssanames.h (range_info_def): New structure.
4411 (value_range_type): Move definition here.
4412 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4413 Declare.
4414 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4415 initialization.
4416 (set_range_info): New function.
4417 (get_range_info): Likewise.
4418 (duplicate_ssa_name_range_info): Likewise.
4419 (duplicate_ssa_name_fn): Check pointer type and call
4420 duplicate_ssa_name_range_info.
4421 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4422 * tree-vrp.c (value_range_type): Remove definition, now in
4423 tree-ssanames.h.
4424 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4425 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4426 (SSA_NAME_RANGE_INFO): New macro.
4427
4428 2013-09-23 Richard Biener <rguenther@suse.de>
4429
4430 PR tree-optimization/58464
4431 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4432 (phi_trans_add): Change to add conditionally on being not present.
4433 (phi_translate_1): Remove recursion detection here.
4434 (phi_translate): Pre-seed the cache with NULL to catch
4435 recursion here in a more generic way.
4436 (bitmap_find_leader): Adjust comment.
4437 (get_representative_for): Dump value-numbers.
4438 (create_expression_by_pieces): Likewise.
4439 (insert_into_preds_of_block): Likewise.
4440
4441 2013-09-23 Christian Bruel <christian.bruel@st.com>
4442
4443 PR target/58475
4444 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4445 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4446
4447 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4448
4449 Revert r202780:
4450 2013-09-20 Renlin Li <renlin.li@arm.com>
4451
4452 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4453 plus_constant.
4454 (aarch64_expand_epilogue): Likewise.
4455 (aarch64_legitimize_reload_address): Likewise.
4456
4457 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4458
4459 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4460 NULL_TREE before pushing them onto the vector. Likewise for labels.
4461
4462 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4463
4464 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4465 comparison operators when -fno-trapping-math is in effect.
4466 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4467 comparison operators in TFmode and assert that unsupported operators
4468 cannot reach here.
4469 (ia64_print_operand): Likewise.
4470
4471 2013-09-21 Jan Hubicka <jh@suse.cz>
4472
4473 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4474 (sse_typeless_stores): Enable for core
4475 (sse_load0_by_pxor): Likewise.
4476 (four_jump_limit): Disable for core.
4477 (pad_returns): Likewise.
4478 (avoid_vector_decode): Likewise.
4479 (fuse_cmp_and_branch): Enable for cores.
4480 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4481
4482 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4483
4484 PR middle-end/56791
4485 * config/pa/pa.c (pa_option_override): Disable auto increment and
4486 decrement instructions until reload is completed.
4487
4488 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4489 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4490 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4491
4492 2013-09-20 DJ Delorie <dj@redhat.com>
4493 Nick Clifton <nickc@redhat.com>
4494
4495 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4496 (need_to_save): Save bank 0 on interrupts.
4497 (characterize_address): Strip far address wrappers.
4498 (rl78_as_legitimate_address): Likewise.
4499 (transcode_memory_rtx): Likewise.
4500 (rl78_peep_movhi_p): Disable this peephole after devirt.
4501 (rl78_propogate_register_origins): Forget all origins when a
4502 CLOBBER is seen.
4503 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4504 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4505 peephole2's.
4506 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4507 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4508 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4509 relocating.
4510 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4511
4512 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4513
4514 * config/pa/pa.md: In "scc" insn patterns, change output template to
4515 handle const0_rtx in reg_or_0_operand operands.
4516
4517 2013-09-20 Martin Husemann <martin@NetBSD.org>
4518
4519 PR target/56875
4520 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4521 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4522
4523 2013-09-20 Richard Biener <rguenther@suse.de>
4524
4525 PR middle-end/58484
4526 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4527 remembering SSA name version and block index.
4528 (new_scev_info_str): Adjust.
4529 (hash_scev_info): Likewise. Also hash the block index.
4530 (eq_scev_info): Adjust.
4531 (find_var_scev_info): Likewise.
4532 (struct instantiate_cache_entry): Remove.
4533 (struct instantiate_cache_type): Use a htab to map name, block
4534 to chrec.
4535 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4536 (get_instantiated_value_entry): Likewise.
4537 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4538 (instantiate_scev_name): Adjust.
4539
4540 2013-09-20 Jeff Law <law@redhat.com>
4541
4542 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4543
4544 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4545
4546 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4547 Call aarch64_simd_expand_args to update op[argc].
4548
4549 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4550
4551 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4552 plugin argument.
4553
4554 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4555
4556 * gengtype.c (file_rules): Added rule for *.cc files.
4557 (get_output_file_with_visibility): Give fatal message when no
4558 rules found.
4559
4560 2013-09-20 Renlin Li <renlin.li@arm.com>
4561
4562 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4563 (aarch64_expand_epilogue): Likewise.
4564 (aarch64_legitimize_reload_address): Likewise.
4565
4566 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4567
4568 PR middle-end/57748
4569 * expr.c (expand_assignment): Remove misalignp code path.
4570
4571 2013-09-20 Marek Polacek <polacek@redhat.com>
4572
4573 PR sanitizer/58413
4574 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4575 TYPE_PRECISION. Add asserts.
4576
4577 2013-09-20 Richard Biener <rguenther@suse.de>
4578
4579 PR tree-optimization/58453
4580 * tree-loop-distribution.c (distribute_loop): Apply the cost
4581 model for -ftree-loop-distribute-patterns, too.
4582
4583 2013-09-20 Richard Biener <rguenther@suse.de>
4584
4585 PR middle-end/58473
4586 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4587 make type comparison less strict.
4588
4589 2013-09-20 Alan Modra <amodra@gmail.com>
4590
4591 * configure: Regenerate.
4592 * aclocal.m4: Regenerate.
4593
4594 2013-09-20 Marek Polacek <polacek@redhat.com>
4595
4596 PR other/58467
4597 * doc/extend.texi: Document that attribute used is meant to be used
4598 on variables with static storage duration.
4599
4600 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4601
4602 PR tree-optimization/58472
4603 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4604 simd_lane_access set inv_p = false.
4605 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4606 the simduid magic VAR_DECL.
4607
4608 2013-09-19 Jan Hubicka <jh@suse.cz>
4609
4610 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4611
4612 2013-09-17 Jeff Law <law@redhat.com>
4613
4614 * tree-ssa-dom.c (record_temporary_equivalences): New function
4615 split out of dom_opt_dom_walker::after_dom_children.
4616 (dom_opt_dom_walker::thread_across_edge): Move common code
4617 in here from dom_opt_dom_walker::after_dom_children.
4618 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4619
4620 2013-09-19 Jan Hubicka <jh@suse.cz>
4621
4622 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4623 (TARGET_GENERIC): Use PROCESOR_GENERIC
4624 (enum processor_type): Unify generic32 and 64.
4625 * i386.md (cpu): Likewise.
4626 * x86-tune.def (use_leave): Enable for generic32.
4627 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4628 * athlon.md: Change generic64 to generic in all occurences.
4629 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4630 (ix86_target_macros_internal): Likewise.
4631 * driver-i386.c (host_detect_local_cpu): Likewise.
4632 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4633 to ..
4634 (generic_memcpy, generic_memset, generic_cost): This one.
4635 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4636 (m_GENERIC32, m_GENERIC64): Remove.
4637 (m_GENERIC): Turn into one flag.
4638 (processor_target): Unify generic tunnings.
4639 (ix86_option_override_internal): Replace generic32/64 by generic.
4640 (ix86_issue_rate): Likewise.
4641 (ix86_adjust_cost): Likewise.
4642
4643 2013-09-19 Jan Hubicka <jh@suse.cz>
4644
4645 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4646 of speculative flag.
4647
4648 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4649
4650 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4651 GOMP_sections_start, even if !exit_reachable.
4652
4653 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4654
4655 * lra-constraints.c (need_for_all_save_p): Use macro
4656 HARD_REGNO_CALL_PART_CLOBBERED.
4657 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4658 set up pseudo conflict hard regs.
4659
4660 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4661
4662 PR target/58452
4663 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4664 operands.
4665
4666 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4667
4668 PR rtl-optimization/58438
4669 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4670 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4671 from previous subpasses.
4672
4673 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4674
4675 * arm.c (arm_get_frame_offsets): Validate architecture supports
4676 LDRD/STRD before accepting the tuning preference.
4677 (arm_expand_prologue): Likewise.
4678 (arm_expand_epilogue): Likewise.
4679
4680 2013-09-18 Richard Biener <rguenther@suse.de>
4681
4682 PR tree-optimization/58417
4683 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4684 have chrecs with symbols defined in the loop as operands.
4685 (chrec_fold_multiply): Likewise.
4686 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4687 parameters before folding binary operations.
4688 (struct instantiate_cache_entry_hasher): Remove.
4689 (struct instantiate_cache_type): Use a pointer-map.
4690 (instantiate_cache_type::instantiate_cache_type): New function.
4691 (instantiate_cache_type::get): Likewise.
4692 (instantiate_cache_type::set): Likewise.
4693 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4694 (get_instantiated_value_entry): Likewise.
4695 (global_cache): New global.
4696 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4697 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4698 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4699 (instantiate_scev_name): Adjust.
4700 (instantiate_scev): Construct global instead of local cache.
4701 (resolve_mixers): Likewise.
4702
4703 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4704 Paolo Carlini <paolo.carlini@oracle.com>
4705
4706 PR c++/58458
4707 * doc/implement-cxx.texi: Fix references to the C++ standards.
4708
4709 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4710
4711 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4712 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4713 simd_lane_access drs, update also DR_ALIGNED_TO.
4714
4715 2013-09-18 Marek Polacek <polacek@redhat.com>
4716
4717 PR sanitizer/58411
4718 * doc/extend.texi: Document no_sanitize_undefined attribute.
4719 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4720 no_sanitize_undefined attribute.
4721
4722 2013-09-18 Nick Clifton <nickc@redhat.com>
4723
4724 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4725 (ASM_DECLARE_FUNCTION_NAME): Define.
4726
4727 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4728
4729 * compare-elim.c (find_comparison_dom_walker): New class
4730 (find_comparisons_in_bb): Rename to
4731 find_comparison_dom_walker::before_dom_children
4732 (find_comparisons): Adjust
4733 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4734 adjust.
4735 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4736 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4737 (init_walk_dominator_tree): Remove declaration.
4738 (fini_walk_dominator_tree): Remove declaration.
4739 * fwprop.c (single_def_use_dom_walker): New class
4740 (single_def_use_enter_block): Convert to
4741 single_def_use_dom_walker::before_dom_children.
4742 (single_def_use_leave_block): Convert to
4743 single_def_use_dom_walker::after_dom_children.
4744 (build_single_def_use_links): Adjust.
4745 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4746 class.
4747 (find_candidates_in_block): Convert to
4748 find_candidates_dom_walker::before_dom_children.
4749 (execute_strength_reduction): Adjust.
4750 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4751 (sese_dom_walker): New class.
4752 (sese_dom_walker::sese_dom_walker): New constructor.
4753 (sese_dom_walker::~sese_dom_walker): New destructor.
4754 (build_sese_conditions_before): Convert to
4755 sese_dom_walker::before_dom_children.
4756 (build_sese_conditions_after): Convert to
4757 sese_dom_walker::after_dom_children.
4758 (build_poly_scop): Adjust
4759 * tree-into-ssa.c (rewrite_dom_walker): New class
4760 (rewrite_enter_block): Convert to
4761 rewrite_dom_walker::before_dom_children.
4762 (rewrite_leave_block): Convert to
4763 rewrite_dom_walker::after_dom_children.
4764 (rewrite_update_dom_walker): New class.
4765 (rewrite_update_enter_block): Convert to
4766 rewrite_update_dom_walker::before_dom_children.
4767 (rewrite_update_leave_block): Convert to
4768 rewrite_update_dom_walker::after_dom_children.
4769 (rewrite_blocks, rewrite_into_ssa): Adjust.
4770 (mark_def_dom_walker): New class.
4771 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4772 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4773 (mark_def_sites_blocks): Convert to
4774 mark_def_dom_walker::before_dom_children.
4775 (mark_def_site_blocks): Remove.
4776 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4777 (tree_ssa_dominator_optimize): Adjust.
4778 (dom_thread_across_edge): Convert to method
4779 dom_opt_dom_walker::thread_across_edge.
4780 (dom_opt_enter_block): Convert to member function
4781 dom_opt_dom_walker::before_dom_children.
4782 (dom_opt_leave_block): Convert to member function
4783 dom_opt_dom_walker::after_dom_children.
4784 * tree-ssa-dse.c (dse_dom_walker): New class.
4785 (dse_enter_block): Convert to member function
4786 dse_dom_walker::before_dom_children.
4787 (tree_ssa_dse): Adjust.
4788 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4789 (determine_invariantness_stmt): Convert to method
4790 invariantness_dom_walker::before_dom_children.
4791 (determine_invariantness): Remove
4792 (move_computations_dom_walker): New class.
4793 (move_computations_stmt): Convert to method
4794 move_computations_dom_walker::before_dom_children.
4795 (move_computations, tree_ssa_lim): Adjust.
4796 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4797 (nt_init_block): Make method
4798 notrappping_dom_walker::before_dom_children.
4799 (nt_fini_block): Make
4800 method nontrapping_dom_walker::after_dom_children.
4801 (get_non_trapping): Adjust.
4802 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4803 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4804 (eliminate_leave_block): Make method.
4805 eliminate_dom_walker::after_dom_children.
4806 (eliminate): Adjust
4807 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4808 (strlen_enter_block): Make method
4809 strlen_dom_walker::before_dom_children.
4810 (strlen_leave_block): Make
4811 method strlen_dom_walker::after_dom_children.
4812 (tree_ssa_strlen): Adjust.
4813 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4814 (tree_ssa_uncprop): Adjust.
4815 (uncprop_leave_block): Make method
4816 uncprop_dom_walker::after_dom_children.
4817 (uncprop_leave_block): Make method
4818 uncprop_dom_walker::before_dom_children.
4819
4820 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4821
4822 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4823 branch in same basic block. Check both src and dest of the move insn.
4824
4825 2013-09-17 Nick Clifton <nickc@redhat.com>
4826
4827 * config/rl78/rl78-real.md (bf): New pattern.
4828 (bt): New pattern.
4829 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4830 (rl78_print_operand): Do not put a # before a %B.
4831 * config/rl78/rl78.opt: Tweak doc strings.
4832
4833 2013-09-17 DJ Delorie <dj@redhat.com>
4834
4835 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4836 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4837 (rl78_option_override): Likewise, if -mallregs.
4838 (is_virtual_register): Likewise.
4839 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4840 (REGNO_OK_FOR_BASE_P): Likewise.
4841 * config/rl78/rl78.opt (-mallregs): New.
4842
4843 2013-09-17 Nick Clifton <nickc@redhat.com>
4844
4845 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4846 For interrupt functions: save all call clobbered registers if the
4847 interrupt handler is not a leaf function.
4848 (rl78_expand_prologue): Always recompute the frame information.
4849 For interrupt functions: only select bank 0 if one of the bank 0
4850 registers is going to be psuhed.
4851
4852 2013-09-17 DJ Delorie <dj@redhat.com>
4853
4854 * config/rl78/constraints.md: For each W* constraint, rename to C*
4855 and create a W* constraint that checks for an optional ES: prefix
4856 pattern also.
4857 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4858 (es_addr): New. Used to wrap far addresses.
4859 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4860 (rl78_es_base): New.
4861 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4862 wrapped far addresses.
4863 (rl78_print_operand_1): Unwrap far addresses before processing.
4864 (rl78_lo16): Wrap far addresses in unspecs.
4865 (rl78_es_addr): New.
4866 (rl78_es_base): New.
4867 (insn_ok_now): Check for not-yet-wrapped far addresses.
4868 (transcode_memory_rtx): Properly re-wrap far addresses.
4869
4870 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4871
4872 * config/sparc/t-rtems: Add leon3 multilibs.
4873
4874 2013-09-17 Cong Hou <congh@google.com>
4875
4876 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4877 when checking the dot production pattern. The type of rhs operand
4878 of multiply is now checked correctly.
4879
4880 2013-09-17 Jeff Law <law@redhat.com>
4881
4882 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4883 edge implied equivalences into successor phis.
4884 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4885 here from tree-ssa-threadedge.c.
4886 (mark_threaded_blocks): When threading through a joiner, if both
4887 successors of the joiner's clone reach the same block, verify the
4888 PHI arguments are equal. If not, cancel the jump threading request.
4889 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4890 tree-ssa-threadupdate.c
4891 (thread_across_edge): Don't check PHI argument equality when
4892 threading through joiner block here.
4893
4894 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4895
4896 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4897 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4898 (warn_uninit, warn_uninitialized_vars,
4899 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4900 Move to tree-ssa-uninit.c.
4901 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4902 (has_undefined_value_p): New. Pass dependant parts of
4903 ssa_undefined_value_p.
4904 (uninit_undefined_value_p): Use has_undefined_value_p.
4905 (warn_uninit, warn_uninitialized_vars,
4906 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4907 Move from tree-ssa.c.
4908 * tree-ssa.h: Adjust prototypes.
4909
4910 2013-09-17 Jan Hubicka <jh@suse.cz>
4911
4912 PR middle-end/58332
4913 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4914 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4915 FUNCTION_NOT_OPTIMIZED.
4916 * ipa-inline-analysis.c (compute_inline_parameters): Function
4917 not optimized is not inlinable unless it is alwaysinline.
4918 (inline_analyze_function): Force calls in not optimized
4919 function not inlinable.
4920
4921 2013-09-17 Jan Hubicka <jh@suse.cz>
4922
4923 PR middle-end/58329
4924 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4925 to return NULL.
4926 * ipa.c (function_and_variable_visibility): Likewise.
4927 * ipa-profile.c (ipa_profile): Likewise.
4928
4929 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4930
4931 PR ipa/58398
4932 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4933 attribute, and don't inline the resolver in this case.
4934
4935 2013-09-17 Teresa Johnson <tejohnson@google.com>
4936
4937 * coverage.c (get_coverage_counts): Add missing newline.
4938
4939 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4940
4941 PR tree-optimization/58088
4942 * fold-const.c (mask_with_trailing_zeros): New function.
4943 (fold_binary_loc): Make sure we don't recurse infinitely
4944 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4945 Use mask_with_trailing_zeros where appropriate.
4946
4947 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4948
4949 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4950 of 'prev' var to get better distance estimation.
4951
4952 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4953
4954 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4955 * tree-inline.c (is_parameter_of): New predicate.
4956 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4957 a parameter has been remapped.
4958 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4959 (optimize_inline_calls): Initialize transform_parameter.
4960 (copy_gimple_seq_and_replace_locals): Likewise.
4961 (tree_function_versioning): Likewise.
4962 (maybe_inline_call_in_expr): Likewise.
4963
4964 2013-09-17 Nick Clifton <nickc@redhat.com>
4965
4966 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4967 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4968 interrupt handlers.
4969 (is_attr_func): New function.
4970 (msp430_is_interrupt_func): New function.
4971 (is_naked_func): New function.
4972 (is_reentrant_func): New function.
4973 (is_critical_func): New function.
4974 (msp430_start_function): Add annotations for function attributes.
4975 (msp430_attr): New function.
4976 (msp430_attribute_table): New.
4977 (msp430_function_section): New function.
4978 (TARGET_ASM_FUNCTION_SECTION): Define.
4979 (msp430_builtin): New enum.
4980 (msp430_init_builtins): New function.
4981 (msp430_builtin_devl): New function.
4982 (msp430_expand_builtin): New function.
4983 (TARGET_INIT_BUILTINS): Define.
4984 (TARGET_EXPAND_BUILTINS): Define.
4985 (TARGET_BUILTIN_DECL): Define.
4986 (msp430_expand_prologue): Add support for naked, interrupt,
4987 critical and reentrant functions.
4988 (msp430_expand_epilogue): Likewise.
4989 (msp430_print_operand): Handle 'O' character.
4990 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4991 NO_TRAMPOLINES.
4992 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4993 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4994 (pushm): Use a 'n' rather than an 'i' constraint.
4995 (msp_return): Add generation of the interrupt return instruction.
4996 (disable_interrupts): New pattern.
4997 (enable_interrupts): New pattern.
4998 (push_intr_state): New pattern.
4999 (pop_intr_state): New pattern.
5000 (bic_SR): New pattern.
5001 (bis_SR): New pattern.
5002 * doc/extend.texi: Document MSP430 function attributes and builtin
5003 functions.
5004
5005 2013-09-17 Richard Biener <rguenther@suse.de>
5006
5007 PR tree-optimization/58432
5008 * tree-loop-distribution.c (tree_loop_distribution): Also
5009 scan PHIs for outside loop uses and seed a partition from them.
5010
5011 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5012
5013 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5014 (restructure_reference): Call backtrace_base_for_ref.
5015
5016 2013-09-17 Alan Modra <amodra@gmail.com>
5017
5018 PR target/57589
5019 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5020 patch.
5021
5022 2013-09-16 DJ Delorie <dj@redhat.com>
5023
5024 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5025 vregs location for RL78/G10.
5026 (rl78_expand_prologue): Avoid SEL on G10.
5027 (rl78_expand_epilogue): Likewise.
5028 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5029 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5030 __RL78_G10__ when appropriate.
5031 (ASM_SPEC): Pass -mg10 along to the assembler.
5032 * config/rl78/rl78.md (sel_rb): Disable for G10.
5033 * config/rl78/rl78.opt: Add -mg10 option.
5034 * config/rl78/t-rl78: Add -mg10 multilib.
5035
5036 2013-09-16 Xinliang David Li <davidxl@google.com>
5037
5038 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5039 * omp-low.c (omp_max_vf): Ditto.
5040 (expand_omp_simd): Ditto.
5041 * tree-vectorizer.c (vectorize_loops): Ditto.
5042 (gate_vect_slp): Ditto.
5043 (gate_increase_alignment): Ditto.
5044 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5045 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5046 (gate_tree_vectorize): Name change.
5047 (tree_vectorize): Ditto.
5048 (pass_vectorize::gate): Call new function.
5049 (pass_vectorize::execute): Ditto.
5050 * opts.c: O3 default setting change.
5051 (finish_options): Check new flag.
5052 * doc/invoke.texi: Document new flags.
5053 * common.opt: New flags.
5054
5055 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5056
5057 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5058 (Condition Code): Don't reference it.
5059 * doc/tm.texi: Regenerate.
5060
5061 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5062
5063 PR middle-end/58418
5064 * lra-constraints.c (undo_optional_reloads): Consider all optional
5065 reload even if it did not get a hard reg.
5066
5067 2013-09-16 Teresa Johnson <tejohnson@google.com>
5068
5069 * dumpfile.c (dump_loc): Remove newline emission.
5070 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5071 emission to dump_printf_loc calls where missing.
5072 (vect_mark_for_runtime_alias_test): Ditto.
5073 (vect_analyze_data_ref_dependence): Ditto.
5074 (vect_analyze_data_ref_dependences): Ditto.
5075 (vect_slp_analyze_data_ref_dependence): Ditto.
5076 (vect_slp_analyze_data_ref_dependences): Ditto.
5077 (vect_compute_data_ref_alignment): Ditto.
5078 (vect_update_misalignment_for_peel): Ditto.
5079 (vect_verify_datarefs_alignment): Ditto.
5080 (vector_alignment_reachable_p): Ditto.
5081 (vect_get_data_access_cost): Ditto.
5082 (vect_enhance_data_refs_alignment): Ditto.
5083 (vect_find_same_alignment_drs): Ditto.
5084 (vect_analyze_data_refs_alignment): Ditto.
5085 (vect_analyze_group_access): Ditto.
5086 (vect_analyze_data_ref_access): Ditto.
5087 (vect_analyze_data_ref_accesses): Ditto.
5088 (vect_prune_runtime_alias_test_list): Ditto.
5089 (vect_analyze_data_refs): Ditto.
5090 (vect_create_addr_base_for_vector_ref): Ditto.
5091 (vect_create_data_ref_ptr): Ditto.
5092 (vect_grouped_store_supported): Ditto.
5093 (vect_grouped_load_supported): Ditto.
5094 * value-prof.c (check_counter): Ditto.
5095 (check_ic_target): Ditto.
5096 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5097 (vect_recog_widen_mult_pattern): Ditto.
5098 (vect_recog_widen_sum_pattern): Ditto.
5099 (vect_recog_over_widening_pattern): Ditto.
5100 (vect_recog_widen_shift_pattern): Ditto.
5101 (vect_recog_rotate_pattern): Ditto.
5102 (vect_recog_vector_vector_shift_pattern): Ditto.
5103 (vect_recog_divmod_pattern): Ditto.
5104 (vect_recog_mixed_size_cond_pattern): Ditto.
5105 (vect_recog_bool_pattern): Ditto.
5106 (vect_pattern_recog_1): Ditto.
5107 (vect_pattern_recog): Ditto.
5108 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5109 (vect_is_simple_iv_evolution): Ditto.
5110 (vect_analyze_scalar_cycles_1): Ditto.
5111 (vect_get_loop_niters): Ditto.
5112 (vect_analyze_loop_1): Ditto.
5113 (vect_analyze_loop_form): Ditto.
5114 (vect_analyze_loop_operations): Ditto.
5115 (vect_analyze_loop_2): Ditto.
5116 (vect_analyze_loop): Ditto.
5117 (report_vect_op): Ditto.
5118 (vect_is_slp_reduction): Ditto.
5119 (vect_is_simple_reduction_1): Ditto.
5120 (vect_get_known_peeling_cost): Ditto.
5121 (vect_estimate_min_profitable_iters): Ditto.
5122 (vect_model_reduction_cost): Ditto.
5123 (vect_model_induction_cost): Ditto.
5124 (get_initial_def_for_induction): Ditto.
5125 (vect_create_epilog_for_reduction): Ditto.
5126 (vectorizable_reduction): Ditto.
5127 (vectorizable_induction): Ditto.
5128 (vectorizable_live_operation): Ditto.
5129 (vect_loop_kill_debug_uses): Ditto.
5130 (vect_transform_loop): Ditto.
5131 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5132 (vect_stmt_relevant_p): Ditto.
5133 (process_use): Ditto.
5134 (vect_mark_stmts_to_be_vectorized): Ditto.
5135 (vect_model_simple_cost): Ditto.
5136 (vect_model_promotion_demotion_cost): Ditto.
5137 (vect_model_store_cost): Ditto.
5138 (vect_get_store_cost): Ditto.
5139 (vect_model_load_cost): Ditto.
5140 (vect_get_load_cost): Ditto.
5141 (vect_init_vector_1): Ditto.
5142 (vect_get_vec_def_for_operand): Ditto.
5143 (vect_finish_stmt_generation): Ditto.
5144 (vectorizable_call): Ditto.
5145 (vectorizable_conversion): Ditto.
5146 (vectorizable_assignment): Ditto.
5147 (vectorizable_shift): Ditto.
5148 (vectorizable_operation): Ditto.
5149 (vectorizable_store): Ditto.
5150 (vectorizable_load): Ditto.
5151 (vectorizable_condition): Ditto.
5152 (vect_analyze_stmt): Ditto.
5153 (vect_transform_stmt): Ditto.
5154 (vect_is_simple_use): Ditto.
5155 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5156 (vect_can_advance_ivs_p): Ditto.
5157 (vect_update_ivs_after_vectorizer): Ditto.
5158 (vect_do_peeling_for_loop_bound): Ditto.
5159 (vect_gen_niters_for_prolog_loop): Ditto.
5160 (vect_update_inits_of_drs): Ditto.
5161 (vect_create_cond_for_alias_checks): Ditto.
5162 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5163 (vect_build_slp_tree_1): Ditto.
5164 (vect_supported_load_permutation_p): Ditto.
5165 (vect_analyze_slp_instance): Ditto.
5166 (vect_analyze_slp): Ditto.
5167 (vect_make_slp_decision): Ditto.
5168 (vect_detect_hybrid_slp): Ditto.
5169 (vect_bb_vectorization_profitable_p): Ditto.
5170 (vect_slp_analyze_bb_1): Ditto.
5171 (vect_update_slp_costs_according_to_vf): Ditto.
5172 (vect_get_mask_element): Ditto.
5173 (vect_transform_slp_perm_load): Ditto.
5174 (vect_schedule_slp_instance): Ditto.
5175 (vect_schedule_slp): Ditto.
5176 (vect_slp_transform_bb): Ditto.
5177 * profile.c (read_profile_edge_counts): Ditto.
5178 (compute_branch_probabilities): Ditto.
5179 * coverage.c (get_coverage_counts): Ditto.
5180
5181 2013-09-16 Diego Novillo <dnovillo@google.com>
5182
5183 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5184
5185 2013-09-16 Jan Hubicka <jh@suse.cz>
5186
5187 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5188 abstract functions; for static functions check the presence of body.
5189
5190 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5191
5192 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5193 * config/aarch64/aarch64-simd.md
5194 (aarch64_mla_elt<mode>): New.
5195 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5196 (aarch64_mls_elt<mode>): Likewise.
5197 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5198 (aarch64_fma4_elt<mode>): Likewise.
5199 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5200 (aarch64_fma4_elt_to_128v2df): Likewise.
5201 (aarch64_fma4_elt_to_64df): Likewise.
5202 (fnma<mode>4): Likewise.
5203 (aarch64_fnma4_elt<mode>): Likewise.
5204 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5205 (aarch64_fnma4_elt_to_128v2df): Likewise.
5206 (aarch64_fnma4_elt_to_64df): Likewise.
5207 * config/aarch64/iterators.md (VDQSF): New.
5208 * config/aarch64/arm_neon.h
5209 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5210 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5211
5212 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5213
5214 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5215 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5216 (aarch64_mul3_elt_to_128df): Likewise.
5217 (aarch64_mul3_elt_to_64v2df): Likewise.
5218 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5219 (VMUL): New.
5220 (VMUL_CHANGE_NLANES) Likewise.
5221 (h_con): Likewise.
5222 (f): Likewise.
5223 * config/aarch64/arm_neon.h
5224 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5225
5226 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5227
5228 * config/aarch64/arm_neon.h
5229 (vcvtx_high_f32_f64): Fix parameters.
5230
5231 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5232 Uros Bizjak <ubizjak@gmail.com>
5233
5234 * config/alpha.c: Include tree-ssanames.h.
5235
5236 2013-09-16 Richard Biener <rguenther@suse.de>
5237
5238 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5239 (dot_rdg_1): Handle control_dd.
5240 (create_edge_for_control_dependence): New function.
5241 (create_rdg_edges): Add control dependences if asked for.
5242 (build_rdg): Likewise.
5243 (generate_loops_for_partition): If there are not necessary
5244 control stmts remove all their dependencies.
5245 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5246 (distribute_loop): Pass on control dependences.
5247 (tree_loop_distribution): Compute control dependences and remove
5248 restriction on number of loop nodes.
5249
5250 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5251
5252 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5253 for internal calls.
5254
5255 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5256
5257 * cse.c (try_const_anchors): Punt on CC modes.
5258
5259 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5260
5261 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5262
5263 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5264
5265 PR target/58382
5266 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5267 calls to word_mode.
5268
5269 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5270
5271 PR target/48094
5272 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5273 seen.
5274 (darwin_objc1_section): Likewise.
5275 (darwin_file_end): Emit Image Info section when required.
5276
5277 2013-09-14 Jan Hubicka <jh@suse.cz>
5278
5279 * tree-into-ssa.c (gate_into_ssa): New.
5280 (pass_data_build_ssa): Use it.
5281 * cgraph.h (expand_thunk): Update prototype.
5282 * cgraphunit.c (analyze_function): Expand thunks early.
5283 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5284 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5285 set lowered flag; do not add new function.
5286 (assemble_thunks_and_aliases): Update.
5287 * tree-ssa.c (gate_init_datastructures): New gate.
5288 (pass_data_init_datastructures): Use it.
5289
5290 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5291
5292 PR target/58269
5293 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5294 xmm register set described in the psABI.
5295
5296 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5297
5298 * dwarf2out.c (should_emit_struct_debug): Add check
5299 for type_decl variable is not NULL.
5300
5301 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5302
5303 * config.gcc: Use new winnt-c.c target hooks
5304 * config/t-winnt: New file
5305 * config/winnt-c.c: New file
5306 * doc/tm.texi.in: Document new hook
5307 * doc/tm.texi: Regenerated
5308
5309 2013-09-13 Jan Hubicka <jh@suse.cz>
5310
5311 PR middle-end/58094
5312 * ipa-inline.c (check_callers): New function.
5313 (check_caller_edge): Remove.
5314 (want_inline_function_to_all_callers_p): Also permit alises that are
5315 called dirrectly.
5316 (inline_to_all_callers): Terminate the walk when devirtualization
5317 introduce new calls.
5318
5319 2013-09-13 Jan Hubicka <jh@suse.cz>
5320
5321 * ipa-inline-analysis.c (struct growth_data): Add node.
5322 (do_estimate_growth_1): Fix detection of recursion.
5323
5324 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5325
5326 PR tree-optimization/58392
5327 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5328 to avoid shadowing of outer loop variable. If
5329 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5330 replace_by_duplicate_decl simduid of loops that have it set and
5331 set dest_cfun->has_simduid_loops and/or
5332 dest_cfun->has_force_vect_loops.
5333 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5334 instead of maybe_lookup_decl.
5335 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5336 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5337 call. Copy over force_vect and copy and remap simduid. Set
5338 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5339 (copy_cfg_body): Remove blocks_to_copy argument. Use
5340 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5341 caller. Don't set cfun->has_simduid_loops and/or
5342 cfun->has_force_vect_loops here.
5343 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5344 caller.
5345 (expand_call_inline, tree_function_versioning): Adjust copy_body
5346 callers.
5347
5348 2013-09-13 Martin Jambor <mjambor@suse.cz>
5349
5350 PR bootstrap/58388
5351 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5352 the assert if the edge was a speculative one.
5353
5354 2013-09-13 Richard Biener <rguenther@suse.de>
5355
5356 * tree-data-ref.h (known_dependences_p): Move here ...
5357 * tree-loop-distribution.c (known_dependences_p): ... from here.
5358 (dump_rdg_component, debug_rdg_component): Remove.
5359 (dump_rdg): Adjust.
5360 (generate_loops_for_partition): Use gimple_uid instead of
5361 relying on matching stmt visit order.
5362 (rdg_build_partitions): Take starting stmt vector.
5363 (ldist_gen): Merge into ...
5364 (distribute_loop): ... this function. Do not compute starting
5365 vertices vector.
5366 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5367
5368 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5369
5370 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5371 Set type attribute correctly. Set predicable_short_it attribute.
5372 (cmpsi_shiftsi): Remove %? from output template.
5373
5374 2013-09-13 Richard Biener <rguenther@suse.de>
5375
5376 * tree-loop-distribution.c (struct rdg_component,
5377 rdg_defs_used_in_other_loops_p, free_rdg_components,
5378 rdg_build_components): Remove.
5379 (stmts_from_loop): Do not record virtual PHIs.
5380 (generate_loops_for_partition): Skip virtual PHIs.
5381 (build_rdg_partition_for_component): Rename to ...
5382 (build_rdg_partition_for_vertex): ... this and adjust.
5383 (rdg_build_partitions): Take a vector of starting vertices
5384 instead of components. Remove unnecessary leftover handling.
5385 (ldist_gen): Do not build components or record other stores.
5386 (distribute_loop): Do not distribute loops containing stmts
5387 with side-effects.
5388
5389 2013-09-13 Christian Bruel <christian.bruel@st.com>
5390
5391 PR target/58314
5392 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5393
5394 2013-09-13 Kai Tietz <ktietz@redhat.com>
5395
5396 * config.gcc: Separate cases for mingw and cygwin targets,
5397 and add 64-bit cygwin target case.
5398
5399 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5400 dll-export inline-functions.
5401 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5402
5403
5404 2013-09-13 Jeff Law <law@redhat.com>
5405
5406 PR middle-end/58387
5407 Revert:
5408 2013-09-06 Jeff Law <law@redhat.com>
5409
5410 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5411 edge implied equivalences into successor phis.
5412
5413 2013-09-12 DJ Delorie <dj@redhat.com>
5414
5415 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5416 separators.
5417
5418 2013-09-12 Brooks Moses <bmoses@google.com>
5419
5420 PR driver/42955
5421 * Makefile.in: Do not install driver binaries in $(target)/bin.
5422
5423 2013-09-12 DJ Delorie <dj@redhat.com>
5424
5425 * config/rl78/rl78.opt (mrelax): New.
5426 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5427 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5428
5429 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5430 between SP and FP.
5431 (rl78_expand_epilogue): Likewise.
5432
5433 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5434
5435 PR middle-end/58335
5436 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5437 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5438 pointer elimination with using remove_reg_equal_offset_note.
5439
5440 2013-09-12 DJ Delorie <dj@redhat.com>
5441
5442 * config/msp430/: New port.
5443 * config.gcc (msp430): Added.
5444 * doc/invoke.texi: Document MSP430 options.
5445 * doc/install.texi: Document msp430-elf
5446 * doc/md.texi: Document msp430-elf
5447 * doc/contrib.texi: Document msp430-elf
5448
5449 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5450 PSImode.
5451
5452 2013-09-12 Martin Jambor <mjambor@suse.cz>
5453
5454 PR ipa/58389
5455 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5456 reference descriptor is NULL.
5457 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5458 edge to NULL.
5459
5460 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5461
5462 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5463 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5464 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5465 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5466 make_temp_ssa_name): move to tree-ssanames.h
5467 * tree-ssa-alias.h: Move prototype.
5468 * tree-ssa.h: Include tree-ssanames.h.
5469 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5470 * tree-ssanames.h: New. Move items from tree-flow*.h
5471 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5472
5473 2013-09-12 Richard Biener <rguenther@suse.de>
5474
5475 PR tree-optimization/58404
5476 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5477 propagate non-invariant addresses into dereferences wrapped
5478 in component references.
5479
5480 2013-09-12 Richard Biener <rguenther@suse.de>
5481
5482 PR tree-optimization/58402
5483 * passes.def: Move pass_late_warn_uninitialized later.
5484
5485 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5486
5487 * tree-ssa.h: New. Move content from tree-flow.h and
5488 tree-flow-inline.h.
5489 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5490 Move prototypes belonging to tree-ssa.c.
5491 * tree-flow-inline.h (redirect_edge_var_map_def,
5492 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5493 tree-ssa.h.
5494 * gimple.h: Adjust prototypes.
5495 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5496 to...
5497 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5498 * tree.h: Move prototype to tree-ssa.h.
5499 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5500 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5501 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5502 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5503 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5504 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5505 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5506 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5507 graphite-blocking.c, graphite-clast-to-gimple.c,
5508 graphite-dependences.c, graphite-interchange.c,
5509 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5510 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5511 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5512 ipa-reference.c, ipa-split.c, ipa-utils.c,
5513 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5514 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5515 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5516 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5517 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5518 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5519 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5520 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5521 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5522 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5523 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5524 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5525 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5526 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5527 value-prof.c, var-tracking.c,
5528 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5529
5530 2013-09-12 Richard Biener <rguenther@suse.de>
5531
5532 PR tree-optimization/58396
5533 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5534 (build_rdg): Take a loop-nest parameter, fix memleaks.
5535 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5536
5537 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5538
5539 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5540 for SLM.
5541
5542 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5543
5544 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5545 The FMA instruction names should have a 'v' prefix.
5546
5547 2013-09-12 Richard Biener <rguenther@suse.de>
5548
5549 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5550 (dot_rdg): Use popen instead of system in optional code.
5551 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5552 (already_processed_vertex_p): Adjust.
5553 (has_anti_or_output_dependence, predecessor_has_mem_write,
5554 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5555 rdg_flag_uses): Remove.
5556 (rdg_flag_vertex): Simplify.
5557 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5558 remove recursion.
5559 (build_rdg_partition_for_component): Process the first vertex
5560 of a component only.
5561 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5562
5563 2013-09-12 Alan Modra <amodra@gmail.com>
5564
5565 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5566
5567 2013-09-11 DJ Delorie <dj@redhat.com>
5568 Nick Clifton <nickc@redhat.com>
5569
5570 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5571 (rl78_stack_based_mem): New.
5572 * config/rl78/constraints.md (Iv08): New.
5573 (Iv16): New.
5574 (Iv24): New.
5575 (Is09): New.
5576 (Is17): New.
5577 (Is25): New.
5578 (ISsi): New.
5579 (IShi): New.
5580 (ISqi): New.
5581 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5582 (movhi): Likewise.
5583 (movsi): Change from expand to insn-and-split.
5584 (ashrsi3): Clobber AX.
5585 (lshrsi3): New.
5586 (ashlsi3): New.
5587 (cbranchsi4): New.
5588 * config/rl78/rl78.md (CC_REG): Fix.
5589 (addsi3): Allow memory and immediate operands.
5590 (addsi3_internal): Split into...
5591 (addsi3_internal_virt): ...new, and ...
5592 (addsi3_internal_real): ...new.
5593 (subsi): New.
5594 (subsi3_internal_virt): New.
5595 (subsi3_internal_real): New.
5596 (mulsi3): Add memory operand.
5597 (mulsi3_rl78): Likewise.
5598 (mulsi3_g13): Likewise.
5599 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5600 (cbranchqi4_real): Add more constraint options.
5601 (cbranchhi4_real): Expand pattern.
5602 (cbranchhi4_real_signed): New.
5603 (cbranchhi4_real_inverted): New.
5604 (cbranchsi4_real_lt): New.
5605 (cbranchsi4_real_ge): New.
5606 (cbranchsi4_real_signed): New.
5607 (cbranchsi4_real): New.
5608 (peephole2): New.
5609 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5610 constant shifts.
5611 (lshrsi3_virt): Likewise.
5612 (ashlsi3_virt): Likewise.
5613 (cbranchqi4_virt_signed): New.
5614 (cbranchhi4_virt_signed): New.
5615 (cbranchsi4_virt): New.
5616 * config/rl78/rl78.c: Whitespace fixes throughout.
5617 (move_elim_pass): New.
5618 (pass_data_rl78_move_elim): New.
5619 (pass_rl78_move_elim): New.
5620 (make_pass_rl78_move_elim): New.
5621 (rl78_devirt_info): Run devirt earlier.
5622 (rl78_move_elim_info): New.
5623 (rl78_asm_file_start): Register it.
5624 (rl78_split_movsi): New.
5625 (rl78_as_legitimate_address): Allow virtual base registers when
5626 appropriate.
5627 (rl78_addr_space_convert): Remove spurious debug stuff.
5628 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5629 (rl78_print_operand): More cases for not printing '#'.
5630 (rl78_expand_compare): Remove most of the logic.
5631 (content_memory): New.
5632 (clear_content_memory): New.
5633 (get_content_index): New.
5634 (get_content_name): New.
5635 (display_content_memory): New.
5636 (update_content): New.
5637 (record_content): New.
5638 (already_contains): New.
5639 (insn_ok_now): Re-recog insns with virtual registers.
5640 (add_postponed_content_update): New.
5641 (process_postponed_content_update): New.
5642 (gen_and_emit_move): New.
5643 (transcode_memory_rtx): Record new location content.
5644 Use gen_and_emit_move.
5645 (force_into_acc): New.
5646 (move_to_acc): Use gen_and_emit_move.
5647 (move_from_acc): Likewise.
5648 (move_acc_to_reg): Likewise.
5649 (move_to_x): Likewise.
5650 (move_to_hl): Likewise.
5651 (move_to_de): Likewise.
5652 (rl78_alloc_physical_registers_op1): Record location content.
5653 (has_constraint): New.
5654 (rl78_alloc_physical_registers_op2): Record location content.
5655 Optimize use of HL.
5656 (rl78_alloc_physical_registers_ro1): Likewise.
5657 (rl78_alloc_physical_registers_cmp): Likewise.
5658 (rl78_alloc_physical_registers_umul): Likewise.
5659 (rl78_alloc_address_registers_macax): New.
5660 (rl78_alloc_physical_registers): Initialize and set location
5661 content memory as needed.
5662 (rl78_reorg): Make sure split2 is called.
5663 (rl78_rtx_costs): New.
5664
5665 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5666
5667 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5668 for (not (neg ...)) and (neg (not ...)) cases.
5669
5670 2013-09-11 Richard Biener <rguenther@suse.de>
5671
5672 PR middle-end/58377
5673 * passes.def: Split critical edges before late uninit warning passes.
5674 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5675
5676 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5677
5678 PR tree-optimization/58385
5679 * fold-const.c (build_range_check): If both low and high are NULL,
5680 use omit_one_operand_loc to preserve exp side-effects.
5681
5682 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5683
5684 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5685
5686 2013-09-11 Richard Biener <rguenther@suse.de>
5687
5688 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5689 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5690 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5691 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5692 stmts_from_loop, known_dependences_p, build_empty_rdg,
5693 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5694 * tree-loop-distribution.c: ... here.
5695 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5696 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5697 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5698 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5699 * tree-loop-distribution.c: ... here.
5700 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5701 (struct partition_s): Add loops member.
5702 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5703 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5704 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5705
5706 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5707 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5708 Sergey Lega <sergey.s.lega@intel.com>
5709 Anna Tikhonova <anna.tikhonova@intel.com>
5710 Ilya Tocar <ilya.tocar@intel.com>
5711 Andrey Turetskiy <andrey.turetskiy@intel.com>
5712 Ilya Verbin <ilya.verbin@intel.com>
5713 Kirill Yukhin <kirill.yukhin@intel.com>
5714 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5715
5716 * config/i386/constraints.md (k): New.
5717 (Yk): Ditto.
5718 * config/i386/i386.c (const regclass_map): Add new mask registers.
5719 (dbx_register_map): Ditto.
5720 (dbx64_register_map): Ditto.
5721 (svr4_dbx_register_map): Ditto.
5722 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5723 disabled.
5724 (ix86_preferred_reload_class): Disable constants for mask registers.
5725 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5726 (ix86_hard_regno_mode_ok): Support new mask registers.
5727 (x86_order_regs_for_local_alloc): Ditto.
5728 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5729 (FIXED_REGISTERS): Add new mask registers.
5730 (CALL_USED_REGISTERS): Ditto.
5731 (REG_ALLOC_ORDER): Ditto.
5732 (VALID_MASK_REG_MODE): New.
5733 (FIRST_MASK_REG): Ditto.
5734 (LAST_MASK_REG): Ditto.
5735 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5736 (MAYBE_MASK_CLASS_P): New.
5737 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5738 (REG_CLASS_CONTENTS): Ditto.
5739 (MASK_REGNO_P): New.
5740 (ANY_MASK_REG_P): Ditto.
5741 (HI_REGISTER_NAMES): Add new mask registers.
5742 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5743 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5744 mask registers.
5745 (attribute "type"): Add mskmov, msklog.
5746 (attribute "length_immediate"): Support them.
5747 (attribute "memory"): Ditto.
5748 (attribute "prefix_0f"): Ditto.
5749 (*movhi_internal): Support new mask registers.
5750 (*movqi_internal): Ditto.
5751 (define_split): Split out clobber pattern is a logic
5752 insn on mask registers.
5753 (*k<logic><mode>): New.
5754 (*andhi_1): Extend to support mask regs.
5755 (*andqi_1): Extend to support mask regs.
5756 (kandn<mode>): New.
5757 (define_split): Split and-not to and and not if operands
5758 are not mask regs.
5759 (*<code><mode>_1): Separate HI mode to new pattern...
5760 (*<code>hi_1): This.
5761 (*<code>qi_1): Extend to support mask regs.
5762 (kxnor<mode>): New.
5763 (kortestzhi): Ditto.
5764 (kortestchi): Ditto.
5765 (kunpckhi): Ditto.
5766 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5767 (*one_cmplhi2_1): ...Here, now with mask registers support.
5768 (*one_cmplqi2_1): Support new mask registers.
5769 (HI/QImode arithmetics splitter): Don't split if mask registers
5770 are used.
5771 (HI/QImode not splitter): Ditto.
5772 * config/i386/predicated.md (mask_reg_operand): New.
5773 (general_reg_operand): Ditto.
5774
5775 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5776
5777 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5778 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5779
5780 2013-09-10 Jeff Law <law@redhat.com>
5781
5782 PR tree-optimization/58380
5783 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5784 of threading through a buried loop header.
5785
5786 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5787 return value for single successor case.
5788
5789 2013-09-10 Jan Hubicka <jh@suse.cz>
5790
5791 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5792
5793 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5794
5795 PR target/58361
5796 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5797 support conditional execution.
5798 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5799
5800 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5801
5802 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5803 constraint pass.
5804 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5805 Check destination too to check move insn.
5806 (undo_optional_reloads): Add check that the original peudo did not
5807 changed its allocation and the optional reload was inherited on last
5808 inheritance pass. Break loop after deciding to keep optional reload.
5809 (lra_undo_inheritance): Add check that inherited pseudo still in
5810 memory.
5811
5812 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5813
5814 * config/aarch64/aarch64.md (generic_sched): New.
5815 * config/aarch64/aarch64-generic.md (load): Make conditional
5816 on generic_sched attribute.
5817 (nonload): Likewise.
5818
5819 2013-09-10 Jan Hubicka <jh@suse.cz>
5820
5821 * lto-cgraph.c: Include ipa-utils.h.
5822 (compute_ltrans_boundary): Also add possible targets into the boundary.
5823
5824 2013-09-10 Jan Hubicka <jh@suse.cz>
5825
5826 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5827 VAR_DECL of vtable rather than full expression.
5828
5829 2013-09-10 Jan Hubicka <jh@suse.cz>
5830 Paolo Carlini <paolo.carlini@oracle.com>
5831
5832 * cgraphunit.c (analyze_functions): Save input_location, set it
5833 to UNKNOWN_LOCATION and restore it at the end.
5834
5835 2013-09-10 Martin Jambor <mjambor@suse.cz>
5836
5837 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5838 represented by a thunk.
5839
5840 2013-09-10 Jeff Law <law@redhat.com>
5841
5842 PR tree-optimization/58343
5843 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5844 jump threading requests through loop headers buried in the middle
5845 of a jump threading path.
5846
5847 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5848 in return value/type.
5849
5850 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5851
5852 PR rtl-optimization/58365
5853 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5854 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5855 it differs.
5856
5857 2013-09-10 Richard Biener <rguenther@suse.de>
5858
5859 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5860 * tree-data-ref.c (create_rdg_vertices): Collect all data
5861 references, signal failure to the caller, use data-ref API.
5862 (build_rdg): Compute data references only once. Maintain lifetime
5863 of data references and data dependences from within RDG.
5864 (free_rdg): Free dependence relations.
5865 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5866 inventing extra dependences.
5867 (distribute_loop): Update for RDG API changes.
5868
5869 2013-09-10 Kai Tietz <ktietz@redhat.com>
5870
5871 * doc/invoke.texi (fms-extensions): Document changed
5872 behavior for ms-abi targets.
5873 * config/i386/i386.c (ix86_option_override_internal):
5874 Set default value of option -fms-extension for ms-abi targets.
5875
5876 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5877
5878 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5879
5880 2013-09-10 Alan Modra <amodra@gmail.com>
5881
5882 PR target/58330
5883 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5884
5885 2013-09-10 Alan Modra <amodra@gmail.com>
5886
5887 * config/rs6000/predicates.md (add_cint_operand): New.
5888 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5889 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5890 using add_cint_operand.
5891 (largetoc_high_plus_aix): Likewise.
5892
5893 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5894
5895 PR tree-optimization/58364
5896 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5897 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5898 the current range can't be an unconditional true or false.
5899
5900 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5901
5902 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5903
5904 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5905
5906 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5907
5908 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5909
5910 PR c++/43452
5911 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5912
5913 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5914
5915 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5916 NO_REGS for immediate that can't be moved directly into FP_REGS.
5917
5918 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5919
5920 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5921 comparison with negated operand.
5922 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5923 RTL form.
5924
5925 2013-09-09 Richard Biener <rguenther@suse.de>
5926
5927 PR middle-end/58326
5928 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5929 of a subloop record all its block as affecting loop-closed SSA form.
5930
5931 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5932
5933 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5934 of an rtx/bitpos pair.
5935 (store_fixed_bit_field): Update accordingly.
5936
5937 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5938
5939 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5940 GEN_INT.
5941 * builtins.c (expand_errno_check): Likewise.
5942 * dwarf2cfi.c (init_return_column_size): Likewise.
5943 * except.c (sjlj_mark_call_sites): Likewise.
5944 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5945 * lra-constraints.c (emit_inc): Likewise.
5946 * ree.c (combine_set_extension): Likewise.
5947 * regmove.c (fixup_match_2): Likewise.
5948 * reload1.c (inc_for_reload): Likewise.
5949
5950 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5951
5952 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5953 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5954 Use gen_int_mode with the mode of the associated simplify_* call.
5955 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5956 * expmed.c (expand_shift_1): Likewise.
5957 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5958 * loop-iv.c (iv_number_of_iterations): Likewise.
5959 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5960 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5961
5962 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5963
5964 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5965 of the associated expand_* call.
5966 (asan_emit_stack_protection): Likewise.
5967 * builtins.c (round_trampoline_addr): Likewise.
5968 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5969 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5970 (emit_store_flag): Likewise.
5971 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5972 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5973 Likewise.
5974 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5975 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5976 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5977 * modulo-sched.c (generate_prolog_epilog): Likewise.
5978 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5979 (expand_ctz, expand_ffs, expand_unop): Likewise.
5980
5981 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5982
5983 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5984 of the associated gen_rtx_* call.
5985 * caller-save.c (init_caller_save): Likewise.
5986 * combine.c (find_split_point, make_extraction): Likewise.
5987 (make_compound_operation): Likewise.
5988 * dwarf2out.c (mem_loc_descriptor): Likewise.
5989 * explow.c (plus_constant, probe_stack_range): Likewise.
5990 * expmed.c (expand_mult_const): Likewise.
5991 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5992 * reload1.c (init_reload): Likewise.
5993 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5994 * var-tracking.c (adjust_mems): Likewise.
5995 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5996 rather than gen_rtx_fmt_ee.
5997
5998 2013-09-09 Jan Hubicka <jh@suse.cz>
5999
6000 PR middle-end/58294
6001 * value-prof.c (gimple_ic): Copy also abnormal edges.
6002
6003 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6004
6005 * asan.c (asan_shadow_cst): Use gen_int_mode.
6006
6007 2013-09-08 Jan Hubicka <jh@suse.cz>
6008
6009 * ipa-profile.c: Add toplevel comment.
6010 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6011 (contains_hot_call_p): New function.
6012 (ipa_propagate_frequency): Set frequencies based on counts when
6013 profile is read.
6014 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6015 profile; do not tamper with profile after inlining if it is read.
6016
6017 2013-09-08 Jan Hubicka <jh@suse.cz>
6018
6019 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6020 speculative edges.
6021
6022 2013-09-08 Jan Hubicka <jh@suse.cz>
6023
6024 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6025 summary generation.
6026
6027 2013-09-08 Jeff Law <law@redhat.com>
6028
6029 PR bootstrap/58340
6030 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6031 of 'found'.
6032
6033 2013-09-08 Andi Kleen <ak@linux.intel.com>
6034
6035 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6036
6037 2013-09-08 Jan Hubicka <jh@suse.cz>
6038
6039 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6040
6041 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6042
6043 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6044 for non-debug insns.
6045 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6046 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6047 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6048 accordingly.
6049
6050 2013-09-08 Jan Hubicka <jh@suse.cz>
6051
6052 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6053 targets and devirtualize to BUILT_IN_UNREACHABLE.
6054 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6055 * ipa.c (walk_polymorphic_call_targets): New function.
6056 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6057 functions; use the new timevar.
6058 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6059 was removed from the program.
6060 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6061 not consider it in the walk when its vtable is dead.
6062 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6063 record_binfo.
6064 (devirt_variable_node_removal_hook): New function.
6065 (possible_polymorphic_call_targets): Also register
6066 devirt_variable_node_removal_hook.
6067 (ipa_devirt): Do not do non-speculative devirtualization.
6068 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6069
6070 2013-09-08 Jan Hubicka <jh@suse.cz>
6071
6072 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6073 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6074 varpool_remove_variable_insertion_hook): Declare.
6075 * varpool.c (varpool_node_hook_list): New structure.
6076 (first_varpool_node_removal_hook,
6077 first_varpool_variable_insertion_hook): New variables.
6078 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6079 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6080 varpool_remove_variable_insertion_hook,
6081 varpool_call_variable_insertion_hooks): New functions.
6082 (varpool_remove_node): Use it.
6083
6084 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6085
6086 PR c++/54941
6087 * diagnostic.c (diagnostic_build_prefix): When s.file is
6088 "<built-in>" don't output line and column numbers.
6089
6090 2013-09-06 Jan Hubicka <jh@suse.cz>
6091
6092 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6093 * lto-streamer-out.c: Stream thunks, too.
6094 * lto-streamer-in.c (input_function): Pop cfun here
6095 (lto_read_body): Instead of here.
6096
6097 2013-09-06 Caroline Tice <cmtice@google.com>
6098
6099 * doc/install.texi: Add documentation for the --enable-vtable-verify
6100 and the --disable-libvtv configure options.
6101
6102 2013-09-06 Jeff Law <law@redhat.com>
6103
6104 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6105 edge implied equivalences into successor phis.
6106
6107 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6108
6109 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6110
6111 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6112
6113 * resource.c (mark_target_live_regs): Compute resources taking
6114 into account if a call is predicated or not.
6115
6116 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6117
6118 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6119 the compiler in the function names.
6120
6121 2013-09-06 Jan Hubicka <jh@suse.cz>
6122
6123 PR middle-end/58094
6124 * ipa-inline.c (has_caller_p): New function.
6125 (want_inline_function_to_all_callers_p): Use it.
6126 (sum_callers, inline_to_all_callers): Break out from ...
6127 (ipa_inline): ... here.
6128
6129 2013-09-06 Jan Hubicka <jh@suse.cz>
6130
6131 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6132 only when AVX is enabled.
6133
6134 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6135
6136 * config/aarch64/aarch64.md
6137 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6138 is fpsimd_<load/store>2.
6139 (load_pair<mode>): Likewise.
6140 (store_pair<mode>): Likewise.
6141
6142 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6143
6144 * config/arm/types.md (type): Add "mrs" type.
6145 * config/aarch64/aarch64.md
6146 (aarch64_load_tp_hard): Make type "mrs".
6147 * config/arm/arm.md
6148 (load_tp_hard): Make type "mrs".
6149 * config/arm/cortex-a15.md: Update with new attributes.
6150 * config/arm/cortex-a5.md: Update with new attributes.
6151 * config/arm/cortex-a53.md: Update with new attributes.
6152 * config/arm/cortex-a7.md: Update with new attributes.
6153 * config/arm/cortex-a8.md: Update with new attributes.
6154 * config/arm/cortex-a9.md: Update with new attributes.
6155 * config/arm/cortex-m4.md: Update with new attributes.
6156 * config/arm/cortex-r4.md: Update with new attributes.
6157 * config/arm/fa526.md: Update with new attributes.
6158 * config/arm/fa606te.md: Update with new attributes.
6159 * config/arm/fa626te.md: Update with new attributes.
6160 * config/arm/fa726te.md: Update with new attributes.
6161
6162 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6163
6164 * config/aarch64/aarch64.md
6165 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6166 (*movtf_aarch64): Likewise.
6167 * config/arm/arm.md
6168 (thumb1_movdi_insn): Use "multiple" for type where more than one
6169 instruction is used for a move.
6170 (*arm32_movhf): Likewise.
6171 (*thumb_movdf_insn): Likewise.
6172
6173 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6174
6175 * config/arm/types.md (type): Rename fcpys to fmov.
6176 * config/arm/vfp.md
6177 (*arm_movsi_vfp): Rename type fcpys as fmov.
6178 (*thumb2_movsi_vfp): Likewise
6179 (*movhf_vfp_neon): Likewise
6180 (*movhf_vfp): Likewise
6181 (*movsf_vfp): Likewise
6182 (*thumb2_movsf_vfp): Likewise
6183 (*movsfcc_vfp): Likewise
6184 (*thumb2_movsfcc_vfp): Likewise
6185 * config/aarch64/aarch64-simd.md
6186 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6187 * config/aarch64/aarch64.md
6188 (*movsi_aarch64): Replace type mov_reg with fmovs.
6189 (*movdi_aarch64): Likewise
6190 (*movsf_aarch64): Likewise
6191 (*movdf_aarch64): Likewise
6192 * config/arm/arm.c
6193 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6194 * config/arm/iwmmxt.md
6195 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6196 * config/arm/arm1020e.md: Update with new attributes.
6197 * config/arm/cortex-a15-neon.md: Update with new attributes.
6198 * config/arm/cortex-a5.md: Update with new attributes.
6199 * config/arm/cortex-a53.md: Update with new attributes.
6200 * config/arm/cortex-a7.md: Update with new attributes.
6201 * config/arm/cortex-a8-neon.md: Update with new attributes.
6202 * config/arm/cortex-a9.md: Update with new attributes.
6203 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6204 * config/arm/cortex-r4f.md: Update with new attributes.
6205 * config/arm/marvell-pj4.md: Update with new attributes.
6206 * config/arm/vfp11.md: Update with new attributes.
6207
6208 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6209
6210 * config/aarch64/aarch64.md
6211 (*madd<mode>): Fix type attribute.
6212 (*maddsi_uxtw): Likewise.
6213 (*msub<mode>): Likewise.
6214 (*msubsi_uxtw): Likewise.
6215 (<su_optab>maddsidi4): Likewise.
6216 (<su_optab>msubsidi4): Likewise.
6217
6218 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6219
6220 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6221 * config/arm/arm.md (core_cycles): Remove fdiv.
6222 * config/arm/vfp.md:
6223 (*sqrtsf2_vfp): Update for attribute changes.
6224 (*sqrtdf2_vfp): Likewise.
6225 * config/aarch64/aarch64.md:
6226 (sqrt<mode>2): Update for attribute changes.
6227 * config/arm/arm1020e.md: Update with new attributes.
6228 * config/arm/cortex-a15-neon.md: Update with new attributes.
6229 * config/arm/cortex-a5.md: Update with new attributes.
6230 * config/arm/cortex-a53.md: Update with new attributes.
6231 * config/arm/cortex-a7.md: Update with new attributes.
6232 * config/arm/cortex-a8-neon.md: Update with new attributes.
6233 * config/arm/cortex-a9.md: Update with new attributes.
6234 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6235 * config/arm/cortex-r4f.md: Update with new attributes.
6236 * config/arm/marvell-pj4.md: Update with new attributes.
6237 * config/arm/vfp11.md: Update with new attributes.
6238
6239 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6240
6241 * config/arm/types.md
6242 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6243 * config/aarch64/aarch64.md
6244 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6245 new attributes.
6246 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6247 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6248 (float<GPI:mode><GPF:mode>2): Likewise.
6249 * config/arm/vfp.md
6250 (*truncsisf2_vfp): Update with new attributes.
6251 (*truncsidf2_vfp): Likewise.
6252 (fixuns_truncsfsi2): Likewise.
6253 (fixuns_truncdfsi2): Likewise.
6254 (*floatsisf2_vfp): Likewise.
6255 (*floatsidf2_vfp): Likewise.
6256 (floatunssisf2): Likewise.
6257 (floatunssidf2): Likewise.
6258 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6259 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6260 * config/arm/arm1020e.md: Update with new attributes.
6261 * config/arm/cortex-a15-neon.md: Update with new attributes.
6262 * config/arm/cortex-a5.md: Update with new attributes.
6263 * config/arm/cortex-a53.md: Update with new attributes.
6264 * config/arm/cortex-a7.md: Update with new attributes.
6265 * config/arm/cortex-a8-neon.md: Update with new attributes.
6266 * config/arm/cortex-a9.md: Update with new attributes.
6267 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6268 * config/arm/cortex-r4f.md: Update with new attributes.
6269 * config/arm/marvell-pj4.md: Update with new attributes.
6270 * config/arm/vfp11.md: Update with new attributes.
6271
6272 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6273
6274 * config/aarch64/arm_neon.h
6275 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6276 (vqtbx<1,2,3,4><q>_s8): Likewise.
6277
6278 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6279
6280 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6281 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6282 (add<mode>3): Add type attribute.
6283 (add<mode>3): Likewise.
6284 (usadd<mode>3): Likewise.
6285 (ssadd<mode>3): Likewise.
6286 (sub<mode>3): Likewise.
6287 (sub<mode>3): Likewise.
6288 (ussub<mode>3): Likewise.
6289 (sssub<mode>3): Likewise.
6290 (ssmulsa3): Likewise.
6291 (usmulusa3): Likewise.
6292 (arm_usatsihi): Likewise.
6293 * config/arm/vfp.md
6294 (*movdi_vfp): Add types for all instructions.
6295 (*movdi_vfp_cortexa8): Likewise.
6296 (*movhf_vfp_neon): Likewise.
6297 (*movhf_vfp): Likewise.
6298 (*movdf_vfp): Likewise.
6299 (*thumb2_movdf_vfp): Likewise.
6300 (*thumb2_movdfcc_vfp): Likewise.
6301 * config/arm/arm.md: Add type attribute to all insn patterns.
6302 (*thumb1_adddi3): Add type attribute.
6303 (*arm_adddi3): Likewise.
6304 (*adddi_sesidi_di): Likewise.
6305 (*adddi_zesidi_di): Likewise.
6306 (*thumb1_addsi3): Likewise.
6307 (addsi3_compare0): Likewise.
6308 (*addsi3_compare0_scratch): Likewise.
6309 (*compare_negsi_si): Likewise.
6310 (cmpsi2_addneg): Likewise.
6311 (*addsi3_carryin_<optab>): Likewise.
6312 (*addsi3_carryin_alt2_<optab>): Likewise.
6313 (*addsi3_carryin_clobercc_<optab>): Likewise.
6314 (*subsi3_carryin): Likewise.
6315 (*subsi3_carryin_const): Likewise.
6316 (*subsi3_carryin_compare): Likewise.
6317 (*subsi3_carryin_compare_const): Likewise.
6318 (*arm_subdi3): Likewise.
6319 (*thumb_subdi3): Likewise.
6320 (*subdi_di_zesidi): Likewise.
6321 (*subdi_di_sesidi): Likewise.
6322 (*subdi_zesidi_di): Likewise.
6323 (*subdi_sesidi_di): Likewise.
6324 (*subdi_zesidi_ze): Likewise.
6325 (thumb1_subsi3_insn): Likewise.
6326 (*arm_subsi3_insn): Likewise.
6327 (*anddi3_insn): Likewise.
6328 (*anddi_zesidi_di): Likewise.
6329 (*anddi_sesdi_di): Likewise.
6330 (*ne_zeroextracts): Likewise.
6331 (*ne_zeroextracts): Likewise.
6332 (*ite_ne_zeroextr): Likewise.
6333 (*ite_ne_zeroextr): Likewise.
6334 (*anddi_notdi_di): Likewise.
6335 (*anddi_notzesidi): Likewise.
6336 (*anddi_notsesidi): Likewise.
6337 (andsi_notsi_si): Likewise.
6338 (thumb1_bicsi3): Likewise.
6339 (*iordi3_insn): Likewise.
6340 (*iordi_zesidi_di): Likewise.
6341 (*iordi_sesidi_di): Likewise.
6342 (*thumb1_iorsi3_insn): Likewise.
6343 (*xordi3_insn): Likewise.
6344 (*xordi_zesidi_di): Likewise.
6345 (*xordi_sesidi_di): Likewise.
6346 (*arm_xorsi3): Likewise.
6347 (*andsi_iorsi3_no): Likewise.
6348 (*smax_0): Likewise.
6349 (*smax_m1): Likewise.
6350 (*arm_smax_insn): Likewise.
6351 (*smin_0): Likewise.
6352 (*arm_smin_insn): Likewise.
6353 (*arm_umaxsi3): Likewise.
6354 (*arm_uminsi3): Likewise.
6355 (*minmax_arithsi): Likewise.
6356 (*minmax_arithsi_): Likewise.
6357 (*satsi_<SAT:code>): Likewise.
6358 (arm_ashldi3_1bit): Likewise.
6359 (arm_ashrdi3_1bit): Likewise.
6360 (arm_lshrdi3_1bit): Likewise.
6361 (*arm_negdi2): Likewise.
6362 (*thumb1_negdi2): Likewise.
6363 (*arm_negsi2): Likewise.
6364 (*thumb1_negsi2): Likewise.
6365 (*negdi_extendsid): Likewise.
6366 (*negdi_zero_extend): Likewise.
6367 (*arm_abssi2): Likewise.
6368 (*thumb1_abssi2): Likewise.
6369 (*arm_neg_abssi2): Likewise.
6370 (*thumb1_neg_abss): Likewise.
6371 (one_cmpldi2): Likewise.
6372 (extend<mode>di2): Likewise.
6373 (*compareqi_eq0): Likewise.
6374 (*arm_extendhisi2addsi): Likewise.
6375 (*arm_movdi): Likewise.
6376 (*thumb1_movdi_insn): Likewise.
6377 (*arm_movt): Likewise.
6378 (*thumb1_movsi_insn): Likewise.
6379 (pic_add_dot_plus_four): Likewise.
6380 (pic_add_dot_plus_eight): Likewise.
6381 (tls_load_dot_plus_eight): Likewise.
6382 (*thumb1_movhi_insn): Likewise.
6383 (*thumb1_movsf_insn): Likewise.
6384 (*movdf_soft_insn): Likewise.
6385 (*thumb_movdf_insn): Likewise.
6386 (cbranchsi4_insn): Likewise.
6387 (cbranchsi4_scratch): Likewise.
6388 (*negated_cbranchsi4): Likewise.
6389 (*tbit_cbranch): Likewise.
6390 (*tlobits_cbranch): Likewise.
6391 (*tstsi3_cbranch): Likewise.
6392 (*cbranchne_decr1): Likewise.
6393 (*addsi3_cbranch): Likewise.
6394 (*addsi3_cbranch_scratch): Likewise.
6395 (*arm_cmpdi_insn): Likewise.
6396 (*arm_cmpdi_unsig): Likewise.
6397 (*arm_cmpdi_zero): Likewise.
6398 (*thumb_cmpdi_zero): Likewise.
6399 (*deleted_compare): Likewise.
6400 (*mov_scc): Likewise.
6401 (*mov_negscc): Likewise.
6402 (*mov_notscc): Likewise.
6403 (*cstoresi_eq0_thumb1_insn): Likewise.
6404 (cstoresi_nltu_thumb1): Likewise.
6405 (cstoresi_ltu_thu): Likewise.
6406 (thumb1_addsi3_addgeu): Likewise.
6407 (*arm_jump): Likewise.
6408 (*thumb_jump): Likewise.
6409 (*check_arch2): Likewise.
6410 (arm_casesi_internal): Likewise.
6411 (thumb1_casesi_dispatch): Likewise.
6412 (*arm_indirect_jump): Likewise.
6413 (*thumb1_indirect_jump): Likewise.
6414 (nop): Likewise.
6415 (*and_scc): Likewise.
6416 (*ior_scc): Likewise.
6417 (*compare_scc): Likewise.
6418 (*cond_move): Likewise.
6419 (*cond_arith): Likewise.
6420 (*cond_sub): Likewise.
6421 (*cmp_ite0): Likewise.
6422 (*cmp_ite1): Likewise.
6423 (*cmp_and): Likewise.
6424 (*cmp_ior): Likewise.
6425 (*ior_scc_scc): Likewise.
6426 (*ior_scc_scc_cmp): Likewise.
6427 (*and_scc_scc): Likewise.
6428 (*and_scc_scc_cmp): Likewise.
6429 (*and_scc_scc_nod): Likewise.
6430 (*negscc): Likewise.
6431 (movcond_addsi): Likewise.
6432 (movcond): Likewise.
6433 (*ifcompare_plus_move): Likewise.
6434 (*if_plus_move): Likewise.
6435 (*ifcompare_move_plus): Likewise.
6436 (*if_move_plus): Likewise.
6437 (*ifcompare_arith_arith): Likewise.
6438 (*if_arith_arith): Likewise.
6439 (*ifcompare_arith_move): Likewise.
6440 (*if_arith_move): Likewise.
6441 (*ifcompare_move_arith): Likewise.
6442 (*if_move_arith): Likewise.
6443 (*ifcompare_move_not): Likewise.
6444 (*if_move_not): Likewise.
6445 (*ifcompare_not_move): Likewise.
6446 (*if_not_move): Likewise.
6447 (*ifcompare_shift_move): Likewise.
6448 (*if_shift_move): Likewise.
6449 (*ifcompare_move_shift): Likewise.
6450 (*if_move_shift): Likewise.
6451 (*ifcompare_shift_shift): Likewise.
6452 (*ifcompare_not_arith): Likewise.
6453 (*ifcompare_arith_not): Likewise.
6454 (*if_arith_not): Likewise.
6455 (*ifcompare_neg_move): Likewise.
6456 (*if_neg_move): Likewise.
6457 (*ifcompare_move_neg): Likewise.
6458 (*if_move_neg): Likewise.
6459 (prologue_thumb1_interwork): Likewise.
6460 (*cond_move_not): Likewise.
6461 (*sign_extract_onebit): Likewise.
6462 (*not_signextract_onebit): Likewise.
6463 (stack_tie): Likewise.
6464 (align_4): Likewise.
6465 (align_8): Likewise.
6466 (consttable_end): Likewise.
6467 (consttable_1): Likewise.
6468 (consttable_2): Likewise.
6469 (consttable_4): Likewise.
6470 (consttable_8): Likewise.
6471 (consttable_16): Likewise.
6472 (*thumb1_tablejump): Likewise.
6473 (prefetch): Likewise.
6474 (force_register_use): Likewise.
6475 (thumb_eh_return): Likewise.
6476 (load_tp_hard): Likewise.
6477 (load_tp_soft): Likewise.
6478 (tlscall): Likewise.
6479 (*arm_movtas_ze): Likewise.
6480 (*arm_rev): Likewise.
6481 (*arm_revsh): Likewise.
6482 (*arm_rev16): Likewise.
6483 * config/arm/thumb2.md
6484 (*thumb2_smaxsi3): Likewise.
6485 (*thumb2_sminsi3): Likewise.
6486 (*thumb32_umaxsi3): Likewise.
6487 (*thumb2_uminsi3): Likewise.
6488 (*thumb2_negdi2): Likewise.
6489 (*thumb2_abssi2): Likewise.
6490 (*thumb2_neg_abss): Likewise.
6491 (*thumb2_movsi_insn): Likewise.
6492 (tls_load_dot_plus_four): Likewise.
6493 (*thumb2_movhi_insn): Likewise.
6494 (*thumb2_mov_scc): Likewise.
6495 (*thumb2_mov_negs): Likewise.
6496 (*thumb2_mov_negs): Likewise.
6497 (*thumb2_mov_nots): Likewise.
6498 (*thumb2_mov_nots): Likewise.
6499 (*thumb2_movsicc_): Likewise.
6500 (*thumb2_movsfcc_soft_insn): Likewise.
6501 (*thumb2_indirect_jump): Likewise.
6502 (*thumb2_and_scc): Likewise.
6503 (*thumb2_ior_scc): Likewise.
6504 (*thumb2_ior_scc_strict_it): Likewise.
6505 (*thumb2_cond_move): Likewise.
6506 (*thumb2_cond_arith): Likewise.
6507 (*thumb2_cond_ari): Likewise.
6508 (*thumb2_cond_sub): Likewise.
6509 (*thumb2_negscc): Likewise.
6510 (*thumb2_movcond): Likewise.
6511 (thumb2_casesi_internal): Likewise.
6512 (thumb2_casesi_internal_pic): Likewise.
6513 (*thumb2_alusi3_short): Likewise.
6514 (*thumb2_mov<mode>_shortim): Likewise.
6515 (*thumb2_addsi_short): Likewise.
6516 (*thumb2_subsi_short): Likewise.
6517 (thumb2_addsi3_compare0): Likewise.
6518 (*thumb2_cbz): Likewise.
6519 (*thumb2_cbnz): Likewise.
6520 (*thumb2_one_cmplsi2_short): Likewise.
6521 (*thumb2_negsi2_short): Likewise.
6522 (*orsi_notsi_si): Likewise.
6523 * config/arm/arm1020e.md: Update with new attributes.
6524 * config/arm/arm1026ejs.md: Update with new attributes.
6525 * config/arm/arm1136jfs.md: Update with new attributes.
6526 * config/arm/arm926ejs.md: Update with new attributes.
6527 * config/arm/cortex-a15.md: Update with new attributes.
6528 * config/arm/cortex-a5.md: Update with new attributes.
6529 * config/arm/cortex-a53.md: Update with new attributes.
6530 * config/arm/cortex-a7.md: Update with new attributes.
6531 * config/arm/cortex-a8.md: Update with new attributes.
6532 * config/arm/cortex-a9.md: Update with new attributes.
6533 * config/arm/cortex-m4.md: Update with new attributes.
6534 * config/arm/cortex-r4.md: Update with new attributes.
6535 * config/arm/fa526.md: Update with new attributes.
6536 * config/arm/fa606te.md: Update with new attributes.
6537 * config/arm/fa626te.md: Update with new attributes.
6538 * config/arm/fa726te.md: Update with new attributes.
6539
6540 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6541
6542 * config/aarch64/aarch64-simd.md
6543 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6544 <vwx> iterator to ensure correct register choice.
6545 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6546 (aarch64_sqdmull_n<mode>): Likewise.
6547 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6548 * config/aarch64/arm_neon.h
6549 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6550 (vml<as><q>_n_<su>16): Likewise.
6551 (vml<as>l_high_lane<q>_<su>16): Likewise.
6552 (vml<as>l_high_n_<su>16): Likewise.
6553 (vml<as>l_lane<q>_<su>16): Likewise.
6554 (vml<as>l_n_<su>16): Likewise.
6555 (vmul<q>_lane<q>_<su>16): Likewise.
6556 (vmul<q>_n_<su>16): Likewise.
6557 (vmull_lane<q>_<su>16): Likewise.
6558 (vmull_n_<su>16): Likewise.
6559 (vmull_high_lane<q>_<su>16): Likewise.
6560 (vmull_high_n_<su>16): Likewise.
6561 (vqrdmulh<q>_n_s16): Likewise.
6562
6563 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6564
6565 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6566 have the correct lane parameter.
6567
6568 2013-09-06 Richard Biener <rguenther@suse.de>
6569
6570 * cfganal.c (control_dependences::~control_dependences):
6571 Properly free all of the vector.
6572
6573 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6574
6575 PR target/58269
6576 * config/i386/i386.c (ix86_conditional_register_usage):
6577 Proper initialize extended SSE registers.
6578
6579 2013-09-06 Jan Hubicka <jh@suse.cz>
6580
6581 PR tree-optimization/58311
6582 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6583
6584 2013-09-06 Jan Hubicka <jh@suse.cz>
6585
6586 * Makefile.in (tree-sra.o): Update dependencies.
6587 * tree-sra.c: Include ipa-utils.h
6588 (scan_function): Use recursive_call_p.
6589 (has_caller_p): New function.
6590 (cgraph_for_node_and_aliases): Count also callers of aliases.
6591
6592 2013-09-06 Jan Hubicka <jh@suse.cz>
6593
6594 PR middle-end/58094
6595 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6596 * tree-tailcall.c: Include ipa-utils.h.
6597 (find_tail_calls): Use it.
6598 * ipa-pure-const.c (check_call): Likewise.
6599 * ipa-utils.c (recursive_call_p): New function.
6600 * ipa-utils.h (recursive_call_p): Dclare.
6601 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6602 (symtab_semantically_equivalent_p): New function.
6603 * Makefile.in (tree-tailcall.o): Update dependencies.
6604
6605 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6606
6607 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6608 non-inlinable part.
6609
6610 2013-09-06 Richard Biener <rguenther@suse.de>
6611
6612 * lto-streamer.h (lto_global_var_decls): Remove.
6613 * Makefile.in (OBJS): Remove lto-symtab.o.
6614 (lto-symtab.o): Remove.
6615 (GTFILES): Remove lto-symtab.c
6616 * lto-symtab.c: Move to lto/
6617
6618 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6619
6620 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6621 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6622 (UNSPEC_FPINT_RINT): New constant definitions.
6623 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6624 definition with 2 attributes.
6625 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6626 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6627 definitions.
6628
6629 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6630
6631 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6632 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6633 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6634 attribute to "z196_alone".
6635 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6636 "zEC12_simple".
6637
6638 2013-09-06 Richard Biener <rguenther@suse.de>
6639
6640 * basic-block.h (class control_dependences): New.
6641 * tree-ssa-dce.c (control_dependence_map): Remove.
6642 (cd): New global.
6643 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6644 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6645 find_pdom, find_control_dependence, find_all_control_dependences):
6646 Move to cfganal.c.
6647 (mark_control_dependent_edges_necessary,
6648 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6649 tree_dce_done, perform_tree_ssa_dce): Adjust.
6650 * cfganal.c (set_control_dependence_map_bit,
6651 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6652 find_all_control_dependences): Move from tree-ssa-dce.c and
6653 implement as methods of control_dependences class.
6654 (control_dependences::control_dependences): New.
6655 (control_dependences::~control_dependences): Likewise.
6656 (control_dependences::get_edges_dependent_on): Likewise.
6657 (control_dependences::get_edge): Likewise.
6658
6659 2013-09-04 Jan Hubicka <jh@suse.cz>
6660
6661 * tree.c (types_same_for_odr): Drop overactive check.
6662 * ipa-devirt.c (hash_type_name): Likewise.
6663
6664 2013-09-04 Jan Hubicka <jh@suse.cz>
6665
6666 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6667 (analyze_functions): ... here.
6668
6669 2013-09-04 Jan Hubicka <jh@suse.cz>
6670
6671 PR middle-end/58201
6672 * cgraphunit.c (analyze_functions): Clear AUX fields
6673 after processing; initialize assembler name has.
6674
6675 2013-09-05 Jeff Law <law@redhat.com>
6676
6677 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6678 from thread_around_empty_block. Record threading path into PATH.
6679 Recurse if threading through the initial block is successful.
6680 (thread_across_edge): Corresponding changes to slightly simplify.
6681
6682 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6683
6684 * config/aarch64/aarch64.md
6685 (type): Remove frecpe, frecps, frecpx.
6686 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6687 fix to be a TARGET_SIMD instruction.
6688 (aarch64_frecps): Remove.
6689 * config/aarch64/aarch64-simd.md
6690 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6691 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6692
6693 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6694 Sofiane Naci <sofiane.naci@arm.com>
6695
6696 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6697 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6698 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6699 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6700 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6701 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6702 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6703 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6704 into "clz, "rbit". Rename "shift" to "shift_imm".
6705 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6706 changes. Update for attribute changes all occurrences of arlo_* and
6707 shift* types.
6708 * config/arm/arm-fixed.md: Update for attribute changes
6709 all occurrences of arlo_* types.
6710 * config/arm/thumb2.md: Update for attribute changes all occurrences
6711 of arlo_* types.
6712 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6713 (cortexa7_older_only): Likewise.
6714 (cortexa7_younger): Likewise.
6715 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6716 (1020alu_shift_op): Likewise.
6717 (1020alu_shift_reg_op): Likewise.
6718 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6719 (alu_shift_op): Likewise.
6720 (alu_shift_reg_op): Likewise.
6721 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6722 (11_alu_shift_op): Likewise.
6723 (11_alu_shift_reg_op): Likewise.
6724 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6725 (9_alu_shift_reg_op): Likewise.
6726 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6727 attribute changes.
6728 (cortex_a15_alu_shift): Likewise.
6729 (cortex_a15_alu_shift_reg): Likewise.
6730 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6731 attribute changes.
6732 (cortex_a5_alu_shift): Likewise.
6733 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6734 attribute changes.
6735 (cortex_a53_alu_shift): Likewise.
6736 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6737 attribute changes.
6738 (cortex_a7_alu_reg): Likewise.
6739 (cortex_a7_alu_shift): Likewise.
6740 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6741 attribute changes.
6742 (cortex_a8_alu_shift): Likewise.
6743 (cortex_a8_alu_shift_reg): Likewise.
6744 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6745 (cortex_a9_dp_shift): Likewise.
6746 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6747 attribute changes.
6748 * config/arm/cortex-r4.md
6749 (cortex_r4_alu): Update for attribute changes.
6750 (cortex_r4_mov): Likewise.
6751 (cortex_r4_alu_shift_reg): Likewise.
6752 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6753 (526_alu_shift_op): Likewise.
6754 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6755 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6756 (626te_alu_shift_op): Likewise.
6757 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6758 (726te_alu_shift_op): Likewise.
6759 (726te_alu_shift_reg_op): Likewise.
6760 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6761 (mp626_alu_shift_op): Likewise.
6762 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6763 (pj4_alu_conds): Likewise.
6764 (pj4_shift): Likewise.
6765 (pj4_shift_conds): Likewise.
6766 (pj4_alu_shift): Likewise.
6767 (pj4_alu_shift_conds): Likewise.
6768 * config/aarch64/aarch64.md: Update for attribute change
6769 all occurrences of arlo_* and shift* types.
6770
6771 2013-09-05 Mike Stump <mikestump@comcast.net>
6772
6773 * tree.h: Move documentation for tree_function_decl to tree-core.h
6774 with the declaration.
6775
6776 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6777
6778 PR target/58139
6779 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6780 looking for widest mode.
6781
6782 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6783
6784 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6785
6786 2013-09-05 Richard Biener <rguenther@suse.de>
6787
6788 PR tree-optimization/58137
6789 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6790 Do not create vectors of pointers.
6791 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6792 types for the components of the vector initializer.
6793 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6794 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6795
6796 2013-09-05 Martin Jambor <mjambor@suse.cz>
6797
6798 * ipa-prop.c (remove_described_reference): Accept missing references,
6799 return false if that hppens, otherwise return true.
6800 (cgraph_node_for_jfunc): New function.
6801 (try_decrement_rdesc_refcount): Likewise.
6802 (try_make_edge_direct_simple_call): Use them.
6803 (ipa_edge_removal_hook): Remove references from rdescs.
6804 (ipa_edge_duplication_hook): Clone rdescs and their references
6805 when the new edge has the same caller as the old one.
6806 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6807 reference before removing any edges.
6808
6809 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6810
6811 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6812 initial store.
6813 * thumb2.md (thumb2_storewb_parisi): New pattern.
6814
6815 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6816
6817 * config/aarch64/aarch64-option-extensions.def: Add
6818 AARCH64_OPT_EXTENSION of 'crc'.
6819 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6820 (AARCH64_ISA_CRC): Ditto.
6821 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6822 description of the CRC extension.
6823
6824 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6825
6826 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6827 * config/rs6000/linux.h: Ditto.
6828 * alpha/linux.h: Ditto.
6829 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6830 no_c99_libc_has_function.
6831 * config/c6x/uclinux-elf.h: Ditto.
6832 * config/lm32/uclinux-elf.h: Ditto.
6833 * config/m68k/uclinux.h: Ditto.
6834 * config/moxie/uclinux.h: Ditto.
6835 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6836 (crisv32-*-linux*, cris-*-linux*): Ditto.
6837 * config/bfin/bfin.c: Include "tm_p.h".
6838
6839 2013-09-05 Richard Biener <rguenther@suse.de>
6840
6841 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6842 check for a definition without a basic-block.
6843
6844 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6845 Sofiane Naci <sofiane.naci@arm.com>
6846
6847 * config/aarch64/aarch64.md
6848 (*movti_aarch64): Rename r_2_f and f_2_r.
6849 (*movsf_aarch64): Likewise.
6850 (*movdf_aarch64): Likewise.
6851 (*movtf_aarch64): Likewise.
6852 (aarch64_movdi_<mode>low): Likewise.
6853 (aarch64_movdi_<mode>high): Likewise.
6854 (aarch64_mov<mode>high_di): Likewise.
6855 (aarch64_mov<mode>low_di): Likewise.
6856 (aarch64_movtilow_tilow): Likewise.
6857 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6858 values to config/arm/types.md
6859 (attribute "conds"): Update for attribute change.
6860 (anddi3_insn): Likewise.
6861 (iordi3_insn): Likewise.
6862 (xordi3_insn): Likewise.
6863 (one_cmpldi2): Likewise.
6864 * config/arm/types.md (type): Add Neon types.
6865 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6866 use "type" attribute.
6867 (movmisalign<mode>_neon_store): Likewise.
6868 (movmisalign<mode>_neon_load): Likewise.
6869 (vec_set<mode>_internal): Likewise.
6870 (vec_setv2di_internal): Likewise.
6871 (vec_extract<mode>): Likewise.
6872 (vec_extractv2di): Likewise.
6873 (add<mode>3_neon): Likewise.
6874 (adddi3_neon): Likewise.
6875 (sub<mode>3_neon): Likewise.
6876 (subdi3_neon): Likewise.
6877 (mul<mode>3_neon): Likewise.
6878 (mul<mode>3add<mode>_neon): Likewise.
6879 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6880 (fma<VCVTF:mode>4)): Likewise.
6881 (fma<VCVTF:mode>4_intrinsic): Likewise.
6882 (fmsub<VCVTF:mode>4)): Likewise.
6883 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6884 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6885 (ior<mode>3): Likewise.
6886 (and<mode>3): Likewise.
6887 (anddi3_neon): Likewise.
6888 (orn<mode>3_neon): Likewise.
6889 (orndi3_neon): Likewise.
6890 (bic<mode>3_neon): Likewise.
6891 (bicdi3_neon): Likewise.
6892 (xor<mode>3): Likewise.
6893 (one_cmpl<mode>2): Likewise.
6894 (abs<mode>2): Likewise.
6895 (neg<mode>2): Likewise.
6896 (umin<mode>3_neon): Likewise.
6897 (umax<mode>3_neon): Likewise.
6898 (smin<mode>3_neon): Likewise.
6899 (smax<mode>3_neon): Likewise.
6900 (vashl<mode>3): Likewise.
6901 (vashr<mode>3_imm): Likewise.
6902 (vlshr<mode>3_imm): Likewise.
6903 (ashl<mode>3_signed): Likewise.
6904 (ashl<mode>3_unsigned): Likewise.
6905 (neon_load_count): Likewise.
6906 (ashldi3_neon_noclobber): Likewise.
6907 (signed_shift_di3_neon): Likewise.
6908 (unsigned_shift_di3_neon): Likewise.
6909 (ashrdi3_neon_imm_noclobber): Likewise.
6910 (lshrdi3_neon_imm_noclobber): Likewise.
6911 (widen_ssum<mode>3): Likewise.
6912 (widen_usum<mode>3): Likewise.
6913 (quad_halves_<code>v4si): Likewise.
6914 (quad_halves_<code>v4sf): Likewise.
6915 (quad_halves_<code>v8hi): Likewise.
6916 (quad_halves_<code>v16qi): Likewise.
6917 (reduc_splus_v2di): Likewise.
6918 (neon_vpadd_internal<mode>): Likewise.
6919 (neon_vpsmin<mode>): Likewise.
6920 (neon_vpsmax<mode>): Likewise.
6921 (neon_vpumin<mode>): Likewise.
6922 (neon_vpumax<mode>): Likewise.
6923 (ss_add<mode>_neon): Likewise.
6924 (us_add<mode>_neon): Likewise.
6925 (ss_sub<mode>_neon): Likewise.
6926 (us_sub<mode>_neon): Likewise.
6927 (neon_vadd<mode>_unspec): Likewise.
6928 (neon_vaddl<mode>): Likewise.
6929 (neon_vaddw<mode>): Likewise.
6930 (neon_vhadd<mode>): Likewise.
6931 (neon_vqadd<mode>): Likewise.
6932 (neon_vaddhn<mode>): Likewise.
6933 (neon_vmul<mode>): Likewise.
6934 (neon_vmla<mode>): Likewise.
6935 (neon_vmlal<mode>): Likewise.
6936 (neon_vmls<mode>): Likewise.
6937 (neon_vmlsl<mode>): Likewise.
6938 (neon_vqdmulh<mode>): Likewise.
6939 (neon_vqdmlal<mode>): Likewise.
6940 (neon_vqdmlsl<mode>): Likewise.
6941 (neon_vmull<mode>): Likewise.
6942 (neon_vqdmull<mode>): Likewise.
6943 (neon_vsub<mode>_unspec): Likewise.
6944 (neon_vsubl<mode>): Likewise.
6945 (neon_vsubw<mode>): Likewise.
6946 (neon_vqsub<mode>): Likewise.
6947 (neon_vhsub<mode>): Likewise.
6948 (neon_vsubhn<mode>): Likewise.
6949 (neon_vceq<mode>): Likewise.
6950 (neon_vcge<mode>): Likewise.
6951 (neon_vcgeu<mode>): Likewise.
6952 (neon_vcgt<mode>): Likewise.
6953 (neon_vcgtu<mode>): Likewise.
6954 (neon_vcle<mode>): Likewise.
6955 (neon_vclt<mode>): Likewise.
6956 (neon_vcage<mode>): Likewise.
6957 (neon_vcagt<mode>): Likewise.
6958 (neon_vtst<mode>): Likewise.
6959 (neon_vabd<mode>): Likewise.
6960 (neon_vabdl<mode>): Likewise.
6961 (neon_vaba<mode>): Likewise.
6962 (neon_vabal<mode>): Likewise.
6963 (neon_vmax<mode>): Likewise.
6964 (neon_vmin<mode>): Likewise.
6965 (neon_vpaddl<mode>): Likewise.
6966 (neon_vpadal<mode>): Likewise.
6967 (neon_vpmax<mode>): Likewise.
6968 (neon_vpmin<mode>): Likewise.
6969 (neon_vrecps<mode>): Likewise.
6970 (neon_vrsqrts<mode>): Likewise.
6971 (neon_vqabs<mode>): Likewise.
6972 (neon_vqneg<mode>): Likewise.
6973 (neon_vcls<mode>): Likewise.
6974 (clz<mode>2): Likewise.
6975 (popcount<mode>2): Likewise.
6976 (neon_vrecpe): Likewise.
6977 (neon_vrsqrte): Likewise.
6978 (neon_vget_lane<mode>_sext_internal): Likewise.
6979 (neon_vget_lane<mode>_zext_internal): Likewise.
6980 (neon_vdup_n<mode>): Likewise.
6981 (neon_vdup_nv2di): Likewise.
6982 (neon_vdpu_lane<mode>_internal): Likewise.
6983 (neon_vswp<mode>): Likewise.
6984 (float<mode><V_cvtto>2): Likewise.
6985 (floatuns<mode><V_cvtto>2): Likewise.
6986 (fix_trunc<mode><V_cvtto>)2): Likewise
6987 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6988 (neon_vcvt<mode>): Likewise.
6989 (neon_vcvtv4sfv4hf): Likewise.
6990 (neon_vcvtv4hfv4sf): Likewise.
6991 (neon_vcvt_n<mode>): Likewise.
6992 (neon_vmovn<mode>): Likewise.
6993 (neon_vqmovn<mode>): Likewise.
6994 (neon_vqmovun<mode>): Likewise.
6995 (neon_vmovl<mode>): Likewise.
6996 (neon_vmul_lane<mode>): Likewise.
6997 (neon_vmull_lane<mode>): Likewise.
6998 (neon_vqdmull_lane<mode>): Likewise.
6999 (neon_vqdmulh_lane<mode>): Likewise.
7000 (neon_vmla_lane<mode>): Likewise.
7001 (neon_vmlal_lane<mode>): Likewise.
7002 (neon_vqdmlal_lane<mode>): Likewise.
7003 (neon_vmls_lane<mode>): Likewise.
7004 (neon_vmlsl_lane<mode>): Likewise.
7005 (neon_vqdmlsl_lane<mode>): Likewise.
7006 (neon_vext<mode>): Likewise.
7007 (neon_vrev64<mode>): Likewise.
7008 (neon_vrev32<mode>): Likewise.
7009 (neon_vrev16<mode>): Likewise.
7010 (neon_vbsl<mode>_internal): Likewise.
7011 (neon_vshl<mode>): Likewise.
7012 (neon_vqshl<mode>): Likewise.
7013 (neon_vshr_n<mode>): Likewise.
7014 (neon_vshrn_n<mode>): Likewise.
7015 (neon_vqshrn_n<mode>): Likewise.
7016 (neon_vqshrun_n<mode>): Likewise.
7017 (neon_vshl_n<mode>): Likewise.
7018 (neon_vqshl_n<mode>): Likewise.
7019 (neon_vqshlu_n<mode>): Likewise.
7020 (neon_vshll_n<mode>): Likewise.
7021 (neon_vsra_n<mode>): Likewise.
7022 (neon_vsri_n<mode>): Likewise.
7023 (neon_vsli_n<mode>): Likewise.
7024 (neon_vtbl1v8qi): Likewise.
7025 (neon_vtbl2v8qi): Likewise.
7026 (neon_vtbl3v8qi): Likewise.
7027 (neon_vtbl4v8qi): Likewise.
7028 (neon_vtbx1v8qi): Likewise.
7029 (neon_vtbx2v8qi): Likewise.
7030 (neon_vtbx3v8qi): Likewise.
7031 (neon_vtbx4v8qi): Likewise.
7032 (neon_vtrn<mode>_internal): Likewise.
7033 (neon_vzip<mode>_internal): Likewise.
7034 (neon_vuzp<mode>_internal): Likewise.
7035 (neon_vld1<mode>): Likewise.
7036 (neon_vld1_lane<mode>): Likewise.
7037 (neon_vld1_dup<mode>): Likewise.
7038 (neon_vld1_dupv2di): Likewise.
7039 (neon_vst1<mode>): Likewise.
7040 (neon_vst1_lane<mode>): Likewise.
7041 (neon_vld2<mode>): Likewise.
7042 (neon_vld2_lane<mode>): Likewise.
7043 (neon_vld2_dup<mode>): Likewise.
7044 (neon_vst2<mode>): Likewise.
7045 (neon_vst2_lane<mode>): Likewise.
7046 (neon_vld3<mode>): Likewise.
7047 (neon_vld3qa<mode>): Likewise.
7048 (neon_vld3qb<mode>): Likewise.
7049 (neon_vld3_lane<mode>): Likewise.
7050 (neon_vld3_dup<mode>): Likewise.
7051 (neon_vst3<mode>): Likewise.
7052 (neon_vst3qa<mode>): Likewise.
7053 (neon_vst3qb<mode>): Likewise.
7054 (neon_vst3_lane<mode>): Likewise.
7055 (neon_vld4<mode>): Likewise.
7056 (neon_vld4qa<mode>): Likewise.
7057 (neon_vld4qb<mode>): Likewise.
7058 (neon_vld4_lane<mode>): Likewise.
7059 (neon_vld4_dup<mode>): Likewise.
7060 (neon_vst4<mode>): Likewise.
7061 (neon_vst4qa<mode>): Likewise.
7062 (neon_vst4qb<mode>): Likewise.
7063 (neon_vst4_lane<mode>): Likewise.
7064 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7065 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7066 (neon_vec_<US>mult_lo_<mode>): Likewise.
7067 (neon_vec_<US>mult_hi_<mode>): Likewise.
7068 (neon_vec_<US>shiftl_<mode>): Likewise.
7069 (neon_unpack<US>_<mode>): Likewise.
7070 (neon_vec_<US>mult_<mode>): Likewise.
7071 (vec_pack_trunc_<mode>): Likewise.
7072 (neon_vec_pack_trunk_<mode>): Likewise.
7073 (neon_vabd<mode>_2): Likewise.
7074 (neon_vabd<mode>_3): Likewise.
7075 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7076 (thumb2_movsi_vfp): Likewise.
7077 (movdi_vfp): Likewise.
7078 (movdi_vfp_cortexa8): Likewise.
7079 (movhf_vfp_neon): Likewise.
7080 (movhf_vfp): Likewiwse.
7081 (movsf_vfp): Likewiwse.
7082 (thumb2_movsf_vfp): Likewiwse.
7083 (movdf_vfp): Likewise.
7084 (thumb2_movdf_vfp): Likewise.
7085 (movsfcc_vfp): Likewise.
7086 (thumb2_movsfcc_vfp): Likewise.
7087 (movdfcc_vfp): Likewise.
7088 (thumb2_movdfcc_vfp): Likewise.
7089 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7090 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7091 (v10_v2c): Likewise.
7092 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7093 attribute change.
7094 (cortex_a15_neon_int_2): Likewise.
7095 (cortex_a15_neon_int_3): Likewise.
7096 (cortex_a15_neon_int_4): Likewise.
7097 (cortex_a15_neon_int_5): Likewise.
7098 (cortex_a15_neon_vqneg_vqabs): Likewise.
7099 (cortex_a15_neon_vmov): Likewise.
7100 (cortex_a15_neon_vaba): Likewise.
7101 (cortex_a15_neon_vaba_qqq): Likewise.
7102 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7103 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7104 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7105 scalar_64_32_long_scalar): Likewise.
7106 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7107 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7108 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7109 lotype_qdd_64_32_long): Likewise.
7110 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7111 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7112 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7113 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7114 (cortex_a15_neon_shift_1): Likewise.
7115 (cortex_a15_neon_shift_2): Likewise.
7116 (cortex_a15_neon_shift_3): Likewise.
7117 (cortex_a15_neon_vshl_ddd): Likewise.
7118 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7119 (cortex_a15_neon_vsra_vrsra): Likewise.
7120 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7121 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7122 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7123 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7124 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7125 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7126 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7127 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7128 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7129 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7130 (cortex_a15_neon_bp_simple): Likewise.
7131 (cortex_a15_neon_bp_2cycle): Likewise.
7132 (cortex_a15_neon_bp_3cycle): Likewise.
7133 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7134 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7135 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7136 (cortex_a15_neon_vld2_4_regs): Likewise.
7137 (cortex_a15_neon_vld3_vld4): Likewise.
7138 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7139 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7140 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7141 (cortex_a15_neon_vst3_vst4): Likewise.
7142 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7143 (cortex_a15_neon_vld3_vld4_lane" 10
7144 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7145 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7146 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7147 (cortex_a15_neon_ldm_2): Likewise.0
7148 (cortex_a15_neon_stm_2): Likewise.
7149 (cortex_a15_neon_mcr): Likewise.
7150 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7151 (cortex_a15_neon_mrc): Likewise.
7152 (cortex_a15_neon_mrrc): Likewise.
7153 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7154 change.
7155 (cortex_a15_alu_shift): Likewise.
7156 (cortex_a15_alu_shift_reg): Likewise.
7157 (cortex_a15_mult32): Likewise.
7158 (cortex_a15_mult64): Likewise.
7159 (cortex_a15_block): Likewise.
7160 (cortex_a15_branch): Likewise.
7161 (cortex_a15_load1): Likewise.
7162 (cortex_a15_load3): Likewise.
7163 (cortex_a15_store1): Likewise.
7164 (cortex_a15_store3): Likewise.
7165 (cortex_a15_call): Likewise.
7166 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7167 (cortex_a5_f2r): Likewise.
7168 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7169 change.
7170 (cortex_a53_f2r): Likewise.
7171 * config/arm/cortex-a7.md
7172 (cortex_a7_branch): Update for attribute change.
7173 (cortex_a7_call): Likewise.
7174 (cortex_a7_alu_imm): Likewise.
7175 (cortex_a7_alu_reg): Likewise.
7176 (cortex_a7_alu_shift): Likewise.
7177 (cortex_a7_mul): Likewise.
7178 (cortex_a7_load1): Likewise.
7179 (cortex_a7_store1): Likewise.
7180 (cortex_a7_load2): Likewise.
7181 (cortex_a7_store2): Likewise.
7182 (cortex_a7_load3): Likewise.
7183 (cortex_a7_store3): Likewise.
7184 (cortex_a7_load4): Likewise.
7185 (cortex_a7_store4): Likewise.
7186 (cortex_a7_fpalu): Likewise.
7187 (cortex_a7_fconst): Likewise.
7188 (cortex_a7_fpmuls): Likewise.
7189 (cortex_a7_neon_mul): Likewise.
7190 (cortex_a7_fpmacs): Likewise.
7191 (cortex_a7_neon_mla: Likewise.
7192 (cortex_a7_fpmuld: Likewise.
7193 (cortex_a7_fpmacd: Likewise.
7194 (cortex_a7_fpfmad: Likewise.
7195 (cortex_a7_fdivs: Likewise.
7196 (cortex_a7_fdivd: Likewise.
7197 (cortex_a7_r2f: Likewise.
7198 (cortex_a7_f2r: Likewise.
7199 (cortex_a7_f_flags: Likewise.
7200 (cortex_a7_f_loads: Likewise.
7201 (cortex_a7_f_loadd: Likewise.
7202 (cortex_a7_f_stores: Likewise.
7203 (cortex_a7_f_stored: Likewise.
7204 (cortex_a7_neon): Likewise.
7205 * config/arm/cortex-a8-neon.md
7206 (cortex_a8_neon_mrc): Update for attribute change.
7207 (cortex_a8_neon_mrrc): Likewise.
7208 (cortex_a8_neon_int_1): Likewise.
7209 (cortex_a8_neon_int_2): Likewise.
7210 (cortex_a8_neon_int_3): Likewise.
7211 (cortex_a8_neon_int_4): Likewise.
7212 (cortex_a8_neon_int_5): Likewise.
7213 (cortex_a8_neon_vqneg_vqabs): Likewise.
7214 (cortex_a8_neon_vmov): Likewise.
7215 (cortex_a8_neon_vaba): Likewise.
7216 (cortex_a8_neon_vaba_qqq): Likewise.
7217 (cortex_a8_neon_vsma): Likewise.
7218 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7219 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7220 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7221 long_scalar): Likewise.
7222 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7223 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7224 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7225 long_scalar_qdd_64_32_long): Likewise.
7226 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7227 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7228 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7229 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7230 (cortex_a8_neon_shift_1): Likewise.
7231 (cortex_a8_neon_shift_2): Likewise.
7232 (cortex_a8_neon_shift_3): Likewise.
7233 (cortex_a8_neon_vshl_ddd): Likewise.
7234 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7235 (cortex_a8_neon_vsra_vrsra): Likewise.
7236 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7237 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7238 (cortex_a8_neon_fp_vsum): Likewise.
7239 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7240 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7241 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7242 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7243 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7244 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7245 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7246 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7247 (cortex_a8_neon_bp_simple): Likewise.
7248 (cortex_a8_neon_bp_2cycle): Likewise.
7249 (cortex_a8_neon_bp_3cycle): Likewise.
7250 (cortex_a8_neon_ldr): Likewise.
7251 (cortex_a8_neon_str): Likewise.
7252 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7253 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7254 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7255 (cortex_a8_neon_vld2_4_regs): Likewise.
7256 (cortex_a8_neon_vld3_vld4): Likewise.
7257 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7258 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7259 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7260 (cortex_a8_neon_vst3_vst4): Likewise.
7261 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7262 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7263 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7264 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7265 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7266 (cortex_a8_neon_mcr): Likewise.
7267 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7268 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7269 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7270 change.
7271 (ca9_neon_mrrc): Likewise.
7272 (cortex_a9_neon_int_1): Likewise.
7273 (cortex_a9_neon_int_2): Likewise.
7274 (cortex_a9_neon_int_3): Likewise.
7275 (cortex_a9_neon_int_4): Likewise.
7276 (cortex_a9_neon_int_5): Likewise.
7277 (cortex_a9_neon_vqneg_vqabs): Likewise.
7278 (cortex_a9_neon_vmov): Likewise.
7279 (cortex_a9_neon_vaba): Likewise.
7280 (cortex_a9_neon_vaba_qqq): Likewise.
7281 (cortex_a9_neon_vsma): Likewise.
7282 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7283 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7284 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7285 long_scalar): Likewise.
7286 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7287 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7288 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7289 long_scalar_qdd_64_32_long): Likewise.
7290 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7291 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7292 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7293 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7294 (cortex_a9_neon_shift_1): Likewise.
7295 (cortex_a9_neon_shift_2): Likewise.
7296 (cortex_a9_neon_shift_3): Likewise.
7297 (cortex_a9_neon_vshl_ddd): Likewise.
7298 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7299 (cortex_a9_neon_vsra_vrsra): Likewise.
7300 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7301 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7302 (cortex_a9_neon_fp_vsum): Likewise.
7303 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7304 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7305 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7306 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7307 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7308 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7309 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7310 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7311 (cortex_a9_neon_bp_simple): Likewise.
7312 (cortex_a9_neon_bp_2cycle): Likewise.
7313 (cortex_a9_neon_bp_3cycle): Likewise.
7314 (cortex_a9_neon_ldr): Likewise.
7315 (cortex_a9_neon_str): Likewise.
7316 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7317 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7318 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7319 (cortex_a9_neon_vld2_4_regs): Likewise.
7320 (cortex_a9_neon_vld3_vld4): Likewise.
7321 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7322 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7323 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7324 (cortex_a9_neon_vst3_vst4): Likewise.
7325 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7326 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7327 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7328 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7329 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7330 (cortex_a9_neon_mcr): Likewise.
7331 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7332 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7333 (cortex_a9_fps): Likewise.
7334 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7335 change.
7336 (cortex_m4_fmuls): Likewise.
7337 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7338 change.
7339 (cortex_r4_mrc): Likewise.
7340 * config/arm/iterators.md: Update comment referring to neon_type.
7341 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7342 (iwmmxt_movsi_insn): Likewise.
7343 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7344 attribute change.
7345 (pj4_core_to_vfp): Likewise.
7346 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7347 attribute change.
7348 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7349 (vfp_fstore): Likewise.
7350 * doc/md.texi: Change references to neon_type to refer to type.
7351
7352 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7353
7354 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7355
7356 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7357
7358 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7359 lglobl if not weak.
7360
7361 2013-09-04 Easwaran Raman <eraman@google.com>
7362
7363 PR middle-end/57370
7364 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7365 (build_and_add_sum): Use it.
7366 (appears_later_in_bb): Simplify code.
7367
7368 2013-09-04 Teresa Johnson <tejohnson@google.com>
7369
7370 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7371
7372 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7373
7374 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7375
7376 2013-09-04 Jan Hubicka <jh@suse.cz>
7377
7378 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7379 * ipa-devirt.c: Include diganostic.h
7380 (odr_type_d): Add types and types_set.
7381 (hash_type_name): Work for types with vtables during LTO.
7382 (odr_hasher::remove): Fix comment; destroy types_set.
7383 (add_type_duplicate): New function,
7384 (get_odr_type): Use it.
7385 (dump_type_inheritance_graph): Dump type duplicates.
7386 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7387 graph.
7388 * tree.c (types_same_for_odr): Give exact answers on types with
7389 virtual tables.
7390
7391 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7392
7393 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7394 explaining their differences.
7395
7396 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7397
7398 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7399
7400 2013-09-03 Jeff Law <law@redhat.com>
7401
7402 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7403 when not threading through a joiner block. Pass joiner/no joiner
7404 state to register_jump_thread.
7405 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7406 state from argument rather than implying on path length.
7407 Dump the entire jump thread path into debugging dump.
7408 * tree-flow.h (register_jump_thread): Update prototype.
7409
7410 2013-08-29 Xinliang David Li <davidxl@google.com>
7411
7412 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7413 Remove a trivial gcc_assert.
7414
7415 2013-08-29 Xinliang David Li <davidxl@google.com>
7416
7417 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7418 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7419 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7420 Delay base decl alignment adjustment.
7421 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7422 * tree-vectorizer.h: New data structure.
7423 (set_dr_misalignment): New function.
7424 (dr_misalignment): Ditto.
7425 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7426 (vectorizable_load): Ditto.
7427 (ensure_base_align): New function.
7428 (vectorize_loops): Add dbg_cnt support.
7429 (execute_vect_slp): Ditto.
7430 * dbgcnt.def: New debug counter.
7431 * Makefile: New dependency.
7432
7433 2013-09-03 Meador Inge <meadori@codesourcery.com>
7434
7435 Revert:
7436
7437 2013-08-30 Meador Inge <meadori@codesourcery.com>
7438
7439 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7440
7441 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7442
7443 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7444 function descriptor.
7445
7446 2013-09-03 Richard Biener <rguenther@suse.de>
7447
7448 * tree-affine.c (add_elt_to_tree): Fix association issue,
7449 avoid useless converts and make sure to always return a
7450 properly typed result.
7451
7452 2013-09-03 Richard Biener <rguenther@suse.de>
7453
7454 PR middle-end/57656
7455 * fold-const.c (negate_expr_p): Fix division case.
7456 (negate_expr): Likewise.
7457
7458 2013-09-03 Richard Biener <rguenther@suse.de>
7459
7460 PR lto/58285
7461 * tree-streamer-out.c: Include tm.h.
7462 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7463
7464 2013-09-03 Jan Hubicka <jh@suse.cz>
7465
7466 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7467
7468 2013-09-03 Alan Modra <amodra@gmail.com>
7469
7470 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7471 multilibs to big-endian target and vice versa.
7472 * config/rs6000/t-linux64: Use := assignment on all vars.
7473 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7474 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7475 * config/rs6000/t-linux64le: New file.
7476 * config/rs6000/t-linux64bele: New file.
7477 * config/rs6000/t-linux64lebe: New file.
7478
7479 2013-09-02 Jan Hubicka <jh@suse.cz>
7480
7481 * ipa-inline-transform.c (inline_transform): Do not
7482 optimize_inline_calls when not optimizing.
7483
7484 2013-09-02 Jan Hubicka <jh@suse.cz>
7485
7486 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7487 duplicated nodes for assembler names.
7488 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7489 hard registers from assembler name hash.
7490
7491 2013-09-02 Jan Hubicka <jh@suse.cz>
7492
7493 * ipa-split.c (execute_split_functions): Split externally visible
7494 functions called once.
7495
7496 2013-09-02 Martin Jambor <mjambor@suse.cz>
7497
7498 PR ipa/58106
7499 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7500 linked list. When finding the correct duplicate, also consider also
7501 the caller in additon to its inlined_to node.
7502
7503 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7504
7505 * config/aarch64/aarch64-simd-builtins.def
7506 (dup_lane_scalar): Remove.
7507 * config/aarch64/aarch64-simd.md
7508 (aarch64_simd_dup): Add 'w->w' alternative.
7509 (aarch64_dup_lane<mode>): Allow for VALL.
7510 (aarch64_dup_lane_scalar<mode>): Remove.
7511 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7512 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7513 (aarch64_get_lane_unsigned<mode>): Likewise.
7514 (aarch64_get_lane<mode>): Likewise.
7515 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7516 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7517 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7518 (VCON): Change container of V2SF.
7519 (vswap_width_name): Likewise.
7520 * config/aarch64/arm_neon.h
7521 (__aarch64_vdup_lane_any): New.
7522 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7523 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7524 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7525
7526 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7527
7528 PR middle-end/56382
7529 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7530 the source or the destination is a single hard register.
7531
7532 2013-09-02 Richard Biener <rguenther@suse.de>
7533
7534 PR middle-end/57511
7535 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7536 non-linear SCEVs.
7537
7538 2013-09-02 Richard Biener <rguenther@suse.de>
7539
7540 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7541 arithmetic to sizetype.
7542
7543 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7544
7545 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7546 Find auto-increment use both before and after candidate.
7547
7548 2013-09-02 Marek Polacek <polacek@redhat.com>
7549
7550 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7551
7552 2013-09-01 Jan Hubicka <jh@suse.cz>
7553
7554 * Makefile.in: Add ipa-profile.o
7555 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7556 * cgraph.c (struct cgraph_propagate_frequency_data,
7557 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7558 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7559 * cgraph.h (cgraph_propagate_frequency): Remove.
7560 * ipa-inline-analysis.c: Include ipa-utils.h;
7561 drop duplicated cfgloop.h.
7562 (inline_update_callee_summaries): Update.
7563 * ipa-profile.c: New file.
7564 * ipa-utils.h (ipa_propagate_frequency): Declare.
7565 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7566 data-streamer.h, value-prof.h.
7567 (symtab_remove_unreachable_nodes): Update profile.
7568 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7569 account_time_size, cmp_counts, dump_histogram,
7570 ipa_profile_generate_summary, ipa_profile_write_summary,
7571 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7572 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7573 Move to ipa-profile.c.
7574
7575 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7576
7577 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7578
7579 2013-09-01 Jan Hubicka <jh@suse.cz>
7580
7581 * common.opt (fdevirtualize-speculatively): New function.
7582 * invoke.texi (fdevirtualize-speculatively): Document.
7583 * ipa-devirt.c: Include ipa-inline.h
7584 (likely_target_p): New function.
7585 (ipa_devirt): New function.
7586 (gate_ipa_devirt): New function.
7587 (pass_data_ipa_devirt): New static var.
7588 (pass_ipa_devirt): Likewise.
7589 (make_pass_ipa_devirt): New function.
7590 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7591 (common_handle_option): Disable devirtualization when
7592 value range profiling is available.
7593 * passes.def (pass_ipa_devirt): Add.
7594 * timever.def (TV_IPA_DEVIRT): New timevar.
7595 * tree-pass.h (make_pass_ipa_devirt):
7596
7597 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7598
7599 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7600 include sanitize(undefined).
7601
7602 2013-08-31 Diego Novillo <dnovillo@google.com>
7603
7604 * Makefile.in (TREE_CORE_H): Define.
7605 (TREE_H): Use.
7606 (GTFILES): Add tree-core.h.
7607 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7608 size the array.
7609 * tree-core.h: New file.
7610 Move all data structures, enum, typedefs, global
7611 declarations and constants from ...
7612 * tree.h: ... here.
7613
7614 2013-08-31 Jan Hubicka <jh@suse.cz>
7615
7616 * bulitins.c (expand_builtin): Do not early exit for gcov
7617 instrumented functions.
7618
7619 2013-08-31 Marek Polacek <polacek@redhat.com>
7620
7621 * ubsan.c: Include tm_p.h.
7622
7623 2013-08-31 Jan Hubicka <jh@suse.cz>
7624
7625 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7626 warning.
7627
7628 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7629 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7630 * tree-cfg.c (verify_gimple_label): ... here.
7631 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7632 (ipa_merge_profiles): New function.
7633 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7634 (lto_input_function_body): Likewise.
7635 * ipa-utils.h (ipa_merge_profiles): Declare.
7636 * lto-streamer.h (lto_input_function_body): Update prototype.
7637 (emit_label_in_global_context_p): Remove.
7638 * lto-symtab.c: Include ipa-utils.h
7639 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7640
7641 2013-08-31 Jan Hubicka <jh@suse.cz>
7642
7643 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7644
7645 2013-08-31 Jan Hubicka <jh@suse.cz>
7646
7647 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7648
7649 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7650
7651 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7652 "cmp" RTX before signed_comparison_operator check to account
7653 for "code" changes.
7654
7655 2013-08-30 Jan Hubicka <jh@suse.cz>
7656
7657 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7658 (detect_type_change_1): Rename to ...
7659 (detect_type_change): ... this one; early return on non-polymorphic
7660 types.
7661 (detect_type_change_ssa): Add comp_type parameter; update
7662 use of detect_type_change.
7663 (compute_complex_assign_jump_func): Add param_type parameter;
7664 update use of detect_type_change_ssa.
7665 (compute_complex_ancestor_jump_func): Likewise.
7666 (ipa_get_callee_param_type): New function.
7667 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7668 update calls to the jump function computation functions.
7669
7670 2013-08-30 Teresa Johnson <tejohnson@google.com>
7671 Steven Bosscher <steven@gcc.gnu.org>
7672
7673 * cfgrtl.c (fixup_new_cold_bb): New routine.
7674 (commit_edge_insertions): Invoke fixup_partitions.
7675 (find_partition_fixes): New routine.
7676 (fixup_partitions): Ditto.
7677 (verify_hot_cold_block_grouping): Update comments.
7678 (rtl_verify_edges): Invoke find_partition_fixes.
7679 (rtl_verify_bb_pointers): Update comments.
7680 (rtl_verify_bb_layout): Ditto.
7681 * basic-block.h (probably_never_executed_edge_p): Declare.
7682 (fixup_partitions): Ditto.
7683 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7684 * bb-reorder.c (sanitize_hot_paths): New function.
7685 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7686 sanitize_hot_paths.
7687 * predict.c (probably_never_executed_edge_p): New routine.
7688 * cfg.c (check_bb_profile): Add partition insanity warnings.
7689
7690 2013-08-30 Meador Inge <meadori@codesourcery.com>
7691
7692 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7693
7694 2013-08-30 Marek Polacek <polacek@redhat.com>
7695
7696 * Makefile.in (ubsan.o): Add.
7697 (c-family/c-ubsan.o): Add.
7698 (builtins.o): Add ubsan.h dependency.
7699 * ubsan.h: New file.
7700 * ubsan.c: New file.
7701 * common.opt: Add -fsanitize=undefined option.
7702 (flag_sanitize): Add variable.
7703 (fsanitize=): Add option. Add Driver.
7704 (fsanitize=thread): Remove option.
7705 (fsanitize=address): Likewise.
7706 (static-libubsan): New option.
7707 * doc/invoke.texi: Document the new flag and -static-libubsan.
7708 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7709 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7710 * builtin-attrs.def (ATTR_COLD): Define.
7711 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7712 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7713 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7714 * flag-types.h (sanitize_code): New enum.
7715 * opts.c (common_handle_option): Parse command line arguments
7716 of -fsanitize=. Add -fsanitize=unreachable option.
7717 * varasm.c (get_variable_section): Adjust.
7718 (assemble_noswitch_variable): Likewise.
7719 (assemble_variable): Likewise.
7720 (output_constant_def_contents): Likewise.
7721 (categorize_decl_for_section): Likewise.
7722 (place_block_symbol): Likewise.
7723 (output_object_block): Likewise.
7724 * builtins.def: Likewise.
7725 * toplev.c (compile_file): Likewise.
7726 (process_options): Likewise.
7727 * cppbuiltin.c: Likewise.
7728 * tsan.c (tsan_pass): Likewise.
7729 (tsan_gate): Likewise.
7730 (tsan_gate_O0): Likewise.
7731 * cfgexpand.c (partition_stack_vars): Likewise.
7732 (expand_stack_vars): Likewise.
7733 (defer_stack_allocation): Likewise.
7734 (expand_used_vars): Likewise.
7735 * cfgcleanup.c (old_insns_match_p): Likewise.
7736 * asan.c (asan_finish_file): Likewise.
7737 (asan_instrument): Likewise.
7738 (gate_asan): Likewise.
7739 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7740 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7741 (asan_global_struct): Use pointer_sized_int_node instead
7742 calling build_nonstandard_integer_type.
7743 (initialize_sanitizer_builtins): Likewise.
7744 (asan_finish_file): Likewise.
7745 * gcc.c: Document %{%:function(args):X}.
7746 (static_spec_functions): Add sanitize.
7747 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7748 store funcval != NULL there.
7749 (do_spec_1): Adjust handle_spec_function caller.
7750 (handle_braces): Allow %:function(args) as condition.
7751 (sanitize_spec_function): New function.
7752 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7753 (LIBUBSAN_SPEC): Likewise.
7754 (LIBUBSAN_EARLY_SPEC): Likewise.
7755 (SANITIZER_SPEC): Handle libubsan.
7756 (SANITIZER_EARLY_SPEC): Likewise.
7757 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7758 instead of fsanitize=address.
7759 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7760 instead of fsanitize=address*.
7761 * builtins.c: Include ubsan.h.
7762 (fold_builtin_0): Instrument __builtin_unreachable.
7763 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7764 instead of flag_asan.
7765 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7766 (pointer_sized_int_node): Define.
7767 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7768
7769 2013-08-30 Mike Stump <mikestump@comcast.net>
7770
7771 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7772 with RE patterns.
7773
7774 2013-08-29 Jan Hubicka <jh@suse.cz>
7775
7776 * cgraph.c (cgraph_function_body_availability): Handle weakref
7777 correctly.
7778 * passes.def: Remove pass_fixup_cfg.
7779 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7780 track when we need to remove functions.
7781 (gate_ipa_inline): Execute inlining always; add comment why.
7782 (pass_data_ipa_inline): Remove TODO_remove_functions.
7783 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7784 do not produce summaries.
7785 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7786 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7787 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7788 constant pool and vtable.
7789
7790 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7791
7792 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7793 New arm_neon.h's internal macros to specify 64-bit constants.
7794 Avoid using stdint.h's macros.
7795
7796 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7797
7798 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7799 before applying REG_P.
7800
7801 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7802
7803 PR tree-optimization/58277
7804 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7805 after seeing too many stmts with vdef in between dombb and current
7806 bb, invalidate everything.
7807
7808 2013-08-30 Richard Biener <rguenther@suse.de>
7809
7810 * fold-const.c (fold_single_bit_test): Fix overflow test.
7811
7812 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7813
7814 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7815 and which can live in a register, always retrieve the value on entry.
7816 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7817 passed by invisible reference specially.
7818 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7819 (vt_add_function_parameter): Correctly deal with a parameter passed by
7820 invisible reference.
7821
7822 2013-08-30 Jan Hubicka <jh@suse.cz>
7823
7824 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7825
7826 2013-08-30 Richard Biener <rguenther@suse.de>
7827
7828 PR tree-optimization/58228
7829 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7830 allow invariant loads in nested loop vectorization.
7831
7832 2013-08-30 Richard Biener <rguenther@suse.de>
7833
7834 PR tree-optimization/58223
7835 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7836 (has_anti_or_output_dependence): ... this and adjust to also
7837 look for output dependences.
7838 (mark_nodes_having_upstream_mem_writes): Adjust.
7839 (rdg_flag_uses): Likewise.
7840
7841 2013-08-30 Richard Biener <rguenther@suse.de>
7842
7843 PR tree-optimization/58010
7844 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7845 assert that we have a loop-closed PHI.
7846
7847 2013-08-29 Jan Hubicka <jh@suse.cz>
7848
7849 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7850 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7851 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7852 (lto_free_function_in_decl_state_for_node): New function.
7853
7854 2013-08-29 Xinliang David Li <davidxl@google.com>
7855
7856 * loop-unroll.c (report_unroll_peel): Minor message change.
7857 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7858 Emit alignment peeling message with default -fopt-info.
7859 (vect_loop_versioning): Emit loop version info message.
7860 * tree-vectorizer.c (vectorize_loops): Minor message change.
7861 (execute_vect_slp): Ditto.
7862
7863 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7864
7865 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7866 of the clone from the DECL_NAME of the original function.
7867
7868 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7869
7870 * passes.c (register_pass): Add overload.
7871 * tree-pass.h (register_pass): Forward declare it. Add comment.
7872
7873 2013-08-29 Jan Hubicka <jh@suse.cz>
7874
7875 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7876 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7877 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7878 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7879 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7880 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7881 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7882 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7883
7884 2013-08-29 Teresa Johnson <tejohnson@google.com>
7885
7886 * dumpfile.c (dump_loc): Output column number.
7887 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7888 * doc/invoke.texi: Document optall -fopt-info flag.
7889 * profile.c (read_profile_edge_counts): Use new dump framework.
7890 (compute_branch_probabilities): Ditto.
7891 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7892 when pass not in any opt group.
7893 * pass_manager.h (pass_manager::get_pass_profile): New method.
7894 * value-prof.c (check_counter): Use new dump framework.
7895 (check_ic_target): Ditto.
7896 * coverage.c (get_coverage_counts): Ditto.
7897 (coverage_init): Setup new dump framework.
7898
7899 2013-08-29 Richard Biener <rguenther@suse.de>
7900
7901 PR tree-optimization/58246
7902 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7903 handle the dominance check inside a basic-block.
7904
7905 2013-08-29 Richard Biener <rguenther@suse.de>
7906
7907 PR middle-end/57287
7908 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7909 of default defs that appear in abnormal PHI nodes.
7910
7911 2013-08-29 Richard Biener <rguenther@suse.de>
7912
7913 PR tree-optimization/57685
7914 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7915 single-use operands to avoid exponential complexity.
7916
7917 2013-08-28 Dehao Chen <dehao@google.com>
7918
7919 * ipa-inline.c (edge_badness): Fix integer underflow.
7920
7921 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7922
7923 * gtm-builtins.def (_ITM_free): Declare leaf.
7924
7925 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7926
7927 PR target/58067
7928 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7929 (*tls_local_dynamic_base_64_largepic): Likewise.
7930 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7931 Remove predicate from call operand.
7932 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7933 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7934
7935 2013-08-28 Jeff Law <law@redhat.com>
7936
7937 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7938 checks for the number of predecessors and successors allowed.
7939 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7940 which require copying a joiner block if there is a request which
7941 is a subpath that requires no joiner block copying.
7942
7943 2013-08-28 Jan Hubicka <jh@suse.cz>
7944
7945 * lto-streamer-out.c (DFS_write_tree_body): Drop
7946 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7947 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7948 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7949 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7950 Do not read DECL_ERROR_ISSUED.
7951 (unpack_ts_decl_with_vis_value_fields): Do not read
7952 DECL_DEFER_OUTPUT.
7953 (lto_input_ts_binfo_tree_pointers): Do not read
7954 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7955 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7956 write DECL_ERROR_ISSUED..
7957 (pack_ts_decl_with_vis_value_fields): Do not write
7958 DECL_DEFER_OUTPUT.
7959 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7960 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7961 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7962 * tree.h (tree_decl_common): Update comment.
7963 (DECL_ERROR_ISSUED): Remove.
7964
7965 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7966
7967 PR middle-end/58257
7968 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7969
7970 2013-08-28 Jan Hubicka <jh@suse.cz>
7971
7972 * builtins.def (free): Declare leaf.
7973
7974 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7975
7976 * gdbhooks.py: New.
7977 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7978 * configure: Regenerate.
7979
7980 2013-08-27 Martin Jambor <mjambor@suse.cz>
7981
7982 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7983 (ipa_ancestor_jf_data): Likewise.
7984 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7985 (ipa_get_jf_pass_through_type_preserved): New function.
7986 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7987 (ipa_get_jf_ancestor_type_preserved): New function.
7988 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7989 (ipa_get_jf_ancestor_result): Likewise.
7990 (propagate_vals_accross_pass_through): Use
7991 ipa_get_jf_pass_through_result to do all the value mappings.
7992 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7993 type_preserved flag.
7994 (ipa_set_jf_cst_copy): New function.
7995 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7996 (ipa_set_jf_arith_pass_through): Likewise.
7997 (ipa_set_ancestor_jf): Likewise.
7998 (compute_complex_assign_jump_func): Set type_preserved instead of
7999 punting.
8000 (ipa_compute_jump_functions_for_edge): Likewise.
8001 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8002 (update_jump_functions_after_inlining): Update type_preserved.
8003 Explicitely create jump functions when combining one with pass_through.
8004 (ipa_write_jump_function): Stream the type_preserved flags.
8005 (ipa_read_jump_function): Likewise.
8006
8007 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8008 Aldy Hernandez <aldyh@redhat.com>
8009
8010 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8011 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8012 * function.h (struct function): Add has_force_vect_loops and
8013 has_simduid_loops.
8014 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8015 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8016 handle it.
8017 * gimple.def: Update CLAUSES comments.
8018 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8019 (gimple_build_omp_for): Add argument to prototype.
8020 (gimple_omp_for_kind): New.
8021 (gimple_omp_for_set_kind): New.
8022 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8023 GOVD_DATA_SHARE_CLASS.
8024 (enum omp_region_type): Add ORT_SIMD.
8025 (gimple_add_tmp_var): Handle ORT_SIMD.
8026 (gimplify_var_or_parm_decl): Same.
8027 (is_gimple_stmt): Same.
8028 (omp_firstprivatize_variable): Same.
8029 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8030 (omp_notice_variable): Handle ORT_SIMD.
8031 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8032 (omp_check_private): Handle ORT_SIMD.
8033 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8034 OMP_CLAUSE_SAFELEN.
8035 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8036 Handle OMP_CLAUSE_LASTPRIVATE.
8037 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8038 OMP_CLAUSE_SAFELEN.
8039 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8040 (gimplify_expr): Handle OMP_SIMD.
8041 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8042 (expand_GOMP_SIMD_VF): New.
8043 (expand_GOMP_SIMD_LAST_LANE): New.
8044 * internal-fn.def (GOMP_SIMD_LANE): New.
8045 (GOMP_SIMD_VF): New.
8046 (GOMP_SIMD_LAST_LANE): New.
8047 * omp-low.c: Include target.h.
8048 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8049 OMP_CLAUSE_SAFELEN.
8050 (check_omp_nesting_restrictions): Same.
8051 (omp_max_vf): New.
8052 (lower_rec_simd_input_clauses): New.
8053 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8054 OMP_CLAUSE_LINEAR.
8055 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8056 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8057 (expand_omp_build_assign): New.
8058 (expand_omp_for_init_counts): New.
8059 (expand_omp_for_init_vars): New.
8060 (extract_omp_for_update_vars): New.
8061 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8062 and rewrite accordingly.
8063 (expand_omp_simd): New.
8064 (expand_omp_for): Use expand_omp_simd.
8065 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8066 (lower_omp_for): Do not lower the body.
8067 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8068 in their own loops.
8069 * tree-flow.h (find_omp_clause): Remove prototype.
8070 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8071 forcing vectorization of the loop, or if flag_tree_vectorize.
8072 (gate_tree_if_conversion): Similarly.
8073 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8074 gimple_build_omp_for.
8075 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8076 * tree-parloops (create_parallel_loop): Pass kind argument to
8077 gimple_build_omp_for.
8078 * tree-pretty-print.c (dump_omp_clause): Add cases for
8079 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8080 OMP_CLAUSE__SIMDUID_.
8081 (dump_generic_node): Handle OMP_SIMD.
8082 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8083 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8084 unroll OMP_SIMD loops here.
8085 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8086 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8087 loop->safelen.
8088 (vect_analyze_data_refs): Handle simd loops.
8089 * tree-vect-loop.c (vectorizable_live_operation): Handle
8090 IFN_GOMP_SIMD*.
8091 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8092 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8093 (vectorizable_load): Same.
8094 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8095 (struct simduid_to_vf): New.
8096 (simduid_to_vf::hash): New.
8097 (simduid_to-vf::equal): New.
8098 (struct simd_array_to_simduid): New.
8099 (simd_array_to_simduid::hash): New.
8100 (simd_array_to_simduid::equal): New.
8101 (adjust_simduid_builtins): New.
8102 (struct note_simd_array_uses_struct): New.
8103 (note_simd_array_uses_cb): New.
8104 (note_simd_array_uses): New.
8105 (vectorize_loops): Handle simd hints and adjust simd builtins
8106 accordingly.
8107 * tree-vectorizer.h (struct _stmt_vec_info): Add
8108 simd_lane_access_p field.
8109 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8110 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8111 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8112 (omp_clause_code_name): Same.
8113 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8114 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8115 * tree.def (OMP_SIMD): New entry.
8116 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8117 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8118 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8119 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8120 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8121 (OMP_CLAUSE_LINEAR_STEP): New.
8122 (OMP_CLAUSE_SAFELEN_EXPR): New.
8123 (OMP_CLAUSE__SIMDUID__DECL): New.
8124 (find_omp_clause): New prototype.
8125
8126 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8127
8128 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8129 Haswell processor detection.
8130
8131 2013-08-27 Christian Widmer <shadow@umbrox.de>
8132
8133 PR target/57927
8134 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8135 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8136 AVX2 capable processors.
8137
8138 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8139
8140 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8141 of vget_low_* with implementations in terms of other intrinsics.
8142
8143 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8144
8145 PR middle-end/57219
8146 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8147 values to -1, 0 and 1.
8148
8149 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8150
8151 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8152 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8153 (<optab><mode>3_insn): Remove.
8154 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8155 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8156 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8157 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8158 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8159 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8160 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8161 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8162 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8163 (ror<mode>3_insn): Likewise.
8164 * config/aarch64/predicates.md (aarch64_simd_register): New.
8165
8166 2013-08-27 Richard Biener <rguenther@suse.de>
8167
8168 PR tree-optimization/57521
8169 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8170 one edge is non-critical.
8171 (find_phi_replacement_condition): Make sure to use a non-critical
8172 edge. Cleanup and remove old bug workarounds.
8173 (bb_postdominates_preds): Remove.
8174 (if_convertible_loop_p_1): Do not compute post-dominators.
8175 (combine_blocks): Do not free post-dominators.
8176 (main_tree_if_conversion): Likewise.
8177 (pass_data_if_conversion): Add TODO_verify_ssa.
8178
8179 2013-08-27 DJ Delorie <dj@redhat.com>
8180
8181 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8182
8183 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8184
8185 * function.c (assign_parm_find_data_types): Set passed_mode and
8186 nominal_mode to the TYPE_MODE of nominal_type for the built
8187 pointer type in case of the struct-pass-by-reference.
8188
8189 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8190
8191 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8192 (UINT16_TYPE): Change default to "unsigned int".
8193
8194 * config/avr/avr.opt (mfract-convert-truncate): New option.
8195 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8196 is set, round negative fractional integers according to n1169
8197 when converting to integer types.
8198
8199 2013-08-26 Jan Hubicka <jh@suse.cz>
8200
8201 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8202 methods can not be called indirectly when their address is not taken.
8203
8204 2013-08-26 Jan Hubicka <jh@suse.cz>
8205
8206 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8207 ctor_for_folding.
8208
8209 2013-08-26 Jan Hubicka <jh@suse.cz>
8210
8211 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8212 can be unshared.
8213
8214 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8215
8216 * reload.c (find_valid_class): Allow classes that do not include
8217 FIRST_PSEUDO_REGISTER - 1.
8218
8219 2013-08-26 Jan Hubicka <jh@suse.cz>
8220
8221 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8222 fix edge count/frequency when speculation failed; fix type check
8223 for the direct call.
8224
8225 2013-08-26 Jan Hubicka <jh@suse.cz>
8226
8227 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8228
8229 2013-08-26 Jan Hubicka <jh@suse.cz>
8230
8231 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8232 to be changed by edge redirection.
8233
8234 2013-08-26 Jan Hubicka <jh@suse.cz>
8235
8236 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8237 formating; add sanity check.
8238 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8239 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8240 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8241 (resolve_noninline_speculation): Update callee keys, too.
8242
8243 2013-08-26 Jan Hubicka <jh@suse.cz>
8244
8245 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8246 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8247
8248 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8249
8250 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8251 into proper place.
8252
8253 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8254
8255 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8256 (x86_64_elf_select_section): Ditto.
8257 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8258 arguments.
8259 (ix86_pass_by_reference): Ditto.
8260 (output_set_got): Ditto.
8261 (ix86_unary_operator_ok): Ditto.
8262 (ix86_expand_builtin): Ditto.
8263
8264 2013-08-23 Jan Hubicka <jh@suse.cz>
8265
8266 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8267
8268 2013-08-23 Jan Hubicka <jh@suse.cz>
8269
8270 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8271 (tree_decl_with_vis): Add FINAL field.
8272
8273 2013-08-23 Jeff Law <law@redhat.com>
8274
8275 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8276 the debugging dump when the expression is fully redundant.
8277
8278 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8279
8280 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8281 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8282 * pretty-print.c (pp_formatted_text_data): Likewise.
8283 (pp_write_text_to_stream): Likewise.
8284 (pp_write_text_as_dot_label_to_stream): Likewise.
8285 (pp_append_r): Likewise.
8286 (pp_format): Likewise.
8287 (pp_flush): Likewise.
8288 (pp_clear_output_area): Likewise.
8289 (pp_append_text): Likewise.
8290 (pp_formatted_text): Likewise.
8291 (pp_remaining_character_count_for_line): Likewise.
8292 (pp_newline): Likewise.
8293 (pp_character): Likewise.
8294 (output_buffer::~output_buffer): Define.
8295 (pretty_printer::~pretty_printer): Destruct output buffer.
8296 * pretty-print.h (output_buffer::~output_buffer): Declare.
8297 (pretty_printer::~pretty_printer): Declare virtual.
8298
8299 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8300
8301 PR other/57324
8302 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8303 HOST_WIDE_INT_M1U): New macros.
8304 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8305 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8306 unsigned -1 for lshift.
8307 * cse.c (cse_insn): Likewise.
8308 * double-int.c (rshift_double, lshift_double): Likewise.
8309 * builtins.c (fold_builtin_bitop): Likewise.
8310 * combine.c (force_to_mode): Likewise.
8311 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8312 * simplify-rtx.c (simplify_const_unary_operation,
8313 simplify_const_binary_operation): Likewise.
8314 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8315 check_va_list_escapes): Likewise.
8316 * rtlanal.c (nonzero_bits1): Likewise.
8317 * expmed.c (expand_smod_pow2): Likewise.
8318 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8319
8320 2013-08-23 Jan Hubicka <jh@suse.cz>
8321
8322 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8323 as having address taken.
8324
8325 2013-08-23 Jan Hubicka <jh@suse.cz>
8326
8327 * ipa-utils.h (method_class_type): Declare.
8328 * ipa-devirt.c (method_class_type): Export.
8329
8330 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8331 do not walk base classes of anonymous types.
8332
8333 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8334
8335 PR rtl-optimization/58220
8336 PR regression/58221
8337 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8338 handle SEQUENCE insns properly.
8339
8340 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8341
8342 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8343 definition.
8344 (pp_newline_and_indent): Likewise.
8345 (pp_separate_with): Likewise.
8346 * pretty-print.c (pp_newline_and_flush): Define.
8347 (pp_newline_and_indent): Likewise.
8348 (pp_separate_with): Likewise.
8349
8350 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8351
8352 PR target/58218
8353 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8354 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8355
8356 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8357
8358 * config/i386/predicates.md (ext_sse_reg_operand): New.
8359 * config/i386/i386.md (*movti_internal): Use
8360 predicate to determine if EVEX is needed.
8361 (*movsi_internal): Ditto.
8362 (*movdf_internal): Ditto.
8363 (*movsf_internal): Ditto.
8364 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8365
8366 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8367
8368 PR tree-optimization/58209
8369 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8370 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8371 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8372 emit POINTER_PLUS_EXPR.
8373 (create_tailcall_accumulator): For pointer result type accumulate in
8374 sizetype type.
8375
8376 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8377
8378 * configure.ac: Add backslashes missing from the last change.
8379 * configure: Regenerate.
8380
8381 2013-08-22 Jan Hubicka <jh@susue.cz>
8382
8383 * ipa.c (function_and_variable_visibility): First remember function
8384 was global and then make it local.
8385
8386 2013-08-22 Julian Brown <julian@codesourcery.com>
8387
8388 * configure.ac: Add aarch64 to list of arches which use "nop" in
8389 debug_line test.
8390 * configure: Regenerate.
8391
8392 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8393
8394 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8395 gnu_libc_has_function.
8396 * config/s390/tpf.h: Likewise.
8397
8398 2013-08-22 Jan Hubicka <jh@susue.cz>
8399
8400 * timevar.c (validate_phases): Add cast.
8401
8402 2013-08-22 Jan Hubicka <jh@susue.cz>
8403
8404 * timevar.c (validate_phases): Use size_t for memory.
8405 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8406
8407 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8408
8409 * pretty-print.h (output_buffer::output_buffer): Declare.
8410 (pretty_printer::pretty_printer): Likewise.
8411 (pp_construct): Remove.
8412 * pretty-print.c (output_buffer::output_buffer): Define.
8413 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8414 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8415 (print_gimple_expr): Likewise.
8416 (print_gimple_seq): Likewise.
8417 (gimple_dump_bb): Likewise.
8418 * sched-vis.c (dump_value_slim): Likewise.
8419 (dump_insn_slim): Likewise.
8420 (dump_rtl_slim): Likewise.
8421 (str_pattern_slim): Likewise.
8422 * tree-mudflap.c (mf_varname_tree): Likewise.
8423 * graph.c (print_graph_cfg): Likewise.
8424 (start_graph_dump): Likewise.
8425 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8426 placement-new.
8427 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8428 pretty printer initialization.
8429 * coretypes.h (diagnostic_context): Remove superflous type alias
8430 declaration.
8431 (pretty_printer): Likewise. Declare directly as a class.
8432 (pretty_print_info): Remove declaration as class.
8433 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8434 and pp_clear_output_area.
8435 (asan_add_global): Likewise.
8436
8437 2013-08-22 Jan Hubicka <jh@suse.cz>
8438
8439 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8440 * ipa-utils.h (update_type_inheritance_graph): Declare.
8441 (possible_polymorphic_call_target_p): Declare.
8442 (possible_polymorphic_call_target_p): New.
8443 * ipa-devirt.c: Update toplevel comments.
8444 (cached_polymorphic_call_targets): Move up.
8445 (odr_type_d): Move ID down.
8446 (polymorphic_type_binfo_p): Update comment.
8447 (odr_hasher::remove): Likewise;
8448 (get_odr_type): Set anonymous_namespace.
8449 (dump_odr_type): Dump it.
8450 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8451 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8452 (record_binfo): Add comment.
8453 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8454 built.
8455 (devirt_node_removal_hook): Do not iCE when cache is freed.
8456 (possible_polymorphic_call_target_p): New predicate.
8457 (update_type_inheritance_graph): New function.
8458
8459 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8460 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8461 Sergey Lega <sergey.s.lega@intel.com>
8462 Anna Tikhonova <anna.tikhonova@intel.com>
8463 Ilya Tocar <ilya.tocar@intel.com>
8464 Andrey Turetskiy <andrey.turetskiy@intel.com>
8465 Ilya Verbin <ilya.verbin@intel.com>
8466 Kirill Yukhin <kirill.yukhin@intel.com>
8467 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8468
8469 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8470 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8471 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8472 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8473 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8474 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8475 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8476 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8477 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8478 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8479 OPT_mavx512pf, OPT_mavx512er cases.
8480 * config/i386/constraints.md (v): New constraint.
8481 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8482 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8483 (bit_AVX512CD): New.
8484 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8485 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8486 * config/i386/i386-c.c (ix86_target_macros_internal):
8487 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8488 __AVX512PF__.
8489 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8490 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8491 * config/i386/i386.c (regclass_map, dbx_register_map)
8492 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8493 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8494 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8495 -mavx512pf options.
8496 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8497 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8498 -mavx512pf options. Fix formatting.
8499 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8500 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8501 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8502 -mavx512cd, -mavx512pf options.
8503 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8504 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8505 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8506 ALL_SSE_REGS.
8507 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8508 (ix86_set_reg_reg_cost): Ditto.
8509 (x86_order_regs_for_local_alloc): Ditto.
8510 (MAX_VECT_LEN): Extend to 64-byte.
8511 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8512 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8513 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8514 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8515 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8516 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8517 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8518 (SSE_REG_MODE_P): Support new modes.
8519 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8520 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8521 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8522 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8523 (REG_CLASS_CONTENTS): Add new registers.
8524 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8525 (EXT_REX_SSE_REGNO_P): New.
8526 (HI_REGISTER_NAMES): Add new registers.
8527 * config/i386/i386.md: Define constants for new registers.
8528 (mode): Add new 512-bit modes.
8529 (prefix): Support evex prefix.
8530 (isa): Support avx512f, noavx512f, fma_avx512f.
8531 (ssemodesuffix): Add new 512-bit modes.
8532 (movxi): New.
8533 (*movxi_internal_avx512f): Ditto.
8534 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8535 Support MODE_XI.
8536 (*movsi_internal): Likewise.
8537 (*movdf_internal): Likewise.
8538 (*movsf_internal): Likewise.
8539 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8540 (<code><mode>3): Likewise.
8541 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8542 New.
8543 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8544 with the new constraint "v".
8545 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8546 modes.
8547 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8548 with the new constraint "v".
8549 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8550 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8551 (*<plusminus_insn><mode>3): Likewise.
8552 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8553 (*mul<mode>3): Likewise.
8554 (<sse>_vmmul<mode>3): Likewise.
8555 (<sse>_div<mode>3): Likewise.
8556 (<sse>_vmdiv<mode>3): Likewise.
8557 (<sse>_sqrt<mode>2): Likewise.
8558 (<sse>_vmsqrt<mode>2): Likewise.
8559 (*<code><mode>3_finite): Likewise.
8560 (*<code><mode>3) <smaxmin>: Likewise.
8561 (<sse>_vm<code><mode>3): Likewise.
8562 (*<code><mode>3) <any_logic>: Likewise.
8563 (*fma_fmadd_<mode>): Likewise.
8564 (*fma_fmsub_<mode>): Likewise.
8565 (*fma_fnmadd_<mode>): Likewise.
8566 (*fma_fnmsub_<mode>): Likewise.
8567 (*fma_fmaddsub_<mode>): Likewise.
8568 (*fma_fmsubadd_<mode>): Likewise.
8569 (*fmai_fmadd_<mode>): Likewise.
8570 (*fmai_fmsub_<mode>): Likewise.
8571 (*fmai_fnmadd_<mode>): Likewise.
8572 (*fmai_fnmsub_<mode>): Likewise.
8573 (sse_cvtsi2ss): Likewise.
8574 (sse_cvtsi2ssq): Likewise.
8575 (sse_cvtss2si): Likewise.
8576 (sse_cvtss2si_2): Likewise.
8577 (sse_cvtss2siq): Likewise.
8578 (sse_cvtss2siq_2): Likewise.
8579 (sse_cvttss2si): Likewise.
8580 (sse_cvtss2siq_2): Likewise.
8581 (float<sseintvecmodelower><mode>2): Likewise.
8582 (sse2_cvtsd2si_2): Likewise.
8583 (sse2_cvtsd2siq_2): Likewise.
8584 (*<plusminus_insn><mode>3): Likewise.
8585 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8586 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8587 (ashr<mode>3): Likewise.
8588 (<shift_insn><mode>3): Likewise.
8589 (avx2_<code><mode>3): Likewise.
8590 (*avx2_<code><mode>3): Likewise.
8591 (*andnot<mode>3): Likewise.
8592 (*<code><mode>3) <any_logic>: Likewise.
8593 (abs<mode>2): Likewise.
8594 (avx2_permvar<mode>): Likewise.
8595 (avx2_perm<mode>_1): Likewise.
8596 (*avx_vpermilp<mode>): Likewise.
8597 (avx_vpermilvar<mode>3): Likewise.
8598 (avx2_ashrv<mode>): Likewise.
8599 (avx2_<shift_insn>v<mode>): Likewise.
8600 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8601 -mavx512cd.
8602 * doc/rtl.texi: Document XImode.
8603
8604 2013-08-21 Jeff Law <law@redhat.com>
8605
8606 * tree-flow.h (register_jump_thread): Pass vector of edges
8607 instead of each important edge.
8608 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8609 thread path into a vector and pass that to register_jump_thread.
8610 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8611 passed in edge vector to the current 3-edge form.
8612
8613 Revert:
8614 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8615
8616 * dce.c (fini_dce): Call df_analyze again just in case
8617 delete_unmarked_insns removed anything.
8618
8619 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8620
8621 * reload.h (struct reg_equivs): Rename to ..
8622 (struct reg_equivs_s): .. this.
8623
8624 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8625
8626 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8627
8628 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8629
8630 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8631
8632 2013-08-21 Jeff Law <law@redhat.com>
8633
8634 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8635 simplify assignments too. If the RHS collapses to a singleton
8636 range, then return the value for the range.
8637
8638 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8639
8640 * config/i386/sse.md (V16): Rename to...
8641 (VMOVE): this.
8642 (mov<mode>): Update iterator name.
8643 (*mov<mode>_internal): Ditto.
8644 (push<mode>1): Ditto.
8645 (movmisalign<mode>): Ditto.
8646
8647 2013-08-20 Jan Hubicka <jh@suse.cz>
8648
8649 PR bootstrap/58186
8650 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8651 entry for direct edges.
8652 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8653
8654 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8655
8656 Revert my last two changes, r201865 and r201864:
8657
8658 Revert r201865:
8659 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8660
8661 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8662 instances can own GC refs.
8663
8664 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8665 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8666 (gcc::context::gt_pch_nx): Likewise.
8667 (gcc::context::gt_pch_nx): Likewise.
8668 * ggc.h (gt_ggc_mx <T>): New.
8669 (gt_pch_nx_with_op <T>): New.
8670 (gt_pch_nx <T>): New.
8671 * passes.c (opt_pass::gt_ggc_mx): New.
8672 (opt_pass::gt_pch_nx): New.
8673 (opt_pass::gt_pch_nx_with_op): New.
8674 (pass_manager::gt_ggc_mx): New.
8675 (pass_manager::gt_pch_nx): New.
8676 (pass_manager::gt_pch_nx_with_op): New.
8677 (pass_manager::operator new): Use
8678 ggc_internal_cleared_alloc_stat rather than xcalloc.
8679 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8680 (pass_manager::gt_ggc_mx): New.
8681 (pass_manager::gt_pch_nx): New.
8682 (pass_manager::gt_pch_nx_with_op): New.
8683 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8684 (opt_pass::operator new): New.
8685 (opt_pass::gt_ggc_mx): New.
8686 (opt_pass::gt_pch_nx): New.
8687 (opt_pass::gt_pch_nx_with_op): New.
8688
8689 Revert r201864:
8690 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8691
8692 * Makefile.in (GTFILES): Add context.h.
8693 * context.c (gcc::context::operator new): New.
8694 (gcc::context::gt_ggc_mx): New.
8695 (gcc::context::gt_pch_nx): New.
8696 (gcc::context::gt_pch_nx): New.
8697 * context.h (gcc::context): Add GTY((user)) marking.
8698 (gcc::context::operator new): New.
8699 (gcc::context::gt_ggc_mx): New.
8700 (gcc::context::gt_pch_nx): New.
8701 (gcc::context::gt_pch_nx): New.
8702 (g): Add GTY marking.
8703 (gt_ggc_mx (gcc::context *)): New.
8704 (gt_pch_nx (gcc::context *)): New.
8705 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8706 void *cookie)): New.
8707 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8708
8709 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8710
8711 * dce.c (fini_dce): Call df_analyze again just in case
8712 delete_unmarked_insns removed anything.
8713
8714 2013-08-20 Teresa Johnson <tejohnson@google.com>
8715
8716 PR rtl-optimizations/57451
8717 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8718 from crossing split section boundaries.
8719
8720 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8721
8722 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8723 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8724 with MULTLIB_DEFAULTS.
8725
8726 2013-08-20 Nick Clifton <nickc@redhat.com>
8727
8728 * target.def (narrow_volatile_bitfield): Note that the default
8729 value is false, not !TARGET_STRICT_ALIGN.
8730 * doc/tm.texi: Regenerate.
8731
8732 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8733
8734 Fix LIB_SPEC for systems without libpthread.
8735
8736 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8737 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8738 for Android.
8739 * config/i386/linux-common.h: Likewise.
8740 * config/mips/linux-common.h: Likewise.
8741
8742 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8743
8744 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8745 checks.
8746
8747 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8748
8749 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8750 instances can own GC refs.
8751
8752 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8753 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8754 (gcc::context::gt_pch_nx): Likewise.
8755 (gcc::context::gt_pch_nx): Likewise.
8756 * ggc.h (gt_ggc_mx <T>): New.
8757 (gt_pch_nx_with_op <T>): New.
8758 (gt_pch_nx <T>): New.
8759 * passes.c (opt_pass::gt_ggc_mx): New.
8760 (opt_pass::gt_pch_nx): New.
8761 (opt_pass::gt_pch_nx_with_op): New.
8762 (pass_manager::gt_ggc_mx): New.
8763 (pass_manager::gt_pch_nx): New.
8764 (pass_manager::gt_pch_nx_with_op): New.
8765 (pass_manager::operator new): Use
8766 ggc_internal_cleared_alloc_stat rather than xcalloc.
8767 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8768 (pass_manager::gt_ggc_mx): New.
8769 (pass_manager::gt_pch_nx): New.
8770 (pass_manager::gt_pch_nx_with_op): New.
8771 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8772 (opt_pass::operator new): New.
8773 (opt_pass::gt_ggc_mx): New.
8774 (opt_pass::gt_pch_nx): New.
8775 (opt_pass::gt_pch_nx_with_op): New.
8776
8777 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8778
8779 * Makefile.in (GTFILES): Add context.h.
8780 * context.c (gcc::context::operator new): New.
8781 (gcc::context::gt_ggc_mx): New.
8782 (gcc::context::gt_pch_nx): New.
8783 (gcc::context::gt_pch_nx): New.
8784 * context.h (gcc::context): Add GTY((user)) marking.
8785 (gcc::context::operator new): New.
8786 (gcc::context::gt_ggc_mx): New.
8787 (gcc::context::gt_pch_nx): New.
8788 (gcc::context::gt_pch_nx): New.
8789 (g): Add GTY marking.
8790 (gt_ggc_mx (gcc::context *)): New.
8791 (gt_pch_nx (gcc::context *)): New.
8792 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8793 void *cookie)): New.
8794 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8795
8796 2013-08-20 Alan Modra <amodra@gmail.com>
8797
8798 PR target/57865
8799 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8800 (rs6000_emit_epilogue): Likewise.
8801
8802 2013-08-19 Dehao Chen <dehao@google.com>
8803
8804 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8805
8806 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8807 Jakub Jelinek <jakub@redhat.com>
8808
8809 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8810 (BUILT_IN_FABSD64): Likewise.
8811 (BUILT_IN_FABSD128): Likewise.
8812 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8813 (fold_builtin_1): Likewise.
8814 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8815 destination and source operands.
8816 (*abstd2_fpr): Likewise.
8817 (*nabstd2_fpr): Likewise.
8818
8819 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8820
8821 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8822 JUMP_P and INSN_P.
8823
8824 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8825
8826 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8827 incomplete.
8828
8829 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8830
8831 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8832 * builtins.c (default_libc_has_function): New.
8833 (gnu_libc_has_function): Ditto.
8834 (no_c99_libc_has_function): Ditto.
8835 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8836 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8837 (fold_builtin_sincos): Likewise.
8838 (fold_builtin_cexp): Likewise.
8839 * builtins.def (DEF_C94_BUILTIN): Likewise.
8840 (DEF_C99_BUILTIN): Likewise.
8841 (DEF_C99_C90RES_BUILTIN): Likewise.
8842 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8843 definitions to using this define.
8844 * config/darwin-protos.h (darwin_libc_has_function): New.
8845 * config/darwin.c (darwin_libc_has_function): Ditto.
8846 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8847 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8848 * config/darwin.h: Ditto.
8849 * config/elfos.h: Ditto.
8850 * config/freebsd.h: Ditto.
8851 * config/i386/cygming.h: Ditto.
8852 * config/i386/djgpp.h: Ditto.
8853 * config/i386/i386-interix.h: Ditto.
8854 * config/microblaze/microblaze.h: Ditto.
8855 * config/mmix/mmix.h: Ditto.
8856 * config/gnu-user.h: Ditto.
8857 * config/ia64/hpux.h: Ditto.
8858 * config/pa/pa-hpux.h: Ditto.
8859 * config/pdp11/pdp11.h: Ditto.
8860 * config/picochip/picochip.h: Ditto.
8861 * config/linux.h: Ditto.
8862 * config/netbsd.h: Ditto.
8863 * config/openbsd.h: Ditto.
8864 * config/rs6000/aix43.h: Ditto.
8865 * config/rs6000/aix51.h: Ditto.
8866 * config/rs6000/aix52.h: Ditto.
8867 * config/rs6000/aix53.h: Ditto.
8868 * config/rs6000/aix61.h: Ditto.
8869 * config/rs6000/darwin.h: Ditto.
8870 * config/rs6000/linux.h: Ditto.
8871 * config/rs6000/linux64.h: Ditto.
8872 * config/s390/tpf.h: Ditto.
8873 * config/sol2-10.h: Ditto.
8874 * config/sol2.h: Ditto.
8875 * config/vms/vms.h: Ditto.
8876 * config/vxworks.h: Ditto.
8877 * config/linux-android.c (linux_android_libc_has_function):
8878 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8879 * config/linux-protos.h (linux_android_libc_has_function):
8880 New declaration.
8881 * config/i386/i386.c (ix86_libc_has_function): New.
8882 * config/i386/i386-protos.h
8883 (ix86_libc_has_function): New declaration.
8884 * config/i386/i386.md
8885 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8886 ("isinf<mode>2): Likewise.
8887 * convert.c (convert_to_integer): Using new target hook
8888 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8889 TARGET_C99_FUNCTIONS.
8890 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8891 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8892 * coretypes.h (function_class): New enum for different
8893 classes of functions.
8894 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8895 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8896 (TARGET_HAS_SINCOS): Likewise.
8897 (TARGET_LIBC_HAS_FUNCTION): New.
8898 * doc/tm.texi: Regenerated.
8899 * targhooks.h (default_libc_has_function): New declaration.
8900 (no_c99_libc_has_function): Ditto.
8901 (gnu_libc_has_function): Ditto.
8902 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8903 and TARGET_HAS_SINCOS.
8904
8905 2013-08-18 Jan Hubicka <jh@suse.cz>
8906
8907 * Makeifle-in (ipa-devirt.o): New.
8908 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8909 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8910 (analyze_functions): Look into possible targets of polymorphic call.
8911 * dumpfile.c (dump_files): Add type-inheritance dump.
8912 * dumpfile.h (TDI_inheritance): New.
8913 * ipa-devirt.c: New file.
8914 * ipa-utils.h (odr_type_d): Forward declare.
8915 (odr_type): New type.
8916 (build_type_inheritance_graph): Declare.
8917 (possible_polymorphic_call_targets): Declare and introduce inline
8918 variant when only edge is pased.
8919 (dump_possible_polymorphic_call_targets): Likewise.
8920 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8921 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8922 (types_same_for_odr): ... here.
8923 * tree.h (type_in_anonymous_namespace_p): Declare.
8924
8925 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8926
8927 PR tree-optimization/58006
8928 * tree-parloops.c (take_address_of): Don't ICE if get_name
8929 returns NULL.
8930 (eliminate_local_variables_stmt): Remove clobber stmts.
8931
8932 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8933
8934 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8935 error message is issued for an alias to undefined symbol.
8936
8937 2013-08-18 Jan Hubicka <jh@suse.cz>
8938
8939 * cgraph.c (cgraph_create_indirect_edge): Discover
8940 polymorphic calls and record basic info into indirect_info.
8941 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8942 devirtualization, ignore objc function calls.
8943 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8944 call with no parm index info.
8945 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8946 * tree.c (virtual_method_call_p): New function.
8947 * tree.h (virtual_method_call_p): Declare.
8948
8949 2013-08-16 Jan Hubicka <jh@suse.cz>
8950
8951 PR middle-end/58179
8952 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8953
8954 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8955
8956 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8957 attribute.
8958
8959 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8960
8961 * gengtype.c (type_for_name): Add special-case support for
8962 locating types within the "gcc::" namespace.
8963 (open_base_files): Emit a "using namespace gcc" directive.
8964
8965 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8966
8967 PR target/58160
8968 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8969 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8970
8971 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8972 array instead of each individual operand as a separate argument.
8973 (emit_fusion_gpr_load): Likewise.
8974 (expand_fusion_gpr_load): Add new function declaration.
8975
8976 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8977 signature to have the operands passed as an array, instead of as
8978 separate arguments. Allow ZERO_EXTEND to be in the memory
8979 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8980 depend on the register live/dead flags when peepholes are run.
8981 (expand_fusion_gpr_load): New function to be called from the
8982 peephole2 pass, to change the register that addis sets to be the
8983 target register.
8984 (emit_fusion_gpr_load): Change the calling signature to have the
8985 operands passed as an array, instead of as separate arguments.
8986 Allow ZERO_EXTEND to be in the memory address, and also
8987 SIGN_EXTEND if -mpower8-fusion-sign.
8988
8989 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8990 unspec enumeration.
8991 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8992 adjust the register addis loads up in the peephole2 pass. Do not
8993 depend on the register live/dead state when the peephole pass is done.
8994
8995 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8996
8997 * gengtype.c (create_user_defined_type): Ensure that the kind
8998 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8999 declaration is seen before the GTY((user)) marking.
9000
9001 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9002
9003 PR target/58105
9004 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9005
9006 2013-08-16 Jan Hubicka <jh@suse.cz>
9007
9008 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9009 arugment expected_type.
9010 (gimple_fold_call): Use it.
9011 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9012 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9013 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9014 (try_make_edge_direct_virtual_call): Likewise.
9015 * tree.c (obj_type_ref_class): New.
9016 * tree.h (obj_type_ref_class): Use it.
9017
9018 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9019
9020 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9021 (rtl_slim_pp): Likewise.
9022 (init_rtl_slim_pretty_print): Likewise.
9023 (dump_value_slim): Don't call it. Use local pretty printer.
9024 (dump_insn_slim): Likewise.
9025 (dump_rtl_slim): Likewise.
9026 (str_pattern_slim): Likewise.
9027 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9028 Simplify.
9029
9030 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9031
9032 PR tree-optimization/58164
9033 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9034 walk gimple_goto_dest of GIMPLE_GOTO.
9035
9036 PR tree-optimization/58165
9037 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9038 bi_call must be the last stmt in a bb, don't split_block, instead
9039 use fallthru edge from it and give up if there is none.
9040 Release conds vector when returning early.
9041
9042 2013-08-14 Xinliang David Li <davidxl@google.com>
9043
9044 * config/i386/i386.c (ix86_option_override_internal):
9045 Remove unused variable and field.
9046
9047 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9048
9049 PR target/57949
9050 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9051 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9052 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9053 and Linux, correct BLKmode alignment when 128-bit alignment is
9054 required and compatibility flag is not set.
9055 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9056 for zero-size arguments when compatibility flag is not set.
9057
9058 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9059
9060 PR tree-optimization/58145
9061 * tree-sra.c (build_ref_for_offset): If prev_base has
9062 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9063
9064 2013-08-14 Xinliang David Li <davidxl@google.com>
9065
9066 * config/i386/i386.c (ix86_option_override_internal):
9067 Fix uninitialized variable error.
9068
9069 2013-08-14 Xinliang David Li <davidxl@google.com>
9070
9071 * config/i386/i386.opt: Define two new options.
9072 * config/i386/x86-tune.def: Add arch selector field in macros.
9073 * config/i386/i386.h: Adjust macro definition.
9074 * config/i386/i386.c (ix86_option_override_internal):
9075 Refactor the code.
9076 (parse_mtune_ctrl_str): New function.
9077 (set_ix86_tune_features): New function.
9078 (ix86_function_specific_restore): Call the new helper function.
9079
9080 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9081
9082 PR rtl-optimization/57662
9083 * sel-sched.c (code_motion_process_successors): When the current insn
9084 is removed after the recursive traversal, break from the loop.
9085 Add comments and debug printouts.
9086
9087 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9088 Alexandre Oliva <aoliva@redhat.com>
9089
9090 PR target/58067
9091 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9092 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9093 there also UNSPEC_PLTOFF.
9094
9095 2013-08-14 Marek Polacek <polacek@redhat.com>
9096
9097 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9098 NUM_CONDITIONS bit positions.
9099
9100 2013-08-13 Cary Coutant <ccoutant@google.com>
9101
9102 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9103 (attr_checksum): Hash vector contents instead of pointer.
9104 (attr_checksum_ordered): Likewise.
9105
9106 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9107
9108 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9109 when Pmode != word_mode. Add length_address attribute.
9110 (sse3_monitor_<mode>): Merge from sse3_monitor and
9111 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9112 Pmode != word_mode. Update insn length attribute.
9113 * config/i386/i386.c (ix86_option_override_internal): Update
9114 ix86_gen_monitor selection for merged sse3_monitor insn.
9115
9116 2013-08-13 Julian Brown <julian@codesourcery.com>
9117
9118 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9119 perform invalid legitimization on greater-than-word-size modes for
9120 TARGET_E500_DOUBLE.
9121
9122 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9123
9124 * ira.c (setup_class_translate_array): Use aclass instead of cl
9125 for classes not fully covered by allocno classes.
9126
9127 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9128
9129 PR tree-optimization/57661
9130 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9131 * tree-inline.c (tree_function_versioning): Initialize it.
9132 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9133 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9134 that is not being copied.
9135
9136 PR sanitizer/56417
9137 * asan.c (instrument_strlen_call): Fix typo in comment.
9138 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9139
9140 2013-08-13 Steve Ellcey <sellcey@mips.com>
9141
9142 * config/mips/mips.md (prefetch): Use lw instead of ld on
9143 loongson in 32bit mode.
9144
9145 2013-08-13 Nick Clifton <nickc@redhat.com>
9146
9147 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9148
9149 2013-08-13 Jan Hubicka <jh@suse.cz>
9150
9151 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9152 introduced edge; fix typo in sanity check.
9153 (cgraph_resolve_speculation): Export; improve diagnostic.
9154 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9155 speculation at type mismatch.
9156 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9157 (cgraph_resolve_speculation): Declare.
9158 (symtab_can_be_discarded): New function.
9159 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9160 * ipa-inline-transform.c (speculation_removed): New global var.
9161 (clone_inlined_nodes): See if speculation can be removed.
9162 (inline_call): If speculations was removed, we growths may not match.
9163 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9164 (speculation_useful_p): New function.
9165 (resolve_noninline_speculation): New function.
9166 (inline_small_functions): Resolve useless speculations.
9167 * ipa-inline.h (speculation_useful_p): Declare
9168 * ipa.c (can_replace_by_local_alias): Simplify.
9169 (ipa_profile): Produce speculative calls in non-lto, too;
9170 add simple cost model; produce local aliases.
9171
9172 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9173
9174 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9175 PASS_MANAGER_H.
9176
9177 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9178
9179 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9180
9181 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9182
9183 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9184 iterator instead of X87MODEF.
9185
9186 2013-08-12 Perez Read <netfirewall@gmail.com>
9187
9188 PR target/58132
9189 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9190 operand 0 for intel asm alternative.
9191 (*movabs<mode>_2): Ditto for operand 1.
9192
9193 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9194
9195 * config/aarch64/arm_none.h
9196 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9197
9198 2013-08-12 Nick Clifton <nickc@redhat.com>
9199
9200 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9201
9202 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9203
9204 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9205 expand for QI/HImode operand to produce more effictive code for
9206 unsigned char(short) --> float(double) conversion.
9207
9208 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9209
9210 * doc/invoke.texi: Mention that -ftls-model does not force the final
9211 model.
9212
9213 2013-08-12 Marek Polacek <polacek@redhat.com>
9214 Marc Glisse <marc.glisse@inria.fr>
9215
9216 PR tree-optimization/57980
9217 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9218 when creating -1 constant.
9219
9220 2013-08-10 Jan Hubicka <jh@suse.cz>
9221
9222 Workaround binutils PR14342.
9223 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9224 (gimple_init_edge_profiler): Likewise.
9225 (gimple_gen_ic_func_profiler): Likewise.
9226
9227 2013-08-09 Jan Hubicka <jh@suse.cz>
9228
9229 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9230
9231 2013-08-09 Xinliang David Li <davidxl@google.com>
9232
9233 * config/i386/stringop.def: New file.
9234 * config/i386/stringop.opt: New file.
9235 * config/i386/i386-opts.h: Include stringopt.def.
9236 * config/i386/i386.opt: Include stringopt.opt.
9237 * config/i386/i386.c (ix86_option_override_internal):
9238 Override default size based stringop inline strategies with options.
9239 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9240 New function.
9241
9242 2013-08-09 Jan Hubicka <jh@suse.cz>
9243
9244 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9245
9246 2013-08-09 Jan Hubicka <jh@suse.cz>
9247
9248 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9249 CGRAPH_FREQ_MAX.
9250 (dump_cgraph_node): Dump profile-id.
9251 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9252 and common_target_probability.
9253 * lto-cgraph.c (lto_output_edge): Stream common targets.
9254 (lto_output_node): Stream profile ids.
9255 (input_node): Stream profile ids.
9256 (input_edge): Stream common targets.
9257 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9258 * ipa.c: Include value-prof.h
9259 (ipa_profile_generate_summary): Turn indirect call statement histograms
9260 into common targets.
9261 (ipa_profile): Turn common targets into speculative edges.
9262
9263 2013-08-09 Jan Hubicka <jh@suse.cz>
9264
9265 * cgraph.h (cgraph_node): Add profile_id.
9266 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9267 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9268 (del_node_map): Update.
9269 (find_func_by_funcdef_no): Replace by ...
9270 (find_func_by_profile_id): ... this one.
9271 (gimple_ic_transform): Do not remove useful histograms when
9272 speculation is not done; dump info when indirect call removal
9273 can happen at LTO.
9274 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9275 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9276 (__gcov_indirect_call_profiler_v2): .. this one.
9277 * profile.h (init_node_map): Update.
9278 * coverage.c (coverage_compute_profile_id): New function.
9279 * coverage.h (coverage_compute_profile_id): Declare.
9280 * tree-profile.c (init_ic_make_global_vars): Make
9281 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9282 (gimple_init_edge_profiler): Update prototype of
9283 __gcov_indirect_call_profiler.
9284 (gimple_gen_ic_func_profiler): Simplify.
9285 (tree_profiling): Use init_node_map
9286
9287 2013-08-09 Jan Hubicka <jh@suse.cz>
9288
9289 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9290 non-speculative refs.
9291 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9292 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9293 (cgraph_set_call_stmt): Likewise.
9294 (cgraph_create_edge_1): Fix release checking compilatoin;
9295 clear lto_stmt_uid.
9296 (cgraph_free_edge): Free indirect info.
9297 (cgraph_turn_edge_to_speculative): New function.
9298 (cgraph_speculative_call_info): New function.
9299 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9300 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9301 (dump_cgraph_node): Dump speculation.
9302 (verify_edge_count_and_frequency): Accept speculative edges.
9303 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9304 (verify_cgraph_node): Handle speculation.
9305 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9306 (cgraph_set_call_stmt): Update prototype.
9307 (cgraph_make_edge_direct): Update prototype.
9308 (cgraph_speculative_call_info): Declare.
9309 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9310 to change; update call of ipa_find_references.
9311 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9312 lto_stmt_uid and speculative flags.
9313 (ipa_dump_references): Dump speculation.
9314 (ipa_clone_references): Clone speculative flag.
9315 (ipa_clone_referring): Likewise.
9316 (ipa_clone_ref): New function.
9317 (ipa_find_reference): Look into lto_stmt_uids
9318 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9319 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9320 (ipa_find_reference): Update declaration.
9321 (ipa_clone_ref): Declare.
9322 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9323 stream speculative flag.
9324 (lto_output_ref): Stream statements uids and speculation.
9325 (input_ref): Likewise.
9326 (input_edge): Stream speuclation.
9327 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9328 (cgraph_set_call_stmt_including_clones): Handle speculation.
9329 * ipa-inline.c (heap_edge_removal_hook): New function.
9330 (inline_small_functions): Register it.
9331 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9332 also initialize refs.
9333 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9334 edge to change.
9335 (try_make_edge_direct_simple_call): Likewise.
9336 (try_make_edge_direct_simple_call): Likewise.
9337 (update_indirect_edges_after_inlining): Likewise.
9338 (remove_described_reference): Look proper lto_stmt_uid.
9339 (propagate_controlled_uses): Likewise.
9340 (propagate_controlled_uses): Liekwise.
9341 * tree-inline.c (copy_bb): Copy speculative edges.
9342 (redirect_all_calls): New function.
9343 (copy_cfg_body): Do redirection after loop info is updated.
9344 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9345
9346 2013-08-09 Jan Hubicka <jh@suse.cz>
9347
9348 * lto-streamer-out.c (output_function): Renumber PHIs.
9349 * lto-streamer-in.c (input_function): Likewise.
9350
9351 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9352
9353 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9354 (get_lane_unsigned): Likewise.
9355 (dup_lane_scalar): Likewise.
9356 (get_lane): enable for VALL.
9357 * config/aarch64/aarch64-simd.md
9358 (aarch64_dup_lane_scalar<mode>): Remove.
9359 (aarch64_get_lane_signed<mode>): Likewise.
9360 (aarch64_get_lane_unsigned<mode>): Likewise.
9361 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9362 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9363 (aarch64_get_lane<mode>): Enable for all vector modes.
9364 (aarch64_get_lanedi): Remove misleading constraints.
9365 * config/aarch64/arm_neon.h
9366 (__aarch64_vget_lane_any): Define.
9367 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9368 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9369 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9370 * config/aarch64/iterators.md (VDQQH): New.
9371 (VDQQHS): Likewise.
9372 (vwcore): Likewise.
9373
9374 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9375
9376 * configure.ac: Add GAS check for LEON instructions on SPARC.
9377 * configure: Regenerate.
9378 * config.in: Likewise.
9379 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9380 sparc*-*-* block.
9381 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9382 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9383 for LEON or LEON3.
9384 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9385 (AS_LEON_FLAG): New macro.
9386 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9387 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9388 Deal with LEON and LEON3 for the memory model.
9389 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9390 (atomic_compare_and_swap<mode>_1): Likewise.
9391 (*atomic_compare_and_swap<mode>_1): Likewise.
9392
9393 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9394
9395 * config/arm/neon.md (vcond): Fix floating-point vector
9396 comparisons against 0.
9397
9398 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9399
9400 * lra-constraints.c (emit_spill_move): Remove assert.
9401 (process_alt_operands): Add more debugging
9402 output. Increase reject for spilling into memory. Decrease
9403 reject for reloading scratch.
9404 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9405
9406 2013-08-08 Steve Ellcey <sellcey@mips.com>
9407
9408 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9409 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9410 micromips incompatible. Add nan2008.
9411 (MULTILIB_DIRNAMES): Add nan2008.
9412 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9413 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9414 and micromips incompatible. Add nan2008.
9415 (MULTILIB_DIRNAMES): Add nan2008.
9416 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9417
9418 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9419
9420 PR rtl-optimization/58079
9421 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9422 simplify_comparison has widened a comparison with an integer.
9423
9424 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9425
9426 * config/arm/neon.md (movmisalign<mode>): Disable when we
9427 don't allow unaligned accesses.
9428 (*movmisalign<mode>_neon_store): Likewise.
9429 (*movmisalign<mode>_neon_load): Likewise.
9430 (*movmisalign<mode>_neon_store): Likewise.
9431 (*movmisalign<mode>_neon_load): Likewise.
9432
9433 2013-08-08 Jan Hubicka <jh@suse.cz>
9434
9435 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9436 (make_pass_rebuild_cgraph_edges): Also clear references.
9437 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9438 * ipa-inline-transform.c (inline_transform): Remove all references
9439 after inlining.
9440 * cgraphunit.c (expand_function): Remove all references after
9441 expansion.
9442 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9443 (ipa_find_reference): Rewrite to iterator.
9444 (remove_stmt_references): Likewise.
9445 (ipa_clear_stmts_in_references): New function.
9446 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9447 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9448 clear references.
9449 * ipa-split.c (split_function): Remove references in split function.
9450
9451 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9452
9453 PR target/57431
9454 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9455 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9456
9457 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9458
9459 PR target/56979
9460 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9461 suggested mode for the assignment isn't compatible with the
9462 registers required.
9463
9464 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9465
9466 PR target/58065
9467 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9468
9469 2013-08-07 Xinliang David Li <davidxl@google.com>
9470
9471 * config/i386/i386.opt: New option -mtune-ctrl=.
9472 * config/i386/x86-tune.def: New file.
9473 * config/i386/i386.h: include x86-tune.def.
9474 * config/i386/i386.c (ix86_option_override_internal):
9475 Parsing -mtune-ctrl= option and set tune features.
9476
9477 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9478
9479 PR other/12081
9480 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9481 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9482 to gen_2arg_fn_t.
9483
9484 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9485
9486 * rtl.h (update_alignments): Declare.
9487 * final.c (grow_label_align): New function extracted from...
9488 (shorten_branches): ...here. Call it.
9489 (update_alignments): New function.
9490 * reorg.c (sibling_labels): New variable.
9491 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9492 the new label along with it onto the sibling_labels vector.
9493 (fill_simple_delay_slots): Adjust call to get_label_before.
9494 (fill_slots_from_thread): Likewise.
9495 (relax_delay_slots): Likewise.
9496 (make_return_insns): Likewise.
9497 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9498
9499 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9500
9501 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9502 document its semantics.
9503 (diagnostic_report_diagnostic): Adjust accordingly.
9504
9505 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9506
9507 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9508 (sparc_option_override): ...and port to new C++ pass API.
9509 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9510
9511 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9512
9513 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9514
9515 2013-08-06 Caroline Tice <cmtice@google.com>
9516
9517 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9518 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9519 * tree-pass.h: Add pass_vtable_verify.
9520 * varasm.c (assemble_variable): Add code to properly set the comdat
9521 section and name for the .vtable_map_vars section.
9522 (assemble_vtyv_preinit_initializer): New function.
9523 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9524 LINK_ONCE flag.
9525 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9526 * vtable-verify.c: New file.
9527 * vtable-verify.h: New file.
9528 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9529 initialiation levels.
9530 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9531 * passes.def: Insert pass_vtable_verify.
9532 * aclocal.m4: Reorder includes.
9533 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9534 -fvtv-counts options.
9535 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9536 as appropriate, if -fvtable-verify=... is used.
9537 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9538 -fvtable-verify=... is used.
9539 * Makefile.in (OBJS): Add vtable-verify.o to list.
9540 (vtable-verify.o): Add new build rule.
9541 (GTFILES): Add vtable-verify.c to list.
9542 * common.opt (fvtable-verify=): New flag.
9543 (vtv_priority): Values for fvtable-verify= flag.
9544 (fvtv-counts): New flag.
9545 (fvtv-debug): New flag.
9546 * tree.h (save_vtable_map_decl): New extern function decl.
9547
9548 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9549
9550 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9551 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9552 (pass_data_rl78_devirt): ...new pass_data instance and...
9553 (make_pass_rl78_devirt): ...new function.
9554 (rl78_asm_file_start): Port pass registration to new C++ API.
9555
9556 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9557
9558 * coretypes.h (rtl_opt_pass): Add.
9559 (gcc::context): Add.
9560 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9561 (epiphany_init): Port to new C++ pass API.
9562 (epiphany_optimize_mode_switching): Likewise.
9563 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9564 (pass_manager::get_pass_mode_switching): New.
9565 (pass_manager::get_pass_peephole2): New.
9566 * mode-switching.c (pass_mode_switching): Add clone method.
9567 * recog.c (pass_peephole2): Add clone method.
9568 (pass_split_all_insns): Add clone method.
9569
9570 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9571
9572 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9573 (mips_option_override): ...here, porting to new C++ API for passes.
9574
9575 2013-08-06 Jan Hubicka <jh@suse.cz>
9576
9577 * cgraph.c (cgraph_get_body): New function based on lto.c
9578 implementation.
9579 * cgraph.h (cgraph_get_body): Declare.
9580 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9581 LTO paths.
9582 * cgraphunit.c (expand_function): Get body prior expanding.
9583 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9584 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9585 really need.
9586 * passes.c (do_per_function_toporder): Get body.
9587 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9588 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9589
9590 2013-08-06 Martin Jambor <mjambor@suse.cz>
9591
9592 PR fortran/57987
9593 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9594 is not re-finalized. Rename second parameter to no_collect.
9595
9596 2013-08-06 Martin Jambor <mjambor@suse.cz>
9597
9598 PR middle-end/58041
9599 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9600 MEM_REF has proper alignment information.
9601
9602 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9603
9604 PR other/12081
9605 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9606 class insn_gen_fn.
9607 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9608 rtx (*) (rtx, ...) with insn_gen_fn.
9609 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9610 insn_gen_fn::stored_funcptr. Add initializer braces.
9611
9612 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9613
9614 Rewrite how instances of passes are cloned to remove assumptions
9615 about their sizes (thus allowing pass subclasses to have
9616 additional data fields, albeit non-GC-managed ones at this point).
9617
9618 * passes.c (make_pass_instance): Now that passes have clone
9619 methods, rewrite this function to eliminate XNEW and memcpy
9620 calls that used hardcoded sizes. Since this function no longer
9621 creates pass instances, rename it to...
9622 (add_pass_instance): ...this. Document the old way that passes were
9623 numbered and flagged, and rework this function to continue using it.
9624 (next_pass_1): Add an initial_pass argument for use by
9625 add_pass_instance.
9626 (position_pass): When adding multiple instances of a pass, use
9627 the pass's clone method, rather than relying on the XNEW/memcpy
9628 within the former make_pass_instance (now add_pass_instance).
9629 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9630 the initial instance of the current pass within the pass manager.
9631
9632 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9633
9634 This is the automated part of the conversion of passes from C
9635 structs to C++ classes.
9636
9637 Patch autogenerated by refactor_passes.py from
9638 https://github.com/davidmalcolm/gcc-refactoring-scripts
9639 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9640
9641 * asan.c (pass_asan): Convert from a global struct to a subclass of
9642 gimple_opt_pass along with...
9643 (pass_data_asan): ...new pass_data instance and...
9644 (make_pass_asan): ...new function.
9645 (pass_asan_O0): Convert from a global struct to a subclass of
9646 gimple_opt_pass along with...
9647 (pass_data_asan_O0): ...new pass_data instance and...
9648 (make_pass_asan_O0): ...new function.
9649 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9650 subclass of rtl_opt_pass along with...
9651 (pass_data_inc_dec): ...new pass_data instance and...
9652 (make_pass_inc_dec): ...new function.
9653 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9654 a subclass of rtl_opt_pass along with...
9655 (pass_data_reorder_blocks): ...new pass_data instance and...
9656 (make_pass_reorder_blocks): ...new function.
9657 (pass_duplicate_computed_gotos): Convert from a global struct to a
9658 subclass of rtl_opt_pass along with...
9659 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9660 (make_pass_duplicate_computed_gotos): ...new function.
9661 (pass_partition_blocks): Convert from a global struct to a subclass of
9662 rtl_opt_pass along with...
9663 (pass_data_partition_blocks): ...new pass_data instance and...
9664 (make_pass_partition_blocks): ...new function.
9665 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9666 struct to a subclass of rtl_opt_pass along with...
9667 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9668 and...
9669 (make_pass_branch_target_load_optimize1): ...new function.
9670 (pass_branch_target_load_optimize2): Convert from a global struct to a
9671 subclass of rtl_opt_pass along with...
9672 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9673 and...
9674 (make_pass_branch_target_load_optimize2): ...new function.
9675 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9676 of rtl_opt_pass along with...
9677 (pass_data_jump): ...new pass_data instance and...
9678 (make_pass_jump): ...new function.
9679 (pass_jump2): Convert from a global struct to a subclass of
9680 rtl_opt_pass along with...
9681 (pass_data_jump2): ...new pass_data instance and...
9682 (make_pass_jump2): ...new function.
9683 * cfgexpand.c (pass_expand): Convert from a global struct to a
9684 subclass of rtl_opt_pass along with...
9685 (pass_data_expand): ...new pass_data instance and...
9686 (make_pass_expand): ...new function.
9687 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9688 of rtl_opt_pass along with...
9689 (pass_data_free_cfg): ...new pass_data instance and...
9690 (make_pass_free_cfg): ...new function.
9691 (pass_into_cfg_layout_mode): Convert from a global struct to a
9692 subclass of rtl_opt_pass along with...
9693 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9694 (make_pass_into_cfg_layout_mode): ...new function.
9695 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9696 subclass of rtl_opt_pass along with...
9697 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9698 (make_pass_outof_cfg_layout_mode): ...new function.
9699 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9700 struct to a subclass of gimple_opt_pass along with...
9701 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9702 (make_pass_build_cgraph_edges): ...new function.
9703 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9704 subclass of gimple_opt_pass along with...
9705 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9706 (make_pass_rebuild_cgraph_edges): ...new function.
9707 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9708 subclass of gimple_opt_pass along with...
9709 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9710 and...
9711 (make_pass_remove_cgraph_callee_edges): ...new function.
9712 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9713 struct to a subclass of rtl_opt_pass along with...
9714 (pass_data_stack_adjustments): ...new pass_data instance and...
9715 (make_pass_stack_adjustments): ...new function.
9716 * combine.c (pass_combine): Convert from a global struct to a subclass
9717 of rtl_opt_pass along with...
9718 (pass_data_combine): ...new pass_data instance and...
9719 (make_pass_combine): ...new function.
9720 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9721 global struct to a subclass of rtl_opt_pass along with...
9722 (pass_data_compare_elim_after_reload): ...new pass_data instance
9723 and...
9724 (make_pass_compare_elim_after_reload): ...new function.
9725 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9726 of rtl_opt_pass along with...
9727 (pass_data_rtl_cprop): ...new pass_data instance and...
9728 (make_pass_rtl_cprop): ...new function.
9729 * cse.c (pass_cse): Convert from a global struct to a subclass of
9730 rtl_opt_pass along with...
9731 (pass_data_cse): ...new pass_data instance and...
9732 (make_pass_cse): ...new function.
9733 (pass_cse2): Convert from a global struct to a subclass of
9734 rtl_opt_pass along with...
9735 (pass_data_cse2): ...new pass_data instance and...
9736 (make_pass_cse2): ...new function.
9737 (pass_cse_after_global_opts): Convert from a global struct to a
9738 subclass of rtl_opt_pass along with...
9739 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9740 (make_pass_cse_after_global_opts): ...new function.
9741 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9742 of rtl_opt_pass along with...
9743 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9744 (make_pass_ud_rtl_dce): ...new function.
9745 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9746 rtl_opt_pass along with...
9747 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9748 (make_pass_fast_rtl_dce): ...new function.
9749 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9750 a subclass of rtl_opt_pass along with...
9751 (pass_data_df_initialize_opt): ...new pass_data instance and...
9752 (make_pass_df_initialize_opt): ...new function.
9753 (pass_df_initialize_no_opt): Convert from a global struct to a
9754 subclass of rtl_opt_pass along with...
9755 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9756 (make_pass_df_initialize_no_opt): ...new function.
9757 (pass_df_finish): Convert from a global struct to a subclass of
9758 rtl_opt_pass along with...
9759 (pass_data_df_finish): ...new pass_data instance and...
9760 (make_pass_df_finish): ...new function.
9761 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9762 rtl_opt_pass along with...
9763 (pass_data_rtl_dse1): ...new pass_data instance and...
9764 (make_pass_rtl_dse1): ...new function.
9765 (pass_rtl_dse2): Convert from a global struct to a subclass of
9766 rtl_opt_pass along with...
9767 (pass_data_rtl_dse2): ...new pass_data instance and...
9768 (make_pass_rtl_dse2): ...new function.
9769 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9770 subclass of rtl_opt_pass along with...
9771 (pass_data_dwarf2_frame): ...new pass_data instance and...
9772 (make_pass_dwarf2_frame): ...new function.
9773 * except.c (pass_set_nothrow_function_flags): Convert from a global
9774 struct to a subclass of rtl_opt_pass along with...
9775 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9776 and...
9777 (make_pass_set_nothrow_function_flags): ...new function.
9778 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9779 subclass of rtl_opt_pass along with...
9780 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9781 and...
9782 (make_pass_convert_to_eh_region_ranges): ...new function.
9783 * final.c (pass_compute_alignments): Convert from a global struct to a
9784 subclass of rtl_opt_pass along with...
9785 (pass_data_compute_alignments): ...new pass_data instance and...
9786 (make_pass_compute_alignments): ...new function.
9787 (pass_final): Convert from a global struct to a subclass of
9788 rtl_opt_pass along with...
9789 (pass_data_final): ...new pass_data instance and...
9790 (make_pass_final): ...new function.
9791 (pass_shorten_branches): Convert from a global struct to a subclass of
9792 rtl_opt_pass along with...
9793 (pass_data_shorten_branches): ...new pass_data instance and...
9794 (make_pass_shorten_branches): ...new function.
9795 (pass_clean_state): Convert from a global struct to a subclass of
9796 rtl_opt_pass along with...
9797 (pass_data_clean_state): ...new pass_data instance and...
9798 (make_pass_clean_state): ...new function.
9799 * function.c (pass_instantiate_virtual_regs): Convert from a global
9800 struct to a subclass of rtl_opt_pass along with...
9801 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9802 (make_pass_instantiate_virtual_regs): ...new function.
9803 (pass_leaf_regs): Convert from a global struct to a subclass of
9804 rtl_opt_pass along with...
9805 (pass_data_leaf_regs): ...new pass_data instance and...
9806 (make_pass_leaf_regs): ...new function.
9807 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9808 subclass of rtl_opt_pass along with...
9809 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9810 and...
9811 (make_pass_thread_prologue_and_epilogue): ...new function.
9812 (pass_match_asm_constraints): Convert from a global struct to a
9813 subclass of rtl_opt_pass along with...
9814 (pass_data_match_asm_constraints): ...new pass_data instance and...
9815 (make_pass_match_asm_constraints): ...new function.
9816 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9817 subclass of rtl_opt_pass along with...
9818 (pass_data_rtl_fwprop): ...new pass_data instance and...
9819 (make_pass_rtl_fwprop): ...new function.
9820 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9821 rtl_opt_pass along with...
9822 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9823 (make_pass_rtl_fwprop_addr): ...new function.
9824 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9825 rtl_opt_pass along with...
9826 (pass_data_rtl_pre): ...new pass_data instance and...
9827 (make_pass_rtl_pre): ...new function.
9828 (pass_rtl_hoist): Convert from a global struct to a subclass of
9829 rtl_opt_pass along with...
9830 (pass_data_rtl_hoist): ...new pass_data instance and...
9831 (make_pass_rtl_hoist): ...new function.
9832 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9833 subclass of gimple_opt_pass along with...
9834 (pass_data_lower_cf): ...new pass_data instance and...
9835 (make_pass_lower_cf): ...new function.
9836 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9837 from a global struct to a subclass of gimple_opt_pass along with...
9838 (pass_data_strength_reduction): ...new pass_data instance and...
9839 (make_pass_strength_reduction): ...new function.
9840 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9841 of rtl_opt_pass along with...
9842 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9843 (make_pass_rtl_ifcvt): ...new function.
9844 (pass_if_after_combine): Convert from a global struct to a subclass of
9845 rtl_opt_pass along with...
9846 (pass_data_if_after_combine): ...new pass_data instance and...
9847 (make_pass_if_after_combine): ...new function.
9848 (pass_if_after_reload): Convert from a global struct to a subclass of
9849 rtl_opt_pass along with...
9850 (pass_data_if_after_reload): ...new pass_data instance and...
9851 (make_pass_if_after_reload): ...new function.
9852 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9853 a subclass of rtl_opt_pass along with...
9854 (pass_data_initialize_regs): ...new pass_data instance and...
9855 (make_pass_initialize_regs): ...new function.
9856 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9857 of ipa_opt_pass_d along with...
9858 (pass_data_ipa_cp): ...new pass_data instance and...
9859 (make_pass_ipa_cp): ...new function.
9860 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9861 global struct to a subclass of gimple_opt_pass along with...
9862 (pass_data_inline_parameters): ...new pass_data instance and...
9863 (make_pass_inline_parameters): ...new function.
9864 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9865 subclass of gimple_opt_pass along with...
9866 (pass_data_early_inline): ...new pass_data instance and...
9867 (make_pass_early_inline): ...new function.
9868 (pass_ipa_inline): Convert from a global struct to a subclass of
9869 ipa_opt_pass_d along with...
9870 (pass_data_ipa_inline): ...new pass_data instance and...
9871 (make_pass_ipa_inline): ...new function.
9872 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9873 struct to a subclass of gimple_opt_pass along with...
9874 (pass_data_local_pure_const): ...new pass_data instance and...
9875 (make_pass_local_pure_const): ...new function.
9876 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9877 ipa_opt_pass_d along with...
9878 (pass_data_ipa_pure_const): ...new pass_data instance and...
9879 (make_pass_ipa_pure_const): ...new function.
9880 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9881 to a subclass of ipa_opt_pass_d along with...
9882 (pass_data_ipa_reference): ...new pass_data instance and...
9883 (make_pass_ipa_reference): ...new function.
9884 * ipa-split.c (pass_split_functions): Convert from a global struct to
9885 a subclass of gimple_opt_pass along with...
9886 (pass_data_split_functions): ...new pass_data instance and...
9887 (make_pass_split_functions): ...new function.
9888 (pass_feedback_split_functions): Convert from a global struct to a
9889 subclass of gimple_opt_pass along with...
9890 (pass_data_feedback_split_functions): ...new pass_data instance and...
9891 (make_pass_feedback_split_functions): ...new function.
9892 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9893 global struct to a subclass of simple_ipa_opt_pass along with...
9894 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9895 instance and...
9896 (make_pass_ipa_function_and_variable_visibility): ...new function.
9897 (pass_ipa_free_inline_summary): Convert from a global struct to a
9898 subclass of simple_ipa_opt_pass along with...
9899 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9900 (make_pass_ipa_free_inline_summary): ...new function.
9901 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9902 subclass of ipa_opt_pass_d along with...
9903 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9904 and...
9905 (make_pass_ipa_whole_program_visibility): ...new function.
9906 (pass_ipa_profile): Convert from a global struct to a subclass of
9907 ipa_opt_pass_d along with...
9908 (pass_data_ipa_profile): ...new pass_data instance and...
9909 (make_pass_ipa_profile): ...new function.
9910 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9911 ipa_opt_pass_d along with...
9912 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9913 (make_pass_ipa_cdtor_merge): ...new function.
9914 * ira.c (pass_ira): Convert from a global struct to a subclass of
9915 rtl_opt_pass along with...
9916 (pass_data_ira): ...new pass_data instance and...
9917 (make_pass_ira): ...new function.
9918 (pass_reload): Convert from a global struct to a subclass of
9919 rtl_opt_pass along with...
9920 (pass_data_reload): ...new pass_data instance and...
9921 (make_pass_reload): ...new function.
9922 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9923 subclass of rtl_opt_pass along with...
9924 (pass_data_cleanup_barriers): ...new pass_data instance and...
9925 (make_pass_cleanup_barriers): ...new function.
9926 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9927 of rtl_opt_pass along with...
9928 (pass_data_loop2): ...new pass_data instance and...
9929 (make_pass_loop2): ...new function.
9930 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9931 rtl_opt_pass along with...
9932 (pass_data_rtl_loop_init): ...new pass_data instance and...
9933 (make_pass_rtl_loop_init): ...new function.
9934 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9935 rtl_opt_pass along with...
9936 (pass_data_rtl_loop_done): ...new pass_data instance and...
9937 (make_pass_rtl_loop_done): ...new function.
9938 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9939 subclass of rtl_opt_pass along with...
9940 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9941 (make_pass_rtl_move_loop_invariants): ...new function.
9942 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9943 rtl_opt_pass along with...
9944 (pass_data_rtl_unswitch): ...new pass_data instance and...
9945 (make_pass_rtl_unswitch): ...new function.
9946 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9947 subclass of rtl_opt_pass along with...
9948 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9949 and...
9950 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9951 (pass_rtl_doloop): Convert from a global struct to a subclass of
9952 rtl_opt_pass along with...
9953 (pass_data_rtl_doloop): ...new pass_data instance and...
9954 (make_pass_rtl_doloop): ...new function.
9955 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9956 a subclass of rtl_opt_pass along with...
9957 (pass_data_lower_subreg): ...new pass_data instance and...
9958 (make_pass_lower_subreg): ...new function.
9959 (pass_lower_subreg2): Convert from a global struct to a subclass of
9960 rtl_opt_pass along with...
9961 (pass_data_lower_subreg2): ...new pass_data instance and...
9962 (make_pass_lower_subreg2): ...new function.
9963 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9964 struct to a subclass of ipa_opt_pass_d along with...
9965 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9966 (make_pass_ipa_lto_gimple_out): ...new function.
9967 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9968 of ipa_opt_pass_d along with...
9969 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9970 (make_pass_ipa_lto_finish_out): ...new function.
9971 * mode-switching.c (pass_mode_switching): Convert from a global struct
9972 to a subclass of rtl_opt_pass along with...
9973 (pass_data_mode_switching): ...new pass_data instance and...
9974 (make_pass_mode_switching): ...new function.
9975 * modulo-sched.c (pass_sms): Convert from a global struct to a
9976 subclass of rtl_opt_pass along with...
9977 (pass_data_sms): ...new pass_data instance and...
9978 (make_pass_sms): ...new function.
9979 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9980 subclass of gimple_opt_pass along with...
9981 (pass_data_expand_omp): ...new pass_data instance and...
9982 (make_pass_expand_omp): ...new function.
9983 (pass_lower_omp): Convert from a global struct to a subclass of
9984 gimple_opt_pass along with...
9985 (pass_data_lower_omp): ...new pass_data instance and...
9986 (make_pass_lower_omp): ...new function.
9987 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9988 of gimple_opt_pass along with...
9989 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9990 (make_pass_diagnose_omp_blocks): ...new function.
9991 * passes.c (pass_early_local_passes): Convert from a global struct to
9992 a subclass of simple_ipa_opt_pass along with...
9993 (pass_data_early_local_passes): ...new pass_data instance and...
9994 (make_pass_early_local_passes): ...new function.
9995 (pass_all_early_optimizations): Convert from a global struct to a
9996 subclass of gimple_opt_pass along with...
9997 (pass_data_all_early_optimizations): ...new pass_data instance and...
9998 (make_pass_all_early_optimizations): ...new function.
9999 (pass_all_optimizations): Convert from a global struct to a subclass
10000 of gimple_opt_pass along with...
10001 (pass_data_all_optimizations): ...new pass_data instance and...
10002 (make_pass_all_optimizations): ...new function.
10003 (pass_all_optimizations_g): Convert from a global struct to a subclass
10004 of gimple_opt_pass along with...
10005 (pass_data_all_optimizations_g): ...new pass_data instance and...
10006 (make_pass_all_optimizations_g): ...new function.
10007 (pass_rest_of_compilation): Convert from a global struct to a subclass
10008 of rtl_opt_pass along with...
10009 (pass_data_rest_of_compilation): ...new pass_data instance and...
10010 (make_pass_rest_of_compilation): ...new function.
10011 (pass_postreload): Convert from a global struct to a subclass of
10012 rtl_opt_pass along with...
10013 (pass_data_postreload): ...new pass_data instance and...
10014 (make_pass_postreload): ...new function.
10015 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10016 subclass of rtl_opt_pass along with...
10017 (pass_data_gcse2): ...new pass_data instance and...
10018 (make_pass_gcse2): ...new function.
10019 * postreload.c (pass_postreload_cse): Convert from a global struct to
10020 a subclass of rtl_opt_pass along with...
10021 (pass_data_postreload_cse): ...new pass_data instance and...
10022 (make_pass_postreload_cse): ...new function.
10023 * predict.c (pass_profile): Convert from a global struct to a subclass
10024 of gimple_opt_pass along with...
10025 (pass_data_profile): ...new pass_data instance and...
10026 (make_pass_profile): ...new function.
10027 (pass_strip_predict_hints): Convert from a global struct to a subclass
10028 of gimple_opt_pass along with...
10029 (pass_data_strip_predict_hints): ...new pass_data instance and...
10030 (make_pass_strip_predict_hints): ...new function.
10031 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10032 of rtl_opt_pass along with...
10033 (pass_data_peephole2): ...new pass_data instance and...
10034 (make_pass_peephole2): ...new function.
10035 (pass_split_all_insns): Convert from a global struct to a subclass of
10036 rtl_opt_pass along with...
10037 (pass_data_split_all_insns): ...new pass_data instance and...
10038 (make_pass_split_all_insns): ...new function.
10039 (pass_split_after_reload): Convert from a global struct to a subclass
10040 of rtl_opt_pass along with...
10041 (pass_data_split_after_reload): ...new pass_data instance and...
10042 (make_pass_split_after_reload): ...new function.
10043 (pass_split_before_regstack): Convert from a global struct to a
10044 subclass of rtl_opt_pass along with...
10045 (pass_data_split_before_regstack): ...new pass_data instance and...
10046 (make_pass_split_before_regstack): ...new function.
10047 (pass_split_before_sched2): Convert from a global struct to a subclass
10048 of rtl_opt_pass along with...
10049 (pass_data_split_before_sched2): ...new pass_data instance and...
10050 (make_pass_split_before_sched2): ...new function.
10051 (pass_split_for_shorten_branches): Convert from a global struct to a
10052 subclass of rtl_opt_pass along with...
10053 (pass_data_split_for_shorten_branches): ...new pass_data instance
10054 and...
10055 (make_pass_split_for_shorten_branches): ...new function.
10056 * ree.c (pass_ree): Convert from a global struct to a subclass of
10057 rtl_opt_pass along with...
10058 (pass_data_ree): ...new pass_data instance and...
10059 (make_pass_ree): ...new function.
10060 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10061 subclass of rtl_opt_pass along with...
10062 (pass_data_stack_regs): ...new pass_data instance and...
10063 (make_pass_stack_regs): ...new function.
10064 (pass_stack_regs_run): Convert from a global struct to a subclass of
10065 rtl_opt_pass along with...
10066 (pass_data_stack_regs_run): ...new pass_data instance and...
10067 (make_pass_stack_regs_run): ...new function.
10068 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10069 subclass of rtl_opt_pass along with...
10070 (pass_data_cprop_hardreg): ...new pass_data instance and...
10071 (make_pass_cprop_hardreg): ...new function.
10072 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10073 subclass of rtl_opt_pass along with...
10074 (pass_data_reginfo_init): ...new pass_data instance and...
10075 (make_pass_reginfo_init): ...new function.
10076 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10077 of rtl_opt_pass along with...
10078 (pass_data_regmove): ...new pass_data instance and...
10079 (make_pass_regmove): ...new function.
10080 * regrename.c (pass_regrename): Convert from a global struct to a
10081 subclass of rtl_opt_pass along with...
10082 (pass_data_regrename): ...new pass_data instance and...
10083 (make_pass_regrename): ...new function.
10084 * reorg.c (pass_delay_slots): Convert from a global struct to a
10085 subclass of rtl_opt_pass along with...
10086 (pass_data_delay_slots): ...new pass_data instance and...
10087 (make_pass_delay_slots): ...new function.
10088 (pass_machine_reorg): Convert from a global struct to a subclass of
10089 rtl_opt_pass along with...
10090 (pass_data_machine_reorg): ...new pass_data instance and...
10091 (make_pass_machine_reorg): ...new function.
10092 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10093 of rtl_opt_pass along with...
10094 (pass_data_sched): ...new pass_data instance and...
10095 (make_pass_sched): ...new function.
10096 (pass_sched2): Convert from a global struct to a subclass of
10097 rtl_opt_pass along with...
10098 (pass_data_sched2): ...new pass_data instance and...
10099 (make_pass_sched2): ...new function.
10100 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10101 to a subclass of rtl_opt_pass along with...
10102 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10103 (make_pass_stack_ptr_mod): ...new function.
10104 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10105 to a subclass of rtl_opt_pass along with...
10106 (pass_data_rtl_store_motion): ...new pass_data instance and...
10107 (make_pass_rtl_store_motion): ...new function.
10108 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10109 of gimple_opt_pass along with...
10110 (pass_data_tracer): ...new pass_data instance and...
10111 (make_pass_tracer): ...new function.
10112 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10113 to a subclass of gimple_opt_pass along with...
10114 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10115 (make_pass_diagnose_tm_blocks): ...new function.
10116 (pass_lower_tm): Convert from a global struct to a subclass of
10117 gimple_opt_pass along with...
10118 (pass_data_lower_tm): ...new pass_data instance and...
10119 (make_pass_lower_tm): ...new function.
10120 (pass_tm_init): Convert from a global struct to a subclass of
10121 gimple_opt_pass along with...
10122 (pass_data_tm_init): ...new pass_data instance and...
10123 (make_pass_tm_init): ...new function.
10124 (pass_tm_mark): Convert from a global struct to a subclass of
10125 gimple_opt_pass along with...
10126 (pass_data_tm_mark): ...new pass_data instance and...
10127 (make_pass_tm_mark): ...new function.
10128 (pass_tm_edges): Convert from a global struct to a subclass of
10129 gimple_opt_pass along with...
10130 (pass_data_tm_edges): ...new pass_data instance and...
10131 (make_pass_tm_edges): ...new function.
10132 (pass_tm_memopt): Convert from a global struct to a subclass of
10133 gimple_opt_pass along with...
10134 (pass_data_tm_memopt): ...new pass_data instance and...
10135 (make_pass_tm_memopt): ...new function.
10136 (pass_ipa_tm): Convert from a global struct to a subclass of
10137 simple_ipa_opt_pass along with...
10138 (pass_data_ipa_tm): ...new pass_data instance and...
10139 (make_pass_ipa_tm): ...new function.
10140 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10141 subclass of gimple_opt_pass along with...
10142 (pass_data_call_cdce): ...new pass_data instance and...
10143 (make_pass_call_cdce): ...new function.
10144 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10145 subclass of gimple_opt_pass along with...
10146 (pass_data_build_cfg): ...new pass_data instance and...
10147 (make_pass_build_cfg): ...new function.
10148 (pass_split_crit_edges): Convert from a global struct to a subclass of
10149 gimple_opt_pass along with...
10150 (pass_data_split_crit_edges): ...new pass_data instance and...
10151 (make_pass_split_crit_edges): ...new function.
10152 (pass_warn_function_return): Convert from a global struct to a
10153 subclass of gimple_opt_pass along with...
10154 (pass_data_warn_function_return): ...new pass_data instance and...
10155 (make_pass_warn_function_return): ...new function.
10156 (pass_warn_function_noreturn): Convert from a global struct to a
10157 subclass of gimple_opt_pass along with...
10158 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10159 (make_pass_warn_function_noreturn): ...new function.
10160 (pass_warn_unused_result): Convert from a global struct to a subclass
10161 of gimple_opt_pass along with...
10162 (pass_data_warn_unused_result): ...new pass_data instance and...
10163 (make_pass_warn_unused_result): ...new function.
10164 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10165 a subclass of gimple_opt_pass along with...
10166 (pass_data_merge_phi): ...new pass_data instance and...
10167 (make_pass_merge_phi): ...new function.
10168 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10169 a subclass of gimple_opt_pass along with...
10170 (pass_data_lower_complex): ...new pass_data instance and...
10171 (make_pass_lower_complex): ...new function.
10172 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10173 gimple_opt_pass along with...
10174 (pass_data_lower_complex_O0): ...new pass_data instance and...
10175 (make_pass_lower_complex_O0): ...new function.
10176 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10177 subclass of gimple_opt_pass along with...
10178 (pass_data_lower_eh): ...new pass_data instance and...
10179 (make_pass_lower_eh): ...new function.
10180 (pass_refactor_eh): Convert from a global struct to a subclass of
10181 gimple_opt_pass along with...
10182 (pass_data_refactor_eh): ...new pass_data instance and...
10183 (make_pass_refactor_eh): ...new function.
10184 (pass_lower_resx): Convert from a global struct to a subclass of
10185 gimple_opt_pass along with...
10186 (pass_data_lower_resx): ...new pass_data instance and...
10187 (make_pass_lower_resx): ...new function.
10188 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10189 of gimple_opt_pass along with...
10190 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10191 (make_pass_lower_eh_dispatch): ...new function.
10192 (pass_cleanup_eh): Convert from a global struct to a subclass of
10193 gimple_opt_pass along with...
10194 (pass_data_cleanup_eh): ...new pass_data instance and...
10195 (make_pass_cleanup_eh): ...new function.
10196 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10197 to a subclass of simple_ipa_opt_pass along with...
10198 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10199 (make_pass_ipa_lower_emutls): ...new function.
10200 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10201 a subclass of gimple_opt_pass along with...
10202 (pass_data_if_conversion): ...new pass_data instance and...
10203 (make_pass_if_conversion): ...new function.
10204 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10205 subclass of gimple_opt_pass along with...
10206 (pass_data_build_ssa): ...new pass_data instance and...
10207 (make_pass_build_ssa): ...new function.
10208 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10209 global struct to a subclass of gimple_opt_pass along with...
10210 (pass_data_loop_distribution): ...new pass_data instance and...
10211 (make_pass_loop_distribution): ...new function.
10212 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10213 subclass of gimple_opt_pass along with...
10214 (pass_data_mudflap_1): ...new pass_data instance and...
10215 (make_pass_mudflap_1): ...new function.
10216 (pass_mudflap_2): Convert from a global struct to a subclass of
10217 gimple_opt_pass along with...
10218 (pass_data_mudflap_2): ...new pass_data instance and...
10219 (make_pass_mudflap_2): ...new function.
10220 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10221 subclass of gimple_opt_pass along with...
10222 (pass_data_mudflap_1): ...new pass_data instance and...
10223 (make_pass_mudflap_1): ...new function.
10224 (pass_mudflap_2): Convert from a global struct to a subclass of
10225 gimple_opt_pass along with...
10226 (pass_data_mudflap_2): ...new pass_data instance and...
10227 (make_pass_mudflap_2): ...new function.
10228 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10229 gimple_opt_pass along with...
10230 (pass_data_nrv): ...new pass_data instance and...
10231 (make_pass_nrv): ...new function.
10232 (pass_return_slot): Convert from a global struct to a subclass of
10233 gimple_opt_pass along with...
10234 (pass_data_return_slot): ...new pass_data instance and...
10235 (make_pass_return_slot): ...new function.
10236 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10237 to a subclass of gimple_opt_pass along with...
10238 (pass_data_object_sizes): ...new pass_data instance and...
10239 (make_pass_object_sizes): ...new function.
10240 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10241 global struct to a subclass of gimple_opt_pass along with...
10242 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10243 and...
10244 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10245 (pass_fixup_cfg): Convert from a global struct to a subclass of
10246 gimple_opt_pass along with...
10247 (pass_data_fixup_cfg): ...new pass_data instance and...
10248 (make_pass_fixup_cfg): ...new function.
10249 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10250 (make_pass_mudflap_1): ...new function.
10251 (pass_mudflap_2): Replace declaration with that of...
10252 (make_pass_mudflap_2): ...new function.
10253 (pass_asan): Replace declaration with that of...
10254 (make_pass_asan): ...new function.
10255 (pass_asan_O0): Replace declaration with that of...
10256 (make_pass_asan_O0): ...new function.
10257 (pass_tsan): Replace declaration with that of...
10258 (make_pass_tsan): ...new function.
10259 (pass_tsan_O0): Replace declaration with that of...
10260 (make_pass_tsan_O0): ...new function.
10261 (pass_lower_cf): Replace declaration with that of...
10262 (make_pass_lower_cf): ...new function.
10263 (pass_refactor_eh): Replace declaration with that of...
10264 (make_pass_refactor_eh): ...new function.
10265 (pass_lower_eh): Replace declaration with that of...
10266 (make_pass_lower_eh): ...new function.
10267 (pass_lower_eh_dispatch): Replace declaration with that of...
10268 (make_pass_lower_eh_dispatch): ...new function.
10269 (pass_lower_resx): Replace declaration with that of...
10270 (make_pass_lower_resx): ...new function.
10271 (pass_build_cfg): Replace declaration with that of...
10272 (make_pass_build_cfg): ...new function.
10273 (pass_early_tree_profile): Replace declaration with that of...
10274 (make_pass_early_tree_profile): ...new function.
10275 (pass_cleanup_eh): Replace declaration with that of...
10276 (make_pass_cleanup_eh): ...new function.
10277 (pass_sra): Replace declaration with that of...
10278 (make_pass_sra): ...new function.
10279 (pass_sra_early): Replace declaration with that of...
10280 (make_pass_sra_early): ...new function.
10281 (pass_early_ipa_sra): Replace declaration with that of...
10282 (make_pass_early_ipa_sra): ...new function.
10283 (pass_tail_recursion): Replace declaration with that of...
10284 (make_pass_tail_recursion): ...new function.
10285 (pass_tail_calls): Replace declaration with that of...
10286 (make_pass_tail_calls): ...new function.
10287 (pass_tree_loop): Replace declaration with that of...
10288 (make_pass_tree_loop): ...new function.
10289 (pass_tree_loop_init): Replace declaration with that of...
10290 (make_pass_tree_loop_init): ...new function.
10291 (pass_lim): Replace declaration with that of...
10292 (make_pass_lim): ...new function.
10293 (pass_tree_unswitch): Replace declaration with that of...
10294 (make_pass_tree_unswitch): ...new function.
10295 (pass_predcom): Replace declaration with that of...
10296 (make_pass_predcom): ...new function.
10297 (pass_iv_canon): Replace declaration with that of...
10298 (make_pass_iv_canon): ...new function.
10299 (pass_scev_cprop): Replace declaration with that of...
10300 (make_pass_scev_cprop): ...new function.
10301 (pass_empty_loop): Replace declaration with that of...
10302 (make_pass_empty_loop): ...new function.
10303 (pass_record_bounds): Replace declaration with that of...
10304 (make_pass_record_bounds): ...new function.
10305 (pass_graphite): Replace declaration with that of...
10306 (make_pass_graphite): ...new function.
10307 (pass_graphite_transforms): Replace declaration with that of...
10308 (make_pass_graphite_transforms): ...new function.
10309 (pass_if_conversion): Replace declaration with that of...
10310 (make_pass_if_conversion): ...new function.
10311 (pass_loop_distribution): Replace declaration with that of...
10312 (make_pass_loop_distribution): ...new function.
10313 (pass_vectorize): Replace declaration with that of...
10314 (make_pass_vectorize): ...new function.
10315 (pass_slp_vectorize): Replace declaration with that of...
10316 (make_pass_slp_vectorize): ...new function.
10317 (pass_complete_unroll): Replace declaration with that of...
10318 (make_pass_complete_unroll): ...new function.
10319 (pass_complete_unrolli): Replace declaration with that of...
10320 (make_pass_complete_unrolli): ...new function.
10321 (pass_parallelize_loops): Replace declaration with that of...
10322 (make_pass_parallelize_loops): ...new function.
10323 (pass_loop_prefetch): Replace declaration with that of...
10324 (make_pass_loop_prefetch): ...new function.
10325 (pass_iv_optimize): Replace declaration with that of...
10326 (make_pass_iv_optimize): ...new function.
10327 (pass_tree_loop_done): Replace declaration with that of...
10328 (make_pass_tree_loop_done): ...new function.
10329 (pass_ch): Replace declaration with that of...
10330 (make_pass_ch): ...new function.
10331 (pass_ccp): Replace declaration with that of...
10332 (make_pass_ccp): ...new function.
10333 (pass_phi_only_cprop): Replace declaration with that of...
10334 (make_pass_phi_only_cprop): ...new function.
10335 (pass_build_ssa): Replace declaration with that of...
10336 (make_pass_build_ssa): ...new function.
10337 (pass_build_alias): Replace declaration with that of...
10338 (make_pass_build_alias): ...new function.
10339 (pass_build_ealias): Replace declaration with that of...
10340 (make_pass_build_ealias): ...new function.
10341 (pass_dominator): Replace declaration with that of...
10342 (make_pass_dominator): ...new function.
10343 (pass_dce): Replace declaration with that of...
10344 (make_pass_dce): ...new function.
10345 (pass_dce_loop): Replace declaration with that of...
10346 (make_pass_dce_loop): ...new function.
10347 (pass_cd_dce): Replace declaration with that of...
10348 (make_pass_cd_dce): ...new function.
10349 (pass_call_cdce): Replace declaration with that of...
10350 (make_pass_call_cdce): ...new function.
10351 (pass_merge_phi): Replace declaration with that of...
10352 (make_pass_merge_phi): ...new function.
10353 (pass_split_crit_edges): Replace declaration with that of...
10354 (make_pass_split_crit_edges): ...new function.
10355 (pass_pre): Replace declaration with that of...
10356 (make_pass_pre): ...new function.
10357 (pass_profile): Replace declaration with that of...
10358 (make_pass_profile): ...new function.
10359 (pass_strip_predict_hints): Replace declaration with that of...
10360 (make_pass_strip_predict_hints): ...new function.
10361 (pass_lower_complex_O0): Replace declaration with that of...
10362 (make_pass_lower_complex_O0): ...new function.
10363 (pass_lower_complex): Replace declaration with that of...
10364 (make_pass_lower_complex): ...new function.
10365 (pass_lower_vector): Replace declaration with that of...
10366 (make_pass_lower_vector): ...new function.
10367 (pass_lower_vector_ssa): Replace declaration with that of...
10368 (make_pass_lower_vector_ssa): ...new function.
10369 (pass_lower_omp): Replace declaration with that of...
10370 (make_pass_lower_omp): ...new function.
10371 (pass_diagnose_omp_blocks): Replace declaration with that of...
10372 (make_pass_diagnose_omp_blocks): ...new function.
10373 (pass_expand_omp): Replace declaration with that of...
10374 (make_pass_expand_omp): ...new function.
10375 (pass_expand_omp_ssa): Replace declaration with that of...
10376 (make_pass_expand_omp_ssa): ...new function.
10377 (pass_object_sizes): Replace declaration with that of...
10378 (make_pass_object_sizes): ...new function.
10379 (pass_strlen): Replace declaration with that of...
10380 (make_pass_strlen): ...new function.
10381 (pass_fold_builtins): Replace declaration with that of...
10382 (make_pass_fold_builtins): ...new function.
10383 (pass_stdarg): Replace declaration with that of...
10384 (make_pass_stdarg): ...new function.
10385 (pass_early_warn_uninitialized): Replace declaration with that of...
10386 (make_pass_early_warn_uninitialized): ...new function.
10387 (pass_late_warn_uninitialized): Replace declaration with that of...
10388 (make_pass_late_warn_uninitialized): ...new function.
10389 (pass_cse_reciprocals): Replace declaration with that of...
10390 (make_pass_cse_reciprocals): ...new function.
10391 (pass_cse_sincos): Replace declaration with that of...
10392 (make_pass_cse_sincos): ...new function.
10393 (pass_optimize_bswap): Replace declaration with that of...
10394 (make_pass_optimize_bswap): ...new function.
10395 (pass_optimize_widening_mul): Replace declaration with that of...
10396 (make_pass_optimize_widening_mul): ...new function.
10397 (pass_warn_function_return): Replace declaration with that of...
10398 (make_pass_warn_function_return): ...new function.
10399 (pass_warn_function_noreturn): Replace declaration with that of...
10400 (make_pass_warn_function_noreturn): ...new function.
10401 (pass_cselim): Replace declaration with that of...
10402 (make_pass_cselim): ...new function.
10403 (pass_phiopt): Replace declaration with that of...
10404 (make_pass_phiopt): ...new function.
10405 (pass_forwprop): Replace declaration with that of...
10406 (make_pass_forwprop): ...new function.
10407 (pass_phiprop): Replace declaration with that of...
10408 (make_pass_phiprop): ...new function.
10409 (pass_tree_ifcombine): Replace declaration with that of...
10410 (make_pass_tree_ifcombine): ...new function.
10411 (pass_dse): Replace declaration with that of...
10412 (make_pass_dse): ...new function.
10413 (pass_nrv): Replace declaration with that of...
10414 (make_pass_nrv): ...new function.
10415 (pass_rename_ssa_copies): Replace declaration with that of...
10416 (make_pass_rename_ssa_copies): ...new function.
10417 (pass_sink_code): Replace declaration with that of...
10418 (make_pass_sink_code): ...new function.
10419 (pass_fre): Replace declaration with that of...
10420 (make_pass_fre): ...new function.
10421 (pass_check_data_deps): Replace declaration with that of...
10422 (make_pass_check_data_deps): ...new function.
10423 (pass_copy_prop): Replace declaration with that of...
10424 (make_pass_copy_prop): ...new function.
10425 (pass_vrp): Replace declaration with that of...
10426 (make_pass_vrp): ...new function.
10427 (pass_uncprop): Replace declaration with that of...
10428 (make_pass_uncprop): ...new function.
10429 (pass_return_slot): Replace declaration with that of...
10430 (make_pass_return_slot): ...new function.
10431 (pass_reassoc): Replace declaration with that of...
10432 (make_pass_reassoc): ...new function.
10433 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10434 (make_pass_rebuild_cgraph_edges): ...new function.
10435 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10436 (make_pass_remove_cgraph_callee_edges): ...new function.
10437 (pass_build_cgraph_edges): Replace declaration with that of...
10438 (make_pass_build_cgraph_edges): ...new function.
10439 (pass_local_pure_const): Replace declaration with that of...
10440 (make_pass_local_pure_const): ...new function.
10441 (pass_tracer): Replace declaration with that of...
10442 (make_pass_tracer): ...new function.
10443 (pass_warn_unused_result): Replace declaration with that of...
10444 (make_pass_warn_unused_result): ...new function.
10445 (pass_diagnose_tm_blocks): Replace declaration with that of...
10446 (make_pass_diagnose_tm_blocks): ...new function.
10447 (pass_lower_tm): Replace declaration with that of...
10448 (make_pass_lower_tm): ...new function.
10449 (pass_tm_init): Replace declaration with that of...
10450 (make_pass_tm_init): ...new function.
10451 (pass_tm_mark): Replace declaration with that of...
10452 (make_pass_tm_mark): ...new function.
10453 (pass_tm_memopt): Replace declaration with that of...
10454 (make_pass_tm_memopt): ...new function.
10455 (pass_tm_edges): Replace declaration with that of...
10456 (make_pass_tm_edges): ...new function.
10457 (pass_split_functions): Replace declaration with that of...
10458 (make_pass_split_functions): ...new function.
10459 (pass_feedback_split_functions): Replace declaration with that of...
10460 (make_pass_feedback_split_functions): ...new function.
10461 (pass_strength_reduction): Replace declaration with that of...
10462 (make_pass_strength_reduction): ...new function.
10463 (pass_ipa_lower_emutls): Replace declaration with that of...
10464 (make_pass_ipa_lower_emutls): ...new function.
10465 (pass_ipa_function_and_variable_visibility): Replace declaration with
10466 that of...
10467 (make_pass_ipa_function_and_variable_visibility): ...new function.
10468 (pass_ipa_tree_profile): Replace declaration with that of...
10469 (make_pass_ipa_tree_profile): ...new function.
10470 (pass_early_local_passes): Replace declaration with that of...
10471 (make_pass_early_local_passes): ...new function.
10472 (pass_ipa_whole_program_visibility): Replace declaration with that
10473 of...
10474 (make_pass_ipa_whole_program_visibility): ...new function.
10475 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10476 (make_pass_ipa_lto_gimple_out): ...new function.
10477 (pass_ipa_increase_alignment): Replace declaration with that of...
10478 (make_pass_ipa_increase_alignment): ...new function.
10479 (pass_ipa_inline): Replace declaration with that of...
10480 (make_pass_ipa_inline): ...new function.
10481 (pass_ipa_free_lang_data): Replace declaration with that of...
10482 (make_pass_ipa_free_lang_data): ...new function.
10483 (pass_ipa_free_inline_summary): Replace declaration with that of...
10484 (make_pass_ipa_free_inline_summary): ...new function.
10485 (pass_ipa_cp): Replace declaration with that of...
10486 (make_pass_ipa_cp): ...new function.
10487 (pass_ipa_reference): Replace declaration with that of...
10488 (make_pass_ipa_reference): ...new function.
10489 (pass_ipa_pure_const): Replace declaration with that of...
10490 (make_pass_ipa_pure_const): ...new function.
10491 (pass_ipa_pta): Replace declaration with that of...
10492 (make_pass_ipa_pta): ...new function.
10493 (pass_ipa_lto_finish_out): Replace declaration with that of...
10494 (make_pass_ipa_lto_finish_out): ...new function.
10495 (pass_ipa_tm): Replace declaration with that of...
10496 (make_pass_ipa_tm): ...new function.
10497 (pass_ipa_profile): Replace declaration with that of...
10498 (make_pass_ipa_profile): ...new function.
10499 (pass_ipa_cdtor_merge): Replace declaration with that of...
10500 (make_pass_ipa_cdtor_merge): ...new function.
10501 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10502 of...
10503 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10504 (pass_init_datastructures): Replace declaration with that of...
10505 (make_pass_init_datastructures): ...new function.
10506 (pass_fixup_cfg): Replace declaration with that of...
10507 (make_pass_fixup_cfg): ...new function.
10508 (pass_expand): Replace declaration with that of...
10509 (make_pass_expand): ...new function.
10510 (pass_instantiate_virtual_regs): Replace declaration with that of...
10511 (make_pass_instantiate_virtual_regs): ...new function.
10512 (pass_rtl_fwprop): Replace declaration with that of...
10513 (make_pass_rtl_fwprop): ...new function.
10514 (pass_rtl_fwprop_addr): Replace declaration with that of...
10515 (make_pass_rtl_fwprop_addr): ...new function.
10516 (pass_jump): Replace declaration with that of...
10517 (make_pass_jump): ...new function.
10518 (pass_jump2): Replace declaration with that of...
10519 (make_pass_jump2): ...new function.
10520 (pass_lower_subreg): Replace declaration with that of...
10521 (make_pass_lower_subreg): ...new function.
10522 (pass_cse): Replace declaration with that of...
10523 (make_pass_cse): ...new function.
10524 (pass_fast_rtl_dce): Replace declaration with that of...
10525 (make_pass_fast_rtl_dce): ...new function.
10526 (pass_ud_rtl_dce): Replace declaration with that of...
10527 (make_pass_ud_rtl_dce): ...new function.
10528 (pass_rtl_dce): Replace declaration with that of...
10529 (make_pass_rtl_dce): ...new function.
10530 (pass_rtl_dse1): Replace declaration with that of...
10531 (make_pass_rtl_dse1): ...new function.
10532 (pass_rtl_dse2): Replace declaration with that of...
10533 (make_pass_rtl_dse2): ...new function.
10534 (pass_rtl_dse3): Replace declaration with that of...
10535 (make_pass_rtl_dse3): ...new function.
10536 (pass_rtl_cprop): Replace declaration with that of...
10537 (make_pass_rtl_cprop): ...new function.
10538 (pass_rtl_pre): Replace declaration with that of...
10539 (make_pass_rtl_pre): ...new function.
10540 (pass_rtl_hoist): Replace declaration with that of...
10541 (make_pass_rtl_hoist): ...new function.
10542 (pass_rtl_store_motion): Replace declaration with that of...
10543 (make_pass_rtl_store_motion): ...new function.
10544 (pass_cse_after_global_opts): Replace declaration with that of...
10545 (make_pass_cse_after_global_opts): ...new function.
10546 (pass_rtl_ifcvt): Replace declaration with that of...
10547 (make_pass_rtl_ifcvt): ...new function.
10548 (pass_into_cfg_layout_mode): Replace declaration with that of...
10549 (make_pass_into_cfg_layout_mode): ...new function.
10550 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10551 (make_pass_outof_cfg_layout_mode): ...new function.
10552 (pass_loop2): Replace declaration with that of...
10553 (make_pass_loop2): ...new function.
10554 (pass_rtl_loop_init): Replace declaration with that of...
10555 (make_pass_rtl_loop_init): ...new function.
10556 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10557 (make_pass_rtl_move_loop_invariants): ...new function.
10558 (pass_rtl_unswitch): Replace declaration with that of...
10559 (make_pass_rtl_unswitch): ...new function.
10560 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10561 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10562 (pass_rtl_doloop): Replace declaration with that of...
10563 (make_pass_rtl_doloop): ...new function.
10564 (pass_rtl_loop_done): Replace declaration with that of...
10565 (make_pass_rtl_loop_done): ...new function.
10566 (pass_web): Replace declaration with that of...
10567 (make_pass_web): ...new function.
10568 (pass_cse2): Replace declaration with that of...
10569 (make_pass_cse2): ...new function.
10570 (pass_df_initialize_opt): Replace declaration with that of...
10571 (make_pass_df_initialize_opt): ...new function.
10572 (pass_df_initialize_no_opt): Replace declaration with that of...
10573 (make_pass_df_initialize_no_opt): ...new function.
10574 (pass_reginfo_init): Replace declaration with that of...
10575 (make_pass_reginfo_init): ...new function.
10576 (pass_inc_dec): Replace declaration with that of...
10577 (make_pass_inc_dec): ...new function.
10578 (pass_stack_ptr_mod): Replace declaration with that of...
10579 (make_pass_stack_ptr_mod): ...new function.
10580 (pass_initialize_regs): Replace declaration with that of...
10581 (make_pass_initialize_regs): ...new function.
10582 (pass_combine): Replace declaration with that of...
10583 (make_pass_combine): ...new function.
10584 (pass_if_after_combine): Replace declaration with that of...
10585 (make_pass_if_after_combine): ...new function.
10586 (pass_ree): Replace declaration with that of...
10587 (make_pass_ree): ...new function.
10588 (pass_partition_blocks): Replace declaration with that of...
10589 (make_pass_partition_blocks): ...new function.
10590 (pass_match_asm_constraints): Replace declaration with that of...
10591 (make_pass_match_asm_constraints): ...new function.
10592 (pass_regmove): Replace declaration with that of...
10593 (make_pass_regmove): ...new function.
10594 (pass_split_all_insns): Replace declaration with that of...
10595 (make_pass_split_all_insns): ...new function.
10596 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10597 (make_pass_fast_rtl_byte_dce): ...new function.
10598 (pass_lower_subreg2): Replace declaration with that of...
10599 (make_pass_lower_subreg2): ...new function.
10600 (pass_mode_switching): Replace declaration with that of...
10601 (make_pass_mode_switching): ...new function.
10602 (pass_sms): Replace declaration with that of...
10603 (make_pass_sms): ...new function.
10604 (pass_sched): Replace declaration with that of...
10605 (make_pass_sched): ...new function.
10606 (pass_ira): Replace declaration with that of...
10607 (make_pass_ira): ...new function.
10608 (pass_reload): Replace declaration with that of...
10609 (make_pass_reload): ...new function.
10610 (pass_clean_state): Replace declaration with that of...
10611 (make_pass_clean_state): ...new function.
10612 (pass_branch_prob): Replace declaration with that of...
10613 (make_pass_branch_prob): ...new function.
10614 (pass_value_profile_transformations): Replace declaration with that
10615 of...
10616 (make_pass_value_profile_transformations): ...new function.
10617 (pass_postreload_cse): Replace declaration with that of...
10618 (make_pass_postreload_cse): ...new function.
10619 (pass_gcse2): Replace declaration with that of...
10620 (make_pass_gcse2): ...new function.
10621 (pass_split_after_reload): Replace declaration with that of...
10622 (make_pass_split_after_reload): ...new function.
10623 (pass_branch_target_load_optimize1): Replace declaration with that
10624 of...
10625 (make_pass_branch_target_load_optimize1): ...new function.
10626 (pass_thread_prologue_and_epilogue): Replace declaration with that
10627 of...
10628 (make_pass_thread_prologue_and_epilogue): ...new function.
10629 (pass_stack_adjustments): Replace declaration with that of...
10630 (make_pass_stack_adjustments): ...new function.
10631 (pass_peephole2): Replace declaration with that of...
10632 (make_pass_peephole2): ...new function.
10633 (pass_if_after_reload): Replace declaration with that of...
10634 (make_pass_if_after_reload): ...new function.
10635 (pass_regrename): Replace declaration with that of...
10636 (make_pass_regrename): ...new function.
10637 (pass_cprop_hardreg): Replace declaration with that of...
10638 (make_pass_cprop_hardreg): ...new function.
10639 (pass_reorder_blocks): Replace declaration with that of...
10640 (make_pass_reorder_blocks): ...new function.
10641 (pass_branch_target_load_optimize2): Replace declaration with that
10642 of...
10643 (make_pass_branch_target_load_optimize2): ...new function.
10644 (pass_leaf_regs): Replace declaration with that of...
10645 (make_pass_leaf_regs): ...new function.
10646 (pass_split_before_sched2): Replace declaration with that of...
10647 (make_pass_split_before_sched2): ...new function.
10648 (pass_compare_elim_after_reload): Replace declaration with that of...
10649 (make_pass_compare_elim_after_reload): ...new function.
10650 (pass_sched2): Replace declaration with that of...
10651 (make_pass_sched2): ...new function.
10652 (pass_stack_regs): Replace declaration with that of...
10653 (make_pass_stack_regs): ...new function.
10654 (pass_stack_regs_run): Replace declaration with that of...
10655 (make_pass_stack_regs_run): ...new function.
10656 (pass_df_finish): Replace declaration with that of...
10657 (make_pass_df_finish): ...new function.
10658 (pass_compute_alignments): Replace declaration with that of...
10659 (make_pass_compute_alignments): ...new function.
10660 (pass_duplicate_computed_gotos): Replace declaration with that of...
10661 (make_pass_duplicate_computed_gotos): ...new function.
10662 (pass_variable_tracking): Replace declaration with that of...
10663 (make_pass_variable_tracking): ...new function.
10664 (pass_free_cfg): Replace declaration with that of...
10665 (make_pass_free_cfg): ...new function.
10666 (pass_machine_reorg): Replace declaration with that of...
10667 (make_pass_machine_reorg): ...new function.
10668 (pass_cleanup_barriers): Replace declaration with that of...
10669 (make_pass_cleanup_barriers): ...new function.
10670 (pass_delay_slots): Replace declaration with that of...
10671 (make_pass_delay_slots): ...new function.
10672 (pass_split_for_shorten_branches): Replace declaration with that of...
10673 (make_pass_split_for_shorten_branches): ...new function.
10674 (pass_split_before_regstack): Replace declaration with that of...
10675 (make_pass_split_before_regstack): ...new function.
10676 (pass_convert_to_eh_region_ranges): Replace declaration with that
10677 of...
10678 (make_pass_convert_to_eh_region_ranges): ...new function.
10679 (pass_shorten_branches): Replace declaration with that of...
10680 (make_pass_shorten_branches): ...new function.
10681 (pass_set_nothrow_function_flags): Replace declaration with that of...
10682 (make_pass_set_nothrow_function_flags): ...new function.
10683 (pass_dwarf2_frame): Replace declaration with that of...
10684 (make_pass_dwarf2_frame): ...new function.
10685 (pass_final): Replace declaration with that of...
10686 (make_pass_final): ...new function.
10687 (pass_rtl_seqabstr): Replace declaration with that of...
10688 (make_pass_rtl_seqabstr): ...new function.
10689 (pass_release_ssa_names): Replace declaration with that of...
10690 (make_pass_release_ssa_names): ...new function.
10691 (pass_early_inline): Replace declaration with that of...
10692 (make_pass_early_inline): ...new function.
10693 (pass_inline_parameters): Replace declaration with that of...
10694 (make_pass_inline_parameters): ...new function.
10695 (pass_update_address_taken): Replace declaration with that of...
10696 (make_pass_update_address_taken): ...new function.
10697 (pass_convert_switch): Replace declaration with that of...
10698 (make_pass_convert_switch): ...new function.
10699 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10700 to a subclass of simple_ipa_opt_pass along with...
10701 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10702 (make_pass_ipa_tree_profile): ...new function.
10703 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10704 subclass of gimple_opt_pass along with...
10705 (pass_data_sra_early): ...new pass_data instance and...
10706 (make_pass_sra_early): ...new function.
10707 (pass_sra): Convert from a global struct to a subclass of
10708 gimple_opt_pass along with...
10709 (pass_data_sra): ...new pass_data instance and...
10710 (make_pass_sra): ...new function.
10711 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10712 gimple_opt_pass along with...
10713 (pass_data_early_ipa_sra): ...new pass_data instance and...
10714 (make_pass_early_ipa_sra): ...new function.
10715 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10716 subclass of gimple_opt_pass along with...
10717 (pass_data_ccp): ...new pass_data instance and...
10718 (make_pass_ccp): ...new function.
10719 (pass_fold_builtins): Convert from a global struct to a subclass of
10720 gimple_opt_pass along with...
10721 (pass_data_fold_builtins): ...new pass_data instance and...
10722 (make_pass_fold_builtins): ...new function.
10723 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10724 subclass of gimple_opt_pass along with...
10725 (pass_data_copy_prop): ...new pass_data instance and...
10726 (make_pass_copy_prop): ...new function.
10727 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10728 global struct to a subclass of gimple_opt_pass along with...
10729 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10730 (make_pass_rename_ssa_copies): ...new function.
10731 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10732 subclass of gimple_opt_pass along with...
10733 (pass_data_dce): ...new pass_data instance and...
10734 (make_pass_dce): ...new function.
10735 (pass_dce_loop): Convert from a global struct to a subclass of
10736 gimple_opt_pass along with...
10737 (pass_data_dce_loop): ...new pass_data instance and...
10738 (make_pass_dce_loop): ...new function.
10739 (pass_cd_dce): Convert from a global struct to a subclass of
10740 gimple_opt_pass along with...
10741 (pass_data_cd_dce): ...new pass_data instance and...
10742 (make_pass_cd_dce): ...new function.
10743 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10744 subclass of gimple_opt_pass along with...
10745 (pass_data_dominator): ...new pass_data instance and...
10746 (make_pass_dominator): ...new function.
10747 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10748 gimple_opt_pass along with...
10749 (pass_data_phi_only_cprop): ...new pass_data instance and...
10750 (make_pass_phi_only_cprop): ...new function.
10751 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10752 subclass of gimple_opt_pass along with...
10753 (pass_data_dse): ...new pass_data instance and...
10754 (make_pass_dse): ...new function.
10755 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10756 a subclass of gimple_opt_pass along with...
10757 (pass_data_forwprop): ...new pass_data instance and...
10758 (make_pass_forwprop): ...new function.
10759 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10760 struct to a subclass of gimple_opt_pass along with...
10761 (pass_data_tree_ifcombine): ...new pass_data instance and...
10762 (make_pass_tree_ifcombine): ...new function.
10763 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10764 subclass of gimple_opt_pass along with...
10765 (pass_data_ch): ...new pass_data instance and...
10766 (make_pass_ch): ...new function.
10767 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10768 subclass of gimple_opt_pass along with...
10769 (pass_data_tree_loop): ...new pass_data instance and...
10770 (make_pass_tree_loop): ...new function.
10771 (pass_tree_loop_init): Convert from a global struct to a subclass of
10772 gimple_opt_pass along with...
10773 (pass_data_tree_loop_init): ...new pass_data instance and...
10774 (make_pass_tree_loop_init): ...new function.
10775 (pass_lim): Convert from a global struct to a subclass of
10776 gimple_opt_pass along with...
10777 (pass_data_lim): ...new pass_data instance and...
10778 (make_pass_lim): ...new function.
10779 (pass_tree_unswitch): Convert from a global struct to a subclass of
10780 gimple_opt_pass along with...
10781 (pass_data_tree_unswitch): ...new pass_data instance and...
10782 (make_pass_tree_unswitch): ...new function.
10783 (pass_predcom): Convert from a global struct to a subclass of
10784 gimple_opt_pass along with...
10785 (pass_data_predcom): ...new pass_data instance and...
10786 (make_pass_predcom): ...new function.
10787 (pass_vectorize): Convert from a global struct to a subclass of
10788 gimple_opt_pass along with...
10789 (pass_data_vectorize): ...new pass_data instance and...
10790 (make_pass_vectorize): ...new function.
10791 (pass_graphite): Convert from a global struct to a subclass of
10792 gimple_opt_pass along with...
10793 (pass_data_graphite): ...new pass_data instance and...
10794 (make_pass_graphite): ...new function.
10795 (pass_graphite_transforms): Convert from a global struct to a subclass
10796 of gimple_opt_pass along with...
10797 (pass_data_graphite_transforms): ...new pass_data instance and...
10798 (make_pass_graphite_transforms): ...new function.
10799 (pass_check_data_deps): Convert from a global struct to a subclass of
10800 gimple_opt_pass along with...
10801 (pass_data_check_data_deps): ...new pass_data instance and...
10802 (make_pass_check_data_deps): ...new function.
10803 (pass_iv_canon): Convert from a global struct to a subclass of
10804 gimple_opt_pass along with...
10805 (pass_data_iv_canon): ...new pass_data instance and...
10806 (make_pass_iv_canon): ...new function.
10807 (pass_scev_cprop): Convert from a global struct to a subclass of
10808 gimple_opt_pass along with...
10809 (pass_data_scev_cprop): ...new pass_data instance and...
10810 (make_pass_scev_cprop): ...new function.
10811 (pass_record_bounds): Convert from a global struct to a subclass of
10812 gimple_opt_pass along with...
10813 (pass_data_record_bounds): ...new pass_data instance and...
10814 (make_pass_record_bounds): ...new function.
10815 (pass_complete_unroll): Convert from a global struct to a subclass of
10816 gimple_opt_pass along with...
10817 (pass_data_complete_unroll): ...new pass_data instance and...
10818 (make_pass_complete_unroll): ...new function.
10819 (pass_complete_unrolli): Convert from a global struct to a subclass of
10820 gimple_opt_pass along with...
10821 (pass_data_complete_unrolli): ...new pass_data instance and...
10822 (make_pass_complete_unrolli): ...new function.
10823 (pass_parallelize_loops): Convert from a global struct to a subclass
10824 of gimple_opt_pass along with...
10825 (pass_data_parallelize_loops): ...new pass_data instance and...
10826 (make_pass_parallelize_loops): ...new function.
10827 (pass_loop_prefetch): Convert from a global struct to a subclass of
10828 gimple_opt_pass along with...
10829 (pass_data_loop_prefetch): ...new pass_data instance and...
10830 (make_pass_loop_prefetch): ...new function.
10831 (pass_iv_optimize): Convert from a global struct to a subclass of
10832 gimple_opt_pass along with...
10833 (pass_data_iv_optimize): ...new pass_data instance and...
10834 (make_pass_iv_optimize): ...new function.
10835 (pass_tree_loop_done): Convert from a global struct to a subclass of
10836 gimple_opt_pass along with...
10837 (pass_data_tree_loop_done): ...new pass_data instance and...
10838 (make_pass_tree_loop_done): ...new function.
10839 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10840 struct to a subclass of gimple_opt_pass along with...
10841 (pass_data_cse_reciprocals): ...new pass_data instance and...
10842 (make_pass_cse_reciprocals): ...new function.
10843 (pass_cse_sincos): Convert from a global struct to a subclass of
10844 gimple_opt_pass along with...
10845 (pass_data_cse_sincos): ...new pass_data instance and...
10846 (make_pass_cse_sincos): ...new function.
10847 (pass_optimize_bswap): Convert from a global struct to a subclass of
10848 gimple_opt_pass along with...
10849 (pass_data_optimize_bswap): ...new pass_data instance and...
10850 (make_pass_optimize_bswap): ...new function.
10851 (pass_optimize_widening_mul): Convert from a global struct to a
10852 subclass of gimple_opt_pass along with...
10853 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10854 (make_pass_optimize_widening_mul): ...new function.
10855 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10856 subclass of gimple_opt_pass along with...
10857 (pass_data_phiopt): ...new pass_data instance and...
10858 (make_pass_phiopt): ...new function.
10859 (pass_cselim): Convert from a global struct to a subclass of
10860 gimple_opt_pass along with...
10861 (pass_data_cselim): ...new pass_data instance and...
10862 (make_pass_cselim): ...new function.
10863 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10864 subclass of gimple_opt_pass along with...
10865 (pass_data_phiprop): ...new pass_data instance and...
10866 (make_pass_phiprop): ...new function.
10867 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10868 subclass of gimple_opt_pass along with...
10869 (pass_data_pre): ...new pass_data instance and...
10870 (make_pass_pre): ...new function.
10871 (pass_fre): Convert from a global struct to a subclass of
10872 gimple_opt_pass along with...
10873 (pass_data_fre): ...new pass_data instance and...
10874 (make_pass_fre): ...new function.
10875 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10876 subclass of gimple_opt_pass along with...
10877 (pass_data_reassoc): ...new pass_data instance and...
10878 (make_pass_reassoc): ...new function.
10879 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10880 subclass of gimple_opt_pass along with...
10881 (pass_data_sink_code): ...new pass_data instance and...
10882 (make_pass_sink_code): ...new function.
10883 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10884 subclass of gimple_opt_pass along with...
10885 (pass_data_strlen): ...new pass_data instance and...
10886 (make_pass_strlen): ...new function.
10887 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10888 struct to a subclass of gimple_opt_pass along with...
10889 (pass_data_build_alias): ...new pass_data instance and...
10890 (make_pass_build_alias): ...new function.
10891 (pass_build_ealias): Convert from a global struct to a subclass of
10892 gimple_opt_pass along with...
10893 (pass_data_build_ealias): ...new pass_data instance and...
10894 (make_pass_build_ealias): ...new function.
10895 (pass_ipa_pta): Convert from a global struct to a subclass of
10896 simple_ipa_opt_pass along with...
10897 (pass_data_ipa_pta): ...new pass_data instance and...
10898 (make_pass_ipa_pta): ...new function.
10899 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10900 subclass of gimple_opt_pass along with...
10901 (pass_data_uncprop): ...new pass_data instance and...
10902 (make_pass_uncprop): ...new function.
10903 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10904 global struct to a subclass of gimple_opt_pass along with...
10905 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10906 (make_pass_late_warn_uninitialized): ...new function.
10907 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10908 to a subclass of gimple_opt_pass along with...
10909 (pass_data_init_datastructures): ...new pass_data instance and...
10910 (make_pass_init_datastructures): ...new function.
10911 (pass_early_warn_uninitialized): Convert from a global struct to a
10912 subclass of gimple_opt_pass along with...
10913 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10914 (make_pass_early_warn_uninitialized): ...new function.
10915 (pass_update_address_taken): Convert from a global struct to a
10916 subclass of gimple_opt_pass along with...
10917 (pass_data_update_address_taken): ...new pass_data instance and...
10918 (make_pass_update_address_taken): ...new function.
10919 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10920 struct to a subclass of gimple_opt_pass along with...
10921 (pass_data_release_ssa_names): ...new pass_data instance and...
10922 (make_pass_release_ssa_names): ...new function.
10923 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10924 subclass of gimple_opt_pass along with...
10925 (pass_data_stdarg): ...new pass_data instance and...
10926 (make_pass_stdarg): ...new function.
10927 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10928 global struct to a subclass of gimple_opt_pass along with...
10929 (pass_data_convert_switch): ...new pass_data instance and...
10930 (make_pass_convert_switch): ...new function.
10931 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10932 to a subclass of gimple_opt_pass along with...
10933 (pass_data_tail_recursion): ...new pass_data instance and...
10934 (make_pass_tail_recursion): ...new function.
10935 (pass_tail_calls): Convert from a global struct to a subclass of
10936 gimple_opt_pass along with...
10937 (pass_data_tail_calls): ...new pass_data instance and...
10938 (make_pass_tail_calls): ...new function.
10939 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10940 struct to a subclass of gimple_opt_pass along with...
10941 (pass_data_lower_vector): ...new pass_data instance and...
10942 (make_pass_lower_vector): ...new function.
10943 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10944 gimple_opt_pass along with...
10945 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10946 (make_pass_lower_vector_ssa): ...new function.
10947 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10948 to a subclass of gimple_opt_pass along with...
10949 (pass_data_slp_vectorize): ...new pass_data instance and...
10950 (make_pass_slp_vectorize): ...new function.
10951 (pass_ipa_increase_alignment): Convert from a global struct to a
10952 subclass of simple_ipa_opt_pass along with...
10953 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10954 (make_pass_ipa_increase_alignment): ...new function.
10955 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10956 gimple_opt_pass along with...
10957 (pass_data_vrp): ...new pass_data instance and...
10958 (make_pass_vrp): ...new function.
10959 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10960 subclass of simple_ipa_opt_pass along with...
10961 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10962 (make_pass_ipa_free_lang_data): ...new function.
10963 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10964 gimple_opt_pass along with...
10965 (pass_data_tsan): ...new pass_data instance and...
10966 (make_pass_tsan): ...new function.
10967 (pass_tsan_O0): Convert from a global struct to a subclass of
10968 gimple_opt_pass along with...
10969 (pass_data_tsan_O0): ...new pass_data instance and...
10970 (make_pass_tsan_O0): ...new function.
10971 * var-tracking.c (pass_variable_tracking): Convert from a global
10972 struct to a subclass of rtl_opt_pass along with...
10973 (pass_data_variable_tracking): ...new pass_data instance and...
10974 (make_pass_variable_tracking): ...new function.
10975 * web.c (pass_web): Convert from a global struct to a subclass of
10976 rtl_opt_pass along with...
10977 (pass_data_web): ...new pass_data instance and...
10978 (make_pass_web): ...new function.
10979 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10980 declaration with that of...
10981 (make_pass_mode_switch_use): ...new function.
10982 (pass_resolve_sw_modes): Replace declaration with that of...
10983 (make_pass_resolve_sw_modes): ...new function.
10984 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10985 from a global struct to a subclass of rtl_opt_pass along with...
10986 (pass_data_mode_switch_use): ...new pass_data instance and...
10987 (make_pass_mode_switch_use): ...new function.
10988 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10989 from a global struct to a subclass of rtl_opt_pass along with...
10990 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10991 (make_pass_resolve_sw_modes): ...new function.
10992 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10993 struct to a subclass of rtl_opt_pass along with...
10994 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10995 (make_pass_insert_vzeroupper): ...new function.
10996 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10997 global struct to a subclass of rtl_opt_pass along with...
10998 (pass_data_work_around_errata): ...new pass_data instance and...
10999 (make_pass_work_around_errata): ...new function.
11000 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11001 struct to a subclass of rtl_opt_pass along with...
11002 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11003 (make_pass_mips_machine_reorg2): ...new function.
11004
11005 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11006
11007 * passes.c (pass_manager::operator new): New.
11008
11009 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11010
11011 Handwritten part of conversion of passes to C++ classes.
11012
11013 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11014 (toplev.o): Add dep on PASS_MANAGER_H.
11015 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11016 of early local pases to reflect this moving from a global to a
11017 member of gcc::pass_manager.
11018 (cgraph_add_new_function): Likewise.
11019 * lto-cgraph.c (lto_output_node): Update for conversion of
11020 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11021 * passes.c (opt_pass::clone): New.
11022 (opt_pass::gate): New.
11023 (opt_pass::execute): New.
11024 (opt_pass::opt_pass): New.
11025 (pass_manager::execute_early_local_passes): New.
11026 (pass_manager::execute_pass_mode_switching): new.
11027 (finish_optimization_passes): Convert to...
11028 (pass_manager::finish_optimization_passes): ...this.
11029 (finish_optimization_passes): Update for conversion of passes to
11030 C++ classes.
11031 (register_dump_files_1): Use has_gate since we cannot portably
11032 check a vtable entry against NULL.
11033 (dump_one_pass): Likewise.
11034 (ipa_write_summaries_2): Likewise.
11035 (ipa_write_optimization_summaries_1): Likewise.
11036 (ipa_read_summaries_1): Likewise.
11037 (ipa_read_optimization_summaries_1): Likewise.
11038 (execute_ipa_stmt_fixups): Likewise.
11039 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11040 pass-creation functions rather than wiring up globals, and
11041 storing the results in fields of pass_manager generated using
11042 pass-instances.def.
11043 (pass_manager::dump_profile_report): Update for conversion of
11044 passes to C++ classes.
11045 (pass_manager::execute_ipa_summary_passes): Likewise.
11046 (execute_one_ipa_transform_pass): Likewise.
11047 (execute_one_pass): Use has_gate and has_execute since we cannot
11048 portably check a vtable entry against NULL.
11049 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11050 (pass_manager): Use pass-instances.def to add fields for the
11051 various pass instances.
11052 * toplev.c (finalize): Update for move of
11053 finish_optimization_passes to a method of gcc::pass_manager.
11054 * toplev.h (finish_optimization_passes): Move to method of class
11055 pass_manager.
11056 * tree-pass.h (struct pass_data): New.
11057 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11058 (opt_pass::gate): Convert to virtual function.
11059 (opt_pass::~opt_pass): New.
11060 (opt_pass::clone): New.
11061 (opt_pass::execute): Convert to virtual function.
11062 (opt_pass::opt_pass): New.
11063 (opt_pass::ctxt_): new.
11064 (gimple_opt_pass): Convert to subclass of opt_pass.
11065 (gimple_opt_pass::gimple_opt_pass): New.
11066 (rtl_opt_pass): Convert to subclass of opt_pass.
11067 (rtl_opt_pass::rtl_opt_pass): New.
11068 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11069 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11070 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11071 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11072 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11073 invocation of pass_mode_switching to reflect this moving from a
11074 global to a member of gcc::pass_manager.
11075 (ix86_option_override): Rework how pass_insert_vzeroupper is
11076 added to the pass_manager to reflect autogenerated changes.
11077 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11078
11079 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11080
11081 PR rtl-optimization/57708
11082 * recog.c (peep2_find_free_register): Validate all regs in a
11083 multi-reg mode.
11084
11085 2013-08-05 Jan Hubicka <jh@suse.cz>
11086
11087 PR lto/57602
11088 * cgraph.c (verify_cgraph_node): Accept local flags from other
11089 partitions.
11090 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11091 (function_and_variable_visibility): Likewise.
11092 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11093
11094 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11095
11096 * graph.c (init_graph_slim_pretty_print): Remove.
11097 (print_graph_cfg): Do not call it. Use local pretty printer.
11098 (start_graph_dump): Likewise.
11099
11100 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11101
11102 * gimple-pretty-print.c (buffer): Remove.
11103 (initialized): Likewise.
11104 (maybe_init_pretty_print): Likewise.
11105 (print_gimple_stmt): Do not call it. Use non-static local
11106 pretty_printer variable.
11107 (print_gimple_expr): Likewise.
11108 (print_gimple_seq): Likewise.
11109 (gimple_dump_bb): Likewise.
11110
11111 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11112
11113 * asan.c (asan_pp): Remove.
11114 (asan_pp_initialized): Likewise.
11115 (asan_pp_initialize): Likewise.
11116 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11117 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11118 (asan_add_global): Likewise.
11119
11120 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11121
11122 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11123 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11124 * pretty-print.c (pp_formatted_text_data): Likewise.
11125 (pp_indent): Rename from pp_base_indent.
11126 (pp_format): Rename from pp_base_format.
11127 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11128 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11129 (pp_flush): Rename from pp_base_flush.
11130 (pp_set_line_maximum_length): Rename from
11131 pp_base_set_line_maximum_length.
11132 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11133 (pp_set_prefix): Rename from pp_base_set_prefix.
11134 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11135 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11136 (pp_append_text): Rename from pp_base_append_text.
11137 (pp_formatted_text): Rename from pp_base_formatted_text.
11138 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11139 (pp_remaining_character_count_for_line): Rename from
11140 pp_base_remaining_character_count_for_line.
11141 (pp_newline): Rename from pp_base_newline.
11142 (pp_character): Rename from pp_base_character.
11143 (pp_string): Rename from pp_base_string.
11144 (pp_maybe_space): Rename from pp_base_maybe_space.
11145 * asan.c (asan_pp_string): Adjust.
11146 (asan_emit_stack_protection): Likewise.
11147 (asan_add_global): Likewise.
11148 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11149 * tree-mudflap.c (mf_varname_tree): Likewise.
11150 * tree-pretty-print.c (pp_tree_identifier): Rename from
11151 pp_base_tree_identifier.
11152 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11153 Declare as function.
11154
11155 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11156
11157 * pretty-print.h (pp_bar_bar): New.
11158 (pp_ampersand_ampersand): Likewise.
11159 (pp_less_equal): Likewise.
11160 (pp_greater_equal): Likewise.
11161 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11162 printer functions instead of pp_string or operators and punctuators.
11163 (dump_gimple_call): Likewise.
11164 (dump_gimple_omp_for): Likewise.
11165 (dump_gimple_transaction): Likewise.
11166 (dump_gimple_phi): Likewise.
11167 (pp_gimple_stmt_1): Likewise.
11168 * sched-vis.c (print_insn): Likewise.
11169 * tree-mudflap.c (mf_varname_tree): Likewise.
11170 * tree-pretty-print.c (dump_block_node): Likewise.
11171 (dump_generic_node): Likewise.
11172
11173 2013-08-02 Jan Hubicka <jh@suse.cz>
11174
11175 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11176 boundaries.
11177 * lto-streamer-out.c (tree_is_indexable): Results decls and
11178 parm decls are not indexable.
11179 (DFS_write_tree_body): Do not follow args and results.
11180 (hash_tree): Likewise.
11181 (output_functions): Rearrange so struct function is needed
11182 only when real body is output; be able to also ouptut abstract
11183 functions; output DECL_ARGUMENTS and DECL_RESULT.
11184 (lto_output): When not in WPA, ale store abstract functions.
11185 (write_symbol): Do not care about RESULT_DECL.
11186 (output_symbol_p): Handle correctly sbtract decls.
11187 * lto-streamer-in.c (input_function): Rearrange so struct
11188 function can be NULL at entry; allow streaming of
11189 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11190 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11191 sanity check during LTO.
11192 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11193 RESULT_DECl and DECL_ARGUMENTS.
11194 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11195 Likewise.
11196
11197 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11198
11199 * pretty-print.h (pp_underscore): New.
11200 (pp_comma): Tidy.
11201 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11202 printer functions instead of pp_character.
11203 (dump_binary_rhs): Likewise.
11204 (dump_ternary_rhs): Likewise.
11205 (dump_gimple_call_args): Likewise.
11206 (pp_points_to_solution): Likewise.
11207 (dump_gimple_call): Likewise.
11208 (dump_gimple_switch): Likewise.
11209 (dump_gimple_cond): Likewise.
11210 (dump_gimple_bind): Likewise.
11211 (dump_gimple_try): Likewise.
11212 (dump_gimple_omp_for): Likewise.
11213 (dump_gimple_omp_continue): Likewise.
11214 (dump_gimple_omp_single): Likewise.
11215 (dump_gimple_omp_sections): Likewise.
11216 (dump_gimple_omp_block): Likewise.
11217 (dump_gimple_omp_critical): Likewise.
11218 (dump_gimple_transaction): Likewise.
11219 (dump_gimple_asm): Likewise.
11220 (dump_gimple_phi): Likewise.
11221 (dump_gimple_omp_parallel): Likewise.
11222 (dump_gimple_omp_task): Likewise.
11223 (dump_gimple_omp_atomic_load): Likewise.
11224 (dump_gimple_omp_atomic_store): Likewise.
11225 (dump_gimple_mem_ops): Likewise.
11226 (pp_gimple_stmt_1): Likewise.
11227 (pp_cfg_jump): Likewise.
11228 (dump_implicit_edges): Likewise.
11229 (gimple_dump_bb_for_graph): Likewise.
11230 * graph.c (draw_cfg_node): Likewise.
11231 * langhooks.c (lhd_print_error_function): Likewise.
11232 * sched-vis.c (print_exp): Likewise.
11233 (print_value): Likewise.
11234 (print_pattern): Likewise.
11235 (print_insn): Likewise.
11236 (rtl_dump_bb_for_graph): Likewise.
11237 * tree-pretty-print.c (dump_function_declaration): Likewise.
11238 (dump_array_domain): Likewise.
11239 (dump_omp_clause): Likewise.
11240 (dump_location): Likewise.
11241 (dump_generic_node): Likewise.
11242 (print_struct_decl): Likewise.
11243 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11244
11245 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11246
11247 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11248 candidate table when replacing a candidate statement.
11249 (replace_rhs_if_not_dup): Likewise.
11250 (replace_one_candidate): Likewise.
11251
11252 2013-08-02 Jan Hubicka <jh@suse.cz>
11253 Martin Liska <marxin.liska@gmail.com>
11254
11255 * cgraphunit.c (add_new_function): Fix logic when adding from
11256 late IPA pass.
11257 (assemble_thunk): Rename to ...
11258 (expand_thunk); .. this one; export; get it working with
11259 general functions; make produced gimple valid.
11260 * cgraph.h (expand_thunk): Declare.
11261
11262 2013-08-02 Jan Hubicka <jh@suse.cz>
11263
11264 * ipa-cp.c (gather_context_independent_values): Use
11265 ipa_get_param_move_cost.
11266 (get_replacement_map): Remove PARAM; move parameter folding
11267 into tree-inline.c
11268 (create_specialized_node): Update.
11269 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11270 assert that we have gimple body; update move_cost.
11271 (count_formal_params): Assert that we have gimple body.
11272 (ipa_dump_param): New function.
11273 (ipa_alloc_node_params): Break out from ...
11274 (ipa_initialize_node_params): ... here.
11275 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11276 (ipa_write_node_info): Stream move costs.
11277 (ipa_read_node_info): Read move costs.
11278 (ipa_update_after_lto_read): Do not recompute node params.
11279 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11280 (ipa_get_param): Check we are not in WPA.
11281 (ipa_get_param_move_cost): New.
11282 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11283 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11284 parm numbers to be present.
11285
11286 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11287
11288 PR rtl-optimization/58048
11289 * lra-constraints.c (process_alt_operands): Don't check asm
11290 operand on register.
11291
11292 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11293
11294 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11295 the implied StoreLoad barrier for atomic operations if before.
11296
11297 2013-08-02 Jan Hubicka <jh@suse.cz>
11298 Martin Liska <marxin.liska@gmail.com>
11299
11300 * cgraph.c (cgraph_function_body_availability): Do not check
11301 cgraph flags.
11302 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11303 symtab_node_availability): Declare.
11304 * ipa.c (can_replace_by_local_alias): New.
11305 (function_and_variable_visibility): Use it.
11306 * symtab.c (symtab_for_node_and_aliases,
11307 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11308
11309 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11310
11311 PR rtl-optimization/57963
11312 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11313 (lra_constraints): Use them.
11314
11315 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11316
11317 * config/arm/types.md (define_attr "type"): Add "load_acq"
11318 and "store_rel".
11319 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11320 changes.
11321 (cortex_a53_store1): Likewise.
11322
11323 2013-08-01 Jan Hubicka <jh@suse.cz>
11324
11325 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11326 partitions are not needed.
11327
11328 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11329
11330 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11331 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11332 MAYBE_NON_Q_CLASS_P where appropriate.
11333
11334 2013-08-01 Jan Hubicka <jh@suse.cz>
11335
11336 * cgraph.h (release_function_body): Declare.
11337 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11338 of unused delcarations.
11339
11340 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11341
11342 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11343 RTL form when subtracting a constant.
11344
11345 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11346
11347 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11348 Generate canonical plus rtx with negated immediate instead of minus
11349 where appropriate.
11350 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11351
11352 2013-08-01 Jan Hubicka <jh@suse.cz>
11353
11354 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11355 (cgraph_release_function_body): Likewise.
11356 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11357 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11358 to used_as_abstract_origin.
11359 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11360 symbols used as abstract origins.
11361 * cgraphunit.c (analyze_functions): Update.
11362 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11363 used_as_abstract_origin.
11364 * tree-inline.c (tree_function_versioning): Update
11365 used_as_abstract_origin; be ready for DECL_RESULT and
11366 DECL_ARGUMENTS to be NULL.
11367
11368 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11369 for abstract functions.
11370 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11371 real symbols.
11372
11373 2013-08-01 Jan Hubicka <jh@suse.cz>
11374
11375 * profile.c (compute_value_histograms): Fix thinko.
11376
11377 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11378
11379 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11380 aarch-common-protos.h to extra_headers.
11381 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11382 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11383 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11384
11385 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11386
11387 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11388 Include "../arm/types.md". Define "type" attribute for all patterns.
11389 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11390 attribute changes.
11391
11392 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11393
11394 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11395 to support power8 load fusion.
11396 (fusion_gpr_mem_load): Likewise.
11397
11398 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11399
11400 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11401 declarations for power8 load fusion.
11402 (emit_fusion_gpr_load): Likewise.
11403
11404 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11405 tuning for power8, turn on fusion mode by default. Turn on sign
11406 extending fusion mode if normal fusion mode is on, and we are at
11407 -O2 or -O3.
11408 (fusion_gpr_load_p): New function, return true if we can fuse an
11409 addis instruction with a dependent load to a GPR.
11410 (emit_fusion_gpr_load): Emit the instructions for power8 load
11411 fusion to GPRs.
11412
11413 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11414 (VSX load fusion peepholes): New peepholes to fuse together an
11415 addi instruction with a VSX load instruction.
11416
11417 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11418 peepholes to fuse an addis instruction with a load to a GPR base
11419 register. If we are supporting sign extending fusions, convert
11420 sign extending loads to zero extending loads and add an explicit
11421 sign extension.
11422
11423 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11424
11425 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11426 aarch-common-protos.h to extra_headers.
11427 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11428 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11429 (arm_early_store_addr_dep): Likewise.
11430 (arm_no_early_alu_shift_dep): Likewise.
11431 (arm_no_early_alu_shift_value_dep): Likewise.
11432 (arm_no_early_mul_dep): Likewise.
11433 (arm_no_early_store_addr_dep): Likewise.
11434 (arm_mac_accumulator_is_mul_result): Likewise.
11435 (arm_mac_accumulator_is_result): Likewise.
11436 * config/arm/aarch-common.c: ... here. New file.
11437 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11438 here to ...
11439 (arm_early_store_addr_dep): Likewise.
11440 (arm_no_early_alu_shift_dep): Likewise.
11441 (arm_no_early_alu_shift_value_dep): Likewise.
11442 (arm_no_early_mul_dep): Likewise.
11443 (arm_no_early_store_addr_dep): Likewise.
11444 (arm_mac_accumulator_is_mul_result): Likewise.
11445 (arm_mac_accumulator_is_result): Likewise.
11446 * config/arm/aarch-common-protos.h: ... here. New file.
11447 * config/arm/t-arm (aarch-common.o): Define.
11448
11449 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11450
11451 * config/arm/arm.md: Include new file "types.md".
11452 (define_attr "type"): Move from here to ...
11453 (define_attr "mul32"): Likewise.
11454 (define_attr "mul64"): Likewise.
11455 * config/arm/types.md: ... here. New file.
11456
11457 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11458
11459 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11460 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11461
11462 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11463
11464 * gen-pass-instances.awk: Fix offset of substr().
11465
11466 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11467
11468 * Makefile.in (pass-instances.def): New.
11469 (passes.o): Replace dependency on passes.def with one on
11470 pass-instances.def
11471
11472 * gen-pass-instances.awk: New.
11473
11474 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11475 rather than passes.def, updating local definition of NEXT_PASS
11476 macro to add an extra NUM parameter (currently unused).
11477
11478 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11479
11480 * Makefile.in (PASS_MANAGER_H): New.
11481 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11482 (passes.o): Likewise.
11483 (statistics.o): Likewise.
11484 (cgraphunit.o): Likewise.
11485 (context.o): Depend on PASS_MANAGER_H.
11486
11487 * pass_manager.h: New.
11488
11489 * cgraphunit.c (cgraph_add_new_function): Update for moves
11490 of globals to fields of pass_manager.
11491 (analyze_function): Likewise.
11492 (expand_function): Likewise.
11493 (ipa_passes): Likewise.
11494 (compile): Likewise.
11495
11496 * context.c (context::context): New.
11497 * context.h (context::context): New.
11498 (context::get_passes): New.
11499 (context::passes_): New.
11500
11501 * lto-cgraph.c (input_node): Update for moves of globals to
11502 fields of pass_manager.
11503
11504 * passes.c (all_passes): Remove, in favor of a field of the
11505 same name within the new class pass_manager.
11506 (all_small_ipa_passes): Likewise.
11507 (all_lowering_passes): Likewise.
11508 (all_regular_ipa_passes): Likewise.
11509 (all_late_ipa_passes): Likewise.
11510 (all_lto_gen_passes): Likewise.
11511 (passes_by_id): Likewise.
11512 (passes_by_id_size): Likewise.
11513 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11514 the new class pass_manager.
11515 (set_pass_for_id): Convert to...
11516 (pass_manager::set_pass_for_id): ...method.
11517 (get_pass_for_id): Convert to...
11518 (pass_manager::get_pass_for_id): ...method.
11519 (register_one_dump_file): Move body of implementation into...
11520 (pass_manager::register_one_dump_file): ...here.
11521 (register_dump_files_1): Convert to...
11522 (pass_manager::register_dump_files_1): ...method.
11523 (register_dump_files): Convert to...
11524 (pass_manager::register_dump_files): ...method.
11525 (create_pass_tab): Update for moves of globals to fields of
11526 pass_manager.
11527 (dump_passes): Move body of implementation into...
11528 (pass_manager::dump_passes): ...here.
11529 (register_pass): Move body of implementation into...
11530 (pass_manager::register_pass): ...here.
11531 (init_optimization_passes): Convert into...
11532 (pass_manager::pass_manager): ...constructor for new
11533 pass_manager class, and initialize the pass_lists array.
11534 (check_profile_consistency): Update for moves of globals to
11535 fields of pass_manager.
11536 (dump_profile_report): Move body of implementation into...
11537 (pass_manager::dump_profile_report): ...here.
11538 (ipa_write_summaries_1): Update for moves of pass lists from
11539 being globals to fields of pass_manager.
11540 (ipa_write_optimization_summaries): Likewise.
11541 (ipa_read_summaries): Likewise.
11542 (ipa_read_optimization_summaries): Likewise.
11543 (execute_all_ipa_stmt_fixups): Likewise.
11544
11545 * statistics.c (statistics_fini): Update for moves of globals to
11546 fields of pass_manager.
11547
11548 * toplev.c (general_init): Replace call to
11549 init_optimization_passes with construction of the pass_manager
11550 instance.
11551
11552 * tree-pass.h (all_passes): Remove, in favor of a field of the
11553 same name within the new class pass_manager.
11554 (all_small_ipa_passes): Likewise.
11555 (all_lowering_passes): Likewise.
11556 (all_regular_ipa_passes): Likewise.
11557 (all_lto_gen_passes): Likewise.
11558 (all_late_ipa_passes): Likewise.
11559 (passes_by_id): Likewise.
11560 (passes_by_id_size): Likewise.
11561 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11562 the new class pass_manager.
11563 (get_pass_for_id): Remove.
11564
11565 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11566
11567 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11568 configs.
11569
11570 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11571
11572 * arm.md (mulhi3): New expand pattern.
11573
11574 2013-07-30 Jan Hubicka <jh@suse.cz>
11575 Martin Liska <marxin.liska@gmail.com>
11576
11577 * profile.c (compute_value_histograms): Do not ICE when
11578 there is mismatch only on some counters.
11579
11580 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11581
11582 PR rtl-optimization/57637
11583 * function.c (move_insn_for_shrink_wrap): Also check the
11584 GEN set of the LIVE problem for the liveness analysis
11585 if it exists, otherwise give up.
11586
11587 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11588
11589 PR tree-optimization/57993
11590 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11591 replaced statement in the candidate table.
11592 (phi_add_costs): Return infinite cost when the hidden basis does
11593 not dominate all phis on which the candidate is dependent.
11594 (replace_one_candidate): Record replaced statement in the
11595 candidate table.
11596
11597 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11598
11599 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11600 (ashlv2si3): New expander.
11601 (*ashlv2si3_i): New define_insn_and_split.
11602 * predicates.md (float_operation): Allow patterns with three
11603 basic sub-patterns.
11604
11605 PR rtl-optimization/58021
11606 * mode-switching.c (create_pre_exit): Always split off preceding
11607 insns if we are not at the basic block head.
11608
11609 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11610
11611 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11612 (UCLIBC_DYNAMIC_LINKER): New macro.
11613 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11614 `-mnan=2008'.
11615 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11616 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11617 `-mnan=2008'.
11618 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11619 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11620 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11621 for SF and DF modes. Use ieee_quad_format for TF mode.
11622 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11623 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11624 (mips_option_override): Handle `-mnan=legacy'.
11625 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11626 `-mabs=2008' and `-mnan=2008'.
11627 (OPTION_DEFAULT_SPECS): Add "nan" default.
11628 (ASM_SPEC): Handle `-mnan='.
11629 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11630 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11631 comment accordingly.
11632 (neg<mode>2): Likewise.
11633 * config/mips/mips.opt (mabs, mnan): New options.
11634 * doc/install.texi (Configuration): Document `--with-nan=' option.
11635 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11636 `-mnan=' options.
11637 (MIPS Options): Document them.
11638 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11639 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11640 * configure: Regenerate.
11641 * config.in: Regenerate.
11642
11643 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11644
11645 * config/i386/i386.md (float post-reload splitters): Do not check
11646 for subregs of SSE registers.
11647
11648 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11649 H.J. Lu <hongjiu.lu@intel.com>
11650
11651 PR target/57954
11652 PR target/57988
11653 * config/i386/i386.md (post-reload splitter
11654 to avoid partial SSE reg dependency stalls): New pattern.
11655
11656 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11657
11658 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11659
11660 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11661
11662 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11663 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11664
11665 2013-07-26 Cary Coutant <ccoutant@google.com>
11666
11667 * dwarf2out.c (die_checksum_ordered): Don't include template
11668 instantiations in signature.
11669 (is_template_parameter): New function.
11670 (is_template_instantiation): New function.
11671 (generate_skeleton_bottom_up): Don't include template instantiations
11672 in type unit DIE.
11673 (generate_skeleton): Likewise.
11674 (break_out_comdat_types): Move recursive call to break out nested
11675 types earlier.
11676 (prune_unused_types_mark_generic_parms_dies): Call
11677 is_template_parameter.
11678
11679 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11680
11681 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11682 uses vector registers.
11683 * config/aarch64/iterators.md: Add attributes rtn and vas.
11684
11685 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11686 Richard Earnshaw <richard.earnshaw@arm.com>
11687
11688 * combine.c (simplify_comparison): Re-canonicalize operands
11689 where appropriate.
11690 * config/arm/arm.md (movcond_addsi): New splitter.
11691
11692 2013-07-25 Sterling Augustine <saugustine@google.com>
11693
11694 * dwarf2out.c (size_of_pubnames): Move code to...
11695 (include_pubname_in_output): ...here. New.
11696 (want_pubnames): Rearrange.
11697 (output_pubnames): Call include_pubname_in_output. Move assertion.
11698
11699 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11700
11701 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11702
11703 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11704
11705 PR target/38836
11706 * doc/extend.texi: Remove obsolete builtins. Fix
11707 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11708
11709 2013-07-25 Jan Hubicka <jh@suse.cz>
11710
11711 * cgraph.c (release_function_body): Break out from ...
11712 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11713 and DECL_ARGUMENTS.
11714 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11715 old_tree in the map.
11716 (create_specialized_node): Update.
11717 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11718 into index.
11719 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11720 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11721 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11722 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11723 DECL_RESULT.
11724
11725 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11726
11727 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11728 addsi3_carryin_alt2_<optab>): Correct output template.
11729
11730 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11731
11732 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11733 Adjust for arm_restrict_it.
11734 Remove trailing whitespace.
11735
11736 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11737
11738 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11739 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11740
11741 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11742
11743 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11744
11745 PR rtl-optimization/57960
11746 * lra-constraints.c (process_alt_operands): Use the right mode
11747 when checking strict_low.
11748
11749 2013-07-25 Jan Hubicka <jh@suse.cz>
11750
11751 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11752 * cgraph.c (cgraph_remove_node): Do not release function body
11753 when in cgraph streaming.
11754 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11755 in other partitions are not considered reachable; fix handling of
11756 clones.
11757
11758 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11759
11760 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11761
11762 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11763
11764 PR target/19599
11765 PR target/57731
11766 PR target/57837
11767 * config/arm/arm.md ("*sibcall_insn): Replace use of
11768 Ss with US. Adjust output for v5 and v4t.
11769 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11770
11771 * config/arm/constraints.md ("Ss"): Rename to US.
11772
11773 2013-07-25 Terry Guo <terry.guo@arm.com>
11774
11775 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11776 shift_add/shift_sub0/shift_sub1 RTXs.
11777
11778 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11779 Anton Blanchard <anton@au1.ibm.com>
11780
11781 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11782 (altivec_vpks<VI_char>ss): Likewise.
11783 (altivec_vpks<VI_char>us): Likewise.
11784 (altivec_vpku<VI_char>us): Likewise.
11785 (altivec_vpku<VI_char>um): Likewise.
11786
11787 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11788
11789 Introduce context class.
11790
11791 * Makefile.in (CONTEXT_H): New.
11792 (OBJS): Add context.o.
11793 (toplev.o): Add CONTEXT_H to dependencies.
11794 (context.o): New.
11795
11796 * toplev.c (general_init): Create the singleton gcc::context instance.
11797
11798 * context.c: New.
11799
11800 * context.h: New.
11801
11802 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11803
11804 PR rtl-optimization/57968
11805 * mode-switching.c (create_pre_exit): Allow instructions that
11806 don't set a return register to need a non-exit mode.
11807
11808 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11809 Anton Blanchard <anton@au1.ibm.com>
11810
11811 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11812 operands to vperm for little endian.
11813 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11814 of lvsl to create the control mask for a vperm for little endian.
11815
11816 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11817 Anton Blanchard <anton@au1.ibm.com>
11818
11819 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11820 two operands for little-endian.
11821
11822 2013-07-23 Steve Ellcey <sellcey@mips.com>
11823
11824 * config/mips/mips.c (mips_case_values_threshold): New.
11825 (TARGET_CASE_VALUES_THRESHOLD): Define.
11826
11827 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11828 Anton Blanchard <anton@au1.ibm.com>
11829
11830 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11831 selection of field for vector splat in little endian mode.
11832
11833 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11834
11835 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11836 expanders to rs6000.md.
11837 (ior<mode>3): Likewise.
11838 (and<mode>3): Likewise.
11839 (one_cmpl<mode>2): Likewise.
11840 (nor<mode>3): Likewise.
11841 (andc<mode>3): Likewise.
11842 (eqv<mode>3): Likewise.
11843 (nand<mode>3): Likewise.
11844 (orc<mode>3): Likewise.
11845
11846 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11847 declaration.
11848
11849 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11850 to split multi-word logical operations.
11851 (rs6000_split_logical_di): Likewise.
11852 (rs6000_split_logical): Likewise.
11853
11854 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11855 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11856 and allow TImode operations in 32-bit.
11857 (vsx_and<mode>3_64bit): Likewise.
11858 (vsx_ior<mode>3_32bit): Likewise.
11859 (vsx_ior<mode>3_64bit): Likewise.
11860 (vsx_xor<mode>3_32bit): Likewise.
11861 (vsx_xor<mode>3_64bit): Likewise.
11862 (vsx_one_cmpl<mode>2_32bit): Likewise.
11863 (vsx_one_cmpl<mode>2_64bit): Likewise.
11864 (vsx_nor<mode>3_32bit): Likewise.
11865 (vsx_nor<mode>3_64bit): Likewise.
11866 (vsx_andc<mode>3_32bit): Likewise.
11867 (vsx_andc<mode>3_64bit): Likewise.
11868 (vsx_eqv<mode>3_32bit): Likewise.
11869 (vsx_eqv<mode>3_64bit): Likewise.
11870 (vsx_nand<mode>3_32bit): Likewise.
11871 (vsx_nand<mode>3_64bit): Likewise.
11872 (vsx_orc<mode>3_32bit): Likewise.
11873 (vsx_orc<mode>3_64bit): Likewise.
11874
11875 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11876 logical types in GPRs.
11877
11878 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11879 logical insns to rs6000.md, and allow TImode operations in
11880 32-bit.
11881 (altivec_ior<mode>3): Likewise.
11882 (altivec_xor<mode>3): Likewise.
11883 (altivec_one_cmpl<mode>2): Likewise.
11884 (altivec_nor<mode>3): Likewise.
11885 (altivec_andc<mode>3): Likewise.
11886
11887 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11888 attributes for moving the 128-bit logical operations into
11889 rs6000.md.
11890 (BOOL_REGS_OUTPUT): Likewise.
11891 (BOOL_REGS_OP1): Likewise.
11892 (BOOL_REGS_OP2): Likewise.
11893 (BOOL_REGS_UNARY): Likewise.
11894 (BOOL_REGS_AND_CR0): Likewise.
11895 (one_cmpl<mode>2): Add support for DI logical operations on
11896 32-bit, splitting the operations to 32-bit.
11897 (anddi3): Likewise.
11898 (iordi3): Likewise.
11899 (xordi3): Likewise.
11900 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11901 changes to combine the 32/64-bit code, allow logical operations on
11902 TI mode in 32-bit, and to use similar match_operator patterns like
11903 scalar mode uses. Combine the Altivec and VSX code for logical
11904 operations, and move it here.
11905 (ior<mode>3, 128-bit types): Likewise.
11906 (xor<mode>3, 128-bit types): Likewise.
11907 (one_cmpl<mode>3, 128-bit types): Likewise.
11908 (nor<mode>3, 128-bit types): Likewise.
11909 (andc<mode>3, 128-bit types): Likewise.
11910 (eqv<mode>3, 128-bit types): Likewise.
11911 (nand<mode>3, 128-bit types): Likewise.
11912 (orc<mode>3, 128-bit types): Likewise.
11913 (and<mode>3_internal): Likewise.
11914 (bool<mode>3_internal): Likewise.
11915 (boolc<mode>3_internal1): Likewise.
11916 (boolc<mode>3_internal2): Likewise.
11917 (boolcc<mode>3_internal1): Likewise.
11918 (boolcc<mode>3_internal2): Likewise.
11919 (eqv<mode>3_internal1): Likewise.
11920 (eqv<mode>3_internal2): Likewise.
11921 (one_cmpl1<mode>3_internal): Likewise.
11922
11923 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11924
11925 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11926 Rename flag_stack_usage to flag_stack_usage_info.
11927
11928 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11929
11930 * config/microblaze/sync.md: New file.
11931 * config/microblaze/microblaze.md: Include sync.md
11932 * config/microblaze/microblaze.c: Add print_operand 'y'.
11933 * config/microblaze/constraints.md: Add memory_contraint
11934 'Q' which is a single register.
11935
11936 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11937
11938 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11939
11940 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11941
11942 * reload.c (find_reloads): Exit loop once we find this operand
11943 cannot be reloaded somehow for this alternative.
11944
11945 * reload.c (find_reloads): Exit loop once we find a hard register.
11946
11947 * rtlanal.c (computed_jump_p): Exit loop once we find label
11948 reference is used.
11949
11950 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11951
11952 * cfgloopmanip.c (remove_path): Exit loop after setting
11953 irred_invalidated.
11954
11955 * gensupport.c (subst_dup): Avoid loop if code is not
11956 MATCH_DUP nor MATCH_OP_DUP.
11957
11958 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11959
11960 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11961
11962 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11963
11964 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11965 true for SP_REGNUM if mode == ptr_mode.
11966 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11967 with value R0_REGNUM + 31.
11968
11969 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11970
11971 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11972 pad pointer-typed argument downward.
11973
11974 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11975
11976 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11977 and __ILP32__ when the ILP32 model is in use.
11978
11979 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11980
11981 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11982 (aarch64_load_symref_appropriately): In the case of
11983 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11984 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11985 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11986 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11987 if 'mode' doesn't equal to 'ptr_mode'.
11988 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11989 'vcall_offset'; change to call aarch64_emit_move differently depending
11990 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11991 to calculate the upper bound of 'vcall_offset'.
11992 (aarch64_cannot_force_const_mem): Change to also return true if
11993 mode != ptr_mode.
11994 (aarch64_legitimize_reload_address): In the case of large
11995 displacements, add new local variable 'xmode' and an assertion
11996 based on it; change to use 'xmode' to generate the new rtx and
11997 reload.
11998 (aarch64_asm_trampoline_template): Change to generate the template
11999 differently depending on TARGET_ILP32 or not; change to use
12000 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12001 (aarch64_trampoline_size): Removed.
12002 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12003 and replace immediate literals with it. Change to use 'ptr_mode'
12004 instead of 'DImode' and call convert_memory_address if the mode
12005 of 'fnaddr' doesn't equal to 'ptr_mode'.
12006 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12007 to output symbol.
12008 (aarch64_elf_asm_destructor): Likewise.
12009 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12010 on TARGET_ILP32 instead of aarch64_trampoline_size.
12011 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12012 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12013 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12014 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12015 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12016 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12017 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12018 depending on the value of 'mode'.
12019 (add_losym_<mode>): New.
12020 (ldr_got_small_<mode>): New, based on ldr_got_small.
12021 (ldr_got_small): Remove.
12022 (ldr_got_small_sidi): New.
12023 * config/aarch64/iterators.md (P): New.
12024 (PTR): Change to 'ptr_mode' in the condition.
12025
12026 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12027
12028 * config.gcc (aarch64*-*-*): Support --with-abi.
12029 (aarch64*-*-elf): Support --with-multilib-list.
12030 (aarch64*-*-linux*): Likewise.
12031 (supported_defaults): Add abi to aarch64*-*-*.
12032 * configure.ac: Mention AArch64 for --with-multilib-list.
12033 * configure: Re-generated.
12034 * config/aarch64/biarchilp32.h: New file.
12035 * config/aarch64/biarchlp64.h: New file.
12036 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12037 (ABI_SPEC): Ditto.
12038 (MULTILIB_DEFAULTS): Ditto.
12039 (DRIVER_SELF_SPECS): Ditto.
12040 (ASM_SPEC): Update to also substitute -mabi.
12041 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12042 file whose name depends on -mabi= and -mbig-endian.
12043 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12044 TARGET_ILP32.
12045 (POINTER_SIZE): New define.
12046 (POINTERS_EXTEND_UNSIGNED): Ditto.
12047 (enum aarch64_abi_type): New enumeration tag.
12048 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12049 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12050 (TARGET_ILP32): New define.
12051 * config/aarch64/aarch64.opt (mabi): New.
12052 (aarch64_abi): New.
12053 (ilp32, lp64): New values for -mabi.
12054 * config/aarch64/t-aarch64 (comma): New define.
12055 (MULTILIB_OPTIONS): Ditto.
12056 (MULTILIB_DIRNAMES): Ditto.
12057 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12058 * doc/invoke.texi: Document -mabi for AArch64.
12059
12060 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12061
12062 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12063
12064 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12065 Anton Blanchard <anton@au1.ibm.com>
12066
12067 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12068 endianness when selecting field to splat.
12069
12070 2013-07-22 Eric Christopher <echristo@gmail.com>
12071
12072 * dwarf2out.c (die_odr_checksum): New function to use
12073 CHECKSUM_ macros and ULEB128 for DIE tag.
12074 (generate_type_signature): Use.
12075
12076 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12077
12078 * config.gcc (sparc*-*-*): Accept leon3 processor.
12079 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12080 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12081 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12082 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12083 (mfix-ut699): Adjust comment.
12084 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12085 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12086 (CPP_CPU_SPEC): Likewise.
12087 (ASM_CPU_SPEC): Likewise.
12088 * config/sparc/sparc.c (leon3_cost): New constant.
12089 (sparc_option_override): Add leon3 support.
12090 (mem_ref): New function.
12091 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12092 (sparc_do_work_around_errata): Look into the instruction in the delay
12093 slot and adjust accordingly. Add fix for the data cache nullify issues
12094 of the UT699. Change insertion position for the NOP.
12095 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12096 (leon3_load): New reservation.
12097 (leon_store): Bump latency to 2.
12098 (grfpu): New automaton.
12099 (grfpu_alu): New unit.
12100 (grfpu_ds): Likewise.
12101 (leon_fp_alu): Adjust.
12102 (leon_fp_mult): Delete.
12103 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12104 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12105 * config/sparc/sparc.md (cpu): Add leon3.
12106 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12107 (swapsi): Likewise.
12108 (atomic_test_and_set): Likewise.
12109 (ldstub): Likewise.
12110
12111 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12112
12113 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12114 default for R5900 targets.
12115 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12116 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12117 * config/mips/mips.c (mips_option_override): Report an error for
12118 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12119 for -march=r5900 -mhard-float.
12120
12121 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12122
12123 * df-problems.c (can_move_insns_across): Exit loop once we
12124 find a non-fixed, non-global register.
12125
12126 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12127 setting can_throw.
12128
12129 * omega.c (omega_eliminate_red): Break after setting red_found.
12130 (omega_problem_has_red_equations): Similarly after setting found.
12131 (omega_query_variable): Similarly after setting coupled.
12132
12133 2013-07-22 Marek Polacek <polacek@redhat.com>
12134
12135 * gimplify.c: Don't include gimple.h twice.
12136
12137 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12138
12139 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12140 instead of TARGET_THUMB1.
12141 (Pz): New constraint.
12142 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12143 encodings.
12144 (compare_negsi_si): Likewise.
12145 (compare_addsi2_op0): Likewise.
12146 (compare_addsi2_op1): Likewise.
12147 (addsi3_carryin_<optab>): Likewise.
12148 (addsi3_carryin_alt2_<optab>): Likewise.
12149 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12150 for arm_restrict_it.
12151 (subsi3_carryin): Likewise.
12152 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12153 (minmax_arithsi): Disable for arm_restrict_it.
12154 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12155 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12156 (satsi_<SAT:code>_shift): Likewise.
12157 (arm_shiftsi3): Add alternative for 16-bit encoding.
12158 (arm32_movhf): Disable for arm_restrict_it.
12159 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12160 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12161
12162 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12163
12164 * config/arm/arm.md (attribute "insn"): Delete.
12165 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12166 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12167 (not_shiftsi): Update for attribute change.
12168 (not_shiftsi_compare0): Likewise.
12169 (not_shiftsi_compare0_scratch): Likewise.
12170 (arm_one_cmplsi2): Likewise.
12171 (thumb1_one_cmplsi2): Likewise.
12172 (notsi_compare0): Likewise.
12173 (notsi_compare0_scratch): Likewise.
12174 (thumb1_movdi_insn): Likewise.
12175 (arm_movsi_insn): Likewise.
12176 (movhi_insn_arch4): Likewise.
12177 (movhi_bytes): Likewise.
12178 (arm_movqi_insn): Likewise.
12179 (thumb1_movqi_insn): Likewise.
12180 (arm32_movhf): Likewise.
12181 (thumb1_movhf): Likewise.
12182 (arm_movsf_soft_insn): Likewise.
12183 (thumb1_movsf_insn): Likewise.
12184 (thumb_movdf_insn): Likewise.
12185 (movsicc_insn): Likewise.
12186 (movsfcc_soft_insn): Likewise.
12187 (and_scc): Likewise.
12188 (cond_move): Likewise.
12189 (if_move_not): Likewise.
12190 (if_not_move): Likewise.
12191 (if_shift_move): Likewise.
12192 (if_move_shift): Likewise.
12193 (if_shift_shift): Likewise.
12194 (if_not_arith): Likewise.
12195 (if_arith_not): Likewise.
12196 (cond_move_not): Likewise.
12197 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12198 (neon_mov<mode>): Likewise.
12199 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12200 (thumb2_movsi_vfp): Likewise.
12201 (movsf_vfp): Likewise.
12202 (thumb2_movsf_vfp): Likewise.
12203 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12204 change.
12205 (cortexa7_older_only): Likewise.
12206 (cortexa7_younger): Likewise.
12207 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12208 (1020alu_shift_op): Likewise.
12209 (1020alu_shift_reg_op): Likewise.
12210 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12211 (alu_shift_op): Likewise.
12212 (alu_shift_reg_op): Likewise.
12213 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12214 (11_alu_shift_op): Likewise.
12215 (11_alu_shift_reg_op): Likewise.
12216 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12217 (9_alu_shift_reg_op): Likewise.
12218 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12219 change.
12220 (cortex_a15_alu_shift): Likewise.
12221 (cortex_a15_alu_shift_reg): Likewise.
12222 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12223 (cortex_a5_alu_shift): Likewise.
12224 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12225 change.
12226 (cortex_a53_alu_shift): Likewise.
12227 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12228 change.
12229 (cortex_a7_alu_reg): Likewise.
12230 (cortex_a7_alu_shift): Likewise.
12231 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12232 (cortex_a8_alu_shift): Likewise.
12233 (cortex_a8_alu_shift_reg): Likewise.
12234 (cortex_a8_mov): Likewise.
12235 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12236 (cortex_a9_dp_shift): Likewise.
12237 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12238 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12239 (cortex_r4_mov): Likewise.
12240 (cortex_r4_alu_shift): Likewise.
12241 (cortex_r4_alu_shift_reg): Likewise.
12242 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12243 (526_alu_shift_op): Likewise.
12244 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12245 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12246 (626te_alu_shift_op): Likewise.
12247 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12248 (726te_alu_op): Likewise.
12249 (726te_alu_shift_op): Likewise.
12250 (726te_alu_shift_reg_op): Likewise.
12251 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12252 (mp626_alu_shift_op): Likewise.
12253 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12254 (pj4_alu_e1_conds): Likewise.
12255 (pj4_alu): Likewise.
12256 (pj4_alu_conds): Likewise.
12257 (pj4_shift): Likewise.
12258 (pj4_shift_conds): Likewise.
12259 (pj4_alu_shift): Likewise.
12260 (pj4_alu_shift_conds): Likewise.
12261
12262 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12263
12264 * config/arm/predicates.md (shiftable_operator_strict_it):
12265 New predicate.
12266 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12267 Disable cond_exec version for arm_restrict_it.
12268 (thumb2_smaxsi3): Convert to generate cond_exec.
12269 (thumb2_sminsi3): Likewise.
12270 (thumb32_umaxsi3): Likewise.
12271 (thumb2_uminsi3): Likewise.
12272 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12273 (thumb2_neg_abssi2): Likewise.
12274 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12275 (thumb2_movsicc_insn): Adjust alternatives.
12276 (thumb2_mov_negscc): Disable for arm_restrict_it.
12277 (thumb2_mov_negscc_strict_it): New pattern.
12278 (thumb2_mov_notscc_strict_it): New pattern.
12279 (thumb2_mov_notscc): Disable for arm_restrict_it.
12280 (thumb2_ior_scc): Likewise.
12281 (thumb2_ior_scc_strict_it): New pattern.
12282 (thumb2_cond_move): Adjust for arm_restrict_it.
12283 (thumb2_cond_arith): Disable for arm_restrict_it.
12284 (thumb2_cond_arith_strict_it): New pattern.
12285 (thumb2_cond_sub): Adjust for arm_restrict_it.
12286 (thumb2_movcond): Likewise.
12287 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12288 (thumb2_zero_extendhisi2_v6): Likewise.
12289 (thumb2_zero_extendqisi2_v6): Likewise.
12290 (orsi_notsi_si): Likewise.
12291 (orsi_not_shiftsi_si): Likewise.
12292
12293 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12294
12295 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12296 instruction sequence is 1 byte shorter.
12297
12298 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12299
12300 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12301 it is not needed after split.
12302
12303 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12304
12305 PR target/51784
12306 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12307 second label for nonlocal goto receivers. Don't output pic base labels
12308 unless we're producing PIC; mark that action unreachable().
12309 (ix86_save_reg): If the function contains a nonlocal label, save the
12310 PIC base reg.
12311 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12312 * config/darwin.c (emitted_pic_label_num): New GTY.
12313 (update_pic_label_number_if_needed): New.
12314 (machopic_output_function_base_name): Adjust for nonlocal receiver
12315 case.
12316 (machopic_should_output_picbase_label): New.
12317 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12318 (nonlocal_goto_receiver): New insn and split.
12319
12320 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12321
12322 * config/aarch64/aarch64-builtins.c
12323 (aarch64_fold_builtin): Fold abs in all modes.
12324 * config/aarch64/aarch64-simd-builtins.def
12325 (abs): Enable for all modes.
12326 * config/aarch64/arm_neon.h
12327 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12328 (vabs_f64): Add missing intrinsic.
12329
12330 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12331
12332 * config/aarch64/arm_neon.h (vabs_s64): New function
12333
12334 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12335
12336 PR target/57516
12337 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12338 * config/avr/avr.md (adjust_len): Add `round'.
12339 * config/avr/avr-protos.h (avr_out_round): New prototype.
12340 (avr_out_plus): Add `out_label' argument.
12341 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12342 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12343 Handle the case where `insn' is just a pattern.
12344 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12345 (avr_out_round): New function.
12346 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12347
12348 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12349
12350 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12351 Add check for flag_stack_usage to handle -fstack-usage support
12352
12353 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12354
12355 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12356 interaction for new Power8 flags and VSX.
12357
12358 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12359
12360 PR middle-end/57698
12361 * tree-inline.c (expand_call_inline): Emit errors during
12362 early_inlining only if optimization is not turned on.
12363
12364 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12365
12366 * passes.def: New.
12367
12368 * passes.c (init_optimization_passes): Move the construction of
12369 the pass hierarchy into a new passes.def file.
12370
12371 * Makefile.in (passes.o): Add dependency on passes.def.
12372
12373 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12374
12375 * passes.c (init_optimization_passes): Introduce macros for
12376 constructing the tree of passes (INSERT_PASSES_AFTER,
12377 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12378 TERMINATE_PASS_LIST).
12379
12380 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12381 Wei Mi <wmi@google.com>
12382
12383 PR rtl-optimization/57878
12384 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12385 top.
12386 (reload_pseudo_compare_func): Check nregs first for reload
12387 pseudos.
12388
12389 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12390
12391 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12392
12393 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12394
12395 * read-rtl.c (validate_const_int): Once an invalid character is
12396 seen, quit the loop.
12397
12398 * gengtype.c (write_roots): Similarly once we find the "deletable"
12399 or "if_marked" option.
12400
12401 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12402
12403 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12404 "xtab" and "sat". Move value "clz" from here to ...
12405 (attriubte "type"): ... here.
12406 (satsi_<SAT:code>): Delete "insn" attribute.
12407 (satsi_<SAT:code>_shift): Likewise.
12408 (arm_zero_extendqisi2addsi): Likewise.
12409 (arm_extendqisi2addsi): Likewise.
12410 (clzsi2): Update for attribute changes.
12411 (rbitsi2): Likewise.
12412 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12413 attribute.
12414 (arm_usatsihi): Likewise.
12415 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12416
12417 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12418
12419 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12420 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12421 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12422 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12423 in alphabetical order.
12424 (attribute "core_cycles"): Update for attribute changes.
12425 (arm_addsi3): Likewise.
12426 (addsi3_compare0): Likewise.
12427 (addsi3_compare0_scratch): Likewise.
12428 (addsi3_compare_op1): Likewise.
12429 (addsi3_compare_op2): Likewise.
12430 (compare_addsi2_op0): Likewise.
12431 (compare_addsi2_op1): Likewise.
12432 (addsi3_carryin_shift_<optab>): Likewise.
12433 (subsi3_carryin_shift): Likewise.
12434 (rsbsi3_carryin_shift): Likewise.
12435 (arm_subsi3_insn): Likewise.
12436 (subsi3_compare0): Likewise.
12437 (subsi3_compare): Likewise.
12438 (arm_andsi3_insn): Likewise.
12439 (thumb1_andsi3_insn): Likewise.
12440 (andsi3_compare0): Likewise.
12441 (andsi3_compare0_scratch): Likewise.
12442 (zeroextractsi_compare0_scratch
12443 (andsi_not_shiftsi_si): Likewise.
12444 (iorsi3_insn): Likewise.
12445 (iorsi3_compare0): Likewise.
12446 (iorsi3_compare0_scratch): Likewise.
12447 (arm_xorsi3): Likewise.
12448 (thumb1_xorsi3_insn): Likewise.
12449 (xorsi3_compare0): Likewise.
12450 (xorsi3_compare0_scratch): Likewise.
12451 (satsi_<SAT:code>_shift): Likewise.
12452 (rrx): Likewise.
12453 (arm_shiftsi3): Likewise.
12454 (shiftsi3_compare0): Likewise.
12455 (not_shiftsi): Likewise.
12456 (not_shiftsi_compare0): Likewise.
12457 (not_shiftsi_compare0_scratch): Likewise.
12458 (arm_one_cmplsi2): Likewise.
12459 (thumb_one_complsi2): Likewise.
12460 (notsi_compare0): Likewise.
12461 (notsi_compare0_scratch): Likewise.
12462 (thumb1_zero_extendhisi2): Likewise.
12463 (arm_zero_extendhisi2): Likewise.
12464 (arm_zero_extendhisi2_v6): Likewise.
12465 (arm_zero_extendhisi2addsi): Likewise.
12466 (thumb1_zero_extendqisi2): Likewise.
12467 (thumb1_zero_extendqisi2_v6): Likewise.
12468 (arm_zero_extendqisi2): Likewise.
12469 (arm_zero_extendqisi2_v6): Likewise.
12470 (arm_zero_extendqisi2addsi): Likewise.
12471 (thumb1_extendhisi2): Likewise.
12472 (arm_extendhisi2): Likewise.
12473 (arm_extendhisi2_v6): Likewise.
12474 (arm_extendqisi): Likewise.
12475 (arm_extendqisi_v6): Likewise.
12476 (arm_extendqisi2addsi): Likewise.
12477 (thumb1_extendqisi2): Likewise.
12478 (thumb1_movdi_insn): Likewise.
12479 (arm_movsi_insn): Likewise.
12480 (movsi_compare0): Likewise.
12481 (movhi_insn_arch4): Likewise.
12482 (movhi_bytes): Likewise.
12483 (arm_movqi_insn): Likewise.
12484 (thumb1_movqi_insn): Likewise.
12485 (arm32_movhf): Likewise.
12486 (thumb1_movhf): Likewise.
12487 (arm_movsf_soft_insn): Likewise.
12488 (thumb1_movsf_insn): Likewise.
12489 (movdf_soft_insn): Likewise.
12490 (thumb_movdf_insn): Likewise.
12491 (arm_cmpsi_insn): Likewise.
12492 (cmpsi_shiftsi): Likewise.
12493 (cmpsi_shiftsi_swp): Likewise.
12494 (arm_cmpsi_negshiftsi_si): Likewise.
12495 (movsicc_insn): Likewise.
12496 (movsfcc_soft_insn): Likewise.
12497 (arith_shiftsi): Likewise.
12498 (arith_shiftsi_compare0
12499 (arith_shiftsi_compare0_scratch
12500 (sub_shiftsi): Likewise.
12501 (sub_shiftsi_compare0
12502 (sub_shiftsi_compare0_scratch
12503 (and_scc): Likewise.
12504 (cond_move): Likewise.
12505 (if_plus_move): Likewise.
12506 (if_move_plus): Likewise.
12507 (if_move_not): Likewise.
12508 (if_not_move): Likewise.
12509 (if_shift_move): Likewise.
12510 (if_move_shift): Likewise.
12511 (if_shift_shift): Likewise.
12512 (if_not_arith): Likewise.
12513 (if_arith_not): Likewise.
12514 (cond_move_not): Likewise.
12515 (thumb1_ashlsi3): Set type attribute.
12516 (thumb1_ashrsi3): Likewise.
12517 (thumb1_lshrsi3): Likewise.
12518 (thumb1_rotrsi3): Likewise.
12519 (shiftsi3_compare0_scratch): Likewise.
12520 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12521 (neon_mov<mode>): Likewise.
12522 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12523 attribute changes.
12524 (thumb2_movsi_insn): Likewise.
12525 (thumb2_cmpsi_neg_shiftsi): Likewise.
12526 (thumb2_extendqisi_v6): Likewise.
12527 (thumb2_zero_extendhisi2_v6): Likewise.
12528 (thumb2_zero_extendqisi2_v6): Likewise.
12529 (thumb2_shiftsi3_short): Likewise.
12530 (thumb2_addsi3_compare0_scratch): Likewise.
12531 (orsi_not_shiftsi_si): Likewise.
12532 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12533 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12534 changes.
12535 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12536 (1020alu_shift_op): Likewise.
12537 (1020alu_shift_reg_op): Likewise.
12538 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12539 (alu_shift_op): Likewise.
12540 (alu_shift_reg_op): Likewise.
12541 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12542 (11_alu_shift_op): Likewise.
12543 (11_alu_shift_reg_op): Likewise.
12544 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12545 (9_alu_shift_reg_op): Likewise.
12546 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12547 changes.
12548 (cortex_a15_alu_shift): Likewise.
12549 (cortex_a15_alu_shift_reg): Likewise.
12550 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12551 changes.
12552 (cortex_a5_alu_shift): Likewise.
12553 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12554 changes.
12555 (cortex_a53_alu_shift): Likewise.
12556 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12557 changes.
12558 (cortex_a7_alu_reg): Likewise.
12559 (cortex_a7_alu_shift): Likewise.
12560 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12561 changes.
12562 (cortex_a8_alu_shift): Likewise.
12563 (cortex_a8_alu_shift_reg): Likewise.
12564 (cortex_a8_mov): Likewise.
12565 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12566 (cortex_a9_dp_shift): Likewise.
12567 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12568 changes.
12569 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12570 changes.
12571 (cortex_r4_mov): Likewise.
12572 (cortex_r4_alu_shift): Likewise.
12573 (cortex_r4_alu_shift_reg): Likewise.
12574 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12575 (526_alu_shift_op): Likewise.
12576 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12577 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12578 (626te_alu_shift_op): Likewise.
12579 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12580 (726te_alu_op): Likewise.
12581 (726te_alu_shift_op): Likewise.
12582 (726te_alu_shift_reg_op): Likewise.
12583 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12584 (mp626_alu_shift_op): Likewise.
12585 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12586 (pj4_alu_e1_conds): Likewise.
12587 (pj4_alu): Likewise.
12588 (pj4_alu_conds): Likewise.
12589 (pj4_shift): Likewise.
12590 (pj4_shift_conds): Likewise.
12591 (pj4_alu_shift): Likewise.
12592 (pj4_alu_shift_conds): Likewise.
12593 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12594 changes.
12595 (cortexa7_older_only): Likewise.
12596 (cortexa7_younger): Likewise.
12597
12598 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12599
12600 * ipa-pure-const.c (generate_summary): Rename to...
12601 (pure_const_generate_summary): ... this.
12602
12603 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12604
12605 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12606
12607 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12608
12609 PR target/57909
12610 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12611 usage in HI mode.
12612
12613 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12614
12615 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12616 enabled without -march=zEC12.
12617 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12618 flags to be set.
12619
12620 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12621
12622 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12623 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12624 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12625 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12626 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12627 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12628 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12629 ISA_HAS_FP4.
12630 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12631 and ISA_HAS_NMADD3_NMSUB3.
12632 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12633 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12634 (nmsub4<mode>, nmsub3<mode>): Likewise.
12635 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12636
12637 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12638
12639 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12640 TARGET_MIPS5400 checking.
12641
12642 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12643 Peter Bergner <bergner@vnet.ibm.com>
12644
12645 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12646 registers in the comment.
12647 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12648 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12649 rather than FIRST_PSEUDO_REGISTERS.
12650
12651 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12652
12653 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12654 enable extra ISA flags with TARGET_HTM.
12655
12656 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12657
12658 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12659 Fix comment typos.
12660
12661 2013-07-15 Cong Hou <congh@google.com>
12662
12663 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12664 in compare function for sorting.
12665
12666 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12667
12668 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12669 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12670 * config/rs6000/rs6000.opt: Add -mhtm option.
12671 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12672 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12673 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12674 __HTM__ if the HTM instructions are available.
12675 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12676 htm_spr_reg_operand): New define_predicates.
12677 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12678 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12679 Include htm.md.
12680 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12681 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12682 HTM builtin functions.
12683 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12684 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12685 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12686 (rs6000_builtin_mask_calculate): Likewise.
12687 (rs6000_option_override_internal): Likewise.
12688 (bdesc_htm): Add new HTM builtin support.
12689 (htm_spr_num): New function.
12690 (htm_spr_regno): Likewise.
12691 (rs6000_htm_spr_icode): Likewise.
12692 (htm_expand_builtin): Likewise.
12693 (htm_init_builtins): Likewise.
12694 (rs6000_expand_builtin): Add support for HTM builtin functions.
12695 (rs6000_init_builtins): Likewise.
12696 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12697 option.
12698 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12699 (TARGET_HTM, MASK_HTM): Define macros.
12700 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12701 (FIXED_REGISTERS): Likewise.
12702 (CALL_USED_REGISTERS): Likewise.
12703 (CALL_REALLY_USED_REGISTERS): Likewise.
12704 (REG_ALLOC_ORDER): Likewise.
12705 (enum reg_class): Likewise.
12706 (REG_CLASS_NAMES): Likewise.
12707 (REG_CLASS_CONTENTS): Likewise.
12708 (REGISTER_NAMES): Likewise.
12709 (ADDITIONAL_REGISTER_NAMES): Likewise.
12710 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12711 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12712 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12713 * config/rs6000/htm.md: New file.
12714 * config/rs6000/htmintrin.h: New file.
12715 * config/rs6000/htmxlintrin.h: New file.
12716
12717 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12718
12719 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12720 Define SYMBOL_TINY_GOT, update comment.
12721 * config/aarch64/aarch64.c
12722 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12723 (aarch64_expand_mov_immediate): Likewise.
12724 (aarch64_print_operand): Likewise.
12725 (aarch64_classify_symbol): Likewise.
12726 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12727 (ldr_got_tiny): Define.
12728
12729 2013-07-13 Tobias Grosser <tobias@grosser.es>
12730
12731 PR tree-optimization/54094
12732 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12733 scheduling dimension for the parallelism check from the polyhedral
12734 information in the AST.
12735 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12736 in 2D + 1 form.
12737
12738 2013-07-13 Jason Merrill <jason@redhat.com>
12739
12740 * print-tree.c (debug_vec_tree): Use debug_raw.
12741 (debug_raw (vec<tree, va_gc> &)): New.
12742 (debug_raw (vec<tree, va_gc> *)): New.
12743 * tree.h: Declare them.
12744
12745 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12746
12747 * ifcvt.c (ifcvt_after_combine): New static variable.
12748 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12749 for size.
12750 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12751 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12752 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12753
12754 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12755
12756 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12757
12758 2013-07-12 Michael Matz <matz@suse.de>
12759
12760 PR middle-end/55771
12761 * convert.c (convert_to_real): Reject non-float inner types.
12762
12763 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12764
12765 * config/aarch64/aarch64-protos.h
12766 (aarch64_simd_immediate_valid_for_move): Remove.
12767 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12768 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12769 cases.
12770 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12771
12772 2013-07-11 Steve Ellcey <sellcey@mips.com>
12773
12774 * config/mips/mips.c (mips_conditional_register_usage): Do not
12775 use t[0-7] registers in MIPS16 mode when optimizing for size.
12776
12777 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12778
12779 * config/i386/i386.c (dispatch_function_versions): Fix array
12780 indexing of function_version_info to match actual_versions.
12781
12782 2013-07-11 Teresa Johnson <tejohnson@google.com>
12783
12784 * vec.h (struct va_gc): Move release out-of-line.
12785 (va_gc::release): Call ggc_free on released vec.
12786
12787 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12788
12789 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12790 Require GOT register as additional operand in UNSPEC.
12791 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12792 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12793 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12794 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12795 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12796 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12797 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12798
12799 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12800
12801 PR target/57631
12802 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12803 name seen by assembler/linker rather if available.
12804
12805 2013-07-11 Andreas Schwab <schwab@suse.de>
12806
12807 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12808
12809 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12810
12811 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12812
12813 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12814
12815 * doc/tm.texi.in: Move hook documentation to ....
12816 * target.def: ... here.
12817
12818 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12819 text on @hook line.
12820 * doc/tm.texi: Regenerate.
12821
12822 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12823
12824 PR c++/57869
12825 * doc/invoke.texi: Document -Wconditionally-supported.
12826
12827 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12828
12829 PR target/57844
12830 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12831 of my_fp.
12832
12833 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12834
12835 PR target/57506
12836 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12837 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12838 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12839 Remove duplicate devices.
12840 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12841 * config/avr/t-multilib: Regenerate.
12842 * config/avr/avr-tables.opt: Regenerate.
12843 * doc/avr-mmcu.texi: Regenerate.
12844
12845 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12846
12847 PR target/56987
12848 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12849
12850 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12851
12852 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12853 the cost of MULT when optimizing for size.
12854
12855 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12856
12857 * config/cr16/cr16-protos.h: Don't include target.h.
12858
12859 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12860
12861 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12862 adjust register size for TDmode and TFmode for VSX registers.
12863
12864 2013-07-08 Kai Tietz <ktietz@redhat.com>
12865
12866 PR target/56892
12867 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12868 hook_bool_const_tree_true.
12869
12870 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12871
12872 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12873 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12874 * config/s390/s390.md: Define FPR*_REGNUM constants.
12875 Fix FPR2_REGNUM constant (18 -> 17).
12876 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12877 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12878 ("trunc<BFP:mode><DFP_ALL:mode>2")
12879 ("trunc<DFP_ALL:mode><BFP:mode>2")
12880 ("*extend<BFP:mode><DFP_ALL:mode>2")
12881 ("*extend<DFP_ALL:mode><BFP:mode>2")
12882 ("extend<BFP:mode><DFP_ALL:mode>2")
12883 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12884 FPR4_REGNUM.
12885
12886 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12887
12888 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12889
12890 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12891
12892 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12893 and cfun_fpr_bit_p to cfun_fpr_save_p.
12894 (s390_frame_area, s390_register_info, s390_frame_info)
12895 (s390_emit_prologue, s390_emit_epilogue)
12896 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12897 register numbers.
12898 * config/s390/s390.h: Define *_REGNUM macros for floating point
12899 register numbers.
12900
12901 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12902
12903 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12904
12905 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12906
12907 PR rtl-optimization/57786
12908 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12909 and break out of the loop when it is set to false.
12910
12911 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12912
12913 PR target/57819
12914 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12915 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12916 (const_int 63)) 0)).
12917 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12918 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12919 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12920
12921 PR rtl-optimization/57829
12922 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12923 mask bits outside of mode are just sign-extension from mode to HWI.
12924
12925 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12926
12927 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12928 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12929 adjust_address instead of change_address to keep info about alignment.
12930 (emit_strmov): Remove.
12931 (emit_memmov): New function.
12932 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12933 (expand_movmem_epilogue): Likewise and return updated rtx for
12934 destination.
12935 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12936 destination and source.
12937 (decide_alignment): Refactor, handle vector_loop.
12938 (ix86_expand_movmem): Likewise.
12939 (ix86_expand_setmem): Likewise.
12940 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12941
12942 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12943
12944 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12945 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12946
12947 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12948
12949 * config/i386/sse.md (sse_movlhps): Change alternative 3
12950 of operand 2 to "m".
12951
12952 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12953
12954 PR target/57807
12955 * config/i386/sse.md (iptr): New mode attribute.
12956 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12957 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12958 (<sse>_vmmul<mode>3): Ditto.
12959 (<sse>_vmdiv<mode>3): Ditto.
12960 (sse_vmrcpv4sf2): Ditto.
12961 (<sse>_vmsqrt<mode>2): Ditto.
12962 (sse_vmrsqrtv4sf2): Ditto.
12963 (<sse>_vm<code><mode>3): Ditto.
12964 (avx_vmcmp<mode>3): Ditto.
12965 (<sse>_vmmaskcmp<mode>3): Ditto.
12966 (<sse>_comi): Ditto.
12967 (<sse>_ucomi): Ditto.
12968 (*xop_vmfrcz_<mode>): Ditto.
12969 (*fmai_fmadd_<mode>): Ditto.
12970 (*fmai_fmsub_<mode>): Ditto.
12971 (*fmai_fnmadd_<mode>): Ditto.
12972 (*fmai_fnmsub_<mode>): Ditto.
12973 (*fma4i_vmfmadd_<mode>): Ditto.
12974 (*fma4i_vmfmsub_<mode>): Ditto.
12975 (*fma4i_vmfnmadd_<mode>): Ditto.
12976 (*fma4i_vmfnmsub_<mode>): Ditto.
12977 (*xop_vmfrcz_<mode>): Ditto.
12978 (sse_cvtps2pi): Ditto.
12979 (sse_cvttps2pi): Ditto.
12980 (sse_cvtss2si): Ditto.
12981 (sse_cvtss2si_2): Ditto.
12982 (sse_cvtss2siq_2): Ditto.
12983 (sse_cvttss2si): Ditto.
12984 (sse_cvttss2siq): Ditto.
12985 (sse_cvtsd2si): Ditto.
12986 (sse_cvtsd2si_2): Ditto.
12987 (sse_cvtsd2siq_2): Ditto.
12988 (sse_cvttsd2si): Ditto.
12989 (sse_cvttsd2siq): Ditto.
12990 (sse_cvtsd2ss): Ditto.
12991 (sse_cvtss2sd): Ditto.
12992 (avx2_pbroadcast<mode>): Ditto.
12993 (avx2_pbroadcast<mode>_1): Ditto.
12994 (*avx_vperm_broadcast_v4sf): Ditto.
12995
12996 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12997 (sse_movlhps): Ditto.
12998 (sse_storehps): Ditto.
12999 (sse_loadhps): Ditto.
13000 (sse_storelps): Ditto.
13001 (sse_loadlps): Ditto.
13002 (*vec_concatv4sf): Ditto.
13003 (*vec_interleave_highv2df): Ditto.
13004 (*vec_interleave_lowv2df): Ditto.
13005 (*vec_extractv2df_1_sse): Ditto.
13006 (*vec_extractv2df_0_sse): Ditto.
13007 (sse2_storelpd): Ditto.
13008 (sse2_loadlpd): Ditto.
13009 (sse2_movsd): Ditto.
13010 (*vec_concatv4si): Ditto.
13011 (vec_concatv2di): Ditto.
13012
13013 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13014 for Intel asm dialect.
13015 (mmx_punpcklwd): Ditto.
13016 (mmx_punpckldq): Ditto.
13017
13018 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13019 for intel assembler dialect.
13020
13021 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13022
13023 PR target/29776
13024 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13025 for BUILT_IN_C{LZ,LRSB}*.
13026 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13027 * tree-vrp.c (extract_range_basic): Handle
13028 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13029 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13030 fall thru to code calling set_value*.
13031 * builtins.c (expand_builtin): Remove *IMAX cases.
13032 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13033 if width is bigger than 2*HWI.
13034
13035 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13036
13037 PR rtl-optimization/55342
13038 * lra-int.h (lra_subreg_reload_pseudos): New.
13039 * lra.c: Add undoing optional reloads to the block diagram.
13040 (lra_subreg_reload_pseudos): New.
13041 (lra_optional_reload_pseudos): Change comments.
13042 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13043 lra_optional_reload_pseudos after undo transformations.
13044 * lra-assigns.c (pseudo_prefix_title): New.
13045 (lra_setup_reg_renumber): Use it.
13046 (spill_for): Ditto. Check subreg reload pseudos too.
13047 (assign_by_spills): Consider subreg reload pseudos too.
13048 * lra-constraints.c (simplify_operand_subreg): Use
13049 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13050 (curr_insn_transform): Recognize and do optional reloads.
13051 (undo_optional_reloads): New.
13052 (lra_undo_inheritance): Call undo_optional_reloads.
13053
13054 2013-07-05 Thomas Quinot <quinot@adacore.com>
13055
13056 * tree-complex.c (expand_complex_operations_1): Fix typo.
13057
13058 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13059
13060 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13061 (tune_params): New member 'const vec_costs'.
13062 * config/aarch64/aarch64.c (generic_vector_cost): New.
13063 (generic_tunings): New member 'generic_vector_cost'.
13064 (aarch64_builtin_vectorization_cost): New.
13065 (aarch64_add_stmt_cost): New.
13066 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13067 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13068
13069 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13070
13071 PR target/57777
13072 * config/i386/predicates.md (vsib_address_operand): Disallow
13073 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13074
13075 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13076
13077 PR middle-end/55030
13078 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13079 expand_builtin_setjmp_receiver.
13080 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13081 with NULL for the label parameter.
13082 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13083 the frame-pointer. Adjust comments.
13084 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13085 only if LABEL is non-NULL.
13086
13087 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13088
13089 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13090 (ARM_ABI_AAPCS64): Ditto.
13091 (arm_abi): Ditto.
13092 (ARM_DEFAULT_ABI): Ditto.
13093
13094 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13095
13096 * config/aarch64/aarch64-builtins.c
13097 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13098 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13099 (st1): Likewise.
13100 * config/aarch64/aarch64-simd.md
13101 (aarch64_ld1<VALL:mode>): New.
13102 (aarch64_st1<VALL:mode>): Likewise.
13103 * config/aarch64/arm_neon.h
13104 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13105
13106 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13107
13108 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13109 target ISA is AVX.
13110 (ix86_option_override_internal):Turn on all -mavx target flags by
13111 default as they are dependent on AVX anyway.
13112
13113 2013-07-02 Cary Coutant <ccoutant@google.com>
13114
13115 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13116 deterministic hash.
13117 (loc_checksum_ordered): Likewise.
13118 (hash_loc_operands): Remove inline keyword.
13119
13120 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13121
13122 PR tree-optimization/57741
13123 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13124 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13125 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13126 Allow REAL_CST step_exprs if flag_associative_math.
13127 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13128
13129 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13130
13131 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13132
13133 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13134
13135 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13136
13137 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13138
13139 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13140 encoding.
13141 (iorsi3_insn): Likewise.
13142 (arm_xorsi3): Likewise.
13143
13144 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13145
13146 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13147 to ...
13148 (attribute "type"): ... here, and prefix with "wmmx_".
13149 (attribute "core_cycles"): Update for attribute changes.
13150 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13151 (tbcstv4hi): Likewise.
13152 (tbcstv2si): Likewise.
13153 (iwmmxt_iordi3): Likewise.
13154 (iwmmxt_xordi3): Likewise.
13155 (iwmmxt_anddi3): Likewise.
13156 (iwmmxt_nanddi3): Likewise.
13157 (iwmmxt_arm_movdi): Likewise.
13158 (iwmmxt_movsi_insn): Likewise.
13159 (mov<mode>_internal): Likewise.
13160 (and<mode>3_iwmmxt): Likewise.
13161 (ior<mode>3_iwmmxt): Likewise.
13162 (xor<mode>3_iwmmxt): Likewise.
13163 (add<mode>3_iwmmxt): Likewise.
13164 (ssaddv8qi3): Likewise.
13165 (ssaddv4hi3): Likewise.
13166 (ssaddv2si3): Likewise.
13167 (usaddv8qi3): Likewise.
13168 (usaddv4hi3): Likewise.
13169 (usaddv2si3): Likewise.
13170 (sub<mode>3_iwmmxt): Likewise.
13171 (sssubv8qi3): Likewise.
13172 (sssubv4hi3): Likewise.
13173 (sssubv2si3): Likewise.
13174 (ussubv8qi3): Likewise.
13175 (ussubv4hi3): Likewise.
13176 (ussubv2si3): Likewise.
13177 (mulv4hi3_iwmmxt): Likewise.
13178 (smulv4hi3_highpart): Likewise.
13179 (umulv4hi3_highpart): Likewise.
13180 (iwmmxt_wmacs): Likewise.
13181 (iwmmxt_wmacsz): Likewise.
13182 (iwmmxt_wmacu): Likewise.
13183 (iwmmxt_wmacuz): Likewise.
13184 (iwmmxt_clrdi): Likewise.
13185 (iwmmxt_clrv8qi): Likewise.
13186 (iwmmxt_clr4hi): Likewise.
13187 (iwmmxt_clr2si): Likewise.
13188 (iwmmxt_uavgrndv8qi3): Likewise.
13189 (iwmmxt_uavgrndv4hi3): Likewise.
13190 (iwmmxt_uavgv8qi3): Likewise.
13191 (iwmmxt_uavgv4hi3): Likewise.
13192 (iwmmxt_tinsrb): Likewise.
13193 (iwmmxt_tinsrh): Likewise.
13194 (iwmmxt_tinsrw): Likewise.
13195 (iwmmxt_textrmub): Likewise.
13196 (iwmmxt_textrmsb): Likewise.
13197 (iwmmxt_textrmuh): Likewise.
13198 (iwmmxt_textrmsh): Likewise.
13199 (iwmmxt_textrmw): Likewise.
13200 (iwmxxt_wshufh): Likewise.
13201 (eqv8qi3): Likewise.
13202 (eqv4hi3): Likewise.
13203 (eqv2si3): Likewise.
13204 (gtuv8qi3): Likewise.
13205 (gtuv4hi3): Likewise.
13206 (gtuv2si3): Likewise.
13207 (gtv8qi3): Likewise.
13208 (gtv4hi3): Likewise.
13209 (gtv2si3): Likewise.
13210 (smax<mode>3_iwmmxt): Likewise.
13211 (umax<mode>3_iwmmxt): Likewise.
13212 (smin<mode>3_iwmmxt): Likewise.
13213 (umin<mode>3_iwmmxt): Likewise.
13214 (iwmmxt_wpackhss): Likewise.
13215 (iwmmxt_wpackwss): Likewise.
13216 (iwmmxt_wpackdss): Likewise.
13217 (iwmmxt_wpackhus): Likewise.
13218 (iwmmxt_wpackwus): Likewise.
13219 (iwmmxt_wpackdus): Likewise.
13220 (iwmmxt_wunpckihb): Likewise.
13221 (iwmmxt_wunpckihh): Likewise.
13222 (iwmmxt_wunpckihw): Likewise.
13223 (iwmmxt_wunpckilb): Likewise.
13224 (iwmmxt_wunpckilh): Likewise.
13225 (iwmmxt_wunpckilw): Likewise.
13226 (iwmmxt_wunpckehub): Likewise.
13227 (iwmmxt_wunpckehuh): Likewise.
13228 (iwmmxt_wunpckehuw): Likewise.
13229 (iwmmxt_wunpckehsb): Likewise.
13230 (iwmmxt_wunpckehsh): Likewise.
13231 (iwmmxt_wunpckehsw): Likewise.
13232 (iwmmxt_wunpckelub): Likewise.
13233 (iwmmxt_wunpckeluh): Likewise.
13234 (iwmmxt_wunpckeluw): Likewise.
13235 (iwmmxt_wunpckelsb): Likewise.
13236 (iwmmxt_wunpckelsh): Likewise.
13237 (iwmmxt_wunpckelsw): Likewise.
13238 (ror<mode>3): Likewise.
13239 (ashr<mode>3_iwmmxt): Likewise.
13240 (lshr<mode>3_iwmmxt): Likewise.
13241 (ashl<mode>3_iwmmxt): Likewise.
13242 (ror<mode>3_di): Likewise.
13243 (ashr<mode>3_di): Likewise.
13244 (lshr<mode>3_di): Likewise.
13245 (ashl<mode>3_di): Likewise.
13246 (iwmmxt_wmadds): Likewise.
13247 (iwmmxt_wmaddu): Likewise.
13248 (iwmmxt_tmia): Likewise.
13249 (iwmmxt_tmiaph): Likewise.
13250 (iwmmxt_tmiabb): Likewise.
13251 (iwmmxt_tmiatb): Likewise.
13252 (iwmmxt_tmiabt): Likewise.
13253 (iwmmxt_tmiatt): Likewise.
13254 (iwmmxt_tmovmskb): Likewise.
13255 (iwmmxt_tmovmskh): Likewise.
13256 (iwmmxt_tmovmskw): Likewise.
13257 (iwmmxt_waccb): Likewise.
13258 (iwmmxt_wacch): Likewise.
13259 (iwmmxt_waccw): Likewise.
13260 (iwmmxt_waligni): Likewise.
13261 (iwmmxt_walignr): Likewise.
13262 (iwmmxt_walignr0): Likewise.
13263 (iwmmxt_walignr1): Likewise.
13264 (iwmmxt_walignr2): Likewise.
13265 (iwmmxt_walignr3): Likewise.
13266 (iwmmxt_wsadb): Likewise.
13267 (iwmmxt_wsadh): Likewise.
13268 (iwmmxt_wsadbz): Likewise.
13269 (iwmmxt_wsadhz): Likewise.
13270 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13271 (iwmmxt_wabsdiffb): Likewise.
13272 (iwmmxt_wabsdiffh): Likewise.
13273 (iwmmxt_wabsdiffw): Likewise.
13274 (iwmmxt_waddsubhx): Likewise
13275 (iwmmxt_wsubaddhx): Likewise.
13276 (addc<mode>3): Likewise.
13277 (iwmmxt_avg4): Likewise.
13278 (iwmmxt_avg4r): Likewise.
13279 (iwmmxt_wmaddsx): Likewise.
13280 (iwmmxt_wmaddux): Likewise.
13281 (iwmmxt_wmaddsn): Likewise.
13282 (iwmmxt_wmaddun): Likewise.
13283 (iwmmxt_wmulwsm): Likewise.
13284 (iwmmxt_wmulwum): Likewise.
13285 (iwmmxt_wmulsmr): Likewise.
13286 (iwmmxt_wmulumr): Likewise.
13287 (iwmmxt_wmulwsmr): Likewise.
13288 (iwmmxt_wmulwumr): Likewise.
13289 (iwmmxt_wmulwl): Likewise.
13290 (iwmmxt_wqmulm): Likewise.
13291 (iwmmxt_wqmulwm): Likewise.
13292 (iwmmxt_wqmulmr): Likewise.
13293 (iwmmxt_wqmulwmr): Likewise.
13294 (iwmmxt_waddbhusm): Likewise.
13295 (iwmmxt_waddbhusl): Likewise.
13296 (iwmmxt_wqmiabb): Likewise.
13297 (iwmmxt_wqmiabt): Likewise.
13298 (iwmmxt_wqmiatb): Likewise.
13299 (iwmmxt_wqmiatt): Likewise.
13300 (iwmmxt_wqmiabbn): Likewise.
13301 (iwmmxt_wqmiabtn): Likewise.
13302 (iwmmxt_wqmiatbn): Likewise.
13303 (iwmmxt_wqmiattn): Likewise.
13304 (iwmmxt_wmiabb): Likewise.
13305 (iwmmxt_wmiabt): Likewise.
13306 (iwmmxt_wmiatb): Likewise.
13307 (iwmmxt_wmiatt): Likewise.
13308 (iwmmxt_wmiabbn): Likewise.
13309 (iwmmxt_wmiabtn): Likewise.
13310 (iwmmxt_wmiatbn): Likewise.
13311 (iwmmxt_wmiattn): Likewise.
13312 (iwmmxt_wmiawbb): Likewise.
13313 (iwmmxt_wmiawbt): Likewise.
13314 (iwmmxt_wmiawtb): Likewise.
13315 (iwmmxt_wmiawtt): Likewise.
13316 (iwmmxt_wmiawbbn): Likewise.
13317 (iwmmxt_wmiawbtn): Likewise.
13318 (iwmmxt_wmiawtbn): Likewise.
13319 (iwmmxt_wmiawttn): Likewise.
13320 (iwmmxt_wmerge): Likewise.
13321 (iwmmxt_tandc<mode>3): Likewise.
13322 (iwmmxt_torc<mode>3): Likewise.
13323 (iwmmxt_torvsc<mode>3): Likewise.
13324 (iwmmxt_textrc<mode>3): Likewise.
13325 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13326 (wmmxt_pack): Likewise.
13327 (wmmxt_mult_c1): Likewise.
13328 (wmmxt_mult_c2): Likewise.
13329 (wmmxt_alu_c1): Likewise.
13330 (wmmxt_alu_c2): Likewise.
13331 (wmmxt_alu_c3): Likewise.
13332 (wmmxt_transfer_c1): Likewise.
13333 (wmmxt_transfer_c2): Likewise.
13334 (wmmxt_transfer_c3): Likewise.
13335 (marvell_f_iwmmxt_wstr): Likewise.
13336 (marvell_f_iwmmxt_wldr): Likewise.
13337
13338 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13339
13340 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13341
13342 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13343
13344 Revert:
13345 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13346 * lra-constraints.c (need_for_split_p): Check call used hard regs
13347 living through calls.
13348
13349 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13350 call used regs for call insn.
13351
13352 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13353
13354 PR target/57736
13355 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13356 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13357
13358 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13359
13360 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13361
13362 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13363
13364 * lra-constraints.c (need_for_split_p): Check call used hard regs
13365 living through calls.
13366
13367 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13368
13369 PR target/57744
13370 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13371 to tie with any other modes. Eliminate Altivec vector mode tests,
13372 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13373 code, to return 0 if testing MODE2 for a condition, if we've
13374 already tested MODE1 for the same condition.
13375
13376 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13377
13378 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13379 layout.
13380
13381 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13382
13383 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13384 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13385
13386 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13387
13388 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13389 Define.
13390 (aarch64_symbolic_constant_p): Remove.
13391 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13392 static. Fix line length and white space.
13393 (aarch64_symbolic_constant_p): Remove.
13394 * config/aarch64/predicates.md (aarch64_valid_symref):
13395 Use aarch64_classify_symbol_expression.
13396
13397 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13398
13399 * config/arm/constraints.md (Ts): New constraint.
13400 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13401 16-bit encodings.
13402 (compare_scc): Use "Ts" constraint for operand 0.
13403 (ior_scc_scc): Likewise.
13404 (and_scc_scc): Likewise.
13405 (and_scc_scc_nodom): Likewise.
13406 (ior_scc_scc_cmp): Likewise for operand 7.
13407 (and_scc_scc_cmp): Likewise.
13408 * config/arm/thumb2.md (thumb2_movsi_insn):
13409 Add alternatives for 16-bit encodings.
13410 (thumb2_movhi_insn): Likewise.
13411 (thumb2_movsicc_insn): Likewise.
13412 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13413 (thumb2_negscc): Use "Ts" constraint.
13414 Move mvn instruction outside cond_exec block.
13415 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13416 for 16-bit encodings.
13417
13418 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13419
13420 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13421 encoding.
13422 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13423 (mulsi3subsi): Likewise.
13424 (mulsidi3adddi): Likewise.
13425 (mulsidi3_v6): Likewise.
13426 (umulsidi3_v6): Likewise.
13427 (umulsidi3adddi_v6): Likewise.
13428 (smulsi3_highpart_v6): Likewise.
13429 (umulsi3_highpart_v6): Likewise.
13430 (mulhisi3tb): Likewise.
13431 (mulhisi3bt): Likewise.
13432 (mulhisi3tt): Likewise.
13433 (maddhisi4): Likewise.
13434 (maddhisi4tb): Likewise.
13435 (maddhisi4tt): Likewise.
13436 (maddhidi4): Likewise.
13437 (maddhidi4tb): Likewise.
13438 (maddhidi4tt): Likewise.
13439 (zeroextractsi_compare0_scratch): Likewise.
13440 (insv_zero): Likewise.
13441 (insv_t2): Likewise.
13442 (anddi_notzesidi_di): Likewise.
13443 (anddi_notsesidi_di): Likewise.
13444 (andsi_notsi_si): Likewise.
13445 (iordi_zesidi_di): Likewise.
13446 (xordi_zesidi_di): Likewise.
13447 (andsi_iorsi3_notsi): Likewise.
13448 (smax_0): Likewise.
13449 (smax_m1): Likewise.
13450 (smin_0): Likewise.
13451 (not_shiftsi): Likewise.
13452 (unaligned_loadsi): Likewise.
13453 (unaligned_loadhis): Likewise.
13454 (unaligned_loadhiu): Likewise.
13455 (unaligned_storesi): Likewise.
13456 (unaligned_storehi): Likewise.
13457 (extv_reg): Likewise.
13458 (extzv_t2): Likewise.
13459 (divsi3): Likewise.
13460 (udivsi3): Likewise.
13461 (arm_zero_extendhisi2addsi): Likewise.
13462 (arm_zero_extendqisi2addsi): Likewise.
13463 (compareqi_eq0): Likewise.
13464 (arm_extendhisi2_v6): Likewise.
13465 (arm_extendqisi2addsi): Likewise.
13466 (arm_movt): Likewise.
13467 (thumb2_ldrd): Likewise.
13468 (thumb2_ldrd_base): Likewise.
13469 (thumb2_ldrd_base_neg): Likewise.
13470 (thumb2_strd): Likewise.
13471 (thumb2_strd_base): Likewise.
13472 (thumb2_strd_base_neg): Likewise.
13473 (arm_negsi2): Add alternative for 16-bit encoding.
13474 (arm_one_cmplsi2): Likewise.
13475
13476 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13477
13478 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13479 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13480 (movdfcc): Likewise.
13481 * config/arm/vfp.md (*thumb2_movsf_vfp):
13482 Disable predication for arm_restrict_it.
13483 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13484 (*thumb2_movdfcc_vfp): Likewise.
13485 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13486 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13487 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13488 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13489 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13490 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13491 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13492 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13493 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13494 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13495 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13496 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13497 Disable predication for arm_restrict_it.
13498
13499 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13500
13501 * config/i386/bmiintrin.h (_bextr_u32): New.
13502 (_bextr_u64): Ditto.
13503
13504 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13505
13506 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13507 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13508 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13509 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13510 * config/mips/n32-elf.h: ...this new file.
13511
13512 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13513
13514 PR target/57224
13515 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13516 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13517
13518 2013-06-27 Catherine Moore <clm@codesourcery.com>
13519
13520 * config/mips/mips-tables.opt: Regenerate.
13521 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13522 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13523 * doc/invoke.texi: Add -m14kc.
13524
13525 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13526
13527 PR target/57623
13528 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13529 constraints of operand 1 and 2.
13530
13531 PR target/57623
13532 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13533 to match RTL canonicalization. Swap predicates and
13534 constraints of operand 1 and 2.
13535
13536 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13537
13538 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13539 Process OP_INOUT regs for splitting too.
13540
13541 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13542
13543 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13544 decl before the loop, initialize to NULL.
13545 (vectorizable_load): Initialize ptr_incr to NULL.
13546
13547 2013-06-27 Martin Jambor <mjambor@suse.cz>
13548
13549 PR lto/57208
13550 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13551 * ipa-ref.c (ipa_maybe_record_reference): New function.
13552 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13553 * ipa-cp.c (create_specialized_node): Record potential references from
13554 aggvals.
13555 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13556
13557 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13558
13559 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13560 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13561 'mode' to force_reg.
13562 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13563 (aarch64_expand_mov_immediate): Likewise.
13564
13565 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13566
13567 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13568 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13569
13570 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13571
13572 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13573 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13574 (struct machine_function): Add tbegin_p.
13575 (s390_canonicalize_comparison): Fold CC mode compares to
13576 conditional jump if possible.
13577 (s390_emit_jump): Return the emitted jump.
13578 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13579 Handle CCRAWmode compares.
13580 (s390_option_override): Default to -mhtm if available.
13581 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13582 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13583 FPRs instead of df_regs_ever_live_p.
13584 (s390_optimize_nonescaping_tx): New function.
13585 (s390_init_frame_layout): Extend clobbered_regs array to cover
13586 FPRs as well.
13587 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13588 (s390_expand_tbegin): New function.
13589 (enum s390_builtin): New enum definition.
13590 (code_for_builtin): New array definition.
13591 (s390_init_builtins): New function.
13592 (s390_expand_builtin): New function.
13593 (TARGET_INIT_BUILTINS): Define.
13594 (TARGET_EXPAND_BUILTIN): Define.
13595 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13596 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13597 (s390_alc_comparison): Likewise.
13598 * config/s390/s390-modes.def: Add CCRAWmode.
13599 * config/s390/s390.h (processor_flags): Add PF_TX.
13600 (TARGET_CPU_HTM): Define macro.
13601 (TARGET_HTM): Define macro.
13602 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13603 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13604 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13605 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13606 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13607 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13608 ("*cc_to_int"): Move up.
13609 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13610 constants other than 0.
13611 ("*ccraw_to_int"): New insn and splitter definition.
13612 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13613 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13614 ("tx_assist"): New expander.
13615 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13616 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13617 * config/s390/s390.opt: Add -mhtm option.
13618 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13619 * config/s390/htmxlintrin.h: New file.
13620 * config/s390/htmintrin.h: New file.
13621 * config/s390/s390intrin.h: New file.
13622 * doc/extend.texi: Document htm builtins.
13623 * config.gcc: Add the new header files to extra_headers.
13624
13625 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13626
13627 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13628 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13629
13630 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13631 Pat Haugen <pthaugen@us.ibm.com>
13632 Peter Bergner <bergner@vnet.ibm.com>
13633
13634 * config/rs6000/power8.md: New.
13635 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13636 setting for power8 entry.
13637 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13638 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13639 test for Power4/Power5 only.
13640 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13641 support.
13642 (force_new_group): Adjust comment.
13643 * config/rs6000/rs6000.md: Include power8.md.
13644
13645 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13646
13647 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13648 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13649 declaration.
13650 (tune_params): Update comment.
13651 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13652 (arm_max_conditional_execute): New function.
13653 (thumb2_final_prescan_insn): Use max_insn_skipped and
13654 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13655
13656 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13657
13658 PR tree-optimization/57705
13659 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13660 SSA_NAME step, provided that it is not defined inside the loop.
13661 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13662 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13663
13664 2013-06-25 Martin Jambor <mjambor@suse.cz>
13665
13666 PR middle-end/57670
13667 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13668 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13669 calls in the dump.
13670 (ipa_note_param_call): Initialize member_ptr flag.
13671 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13672 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13673 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13674 (ipa_read_indirect_edge_info): Likewise.
13675
13676 2013-06-25 Richard Biener <rguenther@suse.de>
13677
13678 PR middle-end/56977
13679 * passes.c (init_optimization_passes): Move pass_fold_builtins
13680 and pass_dce earlier with -Og.
13681
13682 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13683
13684 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13685 <BIT_FIELD_REF>: Remove trailing TAB.
13686 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13687 remove blank line.
13688
13689 2013-06-24 Martin Jambor <mjambor@suse.cz>
13690
13691 PR tree-optimization/57358
13692 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13693 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13694 (ipa_analyze_params_uses): Generate pessimistic info when true.
13695
13696 2013-06-24 Martin Jambor <mjambor@suse.cz>
13697
13698 PR tree-optimization/57539
13699 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13700 global.inlined_to of the new node to it. All callers changed.
13701 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13702 inlining_into, pass it to cgraph_clone_node.
13703 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13704 ipa_free_edge_args_substructures.
13705 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13706 rdesc linked list. Do not assert rdesc edges have inlined caller.
13707 Assert we have found an rdesc in the rdesc list.
13708
13709 2013-06-24 Richard Biener <rguenther@suse.de>
13710
13711 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13712 (pointer_set_lookup): Declare.
13713 (class pointer_map): New template class implementing a
13714 generic pointer to T map.
13715 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13716 pointer_map<T>::contains, pointer_map<T>::insert,
13717 pointer_map<T>::traverse): New functions.
13718 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13719 (pointer_set_lookup): New function.
13720 (pointer_set_contains): Use pointer_set_lookup.
13721 (pointer_set_insert): Likewise.
13722 (insert_aux): Remove.
13723 (struct pointer_map_t): Embed a pointer_set_t.
13724 (pointer_map_create): Adjust.
13725 (pointer_map_destroy): Likewise.
13726 (pointer_map_contains): Likewise.
13727 (pointer_map_insert): Likewise.
13728 (pointer_map_traverse): Likewise.
13729 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13730 pointer_map<unsigned> instead of a pointer_map_t.
13731 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13732 (streamer_tree_cache_lookup): Likewise.
13733 (streamer_tree_cache_create): Likewise.
13734 (streamer_tree_cache_delete): Likewise.
13735 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13736 pointer_map<unsigned> instead of a pointer_map_t.
13737 (lto_init_tree_ref_encoder): Adjust.
13738 (lto_destroy_tree_ref_encoder): Likewise.
13739 * lto-section-out.c (lto_output_decl_index): Likewise.
13740 (lto_record_function_out_decl_state): Likewise.
13741 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13742
13743 2013-06-24 Richard Biener <rguenther@suse.de>
13744
13745 PR tree-optimization/57488
13746 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13747
13748 2013-06-24 Alan Modra <amodra@gmail.com>
13749
13750 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13751 (gen_easy_altivec_constant): Likewise.
13752 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13753 easy_vector_constant_msb): Likewise.
13754
13755 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13756
13757 PR target/57688
13758 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13759 add missing return true.
13760
13761 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13762
13763 PR target/52483
13764 * config/sh/predicates.md (general_extend_operand): Invoke
13765 general_movsrc_operand for memory operands.
13766 (general_movsrc_operand): Allow reg+reg addressing, do not use
13767 general_operand for memory operands.
13768
13769 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13770
13771 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13772 when current target options does not apply.
13773 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13774 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13775 * config/i386/bmiintrin.h: Pass appropriate target
13776 attributes to header.
13777 * config/i386/mmintrin.h: Ditto.
13778 * config/i386/nmmintrin.h: Ditto.
13779 * config/i386/avx2intrin.h: Ditto.
13780 * config/i386/fxsrintrin.h: Ditto.
13781 * config/i386/tbmintrin.h: Ditto.
13782 * config/i386/xsaveintrin.h: Ditto.
13783 * config/i386/f16cintrin.h: Ditto.
13784 * config/i386/xtestintrin.h: Ditto.
13785 * config/i386/xsaveoptintrin.h: Ditto.
13786 * config/i386/bmi2intrin.h: Ditto.
13787 * config/i386/lzcntintrin.h: Ditto.
13788 * config/i386/smmintrin.h: Ditto.
13789 * config/i386/wmmintrin.h: Ditto.
13790 * config/i386/x86intrin.h: Remove all header include guards.
13791 * config/i386/prfchwintrin.h: Ditto.
13792 * config/i386/pmmintrin.h: Ditto.
13793 * config/i386/tmmintrin.h: Ditto.
13794 * config/i386/xmmintrin.h: Ditto.
13795 * config/i386/popcntintrin.h: Ditto.
13796 * config/i386/rdseedintrin.h: Ditto.
13797 * config/i386/ammintrin.h: Ditto.
13798 * config/i386/emmintrin.h: Ditto.
13799 * config/i386/immintrin.h: Remove all header include guards.
13800 * config/i386/fma4intrin.h: Ditto.
13801 * config/i386/lwpintrin.h: Ditto.
13802 * config/i386/xopintrin.h: Ditto.
13803 * config/i386/ia32intrin.h: Ditto.
13804 * config/i386/avxintrin.h: Ditto.
13805 * config/i386/rtmintrin.h: Ditto.
13806 * config/i386/fmaintrin.h: Ditto.
13807 * config/i386/mm3dnow.h: Ditto.
13808
13809 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13810
13811 * common/config/i386/i386-common.c: Handle LZCNT.
13812
13813 2013-06-22 Andi Kleen <ak@linux.intel.com>
13814
13815 * doc/extend.texi: Use __atomic_store_n instead of
13816 __atomic_store in HLE example.
13817
13818 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13819
13820 * config/sh/sh.c: Remove <cstdlib> workaround.
13821
13822 2013-06-21 Andi Kleen <ak@linux.intel.com>
13823
13824 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13825
13826 2013-06-21 Andi Kleen <ak@linux.intel.com>
13827
13828 * doc/extend.texi: Document that __atomic_clear and
13829 __atomic_test_and_set should only be used with bool.
13830
13831 2013-06-20 Jan Hubicka <jh@suse.cz>
13832
13833 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13834 types_same_for_odr.
13835 * tree.c (decls_same_for_odr): New function.
13836 (same_for_edr): New function.
13837 (types_same_for_odr): New function.
13838 (get_binfo_at_offset): Use it.
13839 * tree.h (types_same_for_odr): Declare.
13840
13841 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13842 Jason Merrill <jason@redhat.com>
13843
13844 * system.h: Include <cstdlib> as well as <stdlib.h>.
13845
13846 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13847
13848 PR target/57655
13849 * config/i386/i386.c (construct_container): Report error if
13850 long double is used with disabled x87 float returns.
13851
13852 2013-06-20 Jan Hubicka <jh@suse.cz>
13853
13854 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13855
13856 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13857
13858 PR rtl-optimization/57425
13859 PR rtl-optimization/57569
13860 * alias.c (write_dependence_p): Remove parameters mem_mode and
13861 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13862 Changed all callers.
13863 (canon_anti_dependence): Get comments and semantics in sync.
13864 Add parameter mem_canonicalized. Changed all callers.
13865 * rtl.h (canon_anti_dependence): Update prototype.
13866
13867 2013-06-20 Richard Biener <rguenther@suse.de>
13868
13869 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13870 byte case, inline streamer_read_uchar and defer section
13871 overrun check.
13872
13873 2013-06-20 Richard Biener <rguenther@suse.de>
13874
13875 PR tree-optimization/57584
13876 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13877 SSA names into the expanded expression that take part in
13878 abnormal coalescing.
13879
13880 2013-06-19 Sharad Singhai <singhai@google.com>
13881
13882 * gcov.c (print_usage): Handle new option.
13883 (process_args): Ditto.
13884 (get_gcov_intermediate_filename): New function.
13885 (output_intermediate_file): New function.
13886 (output_gcov_file): New function
13887 (generate_results): Handle new option.
13888 (release_function): Relase demangled name.
13889 (read_graph_file): Handle demangled name.
13890 (output_lines): Ditto.
13891 * doc/gcov.texi: Document gcov intermediate format.
13892
13893 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13894
13895 PR bootstrap/57604
13896 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13897 (lra_emit_add): Use the functions. Add comment about Y as an
13898 address segment.
13899
13900 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13901
13902 PR driver/57652
13903 * collect2.c (collect_atexit): New.
13904 (collect_exit): Delete.
13905 (main): Register collect_atexit with atexit.
13906 (collect_wait): Change collect_exit to exit.
13907 (do_wait): Same.
13908 * collect2.h (collect_exit): Delete.
13909 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13910
13911 2013-06-19 Wei Mi <wmi@google.com>
13912
13913 PR rtl-optimization/57518
13914 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13915 if regno is used in paradoxical subreg.
13916 (update_equiv_regs): Check pdx_subregs[regno] before
13917 set a reg to be equivalent with a mem.
13918
13919 2013-06-19 Matthias Klose <doko@ubuntu.com>
13920
13921 PR driver/57651
13922 * file-find.h (find_a_file): Add a mode parameter.
13923 * file-find.c (find_a_file): Likewise.
13924 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13925 with X_OK for the executables.
13926 * collect2.c (main): Call find_a_file with X_OK.
13927
13928 2013-06-19 Steve Ellcey <sellcey@mips.com>
13929
13930 PR target/56942
13931 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13932 Use NEXT_INSN instead of next_real_insn.
13933
13934 2013-06-19 Jan Hubicka <jh@suse.cz>
13935
13936 * cgraph.h (const_value_known_p): Replace by ...
13937 (ctor_for_folding): .. this one.
13938 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13939 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13940 * expr.c (expand_expr_real_1): Likewise.
13941 (string_constant): Likewise.
13942 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13943 * ipa.c (process_references): Likewise.
13944 (symtab_remove_unreachable_nodes): Likewise.
13945 * ipa-inline-analysis.c (param_change_prob): Likewise.
13946 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13947 (get_base_constructor): Likwise.
13948 * varpool.c (varpool_remove_node): Likewise.
13949 (varpool_remove_initializer): LIkewise.
13950 (dump_varpool_node): LIkwise.
13951 (const_value_known_p): Rewrite to ...
13952 (ctor_for_folding): ... this one.
13953
13954 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13955
13956 PR driver/57651
13957 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13958 PERSONALITY in $PATH derived prefixes.
13959
13960 2013-06-19 Jeff Law <law@redhat.com>
13961
13962 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13963 in comment.
13964
13965 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13966 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13967 booleans.
13968
13969 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13970
13971 * config/arm/vfp.md: Move VFP instruction classification documentation
13972 to ...
13973 * config/arm/arm.md: ... here. Update instruction classification
13974 documentation.
13975
13976 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13977
13978 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13979 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13980 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13981
13982 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13983
13984 PR target/57609
13985 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13986 with NEXT_INSN. Use tablejump_p to check for jump table data
13987 insns.
13988
13989 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13990
13991 PR c++/56544
13992 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13993 that now in C++ the value is correct per the C++ standards.
13994
13995 2013-06-19 Richard Biener <rguenther@suse.de>
13996
13997 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13998 for global context.
13999
14000 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14001
14002 Revert:
14003 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14004
14005 PR target/57609
14006 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14007 with next_active_insn.
14008
14009 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14010
14011 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14012 functions are inlined during failures to flag an error.
14013 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14014 in early inline pass.
14015
14016 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14017
14018 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14019 in comments.
14020
14021 2013-06-18 Julian Brown <julian@codesourcery.com>
14022
14023 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14024 Permit virtual register pre-reload if !strict.
14025 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14026 change.
14027 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14028 prototype.
14029 * config/arm/neon.md (movmisalign<mode>): Use
14030 neon_perm_struct_or_reg_operand instead of
14031 neon_struct_or_register_operand.
14032 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14033 neon_permissive_struct_operand instead of neon_struct_operand.
14034 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14035 neon_vector_mem_operand.
14036 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14037 neon_vector_mem_operand.
14038 (neon_permissive_struct_operand): New.
14039 (neon_struct_or_register_operand): Rename to...
14040 (neon_perm_struct_or_reg_operand): This. Adjust call to
14041 neon_vector_mem_operand.
14042
14043 2013-06-18 Richard Biener <rguenther@suse.de>
14044
14045 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14046 * lto-streamer.h: Include pointer-set.h.
14047 (struct lto_decl_slot): Remove.
14048 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14049 Remove next_index entry.
14050 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14051 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14052 (lto_init_tree_ref_encoder): Adjust.
14053 (lto_destroy_tree_ref_encoder): Likewise.
14054 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14055 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14056 (lto_output_decl_index): Adjust.
14057 (lto_new_out_decl_state): Likewise.
14058 (lto_record_function_out_decl_state): Likewise.
14059 * lto-streamer-out.c (copy_function): Likewise.
14060
14061 2013-06-18 Richard Biener <rguenther@suse.de>
14062
14063 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14064 * cgraphunit.c: Include cfgloop.h.
14065 (init_lowered_empty_function): Initialize the loop tree.
14066 (assemble_thunk): Insert new BBs into loops.
14067
14068 2013-06-18 Richard Biener <rguenther@suse.de>
14069
14070 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14071 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14072 the map from cache entry to cache index optional.
14073 (streamer_tree_cache_replace_tree): Adjust accordingly.
14074 (streamer_tree_cache_append): Likewise.
14075 (streamer_tree_cache_delete): Likewise.
14076 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14077 streamer cache map from cache entry to cache index.
14078 * lto-streamer-out.c (create_output_block): Adjust.
14079
14080 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14081
14082 * config/arm/arm.md (attribute "insn"): Move multiplication and
14083 division attributes to...
14084 (attribute "type"): ... here. Remove mult.
14085 (attribute "mul32"): New attribute.
14086 (attribute "mul64"): Add umaal.
14087 (*arm_mulsi3): Update attributes.
14088 (*arm_mulsi3_v6): Likewise.
14089 (*thumb_mulsi3): Likewise.
14090 (*thumb_mulsi3_v6): Likewise.
14091 (*mulsi3_compare0): Likewise.
14092 (*mulsi3_compare0_v6): Likewise.
14093 (*mulsi_compare0_scratch): Likewise.
14094 (*mulsi_compare0_scratch_v6): Likewise.
14095 (*mulsi3addsi): Likewise.
14096 (*mulsi3addsi_v6): Likewise.
14097 (*mulsi3addsi_compare0): Likewise.
14098 (*mulsi3addsi_compare0_v6): Likewise.
14099 (*mulsi3addsi_compare0_scratch): Likewise.
14100 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14101 (*mulsi3subsi): Likewise.
14102 (*mulsidi3adddi): Likewise.
14103 (*mulsi3addsi_v6): Likewise.
14104 (*mulsidi3adddi_v6): Likewise.
14105 (*mulsidi3_nov6): Likewise.
14106 (*mulsidi3_v6): Likewise.
14107 (*umulsidi3_nov6): Likewise.
14108 (*umulsidi3_v6): Likewise.
14109 (*umulsidi3adddi): Likewise.
14110 (*umulsidi3adddi_v6): Likewise.
14111 (*smulsi3_highpart_nov6): Likewise.
14112 (*smulsi3_highpart_v6): Likewise.
14113 (*umulsi3_highpart_nov6): Likewise.
14114 (*umulsi3_highpart_v6): Likewise.
14115 (mulhisi3): Likewise.
14116 (*mulhisi3tb): Likewise.
14117 (*mulhisi3bt): Likewise.
14118 (*mulhisi3tt): Likewise.
14119 (maddhisi4): Likewise.
14120 (*maddhisi4tb): Likewise.
14121 (*maddhisi4tt): Likewise.
14122 (maddhidi4): Likewise.
14123 (*maddhidi4tb): Likewise.
14124 (*maddhidi4tt): Likewise.
14125 (divsi3): Likewise.
14126 (udivsi3): Likewise.
14127 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14128 (thumb2_mulsi_short_compare0): Likewise.
14129 (thumb2_mulsi_short_compare0_scratch): Likewise.
14130 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14131 (1020mult2): Likewise.
14132 (1020mult3): Likewise.
14133 (1020mult4): Likewise.
14134 (1020mult5): Likewise.
14135 (1020mult6): Likewise.
14136 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14137 change.
14138 (cortex_a15_mult64): Likewise.
14139 (cortex_a15_sdiv): Likewise.
14140 (cortex_a15_udiv): Likewise.
14141 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14142 (mult2): Likewise.
14143 (mult3): Likewise.
14144 (mult4): Likewise.
14145 (mult5): Likewise.
14146 (mult6): Likewise.
14147 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14148 (pj4_ir_div): Likewise.
14149 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14150 (11_mult2): Likewise.
14151 (11_mult3): Likewise.
14152 (11_mult4): Likewise.
14153 (11_mult5): Likewise.
14154 (11_mult6): Likewise.
14155 (11_mult7): Likewise.
14156 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14157 (cortex_a8_mla): Likewise.
14158 (cortex_a8_mull): Likewise.
14159 (cortex_a8_smulwy): Likewise.
14160 (cortex_a8_smlald): Likewise.
14161 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14162 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14163 (cortex_r4_mul_3): Likewise.
14164 (cortex_r4_mla_4): Likewise.
14165 (cortex_r4_mla_3): Likewise.
14166 (cortex_r4_smlald): Likewise.
14167 (cortex_r4_mull): Likewise.
14168 (cortex_r4_sdiv): Likewise.
14169 (cortex_r4_udiv): Likewise.
14170 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14171 (cortex_a7_idiv): Likewise.
14172 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14173 (9_mult2): Likewise.
14174 (9_mult3): Likewise.
14175 (9_mult4): Likewise.
14176 (9_mult5): Likewise.
14177 (9_mult6): Likewise.
14178 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14179 (cortex_a53_sdiv): Likewise.
14180 (cortex_a53_udiv): Likewise.
14181 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14182 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14183 (mp626_mult2): Likewise.
14184 (mp626_mult3): Likewise.
14185 (mp626_mult4): Likewise.
14186 * config/arm/fa526.md (526_mult1): Update attribute change.
14187 (526_mult2): Likewise.
14188 * config/arm/arm-generic.md (mult): Update attribute change.
14189 (mult_ldsched_strongarm): Likewise.
14190 (mult_ldsched): Likewise.
14191 (multi_cycle): Likewise.
14192 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14193 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14194 (606te_mult2): Likewise.
14195 (606te_mult3): Likewise.
14196 (606te_mult4): Likewise.
14197 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14198 (cortex_a9_mac16): Likewise.
14199 (cortex_a9_multiply): Likewise.
14200 (cortex_a9_mac): Likewise.
14201 (cortex_a9_multiply_long): Likewise.
14202 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14203 (626te_mult2): Likewise.
14204 (626te_mult3): Likewise.
14205 (626te_mult4): Likewise.
14206
14207 2013-06-18 Richard Biener <rguenther@suse.de>
14208
14209 PR lto/57334
14210 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14211
14212 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14213
14214 PR target/57609
14215 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14216 with next_active_insn.
14217
14218 2013-06-18 Alan Modra <amodra@gmail.com>
14219
14220 * config/rs6000/rs6000.h (enum data_align): New.
14221 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14222 (DATA_ABI_ALIGNMENT): Define.
14223 (CONSTANT_ALIGNMENT): Correct comment.
14224 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14225 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14226
14227 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14228
14229 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14230 ATTRIBUTE_UNUSED marking.
14231
14232 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14233
14234 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14235 alternative and update.
14236 (aarch64_dup_lanedi): Delete.
14237 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14238 * config/aarch64/aarch64-simd-builtins.def: Update.
14239
14240 2013-06-17 Richard Biener <rguenther@suse.de>
14241
14242 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14243 (lto_input_scc): Declare.
14244 (lto_input_tree_1): Likewise.
14245 (struct lto_stats_d): Add num_tree_bodies_output and
14246 num_pickle_refs_output.
14247 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14248 (lto_read_tree_1): Split out from ...
14249 (lto_read_tree): ... this.
14250 (lto_input_scc): New function.
14251 (lto_input_tree_1): Split out from ...
14252 (lto_input_tree): ... this. Handle LTO_tree_scc.
14253 (lto_data_in_create): Create the streamer cache without hashes.
14254 * lto-streamer-out.c (create_output_block): Create the streamer
14255 cache with hashes when not doing WPA.
14256 (lto_write_tree_1): Split out from ...
14257 (lto_write_tree): ... this.
14258 (get_symbol_initial_value): New function.
14259 (lto_output_tree_1): Split out from ...
14260 (lto_output_tree): ... this. Write trees as series of SCCs
14261 using a DFS walk via DFS_write_tree.
14262 (struct sccs, struct scc_entry): New types.
14263 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14264 (DFS_write_tree_body): New function.
14265 (DFS_write_tree): Likewise.
14266 (hash_tree): Likewise.
14267 (scc_entry_compare): Likewise.
14268 (hash_scc): Likewise.
14269 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14270 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14271 TREE_CHAIN as regular reference.
14272 (streamer_read_integer_cst): Remove.
14273 (streamer_get_pickled_tree): Adjust.
14274 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14275 of DECL_EXTERNALs in BLOCK_VARS for now.
14276 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14277 reference.
14278 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14279 Add hash value argument and record that if hashes are recorded
14280 in the cache.
14281 (streamer_tree_cache_insert_1): Adjust.
14282 (streamer_tree_cache_insert): Likewise.
14283 (streamer_tree_cache_insert_at): Rename to ...
14284 (streamer_tree_cache_replace_tree): ... this and adjust.
14285 (streamer_tree_cache_append): Adjust.
14286 (record_common_node): Likewise.
14287 (streamer_tree_cache_create): Add argument whether to
14288 record hash values together with trees.
14289 (streamer_tree_cache_delete): Adjust.
14290 * tree-streamer.h (struct streamer_tree_cache_d): Add
14291 vector of hashes.
14292 (streamer_read_integer_cst): Remove.
14293 (streamer_tree_cache_insert): Adjust.
14294 (streamer_tree_cache_append): Likewise.
14295 (streamer_tree_cache_insert_at): Rename to ...
14296 (streamer_tree_cache_replace_tree): ... this and adjust.
14297 (streamer_tree_cache_create): Add argument whether to record hashes.
14298 (streamer_tree_cache_get): Rename to ...
14299 (streamer_tree_cache_get_tree): ... this.
14300 (streamer_tree_cache_get_hash): New function.
14301 * tree.c (cache_integer_cst): New function.
14302 * tree.h (cache_integer_cst): Declare.
14303 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14304 * lto-symtab.c (lto_varpool_replace_node): Only release
14305 DECL_INITIAL of non-prevailing decls.
14306 * varpool.c (varpool_remove_initializer): Do not release
14307 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14308
14309 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14310
14311 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14312 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14313 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14314 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14315 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14316 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14317 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14318 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14319 instead of TARGET_64BIT.
14320 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14321 Require ISA_HAS_<D>DIV.
14322
14323 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14324
14325 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14326 (mips*-*-linux*): Move default with_llsc setting to where other
14327 defaults are set.
14328 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14329 with_arch block.
14330 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14331 Likewise. Remove default with_tune setting. Move default float
14332 setting to its own block. Handle with_llsc in the same block as above.
14333
14334 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14335
14336 PR rtl-optimization/57425
14337 PR rtl-optimization/57569
14338 * alias.c (write_dependence_p): Add new parameters mem_mode,
14339 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14340 Changed all callers.
14341 (canon_anti_dependence): New function.
14342 * cse.c (check_dependence): Use canon_anti_dependence.
14343 * cselib.c (cselib_invalidate_mem): Likewise.
14344 * rtl.h (canon_anti_dependence): Declare.
14345
14346 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14347
14348 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14349 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14350 ".set mips3" for 64-bit targets.
14351
14352 2013-06-15 Dehao Chen <dehao@google.com>
14353
14354 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14355 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14356 (gimple_check_call_args): Likewise.
14357 * value-prof.c (check_ic_target): Likewise.
14358 * ipa-inline.c (early_inliner): Likewise.
14359 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14360 * cgraph.c (cgraph_create_edge_1): Likewise.
14361 (cgraph_make_edge_direct): Likewise.
14362
14363 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14364
14365 PR target/57615
14366 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14367 rs6000_output_move_128bit to handle emitting quad memory
14368 operations. Set attribute length to 8 bytes.
14369
14370 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14371
14372 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14373 New pattern.
14374 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14375 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14376 (aarch64_<su>mlsl<mode>): Likewise.
14377
14378 2013-06-14 Mike Stump <mikestump@comcast.net>
14379
14380 * Makefile.in (TARGET_H): Add insn-codes.h.
14381
14382 2013-06-14 Alan Modra <amodra@gmail.com>
14383
14384 PR middle-end/57134
14385 PR middle-end/57586
14386 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14387 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14388 bitfield expansion when EXPAND_MEMORY.
14389 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14390
14391 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14392
14393 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14394 test for clearing quad memory on 32-bit later.
14395
14396 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14397
14398 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14399 (fold_negate_expr): Likewise.
14400 (fold_real_zero_addition_p): Handle vectors.
14401 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14402
14403 2013-06-14 Alan Modra <amodra@gmail.com>
14404
14405 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14406
14407 2013-06-13 Jan Hubicka <jh@suse.cz>
14408
14409 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14410 Local comdats are not externally visible.
14411 * symtab.c (dump_symtab_base): Dump externally visible.
14412 (verify_symtab_base): Verify back links in the symtab hash.
14413
14414 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14415
14416 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14417 CONVERT_EXPR as equal nodes.
14418
14419 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14420
14421 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14422
14423 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14424
14425 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14426 Generalize to complex and vector.
14427 * tree.c (build_all_ones_cst): New function.
14428 * tree.h (build_all_ones_cst): Declare it.
14429
14430 2013-06-13 Alan Modra <amodra@gmail.com>
14431
14432 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14433 * config/rs6000/rs6000.md (signbittf2): New insn.
14434 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14435 (abstf2_internal, cmptf_internal2): Likewise.
14436 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14437
14438 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14439 Pat Haugen <pthaugen@us.ibm.com>
14440 Peter Bergner <bergner@vnet.ibm.com>
14441
14442 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14443 power8 byte, half-word, and quad-word atomic instructions.
14444 (emit_store_conditional): Likewise.
14445 (rs6000_expand_atomic_compare_and_swap): Likewise.
14446 (rs6000_expand_atomic_op): Likewise.
14447
14448 * config/rs6000/sync.md (larx): Add new modes for power8.
14449 (stcx): Likewise.
14450 (AINT): New mode iterator to include TImode as well as normal
14451 integer modes on power8.
14452 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14453 that VSX registers are not considered. Use AINT mode iterator
14454 instead of INT1 to allow inclusion of quad word atomic operations
14455 on power8.
14456 (load_locked<mode>): Likewise.
14457 (store_conditional<mode>): Likewise.
14458 (atomic_compare_and_swap<mode>): Likewise.
14459 (atomic_exchange<mode>): Likewise.
14460 (atomic_nand<mode>): Likewise.
14461 (atomic_fetch_<fetchop_name><mode>): Likewise.
14462 (atomic_nand_fetch<mode>): Likewise.
14463 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14464 each type.
14465 (ATOMIC): On power8, add QImode, HImode modes.
14466 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14467 modes that promote to SImode.
14468 (load_lockedti): Convert TImode arguments to PTImode, so that we
14469 get a guaranteed even/odd register pair.
14470 (load_lockedpti): Likewise.
14471 (store_conditionalti): Likewise.
14472 (store_conditionalpti): Likewise.
14473
14474 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14475 atomic load/store instructions.
14476 (HSI): Likewise.
14477
14478 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14479
14480 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14481 loads.
14482 (insn_count): New attribute, with most cases extracted from...
14483 (length): ...here. Redefine most cases in terms of insn_count.
14484 (single_insn): Delete.
14485 (can_delay): Use insn_count to check for single instructions.
14486 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14487 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14488 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14489 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14490 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14491 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14492 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14493 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14494 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14495 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14496 rather than "length".
14497 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14498 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14499 Use "insn_count" rather than "length".
14500 * config/mips/mips-dsp.md
14501 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14502 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14503 length attributes.
14504
14505 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14506
14507 PR tree-optimization/57361
14508 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14509
14510 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14511
14512 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14513 to split.
14514 (aarch64_simd_combine<mode>): New instruction expansion.
14515 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14516 function prototype.
14517 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14518 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14519
14520 2013-06-12 Jan Hubicka <jh@suse.cz>
14521
14522 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14523 decl has when in streaming stage.
14524 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14525 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14526
14527 2013-06-12 Roland Stigge <stigge@antcom.de>
14528
14529 PR target/57578
14530 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14531
14532 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14533
14534 PR tree-optimization/57537
14535 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14536 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14537
14538 2013-06-12 Richard Biener <rguenther@suse.de>
14539
14540 * data-streamer.h (streamer_write_char_stream): CSE
14541 obs->current_pointer.
14542 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14543 streamer_write_char_stream manually and optimize the resulting loop.
14544 (streamer_write_hwi_stream): Likewise.
14545
14546 2013-06-12 Jan Hubicka <jh@suse.cz>
14547
14548 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14549 * cgraph.h (varpool_create_empty_node): Declare.
14550 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14551 duplicated nodes.
14552 * symtab.c (symtab_unregister_node): Be lax about missin entries
14553 in node hash.
14554 (symtab_get_node): Update comment.
14555 * varpool.c (varpool_create_empty_node): Break out from ...
14556 (varpool_node_for_decl): ... here.
14557 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14558
14559 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14560
14561 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14562 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14563 part. Use straight-line flow at the end.
14564 <COMPONENT_REF>: Remove superfluous else.
14565 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14566
14567 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14568
14569 PR target/56564
14570 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14571 target hook even for !TREE_PUBLIC decls. If no resolution info
14572 is available, return false for common and external decls.
14573
14574 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14575
14576 * config/rl78/constraints.md (U): New constraint.
14577 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14578 valloc attribute.
14579
14580 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14581
14582 PR target/57589
14583 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14584 to allow returning address to AT_PLATFORM name.
14585
14586 2013-06-11 Jan Hubicka <jh@suse.cz>
14587
14588 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14589 * cgraph.h (symtab_node_base): Add weakref flag.
14590 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14591 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14592 (output_weakrefs): Use weakref flag.
14593 * fold-const.c (simple_operand_p): Handle WEAK.
14594 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14595 * ipa.c (varpool_externally_visible_p): Drop weakref.
14596 (function_and_variable_visibility): Update comment; fix weakref
14597 sanity checks; do not clear DECL_WEAK on them.
14598 * lto-cgraph.c (lto_output_node): update.
14599 (lto_output_varpool_node): Update.
14600 (input_overwrite_node): Update.
14601 (input_node): Update.
14602 (input_varpool_node): Update.
14603 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14604 (lto_symtab_merge_symbols): Add sanity check.
14605 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14606 * passes.c (rest_of_decl_compilation): Set static flag, too.
14607 * symtab.c (dump_symtab_base): Dump weakref.
14608 (verify_symtab_base): Sanity check weakrefs.
14609 (symtab_make_decl_local): Remove duplicated code.
14610 (symtab_alias_ultimate_target): Simplify.
14611 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14612
14613 2013-06-11 Tom de Vries <tom@codesourcery.com>
14614
14615 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14616 sequence_vect == NULL.
14617
14618 2013-06-11 DJ Delorie <dj@redhat.com>
14619
14620 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14621 (rl78_unwind_word_mode): New.
14622
14623 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14624
14625 * final.c (debug_prefix_maps): Make static.
14626
14627 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14628
14629 * function.c (initial_trampoline): Remove stray copy.
14630
14631 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14632
14633 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14634
14635 2013-06-11 Martin Jambor <mjambor@suse.cz>
14636
14637 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14638 within bounds at the beginning of the function.
14639
14640 2013-06-11 Alan Modra <amodra@gmail.com>
14641
14642 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14643 reporting.
14644 (get_named_section): Don't NULL !DECL_P decl.
14645
14646 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14647
14648 * doc/invoke.texi (core-avx2): Document.
14649 (slm): Likewise.
14650 (atom): Updated with MOVBE.
14651
14652 2013-06-11 Richard Biener <rguenther@suse.de>
14653
14654 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14655
14656 2013-06-11 Anton Blanchard <anton@samba.org>
14657
14658 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14659 correct shift value in little-endian mode.
14660
14661 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14662
14663 PR target/56564
14664 * varasm.c (get_variable_align): Move #endif to the right place.
14665
14666 2013-06-10 Cary Coutant <ccoutant@google.com>
14667
14668 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14669 for hash so that hash table traversal order is deterministic.
14670
14671 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14672 Pat Haugen <pthaugen@us.ibm.com>
14673 Peter Bergner <bergner@vnet.ibm.com>
14674
14675 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14676 of vectors in GPRS if they are direct moves or quad word load or
14677 store moves.
14678
14679 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14680 declaration.
14681 (direct_move_p): Likewise.
14682 (quad_load_store_p): Likewise.
14683
14684 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14685 classes into bins based on the physical register type.
14686 (reg_class_to_reg_type): Likewise.
14687 (IS_STD_REG_TYPE): Likewise.
14688 (IS_FP_VECT_REG_TYPE): Likewise.
14689 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14690 use direct move instructions.
14691 (reload_gpr_vsx): Likewise.
14692 (reload_vsx_gpr): Likewise.
14693 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14694 information that is a simplification of register classes. Also
14695 precalculate direct move reload helpers.
14696 (direct_move_p): New function to return true if the operation can
14697 be done as a direct move instruciton.
14698 (quad_load_store_p): New function to return true if the operation
14699 is a quad memory operation.
14700 (rs6000_legitimize_address): If quad memory, only allow register
14701 indirect for TImode addresses.
14702 (rs6000_legitimate_address_p): Likewise.
14703 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14704 (rs6000_reload_register_type): Likewise.
14705 (register_to_reg_type): Return register type.
14706 (rs6000_secondary_reload_simple_move): New helper function for
14707 secondary reload and secondary memory needed to identify anything
14708 that is a simple move, and does not need reloading.
14709 (rs6000_secondary_reload_direct_move): New helper function for
14710 secondary reload to identify cases that can be done with several
14711 instructions via the direct move instructions.
14712 (rs6000_secondary_reload_move): New helper function for secondary
14713 reload to identify moves between register types that can be done.
14714 (rs6000_secondary_reload): Add support for quad memory operations
14715 and for direct move.
14716 (rs6000_secondary_memory_needed): Likewise.
14717 (rs6000_debug_secondary_memory_needed): Change argument names.
14718 (rs6000_output_move_128bit): New function to return the move to
14719 use for 128-bit moves, including knowing about the various
14720 limitations of quad memory operations.
14721
14722 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14723 memory operations. call rs6000_output_move_128bit for the actual
14724 instruciton(s) to generate.
14725 (vsx_movti_64bit): Likewise.
14726
14727 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14728 (UNSPEC_P8V_MTVSRWZ): Likewise.
14729 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14730 (UNSPEC_P8V_MTVSRD): Likewise.
14731 (UNSPEC_P8V_XXPERMDI): Likewise.
14732 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14733 (UNSPEC_FUSION_GPR): Likewise.
14734 (FMOVE128_GPR): New iterator for direct move.
14735 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14736 (f32_sv): Likewise.
14737 (f32_dm): Likewise.
14738 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14739 loads and direct move instructions.
14740 (zero_extendsidi2_lfiwzx): Likewise.
14741 (extendsidi2_lfiwax): Likewise.
14742 (extendsidi2_nocell): Likewise.
14743 (floatsi<mode>2_lfiwax): Likewise.
14744 (lfiwax): Likewise.
14745 (floatunssi<mode>2_lfiwzx): Likewise.
14746 (lfiwzx): Likewise.
14747 (fix_trunc<mode>_stfiwx): Likewise.
14748 (fixuns_trunc<mode>_stfiwx): Likewise.
14749 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14750 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14751 (parity<mode>2_cmpb): Set length/type attr.
14752 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14753 for 'mr.' to fast_compare.
14754 (bpermd_<mode>): Change type attr to popcnt.
14755 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14756 (p8_mtvsrwz_1): Likewise.
14757 (p8_mtvsrwz_2): Likewise.
14758 (reload_fpr_from_gpr<mode>): Likewise.
14759 (p8_mtvsrd_1): Likewise.
14760 (p8_mtvsrd_2): Likewise.
14761 (p8_xxpermdi_<mode>): Likewise.
14762 (reload_vsx_from_gpr<mode>): Likewise.
14763 (reload_vsx_from_gprsf): Likewise.
14764 (p8_mfvsrd_3_<mode>): LIkewise.
14765 (reload_gpr_from_vsx<mode>): Likewise.
14766 (reload_gpr_from_vsxsf): Likewise.
14767 (p8_mfvsrd_4_disf): Likewise.
14768 (multi-word GPR splits): Do not split direct moves or quad memory
14769 operations.
14770
14771 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14772
14773 * tree-into-ssa.c (interesting_blocks): Make static.
14774
14775 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14776
14777 PR target/56564
14778 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14779 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14780 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14781 (get_variable_align): New function.
14782 (get_variable_section, emit_bss, emit_common,
14783 assemble_variable_contents, place_block_symbol): Use
14784 get_variable_align instead of DECL_ALIGN.
14785 (assemble_noswitch_variable): Add align argument, use it
14786 instead of DECL_ALIGN.
14787 (assemble_variable): Adjust caller. Use get_variable_align
14788 instead of DECL_ALIGN.
14789 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14790 caller.
14791 (DATA_ABI_ALIGNMENT): Define.
14792 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14793 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14794 opt is false, only return the psABI mandated alignment increase.
14795 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14796 (DATA_ABI_ALIGNMENT): ... this.
14797 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14798 (DATA_ABI_ALIGNMENT): ... this.
14799 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14800 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14801 (DATA_ABI_ALIGNMENT): ... this.
14802 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14803 * doc/tm.texi: Regenerated.
14804
14805 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14806
14807 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14808 cmp_code to construct REG_EQUAL note.
14809
14810 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14811
14812 PR target/57568
14813 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14814 that operands[2] doesn't overlap with operands[0].
14815
14816 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14817 Jan Hubicka <jh@suse.cz>
14818
14819 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14820 hack to mark symbols as used.
14821
14822 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14823
14824 PR rtl-optimization/57559
14825 * lra-constraints.c (process_alt_operands): Don't discourage
14826 memory with known offset for offsetable memory constraint.
14827 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14828
14829 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14830
14831 * varasm.c (struct oc_local_state): Reorder fields.
14832 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14833 and adjust accordingly.
14834 (output_constructor): Reorder initialization code and adjust call to
14835 output_constructor_bitfield.
14836
14837 2013-06-07 Jan Hubicka <jh@suse.cz>
14838
14839 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14840
14841 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14842
14843 * tree-object-size.c (unknown): Make const.
14844
14845 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14846
14847 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14848 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14849 for last alternative in the cpu_facility attribute.
14850
14851 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14852
14853 PR target/56315
14854 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14855 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14856 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14857 * config/arm/constraints.md (Dg): New constraint.
14858 * config/arm/neon.md (xordi3_neon): Remove.
14859 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14860 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14861
14862 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14863
14864 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14865 Clean up alternatives.
14866
14867 2013-06-07 Alan Modra <amodra@gmail.com>
14868
14869 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14870 va_list_gpr_size.
14871
14872 2013-06-07 Alan Modra <amodra@gmail.com>
14873
14874 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14875
14876 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14877
14878 * config/arm/constraints.md (Df): New constraint.
14879 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14880 Correct length attribute for last two alternatives.
14881
14882 2013-06-07 Alan Modra <amodra@gmail.com>
14883
14884 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14885 override user -mfp-in-toc.
14886 (offsettable_ok_by_alignment): Consider just the current access
14887 rather than the whole object, unless BLKmode. Handle
14888 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14889 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14890 for -mcmodel=medium.
14891 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14892 override user -mfp-in-toc or -msum-in-toc. Default to
14893 -mno-fp-in-toc for -mcmodel=medium.
14894
14895 2013-06-06 DJ Delorie <dj@redhat.com>
14896
14897 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14898 TARGET_VALID_POINTER_MODE.
14899
14900 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14901 Pat Haugen <pthaugen@us.ibm.com>
14902 Peter Bergner <bergner@vnet.ibm.com>
14903
14904 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14905 Document new power8 builtins.
14906
14907 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14908 condition code register, to allow 128-bit logical operations to be
14909 done in the VSX or GPR registers.
14910 (nor<mode>3): Use the canonical form for nor.
14911 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14912 vclz*, and vpopcnt* vector instructions.
14913 (nand<mode>3): Likewise.
14914 (orc<mode>3): Likewise.
14915 (clz<mode>2): LIkewise.
14916 (popcount<mode>2): Likewise.
14917
14918 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14919 that only the GPRs are recognized.
14920
14921 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14922 support for new power8 builtins.
14923
14924 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14925 builtin functions.
14926 (xscvdpspn): Likewise.
14927 (vclz): Likewise.
14928 (vclzb): Likewise.
14929 (vclzh): Likewise.
14930 (vclzw): Likewise.
14931 (vclzd): Likewise.
14932 (vpopcnt): Likewise.
14933 (vpopcntb): Likewise.
14934 (vpopcnth): Likewise.
14935 (vpopcntw): Likewise.
14936 (vpopcntd): Likewise.
14937 (vgbbd): Likewise.
14938 (vmrgew): Likewise.
14939 (vmrgow): Likewise.
14940 (eqv): Likewise.
14941 (eqv_v16qi3): Likewise.
14942 (eqv_v8hi3): Likewise.
14943 (eqv_v4si3): Likewise.
14944 (eqv_v2di3): Likewise.
14945 (eqv_v4sf3): Likewise.
14946 (eqv_v2df3): Likewise.
14947 (nand): Likewise.
14948 (nand_v16qi3): Likewise.
14949 (nand_v8hi3): Likewise.
14950 (nand_v4si3): Likewise.
14951 (nand_v2di3): Likewise.
14952 (nand_v4sf3): Likewise.
14953 (nand_v2df3): Likewise.
14954 (orc): Likewise.
14955 (orc_v16qi3): Likewise.
14956 (orc_v8hi3): Likewise.
14957 (orc_v4si3): Likewise.
14958 (orc_v2di3): Likewise.
14959 (orc_v4sf3): Likewise.
14960 (orc_v2df3): Likewise.
14961
14962 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14963 allow power8 quad mode in 64-bit.
14964 (rs6000_builtin_vectorized_function): Add support to vectorize
14965 ISA 2.07 count leading zeros, population count builtins.
14966 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14967 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14968 (builtin_function_type): Add vgbbd builtin function which takes an
14969 unsigned argument.
14970 (altivec_expand_vec_perm_const): Add support for new power8 merge
14971 instructions.
14972
14973 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14974 that does not include TImdoe for use with 32-bit.
14975 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14976 instructions.
14977 (UNSPEC_VSX_CVDPSPN): Likewise.
14978 (vsx_xscvdpspn): Likewise.
14979 (vsx_xscvspdpn): Likewise.
14980 (vsx_xscvdpspn_scalar): Likewise.
14981 (vsx_xscvspdpn_directmove): Likewise.
14982 (vsx_and<mode>3): Split logical operations into 32-bit and
14983 64-bit. Add support to do logical operations on TImode as well as
14984 VSX vector types. Allow logical operations to be done in either
14985 VSX registers or in general purpose registers in 64-bit mode. Add
14986 splitters if GPRs were used. For AND, add clobber of CCmode to
14987 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14988 encoding.
14989 (vsx_and<mode>3_32bit): Likewise.
14990 (vsx_and<mode>3_64bit): Likewise.
14991 (vsx_ior<mode>3): Likewise.
14992 (vsx_ior<mode>3_32bit): Likewise.
14993 (vsx_ior<mode>3_64bit): Likewise.
14994 (vsx_xor<mode>3): Likewise.
14995 (vsx_xor<mode>3_32bit): Likewise.
14996 (vsx_xor<mode>3_64bit): Likewise.
14997 (vsx_one_cmpl<mode>2): Likewise.
14998 (vsx_one_cmpl<mode>2_32bit): Likewise.
14999 (vsx_one_cmpl<mode>2_64bit): Likewise.
15000 (vsx_nor<mode>3): Likewise.
15001 (vsx_nor<mode>3_32bit): Likewise.
15002 (vsx_nor<mode>3_64bit): Likewise.
15003 (vsx_andc<mode>3): Likewise.
15004 (vsx_andc<mode>3_32bit): Likewise.
15005 (vsx_andc<mode>3_64bit): Likewise.
15006 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15007 and xxlorc instructions.
15008 (vsx_eqv<mode>3_64bit): Likewise.
15009 (vsx_nand<mode>3_32bit): Likewise.
15010 (vsx_nand<mode>3_64bit): Likewise.
15011 (vsx_orc<mode>3_32bit): Likewise.
15012 (vsx_orc<mode>3_64bit): Likewise.
15013
15014 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15015
15016 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15017 instruction.
15018 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15019 (p8_vmrgow): Likewise.
15020 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15021 GPRs to be split under VSX.
15022 (p8v_clz<mode>2): Add power8 count leading zero support.
15023 (p8v_popcount<mode>2): Add power8 population count support.
15024 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15025 support.
15026
15027 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15028 instruction.
15029
15030 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15031 builtin functions.
15032 (vec_nand): Likewise.
15033 (vec_vclz): Likewise.
15034 (vec_vclzb): Likewise.
15035 (vec_vclzd): Likewise.
15036 (vec_vclzh): Likewise.
15037 (vec_vclzw): Likewise.
15038 (vec_vgbbd): Likewise.
15039 (vec_vmrgew): Likewise.
15040 (vec_vmrgow): Likewise.
15041 (vec_vpopcnt): Likewise.
15042 (vec_vpopcntb): Likewise.
15043 (vec_vpopcntd): Likewise.
15044 (vec_vpopcnth): Likewise.
15045 (vec_vpopcntw): Likewise.
15046
15047 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15048
15049 PR rtl-optimization/57468
15050 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15051 spilled pseudos.
15052
15053 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15054
15055 PR rtl-optimization/57459
15056 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15057 type when setting live regs.
15058
15059 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15060
15061 * config/s390/s390.opt (mlra): New option.
15062 * config/s390/s390.c (s390_decompose_address): Check displacement
15063 for all registers for LRA.
15064 (s390_secondary_reload): Don't used secondary reloads for LRA.
15065 (s390_lra_p): New function.
15066 (TARGET_LRA_P): Define.
15067 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15068 of attribute cpu_facility to zarch for the last alternative.
15069 (*cmpmem_short): Ditto.
15070
15071 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15072
15073 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15074 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15075 (arm_expand_prologue): Likewise.
15076
15077 2013-06-06 Teresa Johnson <tejohnson@google.com>
15078
15079 PR c++/53743
15080 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15081 as this is now done by redirect_edge_and_branch_force.
15082 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15083 barriers, and fix interaction with splitting.
15084 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15085 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15086 reflect changes made in the routine.
15087 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15088 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15089 since this is called in cfglayout mode, and replace partition fixup
15090 with assert as that is now done by force_nonfallthru_and_redirect.
15091 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15092 already be marked with region crossing note.
15093 (insert_section_boundary_note): Make non-static, gate on flag
15094 has_bb_partition, rewrite to also check for multiple partitions.
15095 (rest_of_handle_reorder_blocks): Remove call to
15096 insert_section_boundary_note, now done later during free_cfg.
15097 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15098 * bb-reorder.h (insert_section_boundary_note): Declare.
15099 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15100 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15101 invoke insert_section_boundary_note.
15102 (try_redirect_by_replacing_jump): Remove unnecessary
15103 check for region crossing note.
15104 (fixup_partition_crossing): New function.
15105 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15106 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15107 in non-cfglayout mode.
15108 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15109 remove old code that tried to do this. Emit barrier correctly
15110 when we are in cfglayout mode.
15111 (last_bb_in_partition): New function.
15112 (rtl_split_edge): Correctly fixup partition boundaries.
15113 (commit_one_edge_insertion): Remove old code that tried to
15114 fixup region crossing edge since this is now handled in
15115 split_block, and set up insertion point correctly since
15116 block may now end in a jump.
15117 (verify_hot_cold_block_grouping): Guard against checking when not in
15118 linearized RTL mode.
15119 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15120 notes.
15121 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15122 rtl_verify_flow_info, so not called in cfglayout mode.
15123 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15124 (fixup_reorder_chain): Remove old code that attempted to fixup region
15125 crossing note as this is now handled in force_nonfallthru_and_redirect.
15126 (duplicate_insn_chain): Don't duplicate switch section notes.
15127 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15128 note.
15129 * basic-block.h (emit_barrier_after_bb): Declare.
15130
15131 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15132
15133 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15134 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15135 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15136
15137 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15138
15139 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15140 where appropriate.
15141 * config/arm/ldmstm.md: Regenerate.
15142
15143 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15144
15145 * config/arm/sync.md (atomic_loaddi_1):
15146 Disable predication for arm_restrict_it.
15147 (arm_load_exclusive<mode>): Likewise.
15148 (arm_load_exclusivesi): Likewise.
15149 (arm_load_exclusivedi): Likewise.
15150 (arm_load_acquire_exclusive<mode>): Likewise.
15151 (arm_load_acquire_exclusivesi): Likewise.
15152 (arm_load_acquire_exclusivedi): Likewise.
15153 (arm_store_exclusive<mode>): Likewise.
15154 (arm_store_exclusive<mode>): Likewise.
15155 (arm_store_release_exclusivedi): Likewise.
15156 (arm_store_release_exclusive<mode>): Likewise.
15157
15158 2013-06-06 Richard Biener <rguenther@suse.de>
15159
15160 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15161 after LTO_null.
15162 (lto_tag_is_tree_code_p): Adjust.
15163 (lto_tag_is_gimple_code_p): Likewise.
15164 (lto_gimple_code_to_tag): Likewise.
15165 (lto_tag_to_gimple_code): Likewise.
15166 (lto_tree_code_to_tag): Likewise.
15167 (lto_tag_to_tree_code): Likewise.
15168 * data-streamer.h (streamer_write_hwi_in_range): Use
15169 uhwi streaming to stream the normalized range.
15170 (streamer_read_hwi_in_range): Likewise.
15171
15172 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15173
15174 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15175 (predicable_short_it): Likewise.
15176 (predicated): Likewise.
15177 (enabled): Handle above.
15178 (define_cond_exec): Set predicated attribute to yes.
15179
15180 2013-06-05 Mike Stump <mikestump@comcast.net>
15181
15182 * gdbinit.in (__FUNCTION__): Add.
15183
15184 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15185
15186 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15187 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15188
15189 2013-06-05 Jan Hubicka <jh@suse.cz>
15190
15191 * varasm.c (mark_decl_referenced): Revert the removal until targets
15192 are fixed.
15193
15194 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15195
15196 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15197 instead of mark_decl_referenced.
15198
15199 2013-06-05 Jan Hubicka <jh@suse.cz>
15200
15201 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15202 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15203 and symtab_used_from_object_file_p.
15204 (cgraph_make_node_local_1): Clear forced_by_abi.
15205 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15206 * cgraph.h (symtab_node_base): Add forced_by_abi.
15207 (decide_is_variable_needed): Remove.
15208 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15209 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15210 (decide_is_symbol_needed): ... this one; handle symbols in general;
15211 always analyze virtuals; honnor forced_by_abi.
15212 (cgraph_finalize_function): Update.
15213 (varpool_finalize_decl): Update.
15214 (symbol_defined_and_needed): Remove.
15215 (analyze_functions): Update.
15216 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15217 output_refs, input_overwrite_node): Handle forced_by_abi.
15218 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15219 (address_taken_from_non_vtable_p): ... this one.
15220 (comdat_can_be_unshared_p_1): New function.
15221 (cgraph_comdat_can_be_unshared_p): Rename to ...
15222 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15223 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15224 (function_and_variable_visibility): Clear forced_by_abi as needed.
15225 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15226 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15227 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15228 * varpool.c (decide_is_variable_needed): Remove.
15229
15230 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15231
15232 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15233 (arm_option_override): Override arm_restrict_it where appropriate.
15234 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15235 * config/arm/arm.opt (mrestrict-it): New command-line option.
15236 * doc/invoke.texi: Document -mrestrict-it.
15237
15238 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15239
15240 * tsan.c (tsan_atomic_table): Make const.
15241
15242 2013-06-05 Richard Biener <rguenther@suse.de>
15243
15244 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15245 index associated with the tree we are supposed to replace.
15246 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15247 TREE_ASM_WRITTEN as zero for everything but SSA names.
15248
15249 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15250
15251 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15252
15253 2013-06-04 Jan Hubicka <jh@suse.cz>
15254
15255 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15256 (input_node, input_varpool_node): Handle correctly external same
15257 body aliases.
15258 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15259 nodes at ltrans stage.
15260
15261 2013-06-04 Jan Hubicka <jh@suse.cz>
15262
15263 * ipa-inline.c (update_caller_keys): Fix availability test.
15264 (update_callee_keys): Likewise.
15265 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15266 to follow ELF standard.
15267
15268 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15269
15270 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15271 (mips64r5900el-*-elf*): New configurations.
15272 * config/mips/mips-cpus.def (r5900): New processor.
15273 * config/mips/mips-tables.opt: Regenerate.
15274 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15275 (mips_issue_rate): Handle PROCESSOR_R5900.
15276 (mips_reorg_process_insns): Force reorder mode for the R5900.
15277 * config/mips/mips.h (TARGET_MIPS5900): Define.
15278 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15279 TARGET_MIPS5900.
15280 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15281 TARGET_MIPS5900.
15282 * config/mips/mips.md (processor): Add r5900.
15283 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15284
15285 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15286
15287 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15288 into function to generate MOVI instruction.
15289 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15290 (aarch64_preferred_simd_mode): Turn into wrapper.
15291 (aarch64_output_scalar_simd_mov_immediate): New function.
15292 * config/aarch64/aarch64-protos.h: Add prototype for above.
15293
15294 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15295
15296 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15297 element_char member.
15298 (sizetochar): Return signed char.
15299 (aarch64_simd_valid_immediate): Remove elchar and other
15300 unnecessary variables.
15301 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15302 Calculate element_char as required.
15303 * config/aarch64/aarch64-protos.h: Update and move prototype
15304 for aarch64_output_simd_mov_immediate.
15305 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15306 Update arguments.
15307
15308 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15309
15310 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15311 information completed by aarch64_simd_valid_immediate.
15312 (aarch64_legitimate_constant_p): Update arguments.
15313 (aarch64_simd_valid_immediate): Work with struct rather than many
15314 pointers.
15315 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15316 (aarch64_simd_make_constant): Update arguments.
15317 (aarch64_output_simd_mov_immediate): Work with struct rather than
15318 many pointers. Output immediate directly rather than as operand.
15319 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15320 Update prototype.
15321 * config/aarch64/constraints.md (Dn): Update arguments.
15322
15323 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15324
15325 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15326 longer static.
15327 (aarch64_simd_immediate_valid_for_move): Remove.
15328 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15329 (aarch64_simd_make_constant): Update call.
15330 (aarch64_output_simd_mov_immediate): Update call.
15331 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15332 Add prototype.
15333 * config/aarch64/constraints.md (Dn): Update call.
15334
15335 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15336
15337 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15338 return type to bool for prototype.
15339 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15340 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15341 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15342
15343 2013-06-04 Catherine Moore <clm@codesourcery.com>
15344
15345 * config/mips/mips.opt (meva): New.
15346 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15347 (ASM_SPEC): Handle -meva.
15348 * doc/invoke.texi (meva): Document.
15349
15350 2013-06-04 Alan Modra <amodra@gmail.com>
15351
15352 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15353 constant output.
15354
15355 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15356
15357 * rtl.def: Add extra fourth optional field to define_cond_exec.
15358 * gensupport.c (process_one_cond_exec): Process attributes from
15359 define_cond_exec.
15360 * doc/md.texi: Document fourth field in define_cond_exec.
15361
15362 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15363
15364 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15365 out the processing order as in store_bit_field_1.
15366
15367 2013-06-04 Jan Hubicka <jh@suse.cz>
15368
15369 PR middle-end/57500
15370 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15371 non-VAR_DECL node if it does not exist yet.
15372
15373 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15374
15375 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15376 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15377 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15378 target_cpu_default setting.
15379
15380 2013-06-03 Teresa Johnson <tejohnson@google.com>
15381
15382 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15383 default to -fopt-info=optimized instead of all.
15384 * doc/invoke.texi: Ditto.
15385 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15386 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15387 (execute_vect_slp): Emit BB vectorization success under
15388 MSG_OPTIMIZED_LOCATIONS.
15389 * tree-vect-slp.c (vect_slp_transform_bb): Change
15390 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15391 * tree-vect-loop.c (vect_transform_loop): Ditto.
15392
15393 2013-06-03 Jason Merrill <jason@redhat.com>
15394
15395 PR c++/57415
15396 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15397 Use TARGET_EXPR for C++.
15398
15399 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15400
15401 PR rtl-optimization/57268
15402 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15403 if DEBUG_INSN_P (insn).
15404
15405 Reapply
15406 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15407
15408 PR rtl-optimization/57268
15409 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15410 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15411
15412 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15413
15414 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15415 (ix86_avoid_lea_for_addr): Likewise.
15416 (exact_dependency_1): Likewise.
15417 (ix86_adjust_cost): Likewise.
15418 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15419 removed.
15420 (do_reorder_for_imul): Fix typo, formatting and
15421 !reload_completed check removed.
15422 (ix86_sched_reorder): Fix typo and formatting.
15423 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15424 list.
15425
15426 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15427
15428 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15429
15430 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15431
15432 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15433 <STRING_CST>: Likewise.
15434 <VECTOR_CST>: Likewise.
15435
15436 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15437 Mikael Morin <mikael@gcc.gnu.org>
15438
15439 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15440 * config.in: Regenerated.
15441 * configure: Regenerated.
15442
15443 2013-06-01 Jan Hubicka <jh@suse.cz>
15444
15445 PR middle-end/57366
15446 * cgraphunit.c (compile): When weakref is not supported,
15447 set up transparent aliases before final output pass.
15448 * varasm.c (assemble_alias): Do not try to do it here.
15449
15450 2013-06-01 Jan Hubicka <jh@suse.cz>
15451
15452 PR middle-end/57467
15453 * passes.c (for_per_function): Skip unanalyzed functions.
15454
15455 2013-06-01 Jan Hubicka <jh@suse.cz>
15456
15457 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15458 (lto_symtab_merge_symbols_1): ... this one.
15459 (lto_symtab_merge_cgraph_nodes): Rename to ...
15460 (lto_symtab_merge_symbols): ... this one; simplify.
15461 * cgraph.c (same_body_aliases_done): Rename to ...
15462 (cpp_implicit_aliases_done): ... this one.
15463 (cgraph_create_function_alias): Update.
15464 (cgraph_same_body_alias): Update.
15465 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15466 (verify_edge_corresponds_to_fndecl): Simplify.
15467 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15468 (cgraph_node): Remove same_body_alias.
15469 (varpool_node): Remove alias_of and extra_name_alias.
15470 (same_body_aliases_done): Rename to ..
15471 (cpp_implicit_aliases_done): ... this one.
15472 (symtab_alias_ultimate_target): Add default parameter.
15473 (symtab_resolve_alias): New function.
15474 (fixup_same_cpp_alias_visibility): Declare.
15475 (cgraph_function_node): Add default parameter.
15476 (cgraph_node_asm_name): Likewise.
15477 (cgraph_function_or_thunk_node): Add default parameter; do
15478 not ICE when it is NULL.
15479 (varpool_variable_node): Likewise.
15480 * tree-emutls.c (create_emultls_var): Update.
15481 (ipa_lower_emutls): Update.
15482 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15483 (cgraph_reset_node): Reset alias info.
15484 (cgraph_finalize_function): Update.
15485 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15486 (analyze_function): Simplify.
15487 (cgraph_process_same_body_aliases): Simplify.
15488 (analyze_functions): Fixup same body aliases.
15489 (handle_alias_pairs): Simplify.
15490 (assemble_thunk): Update.
15491 (assemble_thunks_and_aliases): Update.
15492 (output_weakrefs): Rewrite.
15493 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15494 (lto_output_varpool_node): Likewise.
15495 (compute_ltrans_boundary): Remve assert.
15496 (get_alias_symbol): New functoin.
15497 (input_node): Rewrite alias handling.
15498 (input_varpool_node): Likewise.
15499 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15500 * ipa.c (process_references): Handle weakrefs correctly.
15501 (symtab_remove_unreachable_nodes): Likewise.
15502 * trans-mem.c (get_cg_data): Update.
15503 (ipa_tm_create_version_alias): Update.
15504 (ipa_tm_execute): Update.
15505 * symtab.c (dump_symtab_base): Dump aliases.
15506 (verify_symtab_base): Verify aliases.
15507 (symtab_node_availability): New function.
15508 (symtab_alias_ultimate_target): Simplify.
15509 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15510 handle all the fixup cases.
15511 (symtab_resolve_alias): New function.
15512 * passes.c (ipa_write_summaries): Handle weakrefs.
15513 * varpool.c (varpool_analyze_node): Simplify.
15514 (assemble_aliases): Update.
15515 (varpool_create_variable_alias): Simplify.
15516 (varpool_extra_name_alias): Simplify.
15517 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15518 (lto_symtab_merge_symbols): ... this one.
15519
15520 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15521
15522 Revert
15523 PR rtl-optimization/57268
15524 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15525 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15526
15527 2013-06-01 Tobias Burnus <burnus@net-b.de>
15528
15529 Partially reverted:
15530 2013-05-31 Tobias Burnus <burnus@net-b.de>
15531
15532 PR middle-end/57073
15533 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15534 further up.
15535
15536 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15537
15538 PR rtl-optimization/57268
15539 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15540 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15541
15542 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15543
15544 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15545 unordered comparison operators when -fno-trapping-math is in effect
15546 on the e500.
15547 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15548 and implement unordered comparison operators properly on the e500.
15549
15550 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15551
15552 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15553 for constant scalar integers.
15554 (simplify_relational_operation_1): Likewise.
15555
15556 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15557
15558 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15559 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15560 Fix comment.
15561
15562 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15563 Igor Zamyatin <igor.zamyatin@intel.com>
15564
15565 Silvermont (SLM) architecture performance tuning.
15566 * config/i386/i386.h (enum ix86_tune_indices): Add
15567 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15568 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15569
15570 * config/i386/i386.c (initial_ix86_tune_features)
15571 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15572 (ix86_lea_outperforms): Handle Silvermont tuning.
15573 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15574 call.
15575 (ix86_use_lea_for_mov): Likewise.
15576 (ix86_avoid_lea_for_addr): Likewise.
15577 (ix86_lea_for_add_ok): Likewise.
15578 (exact_dependency_1): New function.
15579 (exact_store_load_dependency): Likewise.
15580 (ix86_adjust_cost): Handle Silvermont tuning.
15581 (do_reoder_for_imul): Likewise.
15582 (swap_top_of_ready_list): New function.
15583 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15584
15585 * config/i386/i386.md (peepholes that split memory operand in fp
15586 converts): New.
15587
15588 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15589
15590 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15591 Remove un-necessary braces.
15592
15593 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15594
15595 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15596 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15597
15598 2013-05-31 Tobias Burnus <burnus@net-b.de>
15599
15600 PR middle-end/57073
15601 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15602
15603 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15604
15605 PR target/56315
15606 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15607 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15608 * config/arm/neon.md (iordi3_neon): Remove.
15609 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15610 * config/arm/predicates.md (imm_for_neon_logic_operand):
15611 Move to earlier in the file.
15612 (neon_logic_op2): Likewise.
15613 (arm_iordi_operand_neon): New predicate.
15614
15615 2013-05-31 Richard Biener <rguenther@suse.de>
15616
15617 PR tree-optimization/57478
15618 PR tree-optimization/57453
15619 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15620 are life as well.
15621
15622 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15623
15624 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15625 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15626
15627 2013-05-30 Tobias Burnus <burnus@net-b.de>
15628 Thomas Koenig <tkoenig@gcc.gnu.org>
15629
15630 PR middle-end/57073
15631 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15632 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15633
15634 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15635
15636 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15637
15638 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15639
15640 * target.def (register_usage_leveling_p): New hook.
15641 * targhooks.c (default_register_usage_leveling_p): New.
15642 * targhooks.h (default_register_usage_leveling_p): New prototype.
15643 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15644 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15645 * doc/tm.texi: Update.
15646 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15647
15648 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15649
15650 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15651 (*insv_reg<mode>): New define_insn.
15652
15653 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15654
15655 PR rtl-optimization/57439
15656 * postreload.c (move2add_valid_value_p): Check that we have
15657 a zero subreg_regno_offset when accessing the register in
15658 the requested mode.
15659
15660 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15661 Igor Zamyatin <igor.zamyatin@intel.com>
15662
15663 Silvermont (SLM) architecture pipeline model, tuning and
15664 insn selection.
15665 * config.gcc: Add slm config options and target.
15666
15667 * config/i386/slm.md: New.
15668
15669 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15670
15671 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15672 PROCESSOR_SLM.
15673 (ix86_target_macros_internal): Likewise.
15674
15675 * config/i386/i386.c (slm_cost): New cost.
15676 (m_SLM): New macro flag.
15677 (initial_ix86_tune_features): Set m_SLM.
15678 (x86_accumulate_outgoing_args): Likewise.
15679 (x86_arch_always_fancy_math_387): Likewise.
15680 (processor_target_table): Add slm cost.
15681 (cpu_names): Add slm cpu name.
15682 (x86_option_override_internal): Set SLM ISA.
15683 (ix86_issue_rate): New case PROCESSOR_SLM.
15684 (ia32_multipass_dfa_lookahead): Likewise.
15685 (fold_builtin_cpu): Add slm.
15686
15687 * config/i386/i386.h (TARGET_SLM): New target macro.
15688 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15689 (processor_type): Add PROCESSOR_SLM.
15690
15691 * config/i386/i386.md (cpu): Add new value "slm".
15692 (slm.md): Include slm.md.
15693
15694 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15695 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15696
15697 * config/arm/arm-protos.h: Add and update function protos.
15698 * config/arm/arm.c (use_simple_return_p): New added.
15699 (thumb2_expand_return): Check simple_return flag.
15700 * config/arm/arm.md: Add simple_return and conditional simple_return.
15701 * config/arm/iterators.md: Add iterator for return and simple_return.
15702
15703 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15704
15705 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15706 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15707 (arm_emit_vfp_multi_reg_pop): Likewise.
15708 (thumb2_emit_ldrd_pop): Likewise.
15709 (arm_expand_epilogue): Add misc REG_CFA notes.
15710 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15711
15712 2013-05-29 Lawrence Crowl <crowl@google.com>
15713
15714 * config/arm/t-arm: Update for below.
15715
15716 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15717 Change type to hash_table. Update dependent calls and types.
15718
15719 * config/i386/t-cygming: Update for below.
15720
15721 * config/i386/t-interix: Update for below.
15722
15723 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15724 Change type to hash_table. Update dependent calls and types.
15725 (i386_find_on_wrapper_list::wrappers): Likewise.
15726
15727 * config/ia64/t-ia64: Update for below.
15728
15729 * config/ia64/ia64.c (bundle_state_table):
15730 Change type to hash_table. Update dependent calls and types.
15731
15732 * config/mips/mips.c (mips_reorg_process_insns::htab):
15733 Change type to hash_table. Update dependent calls and types.
15734
15735 * config/sol2.c (solaris_comdat_htab):
15736 Change type to hash_table. Update dependent calls and types.
15737
15738 * config/t-sol2: Update for above.
15739
15740 2013-05-29 Teresa Johnson <tejohnson@google.com>
15741
15742 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15743 functions are not yet marked as defined.
15744
15745 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15746 Pat Haugen <pthaugen@us.ibm.com>
15747 Peter Bergner <bergner@vnet.ibm.com>
15748
15749 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15750 instructions.
15751 (VEC_A): Likewise.
15752 (VEC_C): Likewise.
15753 (vrotl<mode>3): Likewise.
15754 (vashl<mode>3): Likewise.
15755 (vlshr<mode>3): Likewise.
15756 (vashr<mode>3): Likewise.
15757
15758 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15759 support for power8 V2DI builtins.
15760
15761 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15762 power8 V2DI builtins.
15763 (vupkhsw): Likewise.
15764 (vupklsw): Likewise.
15765 (vaddudm): Likewise.
15766 (vminsd): Likewise.
15767 (vmaxsd): Likewise.
15768 (vminud): Likewise.
15769 (vmaxud): Likewise.
15770 (vpkudum): Likewise.
15771 (vpksdss): Likewise.
15772 (vpkudus): Likewise.
15773 (vpksdus): Likewise.
15774 (vrld): Likewise.
15775 (vsld): Likewise.
15776 (vsrd): Likewise.
15777 (vsrad): Likewise.
15778 (vsubudm): Likewise.
15779 (vcmpequd): Likewise.
15780 (vcmpgtsd): Likewise.
15781 (vcmpgtud): Likewise.
15782 (vcmpequd_p): Likewise.
15783 (vcmpgtsd_p): Likewise.
15784 (vcmpgtud_p): Likewise.
15785 (vupkhsw): Likewise.
15786 (vupklsw): Likewise.
15787 (vaddudm): Likewise.
15788 (vmaxsd): Likewise.
15789 (vmaxud): Likewise.
15790 (vminsd): Likewise.
15791 (vminud): Likewise.
15792 (vpksdss): Likewise.
15793 (vpksdus): Likewise.
15794 (vpkudum): Likewise.
15795 (vpkudus): Likewise.
15796 (vrld): Likewise.
15797 (vsld): Likewise.
15798 (vsrad): Likewise.
15799 (vsrd): Likewise.
15800 (vsubudm): Likewise.
15801
15802 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15803 support for power8 V2DI instructions.
15804
15805 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15806 power8 V2DI instructions. Combine pack and unpack insns to use an
15807 iterator for each mode. Check whether a particular mode supports
15808 Altivec instructions instead of just checking TARGET_ALTIVEC.
15809 (UNSPEC_VPKUWUM): Likewise.
15810 (UNSPEC_VPKSHSS): Likewise.
15811 (UNSPEC_VPKSWSS): Likewise.
15812 (UNSPEC_VPKUHUS): Likewise.
15813 (UNSPEC_VPKSHUS): Likewise.
15814 (UNSPEC_VPKUWUS): Likewise.
15815 (UNSPEC_VPKSWUS): Likewise.
15816 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15817 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15818 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15819 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15820 (UNSPEC_VUPKHSB): Likewise.
15821 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15822 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15823 (UNSPEC_VUPKHSH): Likewise.
15824 (UNSPEC_VUPKLSB): Likewise.
15825 (UNSPEC_VUPKLSH): Likewise.
15826 (VI2): Likewise.
15827 (VI_char): Likewise.
15828 (VI_scalar): Likewise.
15829 (VI_unit): Likewise.
15830 (VP): Likewise.
15831 (VP_small): Likewise.
15832 (VP_small_lc): Likewise.
15833 (VU_char): Likewise.
15834 (add<mode>3): Likewise.
15835 (altivec_vaddcuw): Likewise.
15836 (altivec_vaddu<VI_char>s): Likewise.
15837 (altivec_vadds<VI_char>s): Likewise.
15838 (sub<mode>3): Likewise.
15839 (altivec_vsubcuw): Likewise.
15840 (altivec_vsubu<VI_char>s): Likewise.
15841 (altivec_vsubs<VI_char>s): Likewise.
15842 (altivec_vavgs<VI_char>): Likewise.
15843 (altivec_vcmpbfp): Likewise.
15844 (altivec_eq<mode>): Likewise.
15845 (altivec_gt<mode>): Likewise.
15846 (altivec_gtu<mode>): Likewise.
15847 (umax<mode>3): Likewise.
15848 (smax<mode>3): Likewise.
15849 (umin<mode>3): Likewise.
15850 (smin<mode>3): Likewise.
15851 (altivec_vpkuhum): Likewise.
15852 (altivec_vpkuwum): Likewise.
15853 (altivec_vpkshss): Likewise.
15854 (altivec_vpkswss): Likewise.
15855 (altivec_vpkuhus): Likewise.
15856 (altivec_vpkshus): Likewise.
15857 (altivec_vpkuwus): Likewise.
15858 (altivec_vpkswus): Likewise.
15859 (altivec_vpks<VI_char>ss): Likewise.
15860 (altivec_vpks<VI_char>us): Likewise.
15861 (altivec_vpku<VI_char>us): Likewise.
15862 (altivec_vpku<VI_char>um): Likewise.
15863 (altivec_vrl<VI_char>): Likewise.
15864 (altivec_vsl<VI_char>): Likewise.
15865 (altivec_vsr<VI_char>): Likewise.
15866 (altivec_vsra<VI_char>): Likewise.
15867 (altivec_vsldoi_<mode>): Likewise.
15868 (altivec_vupkhsb): Likewise.
15869 (altivec_vupkhs<VU_char>): Likewise.
15870 (altivec_vupkls<VU_char>): Likewise.
15871 (altivec_vupkhsh): Likewise.
15872 (altivec_vupklsb): Likewise.
15873 (altivec_vupklsh): Likewise.
15874 (altivec_vcmpequ<VI_char>_p): Likewise.
15875 (altivec_vcmpgts<VI_char>_p): Likewise.
15876 (altivec_vcmpgtu<VI_char>_p): Likewise.
15877 (abs<mode>2): Likewise.
15878 (vec_unpacks_hi_v16qi): Likewise.
15879 (vec_unpacks_hi_v8hi): Likewise.
15880 (vec_unpacks_lo_v16qi): Likewise.
15881 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15882 (vec_unpacks_lo_v8hi): Likewise.
15883 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15884 (vec_pack_trunc_v8h): Likewise.
15885 (vec_pack_trunc_v4si): Likewise.
15886 (vec_pack_trunc_<mode>): Likewise.
15887
15888 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15889 V2DI builtins.
15890 (vec_vmaxsd): Likewise.
15891 (vec_vmaxud): Likewise.
15892 (vec_vminsd): Likewise.
15893 (vec_vminud): Likewise.
15894 (vec_vpksdss): Likewise.
15895 (vec_vpksdus): Likewise.
15896 (vec_vpkudum): Likewise.
15897 (vec_vpkudus): Likewise.
15898 (vec_vrld): Likewise.
15899 (vec_vsld): Likewise.
15900 (vec_vsrad): Likewise.
15901 (vec_vsrd): Likewise.
15902 (vec_vsubudm): Likewise.
15903 (vec_vupkhsw): Likewise.
15904 (vec_vupklsw): Likewise.
15905
15906 2013-05-29 Jan Hubicka <jh@suse.cz>
15907
15908 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15909 flags; reorder rest of fields in more consistent way.
15910 (varpool_node): Remove analyzed, finalized and alias.
15911 (cgraph_ndoe): Likewise.
15912 (symtab_alias_ultimate_target): New function.
15913 (cgraph_function_node): Move offline.
15914 (cgraph_reset_node): Declare.
15915 (cgraph_comdat_can_be_unshared_p): Remove.
15916 (varpool_remove_initializer): Declare.
15917 (varpool_first_defined_variable, varpool_next_defined_variable
15918 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15919 (cgraph_function_with_gimple_body_p): Update.
15920 (varpool_all_refs_explicit_p): Update.
15921 (symtab_alias_target): New function.
15922 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15923 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15924 (cgraph_function_or_thunk_node): Simplify using
15925 symtab_alias_ultimate_target.
15926 (varpool_variable_node): Likewise.
15927 * cgraph.c (cgraph_create_function_alias): Update.
15928 (cgraph_add_thunk): Update.
15929 (cgraph_remove_node): Update.
15930 (dump_cgraph_node): Do not dump removed flags.
15931 (cgraph_function_body_availability): Update.
15932 (cgraph_propagate_frequency): Update.
15933 (verify_cgraph_node): Check sanity of local flag.
15934 (cgraph_function_node): Move here from cgraph.h; revamp for
15935 cgraph_function_or_thunk_node.
15936 * lto-symtab.c (lto_varpool_replace_node): Update.
15937 (lto_symtab_resolve_can_prevail_p): Update.
15938 (lto_symtab_merge_cgraph_nodes): Update.
15939 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15940 propagate_constants_accross_call, devirtualization_time_bonus,
15941 ipcp_propagate_stage): Update.
15942 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15943 * ipa-inline-transform.c (clone_inlined_nodes,
15944 preserve_function_body_p): Update.
15945 * ipa-reference.c (propagate): Update.
15946 (write_node_summary_p): Update.
15947 * toplev.c (wrapup_global_declaration_2): Update.
15948 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15949 (analyze_function) ... this one.
15950 (cgraph_process_new_functions): Update.
15951 (cgraph_reset_node): Export.
15952 (cgraph_finalize_function): Update.
15953 (cgraph_add_new_function): Update.
15954 (process_function_and_variable_attributes): Update.
15955 (varpool_finalize_decl): Update.
15956 (symbol_finalized): Remove.
15957 (symbol_finalized_and_needed): Rename to ...
15958 (symbol_defined_and_needed): ... update.
15959 (cgraph_analyze_functions): Update.
15960 (handle_alias_pairs): Update.
15961 (mark_functions_to_output): Update.
15962 (assemble_thunk): Update.
15963 (output_in_order): Update.
15964 (output_weakrefs): Update.
15965 (finalize_compilation_unit): Update.
15966 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15967 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15968 input_node, input_varpool_node): Update.
15969 * dbxout.c (dbxout_expand_expr): Update.
15970 * cgraphclones.c (cgraph_clone_node): Update.
15971 (cgraph_copy_node_for_versioning): Update.
15972 (cgraph_materialize_clone): Update.
15973 (cgraph_materialize_all_clones): Update.
15974 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15975 propagate_pure_const, propagate_nothrow): Update.
15976 * lto-streamer-out.c (lto_output, write_symbol): Update.
15977 * ipa-utils.c (ipa_reverse_postorder): Update.
15978 * ipa-inline.c (can_inline_edge_p): Update.
15979 (update_caller_keys, ipa_inline): Update.
15980 * dwarf2out.c (reference_to_unused,
15981 premark_types_used_by_global_vars_helper): Update.
15982 * tree-eh.c (tree_could_trap_p): Update.
15983 * ipa-split.c (consider_split, execute_split_functions): Update.
15984 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15985 has_addr_references_p): Update; move ahead in file for better
15986 readability.
15987 (process_references): Simplify.
15988 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15989 bodies are removed.
15990 (cgraph_comdat_can_be_unshared_p): Make static.
15991 (cgraph_externally_visible_p): Update.
15992 (varpool_externally_visible_p): Update.
15993 (function_and_variable_visibility): Update.
15994 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15995 ipa_tm_mark_force_output_node): Update.
15996 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15997 estimate_edge_devirt_benefit, inline_generate_summary,
15998 inline_write_summary): Update.
15999 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16000 * ipa-prop.c (ipa_compute_jump_functions): Update.
16001 (ipa_print_node_params, ipa_prop_read_section,
16002 ipa_update_after_lto_read, read_replacements_section): Update.
16003 * varasm.c (mark_decl_referenced): Update.
16004 (assemble_alias, dump_tm_clone_pairs): Update.
16005 * tree-inline.c (copy_bb): Update.
16006 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16007 Update.
16008 * symtab.c (dump_symtab_base): Print new flags.
16009 (verify_symtab_base): Verify new flags.
16010 (symtab_alias_ultimate_target): New function.
16011 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16012 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16013 Update.
16014 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16015 Update.
16016 * i386.c (ix86_get_function_versions_dispatcher,
16017 ix86_generate_version_dispatcher_body): Update.
16018 (fold_builtin_cpu): Use varpool_add_new_variable.
16019 * varpool.c (varpool_remove_initializer): Break out from ...
16020 (varpool_remove_node): ... this one.
16021 (dump_varpool_node, varpool_node_for_asm,
16022 cgraph_variable_initializer_availability, varpool_analyze_node,
16023 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16024 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16025 Update.
16026
16027 2013-05-29 Jan Hubicka <jh@suse.cz>
16028
16029 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16030
16031 2013-05-29 Easwaran Raman <eraman@google.com>
16032
16033 PR tree-optimization/57442
16034 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16035 when control exits the main loop.
16036
16037 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16038
16039 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16040 and RX600.
16041 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16042 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16043 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16044 option.
16045
16046 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16047
16048 PR tree-optimization/57441
16049 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16050 Don't limit size of incr_vec to number of candidates.
16051
16052 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16053
16054 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16055 and mips16 directories.
16056 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16057 (MULTILIB_DIRNAMES): Ditto.
16058 (MULTILIB_EXCEPTIONS): Add new exceptions.
16059 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16060 (MULTILIB_DIRNAMES): Ditto.
16061 (MULTILIB_EXCEPTIONS): Add new exceptions.
16062
16063 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16064 Marcus Shawcroft <marcus.shawcroft@arm.com>
16065
16066 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16067 SYMBOL_TINY_ABSOLUTE.
16068 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16069 SYMBOL_TINY_ABSOLUTE.
16070 (aarch64_expand_mov_immediate): Likewise.
16071 (aarch64_classify_symbol): Likewise.
16072 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16073 Permit SYMBOL_TINY_ABSOLUTE.
16074 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16075
16076 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16077 Marcus Shawcroft <marcus.shawcroft@arm.com>
16078
16079 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16080 Refactor if/switch. Replace gcc_assert with if.
16081
16082 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16083
16084 * config/i386/i386.c (initial_ix86_tune_features): Enable
16085 FP Reassociation for AMD bdver1 and bdver2.
16086
16087 2013-05-29 Martin Jambor <mjambor@suse.cz>
16088
16089 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16090 and IMAGPART_EXPR do not occur within other handled_components.
16091
16092 2013-05-29 Richard Biener <rguenther@suse.de>
16093
16094 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16095 access on whether the use is in the BB we currently try to
16096 vectorize.
16097 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16098 vectorize to vect_bb_slp_scalar_cost.
16099
16100 2013-05-29 Richard Biener <rguenther@suse.de>
16101
16102 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16103 computing scalar cost offsetted by stmts that are kept live
16104 by scalar uses.
16105 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16106 for computation of scalar cost.
16107
16108 2013-05-28 Steve Ellcey <sellcey@mips.com>
16109
16110 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16111
16112 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16113
16114 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16115 array notation built-in reduction functions.
16116 * doc/passes.texi (Passes): Added documentation about changes done
16117 for Cilk Plus.
16118 * doc/invoke.texi (C Dialect Options): Added documentation about
16119 the -fcilkplus flag.
16120 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16121 (BUILTINS_DEF): Depend on cilkplus.def.
16122 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16123 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16124 * cilkplus.def: New file.
16125
16126 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16127
16128 PR rtl-optimization/57439
16129 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16130
16131 2013-05-28 Easwaran Raman <eraman@google.com>
16132
16133 PR tree-optimization/57337
16134 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16135 (find_insert_point): Correctly identify the insertion point
16136 when two statements with the same UID is compared.
16137
16138 2013-05-28 Richard Biener <rguenther@suse.de>
16139
16140 PR tree-optimization/56787
16141 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16142 from the list of data references.
16143 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16144 clobbers.
16145 (vect_analyze_loop_operations): Likewise.
16146 (vect_transform_loop): Remove clobbers.
16147
16148 2013-05-28 Martin Jambor <mjambor@suse.cz>
16149
16150 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16151 and REALPART_EXPRs have scalar type.
16152
16153 2013-05-28 Richard Biener <rguenther@suse.de>
16154
16155 PR tree-optimization/57411
16156 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16157 virtual operands.
16158 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16159 virtual operand propagation.
16160
16161 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16162
16163 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16164 destination register for bmasksi_vis.
16165 (vector_init_bshuffle): Likewise.
16166 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16167
16168 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16169
16170 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16171 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16172 mode if the instruction isn't available in the original mode.
16173 * config/sparc/sparc.opt (mfix-ut699): New option.
16174 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16175 (divdf3): Turn into expander.
16176 (divdf3_nofix): New insn.
16177 (divdf3_fix): Likewise.
16178 (divsf3): Disable if -mfix-ut699.
16179 (sqrtdf2): Turn into expander.
16180 (sqrtdf2_nofix): New insn.
16181 (sqrtdf2_fix): Likewise.
16182 (sqrtsf2): Disable if -mfix-ut699.
16183
16184 2013-05-27 Richard Biener <rguenther@suse.de>
16185
16186 PR middle-end/57412
16187 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16188 block for the new loop.
16189
16190 2013-05-27 Richard Biener <rguenther@suse.de>
16191
16192 PR tree-optimization/57343
16193 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16194 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16195 (number_of_iterations_cond): Do not build the folded tree.
16196
16197 2013-05-27 Richard Biener <rguenther@suse.de>
16198
16199 Revert
16200 PR middle-end/57381
16201 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16202 OEP_CONSTANT_ADDRESS_OF retained.
16203
16204 PR tree-optimization/57417
16205 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16206 for unchanged base.
16207 (set_ssa_val_to): Compare addresses using
16208 get_addr_base_and_unit_offset.
16209
16210 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16211
16212 PR rtl-optimization/56833
16213 * postreload.c (move2add_record_mode): New function.
16214 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16215 (move2add_use_add2_insn): Use move2add_record_sym_value.
16216 (move2add_use_add3_insn): Likewise.
16217 (reload_cse_move2add): Use move2add_valid_value_p and
16218 move2add_record_mode. Invalidate call-clobbered and REG_INC
16219 affected regs by setting reg_mode to VOIDmode.
16220 (move2add_note_store): Don't pretend the inside of a SUBREG is
16221 the actual destination. Invalidate single/leading registers by
16222 setting reg_mode to VOIDmode.
16223 Use move2add_record_sym_value, move2add_valid_value_p and
16224 move2add_record_mode.
16225
16226 2013-05-27 Richard Biener <rguenther@suse.de>
16227
16228 PR tree-optimization/57396
16229 * tree-affine.c (double_int_constant_multiple_p): Properly
16230 return false for val == 0 and div != 0.
16231
16232 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16233
16234 * config/mips/mips.h: Use #elif in preprocessor conditions.
16235
16236 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16237
16238 PR target/53916
16239 * config/mips/constraints.md (kl): New constraint.
16240 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16241 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16242 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16243 from LO for MIPS16.
16244 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16245
16246 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16247
16248 PR target/55777
16249 * config/mips/mips.c (mips_can_inline_p): New function.
16250 (TARGET_CAN_INLINE_P): Define.
16251
16252 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16253
16254 * sched-int.h (ds_t, dw_t): Make unsigned int.
16255 Fix documentation that describes how all the ds_t bits are used.
16256 Reserve the last bit for delayed-branch scheduling.
16257 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16258 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16259 (gen_dep_weak_1): Remove prototype.
16260 * sched-deps.c (get_dep_weak_1): Make static.
16261 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16262 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16263 * doc/tm.texi: Regenerate.
16264 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16265
16266 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16267
16268 PR debug/56950
16269 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16270
16271 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16272 Sandra Loosemore <sandra@codesourcery.com>
16273
16274 * config.gcc (powerpc-*): Allow native for with-cpu.
16275
16276 2013-05-24 Jeff Law <law@redhat.com>
16277
16278 PR tree-optimization/57124
16279 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16280 conversion feeding a condition if the range has an overflow
16281 if -fstrict-overflow. Add warnings for when we do make the
16282 transformation.
16283
16284 2013-05-24 Dehao Chen <dehao@google.com>
16285
16286 * tree-cfg.c (locus_discrim_map): Fix the typo.
16287 (locus_discrim_hasher): Likewise.
16288 (locus_discrim_hasher::hash): Likewise.
16289 (locus_discrim_hasher::equal): Likewise.
16290
16291 2013-05-24 Martin Jambor <mjambor@suse.cz>
16292
16293 PR tree-optimization/57294
16294 * cgraph.h (ipa_record_stmt_references): Declare.
16295 * cgraphbuild.c (ipa_record_stmt_references): New function.
16296 (build_cgraph_edges): Use ipa_record_stmt_references.
16297 (rebuild_cgraph_edges): Likewise.
16298 (cgraph_rebuild_references): Likewise.
16299 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16300 associated with the old statement and build references from the
16301 newly built statements.
16302 * ipa-ref.c (ipa_remove_stmt_references): New function.
16303 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16304
16305 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16306
16307 * lra-constraints.c (emit_spill_move): Use smaller mode for
16308 mem-mem moves.
16309 (check_and_process_move): Consider mem-reg moves for secondary
16310 too.
16311 (curr_insn_transform): Don't lose insns emitted before for
16312 secondary memory moves.
16313 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16314 reg set up in the current insn.
16315
16316 2013-05-24 Dehao Chen <dehao@google.com>
16317
16318 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16319 hash function.
16320 (locus_descrim_hasher::equal): Likewise.
16321 (build_gimple_cfg): New discriminator assignment algorithm.
16322 (make_edges): Likewise.
16323 (next_discriminator_for_locus): Likewise.
16324 (same_line_p): Likewise.
16325 (assign_discriminators): Likewise.
16326 (make_cond_expr_edges): Likewise.
16327 (make_gimple_switch_edges): Likewise.
16328 (make_goto_expr_edges): Likewise.
16329 (make_gimple_asm_edges): Likewise.
16330
16331 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16332
16333 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16334 X format specifier to only display bottom 16 bits.
16335 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16336 immediate to match for operand 2, since it will be masked.
16337
16338 2013-05-24 Richard Biener <rguenther@suse.de>
16339
16340 PR tree-optimization/57287
16341 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16342 all SSA names that occur in abnormal PHIs.
16343
16344 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16345
16346 PR tree-ssa/57385
16347 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16348 that index is not negative.
16349
16350 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16351
16352 PR rtl-optimization/55177
16353 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16354 (simplify_byte_swapping_operation): New.
16355 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16356 (simplify_relational_operation_1): Deal with BSWAP.
16357
16358 2013-05-23 Richard Henderson <rth@redhat.com>
16359
16360 PR target/56742
16361 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16362 (ix86_reorg): Call it.
16363
16364 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16365
16366 PR target/57379
16367 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16368 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16369 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16370
16371 2013-05-23 Christian Bruel <christian.bruel@st.com>
16372
16373 PR debug/57351
16374 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16375
16376 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16377 Marcus Shawcroft <marcus.shawcroft@arm.com>
16378
16379 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16380 * config/aarch64/constraints.md (Usa): Remove.
16381 * doc/md.texi (AArch64 Usa): Remove.
16382
16383 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16384 Marcus Shawcroft <marcus.shawcroft@arm.com>
16385
16386 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16387 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16388 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16389 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16390
16391 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16392
16393 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16394 instruction (AdvSIMD).
16395 * config/aarch64/aarch64-builtins.c
16396 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16397 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16398
16399 2013-05-23 Martin Jambor <mjambor@suse.cz>
16400
16401 PR middle-end/57347
16402 * tree.h (contains_bitfld_component_ref_p): Declare.
16403 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16404 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16405 caller.
16406 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16407 not access a bit-field. Assert all final offsets are byte-aligned.
16408
16409 2013-05-23 Richard Biener <rguenther@suse.de>
16410
16411 PR tree-optimization/57380
16412 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16413 least one invariant or re-used load.
16414 * passes.c (init_optimization_passes): Move pass_phiprop before
16415 pass_forwprop.
16416
16417 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16418
16419 * config/aarch64/aarch64-simd.md
16420 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16421
16422 2013-05-23 Richard Biener <rguenther@suse.de>
16423
16424 PR middle-end/57381
16425 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16426 OEP_CONSTANT_ADDRESS_OF retained.
16427
16428 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16429
16430 PR middle-end/57344
16431 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16432 don't lower unit. Handle unit not being always BITS_PER_WORD.
16433
16434 2013-05-23 Richard Biener <rguenther@suse.de>
16435
16436 PR rtl-optimization/57341
16437 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16438 instead of true_dependence.
16439
16440 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16441
16442 * bb-reorder.c (branch_threshold): Make const.
16443 (exec_threshold): Ditto.
16444
16445 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16446 Pat Haugen <pthaugen@us.ibm.com>
16447 Peter Bergner <bergner@vnet.ibm.com>
16448
16449 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16450 documentation for the power8 crypto builtins.
16451
16452 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16453
16454 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16455 macros for defining power8 builtin functions.
16456 (BU_P8V_AV_2): Likewise.
16457 (BU_P8V_AV_P): Likewise.
16458 (BU_P8V_VSX_1): Likewise.
16459 (BU_P8V_OVERLOAD_1): Likewise.
16460 (BU_P8V_OVERLOAD_2): Likewise.
16461 (BU_CRYPTO_1): Likewise.
16462 (BU_CRYPTO_2): Likewise.
16463 (BU_CRYPTO_3): Likewise.
16464 (BU_CRYPTO_OVERLOAD_1): Likewise.
16465 (BU_CRYPTO_OVERLOAD_2): Likewise.
16466 (XSCVSPDP): Fix typo, point to the correct instruction.
16467 (VCIPHER): Add power8 crypto builtins.
16468 (VCIPHERLAST): Likewise.
16469 (VNCIPHER): Likewise.
16470 (VNCIPHERLAST): Likewise.
16471 (VPMSUMB): Likewise.
16472 (VPMSUMH): Likewise.
16473 (VPMSUMW): Likewise.
16474 (VPERMXOR_V2DI): Likewise.
16475 (VPERMXOR_V4SI: Likewise.
16476 (VPERMXOR_V8HI: Likewise.
16477 (VPERMXOR_V16QI: Likewise.
16478 (VSHASIGMAW): Likewise.
16479 (VSHASIGMAD): Likewise.
16480 (VPMSUM): Likewise.
16481 (VPERMXOR): Likewise.
16482 (VSHASIGMA): Likewise.
16483
16484 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16485 __CRYPTO__ if the crypto instructions are available.
16486 (altivec_overloaded_builtins): Add support for overloaded power8
16487 builtins.
16488
16489 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16490 support for power8 crypto builtins.
16491 (builtin_function_type): Likewise.
16492 (altivec_init_builtins): Add support for builtins that take vector
16493 long long (V2DI) arguments.
16494
16495 * config/rs6000/crypto.md: New file, define power8 crypto
16496 instructions.
16497
16498 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16499 Pat Haugen <pthaugen@us.ibm.com>
16500 Peter Bergner <bergner@vnet.ibm.com>
16501
16502 * doc/invoke.texi (Option Summary): Add power8 options.
16503 (RS/6000 and PowerPC Options): Likewise.
16504
16505 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16506 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16507 wm, wn, wr documentation.
16508
16509 * config/rs6000/constraints.md (wm): New constraint for VSX
16510 registers if direct move instructions are enabled.
16511 (wn): New constraint for no registers.
16512 (wq): New constraint for quad word even GPR registers.
16513 (wr): New constraint if 64-bit instructions are enabled.
16514 (wv): New constraint if power8 vector instructions are enabled.
16515 (wQ): New constraint for quad word memory locations.
16516
16517 * config/rs6000/predicates.md (const_0_to_15_operand): New
16518 constraint for 0..15 for crypto instructions.
16519 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16520 as GPR and floating point registers.
16521 (int_reg_operand): New predicate to match only GPR registers.
16522 (base_reg_operand): New predicate to match base registers.
16523 (quad_int_reg_operand): New predicate to match even GPR registers
16524 for quad memory operations.
16525 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16526 operations in both GPR and VSX registers.
16527 (quad_memory_operand): New predicate for quad memory operations.
16528 (reg_or_indexed_operand): New predicate for direct move support.
16529
16530 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16531 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16532 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16533 (POWERPC_MASKS): Add power8 options.
16534 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16535 various options.
16536
16537 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16538 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16539
16540 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16541 (-mpower8-fusion): New power8 options.
16542 (-mpower8-fusion-sign): Likewise.
16543 (-mpower8-vector): Likewise.
16544 (-mcrypto): Likewise.
16545 (-mdirect-move): Likewise.
16546 (-mquad-memory): Likewise.
16547
16548 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16549 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16550 registers.
16551 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16552 (rs6000_debug_vector_unit): Add p8_vector.
16553 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16554 definitions. Also print fusion state.
16555 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16556 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16557 (rs6000_option_override_internal): Add support for power8.
16558 (rs6000_common_init_builtins): Add debugging for skipped builtins
16559 if -mdebug=builtin.
16560 (rs6000_adjust_cost): Add power8 support.
16561 (rs6000_issue_rate): Likewise.
16562 (insn_must_be_first_in_group): Likewise.
16563 (insn_must_be_last_in_group): Likewise.
16564 (force_new_group): Likewise.
16565 (rs6000_register_move_cost): Likewise.
16566 (rs6000_opt_masks): Likewise.
16567
16568 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16569 power8 capable assembler, default to power7 options.
16570 (TARGET_DIRECT_MOVE): Likewise.
16571 (TARGET_CRYPTO): Likewise.
16572 (TARGET_P8_VECTOR): Likewise.
16573 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16574 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16575 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16576 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16577 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16578 (TARGET_XSCVDPSPN): Likewise.
16579 (TARGET_XSCVSPDPN): Likewsie.
16580 (TARGET_SYNC_HI_QI): Likewise.
16581 (TARGET_SYNC_TI): Likewise.
16582 (MASK_CRYPTO): Likewise.
16583 (MASK_DIRECT_MOVE): Likewise.
16584 (MASK_P8_FUSION): Likewise.
16585 (MASK_P8_VECTOR): Likewise.
16586 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16587 temporary used by some of the direct move instructions to get two FP
16588 temporary registers does not force creation of a stack frame.
16589 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16590 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16591 that any VSX registers are tieable, even if they are also an
16592 Altivec vector mode.
16593 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16594 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16595 (RS6000_BTM_CRYPTO): Likewise.
16596 (RS6000_BTM_COMMON): Likewise.
16597
16598 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16599 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16600 (enum rs6000_vector): Add power8 vector support.
16601
16602 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16603
16604 PR target/19599
16605 PR target/57340
16606 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16607 (any_sibcall_could_use_r3): this and handle indirect calls.
16608 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16609
16610 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16611
16612 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16613
16614 2013-05-22 Richard Biener <rguenther@suse.de>
16615
16616 PR middle-end/57349
16617 * profile.c (branch_prob): Do not split blocks that are
16618 abnormally receiving from ECF_RETURNS_TWICE functions.
16619
16620 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16621
16622 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16623 address mode. Move pointer mode initialization to the same place.
16624
16625 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16626
16627 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16628 while it has any effect.
16629
16630 2013-05-21 Easwaran Raman <eraman@google.com>
16631
16632 PR tree-optimization/57322
16633 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16634 UID of the statement added to the BB to be 1.
16635
16636 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16637
16638 PR tree-optimization/57331
16639 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16640 of conversion from pointer type to integral type with integer.
16641
16642 2013-05-21 Martin Jambor <mjambor@suse.cz>
16643
16644 PR lto/57289
16645 * ipa-prop.c (ipa_read_node_info): Process param_used and
16646 controlled_uses in the same order as when writing.
16647
16648 2013-05-21 Magnus Granberg <baldrick@free.fr>
16649
16650 PR plugins/56754
16651 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16652
16653 2013-05-21 Richard Biener <rguenther@suse.de>
16654
16655 PR tree-optimization/57318
16656 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16657 estimate stmts with side-effects as likely eliminated.
16658
16659 2013-05-21 Richard Biener <rguenther@suse.de>
16660
16661 PR tree-optimization/57330
16662 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16663 preserve the call stmts fntype.
16664
16665 2013-05-21 Richard Biener <rguenther@suse.de>
16666
16667 PR tree-optimization/57303
16668 * tree-ssa-sink.c (statement_sink_location): Improve killing
16669 stmt detection and properly handle self-assignments.
16670
16671 2013-05-21 Christian Bruel <christian.bruel@st.com>
16672
16673 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16674 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16675 registers. Set register size out of the PARALLEL loop.
16676
16677 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16678
16679 PR target/56547
16680 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16681 (*fmasf4, *fmasf4_media): New insns.
16682
16683 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16684
16685 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16686 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16687 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16688 (mips_idiv_insns): Update the comments to say that the returned
16689 instruction counts are in units of BASE_INSN_LENGTH.
16690 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16691 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16692 using 2 rather than 4 as the length of indirect MIPS16 and
16693 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16694 length of a NOP. Don't divide MIPS16 lengths by 2.
16695 (mips16_split_long_branches): Assume a branch is long if the
16696 length is greater than 4 rather than 8.
16697 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16698 rather than multiplying them by 2. Multiply instruction counts
16699 by BASE_INSN_LENGTH rather than 4.
16700 (*jump_mips16, tls_get_tp_mips16_<mode>)
16701 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16702
16703 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16704
16705 * config/mips/mips.md (extended_mips16): Remove branch case.
16706 (length): Remove duplicated extended_mips16 test.
16707
16708 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16709
16710 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16711
16712 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16713
16714 * recog.h (Recog_data): Rename to...
16715 (recog_data_d): ...this.
16716 (recog_data): Update accordingly.
16717 * recog.c (recog_data): Likewise.
16718 * reload.c (save_recog_data): Likewise.
16719 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16720 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16721
16722 2013-05-17 Julian Brown <julian@codesourcery.com>
16723
16724 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16725 found in a REG_EQUAL note, invalidate it.
16726
16727 2013-05-17 Easwaran Raman <eraman@google.com>
16728
16729 * tree-ssa-reassoc.c (find_insert_point): New function.
16730 (insert_stmt_after): Likewise.
16731 (get_def_stmt): Likewise.
16732 (ensure_ops_are_available): Likewise.
16733 (not_dominated_by): Likewise.
16734 (rewrite_expr_tree): Do not move statements beyond what is
16735 necessary. Remove call to swap_ops_for_binary_stmt...
16736 (reassociate_bb): ... and move it here.
16737 (build_and_add_sum): Assign UIDs for new statements.
16738 (linearize_expr): Likewise.
16739 (do_reassoc): Renumber gimple statement UIDs.
16740
16741 2013-05-17 Jan Hubicka <jh@suse.cz>
16742
16743 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16744 weakrefs.
16745 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16746 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16747 weakrefs.
16748 (output_weakrefs): Update.
16749
16750 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16751 Martin Jambor <mjambor@suse.cz>
16752
16753 PR middle-end/57276
16754 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16755 value that corresponds to the given aggval is found in values vector.
16756
16757 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16758
16759 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16760 sse, sse2, sse3, ssse3 and sse4a flags to options.
16761
16762 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16763
16764 * gengtype-state.c: (s_expr_writer): New class, to handle
16765 prettifying of output layout of s-expressions.
16766 (state_writer): New class, to write out gtype.state.
16767 (state_written_type_count): Move this variable into member data of
16768 state_writer.
16769 (s_expr_writer::s_expr_writer): New code: constructor for new class
16770 (state_writer::state_writer(): ditto
16771 (s_expr_writer::write_new_line): New function
16772 (s_expr_writer::write_any_indent): ditto
16773 (s_expr_writer::begin_s_expr): ditto
16774 (s_expr_writer::end_s_expr): ditto
16775 (write_state_fileloc): convert to method of state_writer...
16776 (state_writer:: write_state_fileloc): ...and use methods of
16777 s_expr_writer to write indentation into the gtype.state output file
16778 to visually represent the hierarchical structure of the list
16779 structures
16780 (write_state_fields): ditto, renaming to...
16781 (state_writer::write_state_fields)
16782 (write_state_a_string): ditto, renaming to...
16783 (state_writer::write_state_a_string)
16784 (write_state_string_option): ditto, renaming to...
16785 (state_writer::write_state_string_option)
16786 (write_state_type_option): ditto, renaming to...
16787 (state_writer::write_state_type_option)
16788 (write_state_nested_option): ditto, renaming to...
16789 (state_writer::write_state_nested_option)
16790 (write_state_option): ditto, renaming to...
16791 (state_writer::write_state_option)
16792 (write_state_options): ditto, renaming to...
16793 (state_writer::write_state_options)
16794 (write_state_lang_bitmap): ditto, renaming to...
16795 (state_writer::write_state_lang_bitmap)
16796 (write_state_version): ditto, renaming to...
16797 (state_writer::write_state_version)
16798 (write_state_scalar_type): ditto, renaming to...
16799 (state_writer::write_state_scalar_type)
16800 (write_state_string_type): ditto, renaming to...
16801 (state_writer::write_state_string_type)
16802 (write_state_undefined_type): ditto, renaming to...
16803 (state_writer::write_state_undefined_type)
16804 (write_state_struct_union_type): ditto, renaming to...
16805 (state_writer::write_state_struct_union_type)
16806 (write_state_struct_type): ditto, renaming to...
16807 (state_writer::write_state_struct_type)
16808 (write_state_user_struct_type): ditto, renaming to...
16809 (state_writer::write_state_user_struct_type)
16810 (write_state_lang_struct_type): ditto, renaming to...
16811 (state_writer::write_state_lang_struct_type)
16812 (write_state_param_struct_type): ditto, renaming to...
16813 (state_writer::write_state_param_struct_type)
16814 (write_state_pointer_type): ditto, renaming to...
16815 (state_writer::write_state_pointer_type)
16816 (write_state_array_type): ditto, renaming to...
16817 (state_writer::write_state_array_type)
16818 (write_state_gc_used): ditto, renaming to...
16819 (state_writer::write_state_gc_used)
16820 (write_state_common_type_content): ditto, renaming to...
16821 (state_writer::write_state_common_type_content)
16822 (write_state_type): ditto, renaming to...
16823 (state_writer::write_state_type)
16824 (write_state_pair_list): ditto, renaming to...
16825 (state_writer::write_state_pair_list)
16826 (write_state_pair): ditto, renaming to...
16827 (state_writer::write_state_pair)
16828 (write_state_typedefs): ditto, renaming to...
16829 (state_writer::write_state_typedefs)
16830 (write_state_structures): ditto, renaming to...
16831 (state_writer::write_state_structures)
16832 (write_state_param_structs): ditto, renaming to...
16833 (state_writer::write_state_param_structs)
16834 (write_state_variables): ditto, renaming to...
16835 (state_writer::write_state_variables)
16836 (write_state_srcdir): ditto, renaming to...
16837 (state_writer::write_state_srcdir)
16838 (write_state_files_list): ditto, renaming to...
16839 (state_writer::write_state_files_list)
16840 (write_state_languages): ditto, renaming to...
16841 (state_writer::write_state_languages)
16842 (write_state): create a state_writer instance and use it when
16843 writing out the state file
16844
16845 2013-05-17 Mike Stump <mikestump@comcast.net>
16846
16847 PR rtl-optimization/57304
16848 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16849 accessing DF_REF_REAL_LOC.
16850
16851 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16852
16853 PR rtl-optimization/57281
16854 PR rtl-optimization/57300
16855 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16856 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16857 what the other splitter did if the registers are dead.
16858
16859 2013-05-17 Richard Biener <rguenther@suse.de>
16860
16861 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16862 MEM_REF offsets.
16863
16864 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16865
16866 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16867 linking.
16868
16869 2013-05-17 Marek Polacek <polacek@redhat.com>
16870
16871 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16872 length when doing non-zero store of storing '\0' to '\0'.
16873
16874 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16875
16876 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16877 vect_external_def oprnd1 with loop_vinfo, try to emit
16878 optional cast, negation and and stmts on the loop preheader
16879 edge instead of into the pattern def seq.
16880
16881 PR tree-optimization/57051
16882 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16883 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16884
16885 2013-05-16 Nick Clifton <nickc@redhat.com>
16886
16887 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16888 (rl78_is_naked_func): New function.
16889 (rl78_expand_prologue): Skip prologue generation for naked functions.
16890 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16891 * doc/extend.texi (naked): Add RL78 to the list of processors
16892 that supports this attribute.
16893
16894 2013-05-16 Jeff Law <law@redhat.com>
16895
16896 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16897
16898 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16899
16900 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16901 cache parameters using detect_caches_amd also for CYRIX,
16902 NSC and TM2 signatures.
16903
16904 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16905 Dzianis Kahanovich <mahatma@eu.by>
16906
16907 PR target/45359
16908 PR target/46396
16909 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16910 VIA/Centaur processors and determine their cache parameters
16911 using detect_caches_amd.
16912
16913 2013-05-16 Teresa Johnson <tejohnson@google.com>
16914
16915 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16916 (rtl_verify_edges): New function.
16917 (rtl_verify_bb_insns): Ditto.
16918 (rtl_verify_bb_pointers): Ditto.
16919 (rtl_verify_bb_insn_chain): Ditto.
16920 (rtl_verify_fallthru): Ditto.
16921 (rtl_verify_bb_layout): Ditto.
16922 (rtl_verify_flow_info_1): Outline checks into new functions.
16923 (rtl_verify_flow_info): Ditto.
16924
16925 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16926
16927 * cfghooks.c (copy_bbs): Add update_dominance argument.
16928 * cfghooks.h (copy_bbs): Update prototype.
16929 * tree-cfg.c (gimple_duplicate_sese_region):
16930 Add update_dominance argument.
16931 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16932 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16933 gimple_duplicate_sese_region call.
16934 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16935 Update copy_bbs call.
16936 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16937 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16938
16939 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16940
16941 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16942 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16943 vect_recog_rotate_pattern.
16944 (vect_recog_rotate_pattern): New function.
16945
16946 2013-05-16 Jason Merrill <jason@redhat.com>
16947
16948 * Makefile.in (LLINKER): New variable.
16949 (mostlyclean): Remove link mutex.
16950 * configure.ac: Handle --enable-link-mutex.
16951 * lock-and-run.sh: New script.
16952
16953 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16954
16955 PR target/19599
16956 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16957 for NULL decl.
16958
16959 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16960
16961 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16962
16963 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16964
16965 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16966 * config/arm/arm.c (next_consecutive_mem): New function.
16967 (gen_movmem_ldrd_strd): Likewise.
16968 * config/arm/arm.md (movmemqi): Update condition and code.
16969 (unaligned_loaddi, unaligned_storedi): New patterns.
16970
16971 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16972
16973 * config.gcc: Obsolete *-*-solaris2.9*.
16974 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16975
16976 2013-05-16 Richard Biener <rguenther@suse.de>
16977
16978 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16979 earlier, after GRAPHITE transforms and IV canonicalization.
16980
16981 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16982
16983 * omp-low.c (extract_omp_for_data): For collapsed loops,
16984 if at least one of the loops is known at compile time to
16985 iterate zero times, set count to 0.
16986 (expand_omp_regimplify_p): New function.
16987 (expand_omp_for_generic): For collapsed loops, if at least
16988 one of the loops isn't known to iterate at least once,
16989 add runtime check with setting count to 0.
16990 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16991 For unsigned types if it isn't known at compile time that
16992 the loop will iterate at least once, add runtime check to bypass
16993 the whole loop if initial condition isn't true.
16994
16995 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16996
16997 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16998
16999 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17000
17001 PR middle-end/57286
17002 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17003 transformations to avoid an infinite loop.
17004
17005 2013-05-16 Marek Polacek <polacek@redhat.com>
17006
17007 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17008
17009 2013-05-15 Leif Ekblad <leif@rdos.net>
17010
17011 * config/i386/i386.c (ix86_decompose_address): Use
17012 DEFAULT_TLS_SEG_REG to access TLS segment register.
17013 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17014 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17015 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17016
17017 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17018
17019 PR target/57260
17020 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17021 sibling calls to functions that would normally be lazily bound,
17022 unless $gp is call-clobbered.
17023
17024 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17025
17026 * config/i386/i386.c (ix86_option_override_internal): Update
17027 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17028 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17029 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17030 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17031 of TARGET_3DNOW.
17032 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17033
17034 2013-05-15 Andreas Schwab <schwab@suse.de>
17035
17036 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17037 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17038 third operand.
17039
17040 2013-05-15 Teresa Johnson <tejohnson@google.com>
17041
17042 * loop-unroll.c (report_unroll_peel): Check decision before
17043 emitting unroll/peel message.
17044
17045 2013-05-15 Teresa Johnson <tejohnson@google.com>
17046
17047 * function.h (has_bb_partition): New rtl_data flag.
17048 (bb_reorder_complete): Ditto.
17049 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17050 instead of flag_reorder_blocks_and_partition.
17051 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17052 with some enhancements.
17053 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17054 * bb-reorder.c (connect_traces): Check for has_bb_partition
17055 instead of flag_reorder_blocks_and_partition.
17056 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17057 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17058 verify_hot_cold_block_grouping.
17059 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17060
17061 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17062
17063 PR target/19599
17064 * config/arm/predicates.md (call_insn_operand): New predicate.
17065 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17066 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17067 if insn is not a tail call.
17068 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17069 registers.
17070 * config/arm/arm.h (enum reg_class): New caller save register class.
17071 (REG_CLASS_NAMES): Likewise.
17072 (REG_CLASS_CONTENTS): Likewise.
17073 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17074 without decls.
17075
17076 2013-05-15 Richard Biener <rguenther@suse.de>
17077
17078 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17079 of MSG_OPTIMIZED_LOCATIONS.
17080 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17081 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17082 message.
17083 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17084 of MSG_OPTIMIZED_LOCATIONS.
17085 (execute_vect_slp): Likewise.
17086 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17087 (vect_create_cond_for_alias_checks): Likewise.
17088 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17089 (vect_recog_widen_mult_pattern): Likewise.
17090 (vect_recog_widen_sum_pattern): Likewise.
17091 (vect_recog_over_widening_pattern): Likewise.
17092 (vect_recog_widen_shift_pattern): Likewise.
17093 (vect_recog_vector_vector_shift_pattern): Likewise.
17094 (vect_recog_divmod_pattern): Likewise.
17095 (vect_recog_mixed_size_cond_pattern): Likewise.
17096 (vect_recog_bool_pattern): Likewise.
17097 (vect_pattern_recog_1): Likewise.
17098
17099 2013-05-15 Martin Jambor <mjambor@suse.cz>
17100
17101 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17102 non-functions to builtin_unreachable.
17103 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17104 correct when new direct edges were discovered.
17105
17106 2013-05-15 Martin Jambor <mjambor@suse.cz>
17107
17108 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17109 header, print symbol order instead of node uid, print more information
17110 about indirect edge targets.
17111 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17112 uids.
17113 (ipa_make_edge_direct_to_target): Likewise.
17114 (remove_described_reference): Likewise.
17115 (propagate_controlled_uses): Likewise.
17116 (ipa_print_node_params): Also print symbol order.
17117 (ipcp_transform_function): Print symbol order instead of node uids.
17118 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17119 (cgraph_get_create_real_symbol_node): Likewise.
17120 * ipa-cp.c (print_lattice): Likewise.
17121 (print_all_lattices): Likewise.
17122 (determine_versionability): Likewise.
17123 (initialize_node_lattices): Likewise.
17124 (estimate_local_effects): Likewise.
17125 (update_profiling_info): Likewise.
17126 (create_specialized_node): Likewise.
17127 (perhaps_add_new_callers): Likewise.
17128 (decide_about_value): Likewise.
17129 (decide_whether_version_node): Likewise.
17130 (identify_dead_nodes): Likewise.
17131 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17132 (dump_inline_summary): Likewise.
17133 (estimate_node_size_and_time): Likewise.
17134 (inline_analyze_function): Likewise.
17135 * ipa-inline.c (report_inline_failed_reason): Likewise.
17136 (want_early_inline_function_p): Likewise.
17137 (edge_badness): Likewise.
17138 (update_edge_key): Likewise.
17139 (inline_small_functions): Likewise. Add dumping of order to two other
17140 dumps.
17141 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17142 instead of node uids.
17143 (propagate_pure_const): Likewise.
17144 (propagate_pure_const): Likewise.
17145 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17146 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17147 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17148 of node uids.
17149 * tree-pretty-print.c (dump_function_header): Likewise.
17150 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17151 Print symbol order instead of node uids.
17152
17153 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17154
17155 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17156 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17157
17158 2013-05-15 Richard Biener <rguenther@suse.de>
17159
17160 PR tree-optimization/57275
17161 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17162 return value for fail to do runtime alias checks for gather loads.
17163
17164 2013-05-15 Jan Hubicka <jh@suse.cz>
17165
17166 PR lto/57038
17167 PR lto/47375
17168 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17169 weakrefs are not external.
17170 (lto_symtab_merge_decls): Fix thinko when dealing with
17171 non-lto_symtab decls.
17172 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17173 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17174 * varpool.c (dump_varpool_node): Dump more flags.
17175
17176 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17177
17178 * config/i386/i386.c (processor_alias_table): Add instruction
17179 FSGSBASE for AMD bdver3 architecture.
17180
17181 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17182
17183 * tree.c (warn_deprecated_use): Print file:line using locus color.
17184 * diagnostic.c (diagnostic_report_current_module): Print file:line
17185 and file:line:column using locus color.
17186
17187 2013-05-14 Mike Stump <mikestump@comcast.net>
17188
17189 * gdbinit.in: Add __null.
17190
17191 2013-05-14 Mike Stump <mikestump@comcast.net>
17192
17193 * recog.h: Rename struct recog_data to Recog_data.
17194 * recog.c: Likewise.
17195 * reload.c (can_reload_into): Likewise.
17196 * config/picochip/picochip.c: Likewise.
17197
17198 2013-05-14 Mike Stump <mikestump@comcast.net>
17199
17200 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17201
17202 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17203
17204 * resource.h (struct resources): Remove unch_memory member.
17205 (CLEAR_RESOURCE): Don't clear unch_memory.
17206 * resource.c (mark_referenced_resources): Don't set it.
17207 (mark_set_resources): Likewise.
17208 (mark_target_live_regs): Don't clear it.
17209 (init_resource_info): Likewise.
17210 * reorg.c (resource_conflicts_p): Don't compare it.
17211 (redundant_insn): Don't set it.
17212
17213 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17214 Remove prototypes.
17215 * emit-rtl.c (next_label): Remove unused function.
17216 (skip_consecutive_labels, link_cc0_insns): Move to ...
17217 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17218 only place where these functions are used, and make them static.
17219
17220 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17221
17222 * fold-const.c (fold_negate_expr): Handle vectors.
17223 (fold_truth_not_expr): Make it static.
17224 (fold_invert_truthvalue): New static function.
17225 (invert_truthvalue_loc): Handle vectors. Do not call
17226 fold_truth_not_expr directly.
17227 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17228 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17229 (fold_comparison): Handle vector constants.
17230 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17231 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17232 * tree.h (fold_truth_not_expr): Remove declaration.
17233
17234 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17235
17236 * config/aarch64/aarch64-simd.md
17237 (aarch64_vcond_internal<mode>): Rename to...
17238 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17239 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17240 float modes. Clarify all iterator modes.
17241 (vcond<mode><mode>): Use new name for vcond expanders.
17242 (vcond<v_cmp_result><mode>): Likewise.
17243 (vcondu<mode><mode>: Likewise.
17244 * config/aarch64/iterators.md (VDQF_COND): New.
17245
17246 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17247
17248 PR bootstrap/57266
17249 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17250 variable for the shift amount. Check that we shift by non-negative
17251 amounts.
17252
17253 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17254
17255 PR target/42017
17256 * config/arm/arm.h (EPILOGUE_USES): Only return true
17257 for LR_REGNUM after epilogue_completed.
17258
17259 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17260
17261 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17262 is error_mark_node.
17263
17264 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17265
17266 PR target/57261
17267 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17268 and Solaris 11+/x86 with gld.
17269 * configure: Regenerate.
17270
17271 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17272
17273 * expmed.c (expand_shift_1): Canonicalize rotates by
17274 constant bitsize / 2 to bitsize - 1.
17275 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17276 case ROTATERT>: Likewise.
17277
17278 Revert:
17279 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17280
17281 * config/i386/i386.md (rotateinv): New code attr.
17282 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17283 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17284 roll $31, %eax, etc.
17285
17286 2013-05-14 Richard Biener <rguenther@suse.de>
17287
17288 PR middle-end/57235
17289 * tree-eh.c (sink_clobbers): Give up for successors with
17290 multiple predecessors and no virtual uses.
17291
17292 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17293
17294 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17295 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17296
17297 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17298
17299 PR middle-end/57251
17300 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17301 the case when both op0 and op1 have VOIDmode.
17302
17303 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17304
17305 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17306 in multiply-accumulate mode.
17307
17308 2013-05-13 Guozhi Wei <carrot@google.com>
17309
17310 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17311
17312 2013-05-13 Kai Tietz <ktietz@redhat.com>
17313
17314 PR target/56975
17315 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17316 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17317 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17318 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17319 (ix86_expand_prologue): Likewise.
17320 (ix86_expand_split_stack_prologue): Likewise.
17321 (legitimate_pic_address_disp_p): Likewise.
17322 (legitimize_pic_address): Likewise.
17323 (legitimize_tls_address): Likewise.
17324 (legitimize_pe_coff_symbol): Likewise.
17325 (output_pic_addr_const): Likewise.
17326 (construct_plt_address): Likewise.
17327 (ix86_expand_call): Likewise.
17328 (x86_output_mi_thunk): Likewise.
17329 (x86_function_profiler): Likewise.
17330
17331 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17332
17333 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17334 similar switch cases.
17335 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17336 (aarch64_simd_mov_to_<mode>low): Delete.
17337 (aarch64_simd_mov_to_<mode>high): Delete.
17338 (move_lo_quad_<mode>): Add w<-r alternative.
17339 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17340 (aarch64_simd_mov_from_*): Update type attribute.
17341 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17342 statement.
17343
17344 2013-05-13 Jan Hubicka <jh@suse.cz>
17345
17346 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17347 * config/i386/i386.c (ix86_compute_frame_layout,
17348 ix86_expand_epilogue, emit_i387_cw_initialization,
17349 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17350 ix86_local_alignment): Fix use of size/speed predicates.
17351
17352 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17353
17354 PR tree-optimization/45216
17355 PR tree-optimization/57157
17356 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17357 the (-Y) & (B - 1) variant if OP is |.
17358 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17359 return shifted. Use (-op1) & (prec - 1) as other_amount
17360 instead of prec - op1.
17361
17362 2013-05-13 Martin Jambor <mjambor@suse.cz>
17363
17364 PR middle-end/42371
17365 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17366 (ipa_constant_data): New type.
17367 (ipa_jump_func): Use ipa_constant_data to hold information about
17368 constant jump functions.
17369 (ipa_get_jf_constant): Adjust to jump function type changes.
17370 (ipa_get_jf_constant_rdesc): New function.
17371 (ipa_param_descriptor): New field controlled_uses.
17372 (ipa_get_controlled_uses): New function.
17373 (ipa_set_controlled_uses): Likewise.
17374 * ipa-ref.h (ipa_find_reference): Declare.
17375 * ipa-prop.c (ipa_cst_ref_desc): New type.
17376 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17377 changes.
17378 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17379 New parameter cs. Adjust all callers.
17380 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17381 (remove_described_reference): New function.
17382 (jfunc_rdesc_usable): Likewise.
17383 (try_make_edge_direct_simple_call): Decrement controlled use count,
17384 attempt to remove reference if it hits zero.
17385 (combine_controlled_uses_counters): New function.
17386 (propagate_controlled_uses): Likewise.
17387 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17388 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17389 (ipa_print_node_params): Print described use counter.
17390 (ipa_write_jump_function): Adjust to jump function type changes.
17391 (ipa_read_jump_function): New parameter CS, pass it to
17392 ipa_set_jf_constant. Adjust caller.
17393 (ipa_write_node_info): Stream controlled use count
17394 (ipa_read_node_info): Likewise.
17395 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17396 asserting.
17397 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17398 count. Remove cloning-added reference if it reaches zero.
17399 * ipa-ref.c (ipa_find_reference): New function.
17400
17401 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17402
17403 * config/i386/i386.c (processor_target_table): Modified default
17404 alignment values for AMD BD and BT architectures.
17405
17406 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17407
17408 * tree-vect-generic.c (uniform_vector_p): Move ...
17409 * tree.c (uniform_vector_p): ... here.
17410 * tree.h (uniform_vector_p): Declare it.
17411 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17412 into a scalar.
17413
17414 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17415
17416 PR tree-optimization/57230
17417 * tree-ssa-strlen.c (handle_char_store): Record length for
17418 array store from STRING_CST.
17419
17420 PR tree-optimization/57230
17421 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17422 check.
17423
17424 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17425
17426 * config/epiphany/epiphany.c (epiphany_init): Check size of
17427 NUM_MODES_FOR_MODE_SWITCHING.
17428 (epiphany_expand_prologue):
17429 Remove CONFIG_REGNUM initial value handling code.
17430 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17431 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17432 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17433 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17434 Don't return 1 for FP_MODE_NONE.
17435 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17436 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17437 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17438 * config/epiphany/epiphany.md (save_config): New pattern.
17439
17440 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17441
17442 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17443
17444 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17445
17446 * config/i386/i386.md (memory): Handle sseishft1.
17447 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17448 (*vec_extractv2di_1): Ditto.
17449
17450 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17451
17452 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17453 saved registers.
17454
17455 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17456
17457 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17458 Add mthumb/march=armv7-a multilib.
17459 Add mthumb/march=armv7-r multilib.
17460 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17461
17462 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17463
17464 * config/v850/t-rtems: Add more multilibs.
17465
17466 2013-05-10 Richard Biener <rguenther@suse.de>
17467
17468 PR tree-optimization/57214
17469 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17470 not propagate from SSA names that occur in abnormal PHI nodes.
17471
17472 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17473
17474 * stor-layout.c (element_precision): New function.
17475 * machmode.h (element_precision): Declare it.
17476 * tree.c (build_minus_one_cst): New function.
17477 (element_precision): Likewise.
17478 * tree.h (build_minus_one_cst): Declare new function.
17479 (element_precision): Likewise.
17480 * fold-const.c (operand_equal_p): Use element_precision.
17481 (fold_binary_loc): Handle vector types.
17482 * convert.c (convert_to_integer): Use element_precision.
17483 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17484 separately.
17485
17486 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17487
17488 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17489 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17490 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17491 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17492 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17493 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17494 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17495 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17496 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17497 (Uuw8): New constraints.
17498 (Usb4): Move into alphabetical order.
17499 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17500 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17501 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17502 previously unnamed patterns.
17503 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17504 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17505 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17506 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17507 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17508 of specific lengths.
17509
17510 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17511
17512 * config/i386/i386.md (rotateinv): New code attr.
17513 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17514 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17515 roll $31, %eax, etc.
17516
17517 PR tree-optimization/45216
17518 PR tree-optimization/57157
17519 * tree-ssa-forwprop.c (simplify_rotate): New function.
17520 (ssa_forward_propagate_and_combine): Call it.
17521
17522 2013-05-10 Richard Biener <rguenther@suse.de>
17523
17524 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17525 disable peeling when we version for aliasing.
17526 (vector_alignment_reachable_p): Honor explicit user alignment.
17527 (vect_supportable_dr_alignment): Likewise.
17528 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17529 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17530 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17531 then peeling to arrange for the cost-model check to come first.
17532
17533 2013-05-10 Alan Modra <amodra@gmail.com>
17534
17535 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17536 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17537 * configure: Regenerate.
17538
17539 2013-05-10 Alan Modra <amodra@gmail.com>
17540
17541 PR target/55033
17542 * varasm.c (default_elf_select_section): Move !DECL_P check..
17543 (get_named_section): ..to here before calling get_section_name.
17544 Adjust assertion.
17545 (default_section_type_flags): Add DECL_P check.
17546 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17547 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17548
17549 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17550
17551 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17552 When using gen_stack_adjust_str with a register offset, add a
17553 REG_FRAME_RELATED_EXPR note.
17554
17555 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17556
17557 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17558 (*vec_extractv4si_zext_mem): Ditto.
17559 (*vec_extractv2di): Add 0->x and x->x alternatives.
17560 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17561 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17562
17563 2013-05-09 Jason Merrill <jason@redhat.com>
17564
17565 N3639 C++1y VLA support
17566 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17567 DECL_VALUE_EXPR.
17568
17569 * tree.c (build_constructor_va): New.
17570 * tree.h: Declare it.
17571
17572 2013-05-09 Martin Jambor <mjambor@suse.cz>
17573
17574 PR lto/57084
17575 * gimple-fold.c (canonicalize_constructor_val): Call
17576 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17577
17578 2013-05-09 Jan Hubicka <jh@suse.cz>
17579 Richard Biener <rguenther@suse.de>
17580
17581 PR lto/54095
17582 * symtab.c (symtab_make_decl_local): Do not add private names.
17583
17584 2013-05-09 Jan Hubicka <jh@suse.cz>
17585
17586 PR lto/54095
17587 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17588 (unlink_from_assembler_name_hash): Likewise.
17589 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17590 symtab_unregister_node, symtab_initialize_asm_name_hash,
17591 change_decl_assembler_name): Update.
17592
17593 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17594
17595 * config/aarch64/aarch64.md: New movtf split.
17596 (*movtf_aarch64): Update.
17597 (aarch64_movdi_tilow): Handle TF modes and rename to
17598 aarch64_movdi_<mode>low.
17599 (aarch64_movdi_tihigh): Handle TF modes and rename to
17600 aarch64_movdi_<mode>high
17601 (aarch64_movtihigh_di): Handle TF modes and rename to
17602 aarch64_mov<mode>high_di
17603 (aarch64_movtilow_di): Handle TF modes and rename to
17604 aarch64_mov<mode>low_di
17605 (aarch64_movtilow_tilow): Remove spurious whitespace.
17606 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17607 splits.
17608 (aarch64_print_operand): Update.
17609
17610 2013-05-09 Alan Modra <amodra@gmail.com>
17611
17612 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17613 powerpc64le.
17614 * configure: Regenerate.
17615
17616 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17617
17618 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17619 splitter preparation statements.
17620 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17621 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17622 adjust_address_nv.
17623
17624 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17625
17626 * gimple-ssa-strength-reduction.c (count_candidates): Change
17627 return value to int.
17628 (analyze_candidates_and_replace): Change type of length to int.
17629
17630 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17631
17632 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17633 (*vec_extract<mode>): Use VI12_128 mode iterator.
17634 (*vec_extract<mode>_mem): Ditto.
17635 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17636 attribute.
17637
17638 2013-05-08 Diego Novillo <dnovillo@google.com>
17639
17640 PR bootstrap/54659
17641
17642 Revert:
17643 2012-08-17 Diego Novillo <dnovillo@google.com>
17644
17645 PR bootstrap/54281
17646 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17647 * config.in: Regenerate.
17648 * configure: Regenerate.
17649 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17650
17651 2013-05-08 Jan Hubicka <jh@suse.cz>
17652
17653 PR lto/54095
17654 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17655 * cgraph.h (symtab_node_base): Add unique_name.
17656 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17657 input_overwrite_node, input_varpool_node): Stream unique_name.
17658 * cgraphclones.c (cgraph_create_virtual_clone,
17659 cgraph_function_versioning): Set unique_name.
17660 * ipa.c (function_and_variable_visibility): Set unique_name.
17661
17662 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17663
17664 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17665 (alloc_cand_and_find_basis): Restrict conditional candidate
17666 processing to CAND_MULTs.
17667
17668 2013-05-08 Jan Hubicka <jh@suse.cz>
17669
17670 PR lto/54095
17671 lto-symtab.c (lto_symtab_symbol_p): New function.
17672 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17673 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17674 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17675 Skip static symbols.
17676
17677 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17678
17679 PR tree-optimization/57200
17680 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17681 Only call inform if the preceding warning_at returns true.
17682
17683 2013-05-07 Han Shen <shenhan@google.com>
17684
17685 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17686 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17687 * common.opt (fstack-protector-strong): New option.
17688 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17689 * doc/invoke.texi (Optimization Options): Document
17690 "-fstack-protector-strong".
17691 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17692
17693 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17694
17695 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17696
17697 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17698
17699 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17700 it is not lhs of the insn.
17701 (setup_reg_equiv): Remove insn having equiv memory which it is not
17702 lhs of the insn.
17703 * lra-constraints.c (process_address): Try to improve generation
17704 code for address base + disp.
17705 (lra_constraints): Make correct the code for checking insn setting
17706 up backward equivalence. Remove insn only if it is in the init
17707 insn list.
17708 * lra-eliminations.c (update_reg_eliminate): Change return value.
17709 (lra_eliminate): Use the result.
17710
17711 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17712
17713 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17714 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17715 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17716 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17717 register target operands.
17718 (*vec_extractv8hi_sse2): New pattern.
17719 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17720 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17721 (*vec_extract<mode>_mem): New insn and split pattern.
17722
17723 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17724
17725 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17726 (TARGET_ASAN_SHADOW_OFFSET): Define.
17727 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17728 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17729
17730 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17731
17732 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17733 (incr_vec_index): Return -1 if increment not found.
17734 (create_add_on_incoming_edge): Assert if increment not found.
17735 (record_increment): Limit number of increments recorded.
17736 (all_phi_incrs_profitable): Return false if an increment not found.
17737 (replace_profitable_candidates): Don't process increments that were
17738 not recorded.
17739 (analyze_candidates_and_replace): Limit size of incr_vec.
17740
17741 2013-05-07 Richard Biener <rguenther@suse.de>
17742
17743 * calls.c (special_function_p): setjmp-like functions are leaf.
17744 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17745 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17746
17747 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17748
17749 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17750 (aarch64_simd_mov<mode>): New expander.
17751 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17752 (aarch64_simd_mov_to_<mode>high): Likewise.
17753 (aarch64_simd_mov_from_<mode>low): Likewise.
17754 (aarch64_simd_mov_from_<mode>high): Likewise.
17755 (aarch64_dup_lane<mode>): Update.
17756 (aarch64_dup_lanedi): New instruction pattern.
17757 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17758 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17759
17760 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17761
17762 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17763 (replace_mult_candidate): Remove unnecessary argument; remove
17764 unnecessary parameter from call to introduce_cast_before_cand.
17765 (replace_unconditional_candidate): Remove unnecessary parameter
17766 from call to replace_mult_candidate.
17767 (replace_conditional_candidate): Likewise.
17768 (insert_initializers): Use make_temp_ssa_name.
17769 (introduce_cast_before_cand): Remove unnecessary argument; use
17770 make_temp_ssa_name.
17771 (replace_one_candidate): Remove unnecessary argument; remove
17772 unnecessary parameter from calls to introduce_cast_before_cand.
17773 (replace_profitable_candidates): Remove unnecessary parameters
17774 from calls to replace_one_candidate.
17775
17776 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17777
17778 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17779 phi def as possibly hiding a basis for a CAND_ADD whose operands
17780 have been commuted in the analysis.
17781 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17782
17783 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17784
17785 * config/aarch64/aarch64.md
17786 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17787 shift value between 0-4.
17788
17789 2013-05-07 Richard Biener <rguenther@suse.de>
17790
17791 * double-int.h (rshift): New overload.
17792 * double-int.c (rshift): New function.
17793 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17794 (create_reference_ops_from_ref): Remove.
17795 (vn_reference_insert): Use shared ops for constructing the
17796 reference and copy it.
17797
17798 2013-05-07 Richard Biener <rguenther@suse.de>
17799
17800 PR middle-end/57190
17801 * tree-eh.c (sink_clobbers): Properly propagate
17802 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17803
17804 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17805
17806 PR tree-optimization/57149
17807 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17808 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17809 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17810 uninit_undefined_value_p instead of ssa_undefined_value_p.
17811
17812 PR debug/57184
17813 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17814 for modifier == EXPAND_INITIALIZER.
17815
17816 2013-05-07 Anton Blanchard <anton@samba.org>
17817
17818 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17819 for powerpc64 little endian.
17820 * configure: Regenerate.
17821
17822 2013-05-06 Graham Stott <grahams@btinternet.com>
17823
17824 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17825 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17826 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17827 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17828
17829 2013-05-06 Graham Stott <grahams@btinternet.com>
17830
17831 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17832 codes which allow non-lvalues.
17833
17834 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17835
17836 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17837 components are all 1s.
17838 (integer_minus_onep): New function.
17839 * tree.h (integer_minus_onep): Declare it.
17840 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17841 integer_minus_onep instead of integer_all_onesp.
17842
17843 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17844
17845 PR target/52933
17846 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17847 variations of these patterns.
17848
17849 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17850
17851 * config/i386/i386.md (isa): Add x64_sse4 member.
17852 (enabled): Handle x64_sse4.
17853 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17854 instruction for 64bit SSE4_1 targets. Update insn attributes.
17855 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17856 instruction for SSE4_1 targets. Update insn attributes.
17857 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17858 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17859 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17860 const_1 selector.
17861 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17862 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17863 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17864
17865 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17866
17867 PR target/57108
17868 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17869
17870 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17871
17872 * final.c (do_assembler_dialects): Don't handle curly braces and
17873 vertical bar escaped by % as dialect delimiters.
17874 (output_asm_insn): Print curly braces and vertical bar if escaped
17875 by % and ASSEMBLER_DIALECT defined.
17876 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17877 * doc/tm.texi: Regenerated.
17878
17879 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17880
17881 * config/mips/mips.c: Include tree-pass.h.
17882 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17883 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17884 (pass_mips_machine_reorg2): New machine specific pass.
17885 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17886 (mips_option_override): Register the new pass.
17887 * rtl.h (cleanup_barriers): Remove prototype.
17888 (dbr_schedule): Likewise.
17889 * jump.c (cleanup_barriers): Make static.
17890 * reorg.c (dbr_schedule): Likewise.
17891
17892 2013-05-06 Richard Biener <rguenther@suse.de>
17893
17894 PR tree-optimization/57185
17895 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17896 SSA names properly.
17897
17898 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17899
17900 PR target/57106
17901 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17902 between "first_arg" and "insn", not "last" and "insn".
17903
17904 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17905
17906 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17907 (find_candidates_in_block): Re-enable slsr_process_phi.
17908 (create_phi_basis): Fix double counting of candidate adjustment.
17909
17910 2013-05-06 Richard Biener <rguenther@suse.de>
17911
17912 PR middle-end/57147
17913 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17914 the edge is also fallthru, preserve it and just clear the
17915 abnormal flag.
17916 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17917 also complex, preserve that and just clear the fallthru flag.
17918 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17919 update virtual operands.
17920
17921 2013-05-06 Alan Modra <amodra@gmail.com>
17922
17923 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17924 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17925 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17926 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17927 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17928 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17929
17930 2013-05-06 Alan Modra <amodra@gmail.com>
17931
17932 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17933 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17934 (DEFAULT_ASM_ENDIAN): Define.
17935 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17936 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17937 Update -K PIC clause from sysv4.h.
17938 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17939 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17940
17941 2013-05-06 Alan Modra <amodra@gmail.com>
17942
17943 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17944 twice for little-endian.
17945 (ashrdi3_no_power, ashrdi3): Support little-endian.
17946
17947 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17948
17949 PR target/55303
17950 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17951 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17952 related expanders.
17953 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17954 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17955 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17956 New predicates.
17957
17958 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17959 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17960
17961 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17962 * config/pa/pa.opt: Make mbig-switch a no-op.
17963 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17964 (CASE_VECTOR_MODE): Always return SImode.
17965 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17966 for the !TARGET_BIG_SWITCH case.
17967 * config/pa/pa-linux.h: Likewise.
17968 * config/pa/pa-openbsd.h: Likewise.
17969 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17970 * config/pa/pa.md (short_jump): Remove define_insn.
17971 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17972 (casesi0): Remove define_insn.
17973 (type): Remove btable_branch.
17974 (pa_combine_type): Likewise.
17975 (in_nullified_branch_delay): Likewise.
17976 (in_call_delay): Likewise.
17977 (define_delay): Likewise.
17978 (define_insn_reservation "Z3"): Likewise.
17979 (define_insn_reservation "Z4"): Likewise.
17980 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17981 (pa_adjust_insn_length): Remove adjustment for btable branches.
17982 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17983 and mno-big-switch
17984
17985 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17986
17987 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17988 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17989 Add m->r,x alternatives.
17990 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17991 splitters using SWI48x mode iterator.
17992 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17993 TARGET_64BIT. Add m->x alternative.
17994 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17995 Add o->x alternative. Enable for TARGET_SSE.
17996 (sse_storeq): Remove expander.
17997 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17998 with memory input operand.
17999 (*vec_extractv2di_1 splitter): New.
18000 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18001 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18002
18003 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18004
18005 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18006 (INT_LOWPART): Delete.
18007 (extract_MB): Adjust.
18008 (extract_ME): Adjust.
18009 (print_operand): Adjust.
18010
18011 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18012
18013 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18014 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18015 (reg_or_logical_cint_operand, easy_fp_constant,
18016 logical_const_operand): Delete "CONST_DOUBLE" case.
18017 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18018 "HOST_BITS_PER_WIDE_INT == 64" test.
18019 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18020 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18021 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18022 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18023 test.
18024 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18025 CONST_DOUBLE DImode/VOIDmode case.
18026 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18027 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18028 CONST_DOUBLE VOIDmode case.
18029 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18030 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18031 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18032 Delete CONST_DOUBLE case.
18033 (splitters for mov FMOVE64 const_double): Delete
18034 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18035 "HOST_BITS_PER_WIDE_INT >= 64" test.
18036 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18037 case.
18038 (mov DI const_double): Delete.
18039
18040 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18041
18042 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18043 on op shows all bits zero in mode of a lowpart subreg, return zero.
18044
18045 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18046
18047 PR target/57150
18048 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18049 to save TFmode registers and DImode to save TImode registers for
18050 caller save operations.
18051 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18052 mark being partially clobbered since they only use the first
18053 double word.
18054
18055 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18056 and TDmode only use the upper 64-bits of each VSX register.
18057
18058 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18059
18060 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18061 (find_candidates_in_block): Disable slsr_process_phi.
18062
18063 2013-05-03 Guozhi Wei <carrot@google.com>
18064
18065 * coverage.c (coverage_obj_init): Move the construction of gcov
18066 constructor to ...
18067 (build_init_ctor): ... here.
18068
18069 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18070
18071 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18072 (slsr_cand_d): Redefine def_phi.
18073 (stride_status, phi_adjust_status, count_phis_status): New enums.
18074 (find_phi_def): New.
18075 (find_basis_for_base_expr): New.
18076 (find_basis_for_candidate): Handle hidden bases.
18077 (alloc_cand_and_find_basis): Handle phi candidates.
18078 (slsr_process_phi): New.
18079 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18080 (create_mul_imm_cand): Likewise.
18081 (create_add_ssa_cand): Exclude phi base candidates.
18082 (create_add_imm_cand): Likewise.
18083 (slsr_process_cast): Likewise.
18084 (slsr_process_copy): Likewise.
18085 (find_candidates_in_block): Handle phi candidates.
18086 (dump_candidate): Likewise.
18087 (unconditional_cands): Delete.
18088 (unconditional_cands_with_known_stride_p): Delete.
18089 (phi_dependent_cand_p): New.
18090 (cand_increment): Handle phi-dependent candidates.
18091 (replace_dependent): Delete.
18092 (replace_mult_candidate): New.
18093 (replace_unconditional_candidate): New.
18094 (incr_vec_index): Move to avoid forward reference.
18095 (create_add_on_incoming_edge): New.
18096 (create_phi_basis): New.
18097 (replace_dependents): Delete.
18098 (replace_conditional_candidate): New.
18099 (phi_add_costs): New.
18100 (replace_uncond_cands_and_profitable_phis): New.
18101 (record_increment): Handle phi adjustments.
18102 (record_phi_increments): New.
18103 (record_increments): Handle phi adjustments.
18104 (phi_incr_cost): New.
18105 (lowest_cost_path): Handle phis.
18106 (total_savings): Likewise.
18107 (analyze_increments): Likewise.
18108 (ncd_with_phi): New.
18109 (ncd_of_cand_and_phis): New.
18110 (nearest_common_dominator_for_cands): Handle phi increments.
18111 (all_phi_incrs_profitable): New.
18112 (replace_profitable_candidates): Handle phi-dependent candidates.
18113 (analyze_candidates_and_replace): Likewise.
18114
18115 2013-05-03 Teresa Johnson <tejohnson@google.com>
18116
18117 PR bootstrap/57154
18118 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18119 do not exceed REG_BR_PROB_BASE.
18120
18121 2013-05-03 Jeff Law <law@redhat.com>
18122
18123 PR tree-optimization/57144
18124 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18125 operand of the condition will bit into the new type when eliminating
18126 a cast feeding a condition.
18127
18128 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18129
18130 PR rtl-optimization/57130
18131 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18132 of COMPARE as in_code to the recursive call if needed.
18133
18134 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18135
18136 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18137 (enabled): Handle new members.
18138 * config/i386/sse.md (*vec_concatv2si): Merge from
18139 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18140 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18141
18142 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18143
18144 PR tree-optimization/57027
18145 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18146 for fnms opportunity, check we got the prerequisite kind
18147 of tree / gimple before using accessor functions.
18148
18149 2013-05-03 Richard Biener <rguenther@suse.de>
18150
18151 * double-int.h (lshift): New overload without precision
18152 and arith argument.
18153 (operator *=, operator +=, operator -=): Move ...
18154 * double-int.c (operator *=, operator +=, operator -=): ... here
18155 and implement more efficiently.
18156 (mul_double_with_sign): Remove.
18157 (lshift_double): Adjust to take unsinged shift argument, push
18158 dispatching code to callers.
18159 (mul_double_wide_with_sign): Add early out for callers that
18160 are not interested in high parts or overflow.
18161 (lshift): New function.
18162 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18163 dispatch code here.
18164 (lrotate, rrotate): Use logical shifts.
18165 * expr.c (get_inner_reference): Use lshift.
18166 * fixed-value.c (do_fixed_divide): Likewise.
18167 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18168 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18169 (indirect_refs_may_alias_p): Likewise.
18170 (stmt_kills_ref_p_1): Likewise.
18171
18172 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18173
18174 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18175
18176 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18177
18178 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18179 scalar form of FABD instruction.
18180
18181 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18182
18183 * lra-constraints.c (process_alt_operands): Add checking alt
18184 number to choose the best alternative.
18185
18186 2013-05-02 Richard Biener <rguenther@suse.de>
18187
18188 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18189 bitmap and its handling.
18190 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18191
18192 2013-05-02 Richard Biener <rguenther@suse.de>
18193
18194 PR middle-end/57140
18195 * tree-inline.c (copy_loops): Properly handle removed loops.
18196 (copy_cfg_body): Mark destination loops for fixup if source
18197 loops needed fixup.
18198
18199 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18200
18201 PR target/56732
18202 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18203 generating simple_return for naked functions.
18204
18205 2013-05-02 Martin Jambor <mjambor@suse.cz>
18206
18207 PR middle-end/56988
18208 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18209 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18210 flags match.
18211 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18212 ipa_agg_replacement_value structures.
18213 (known_aggs_to_agg_replacement_list): Likewise.
18214 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18215 (read_agg_replacement_chain): Likewise.
18216 (ipcp_transform_function): Also check that by_ref flags match.
18217
18218 2013-05-02 Richard Biener <rguenther@suse.de>
18219
18220 * graphds.h (struct graph): Add obstack member.
18221 * graphds.c (new_graph): Initialize obstack and allocate
18222 vertices from it.
18223 (add_edge): Allocate edge from the obstack.
18224 (free_graph): Free the obstack instead of all edges and vertices.
18225
18226 2013-05-02 Teresa Johnson <tejohnson@google.com>
18227
18228 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18229 divides.
18230 * cfg.c (update_bb_profile_for_threading): Ditto.
18231 * tree-inline.c (copy_bb): Ditto.
18232 (copy_edges_for_bb): Ditto.
18233 (initialize_cfun): Ditto.
18234 (copy_cfg_body): Ditto.
18235 (expand_call_inline): Ditto.
18236 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18237 (estimate_node_size_and_time): Ditto.
18238 (inline_merge_summary): Ditto.
18239 * cgraphclones.c (cgraph_clone_edge): Ditto.
18240 (cgraph_clone_node): Ditto.
18241 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18242 (compute_trg_info): Ditto.
18243
18244 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18245
18246 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18247 S reg when fp attribute set.
18248 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18249
18250 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18251
18252 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18253 New pattern.
18254 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18255 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18256 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18257
18258 2013-05-02 Richard Biener <rguenther@suse.de>
18259
18260 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18261 (struct instantiate_cache_entry): New type.
18262 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18263 (struct instantiate_cache_type): New type.
18264 (set_instantiated_value, get_instantiated_value): Remove.
18265 (get_instantiated_value_entry): New function.
18266 (instantiate_scev_name): Use the new cache and adjust.
18267 (instantiate_scev_poly): Adjust.
18268 (instantiate_scev_binary): Likewise.
18269 (instantiate_array_ref): Likewise.
18270 (instantiate_scev_convert): Likewise.
18271 (instantiate_scev_not): Likewise.
18272 (instantiate_scev_3): Likewise.
18273 (instantiate_scev_2): Likewise.
18274 (instantiate_scev_r): Likewise.
18275 (instantiate_scev): Likewise.
18276 (resolve_mixers): Likewise.
18277
18278 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18279
18280 PR target/57091
18281 * lra-constraints.c (best_small_class_operands_num): Remove.
18282 (process_alt_operands): Remove small_class_operands_num. Take
18283 small classes operands into losers and only if the operand is not
18284 matched. Modify debugging output.
18285 (curr_insn_transform): Remove best_small_class_operands_num.
18286 Print insn name.
18287
18288 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18289
18290 * config/aarch64/aarch64-builtins.c
18291 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18292 * config/aarch64/aarch64-simd-builtins.def
18293 (reduc_splus_): Add new modes.
18294 (reduc_uplus_): New.
18295 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18296 (reduc_uplus_v4sf): Likewise.
18297 (reduc_splus_v4sf): Likewise.
18298 (aarch64_addv<mode>): Likewise.
18299 (reduc_uplus_<mode>): Likewise.
18300 (reduc_splus_<mode>): Likewise.
18301 (aarch64_addvv2di): Likewise.
18302 (reduc_uplus_v2di): Likewise.
18303 (reduc_splus_v2di): Likewise.
18304 (aarch64_addvv2si): Likewise.
18305 (reduc_uplus_v2si): Likewise.
18306 (reduc_splus_v2si): Likewise.
18307 (reduc_<sur>plus_<mode>): New.
18308 (reduc_<sur>plus_v2di): Likewise.
18309 (reduc_<sur>plus_v2si): Likewise.
18310 (reduc_<sur>plus_v4sf): Likewise.
18311 (aarch64_addpv4sf): Likewise.
18312 * config/aarch64/arm_neon.h
18313 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18314 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18315 add UNSPEC_SADDV, UNSPEC_UADDV.
18316 (SUADDV): New.
18317 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18318
18319 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18320
18321 * config/aarch64/arm_neon.h
18322 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18323
18324 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18325
18326 * config/aarch64/aarch64-builtins
18327 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18328
18329 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18330
18331 * config/aarch64/aarch64-simd-builtins.def
18332 (reduc_smax_): New.
18333 (reduc_smin_): Likewise.
18334 (reduc_umax_): Likewise.
18335 (reduc_umin_): Likewise.
18336 (reduc_smax_nan_): Likewise.
18337 (reduc_smin_nan_): Likewise.
18338 (fmax): Remove.
18339 (fmin): Likewise.
18340 (smax): Update for V2SF, V4SF and V2DF modes.
18341 (smin): Likewise.
18342 (smax_nan): New.
18343 (smin_nan): Likewise.
18344 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18345 (<su><maxmin><mode>3): ...This, refactor.
18346 (s<maxmin><mode>3): New.
18347 (<maxmin_uns><mode>3): Likewise.
18348 (reduc_<maxmin_uns>_<mode>): Refactor.
18349 (reduc_<maxmin_uns>_v4sf): Likewise.
18350 (reduc_<maxmin_uns>_v2si): Likewise.
18351 (aarch64_<fmaxmin><mode>: Remove.
18352 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18353 new builtin names.
18354 (vmin<q>_f<32,64>): Likewise.
18355 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18356 (FMAXMIN): New.
18357 (su): Add mappings for smax, smin, umax, umin.
18358 (maxmin): New.
18359 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18360 (FMAXMIN): Rename as...
18361 (FMAXMIN_UNS): ...This.
18362 (maxminv): Remove.
18363 (fmaxminv): Likewise.
18364 (fmaxmin): Likewise.
18365 (maxmin_uns): New.
18366 (maxmin_uns_op): Likewise.
18367
18368 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18369
18370 * config/aarch64/arm_neon.h
18371 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18372 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18373 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18374
18375 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18376
18377 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18378 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18379
18380 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18381
18382 * config/aarch64/aarch64-simd.md
18383 (vcond<mode>_internal): Handle special cases for constant masks.
18384 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18385 (vcondu<mode><mode>): Likewise.
18386 (vcond<v_cmp_result><mode>): New.
18387
18388 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18389
18390 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18391 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18392 * config/aarch64/aarch64-simd-builtins.def
18393 (cmeq): Update to BUILTIN_VALLDI.
18394 (cmgt): Likewise.
18395 (cmge): Likewise.
18396 (cmle): Likewise.
18397 (cmlt): Likewise.
18398 * config/aarch64/arm_neon.h
18399 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18400 to builtins or C as appropriate.
18401
18402 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18403
18404 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18405 (cmgeu): ...This.
18406 (cmhi): Rename to...
18407 (cmgtu): ...This.
18408 * config/aarch64/aarch64-simd.md
18409 (simd_mode): Add SF.
18410 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18411 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18412 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18413 (cstore<mode>_neg): ...This.
18414 * config/aarch64/iterators.md
18415 (VALLF): new.
18416 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18417 (COMPARISONS): New.
18418 (UCOMPARISONS): Likewise.
18419 (optab): Add missing comparisons.
18420 (n_optab): New.
18421 (cmp_1): Likewise.
18422 (cmp_2): Likewise.
18423 (CMP): Likewise.
18424 (cmp): Remove.
18425 (VCMP_S): Likewise.
18426 (VCMP_U): Likewise.
18427 (V_cmp_result): Add DF, SF modes.
18428 (v_cmp_result): Likewise.
18429 (v): Likewise.
18430 (vmtype): Likewise.
18431 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18432
18433 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18434
18435 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18436 define_insn to define_insn_and_split.
18437 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18438 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18439 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18440 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18441 (thumb2_negscc): Likewise.
18442
18443 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18444
18445 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18446
18447 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18448
18449 * config/arm/thumb2.md: Remove trailing whitespaces.
18450
18451 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18452
18453 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18454 Use gen_int_mode rather than GEN_INT.
18455
18456 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18457
18458 * value-prof.c (stream_in_histogram_value): Remove the strayed
18459 debug_gimple_stmt.
18460
18461 2013-04-30 Richard Biener <rguenther@suse.de>
18462
18463 PR middle-end/57122
18464 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18465
18466 2013-04-30 Richard Biener <rguenther@suse.de>
18467
18468 PR middle-end/57107
18469 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18470
18471 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18472
18473 PR rtl-optimization/56957
18474 PR rtl-optimization/57105
18475 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18476 variable. Use just INSN_UID for determining whether an insn
18477 should be only disconnected from the insn stream.
18478 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18479
18480 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18481
18482 PR tree-optimization/57104
18483 * tsan.c (instrument_expr): Don't instrument accesses to
18484 DECL_HARD_REGISTER VAR_DECLs.
18485
18486 2013-04-30 Richard Biener <rguenther@suse.de>
18487
18488 * function.h (loops_for_fn): New inline function.
18489 (set_loops_for_fn): Likewise.
18490 * cfgloop.h (place_new_loop): Add struct function parameter.
18491 (get_loop): Likewise.
18492 (get_loops): Likewise.
18493 (number_of_loops): Likewise.
18494 (fel_next): Adjust.
18495 (fel_init): Likewise.
18496 * cfg.c (get_loop_copy): Adjust.
18497 * cfgloop.c (flow_loops_dump): Likewise.
18498 (record_loop_exits): Likewise.
18499 (verify_loop_structure): Likewise.
18500 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18501 (estimate_reg_pressure_cost): Likewise.
18502 (mark_loop_exit_edges): Likewise.
18503 * cfgloopmanip.c (place_new_loop): Likewise.
18504 (add_loop): Likewise.
18505 (duplicate_loop): Likewise.
18506 * graph.c (draw_cfg_nodes): Likewise.
18507 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18508 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18509 (extract_affine_chrec): Likewise.
18510 (build_scop_iteration_domain): Likewise.
18511 * graphite.c (graphite_initialize): Likewise.
18512 * ira-build.c (create_loop_tree_nodes): Likewise.
18513 (more_one_region_p): Likewise.
18514 (rebuild_regno_allocno_maps): Likewise.
18515 (mark_loops_for_removal): Likewise.
18516 (mark_all_loops_for_removal): Likewise.
18517 (remove_unnecessary_regions): Likewise.
18518 (ira_build): Likewise.
18519 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18520 * loop-init.c (fix_loop_structure): Likewise.
18521 (gate_rtl_move_loop_invariants): Likewise.
18522 (gate_rtl_unswitch): Likewise.
18523 (gate_rtl_unroll_and_peel_loops): Likewise.
18524 (rtl_doloop): Likewise.
18525 * lto-streamer-in.c (input_cfg): Likewise.
18526 * lto-streamer-out.c (output_cfg): Likewise.
18527 * modulo-sched.c (sms_schedule): Likewise.
18528 * predict.c (tree_estimate_probability): Likewise.
18529 (tree_estimate_probability_driver): Likewise.
18530 (estimate_loops): Likewise.
18531 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18532 (move_sese_region_to_fn): Likewise.
18533 (debug_loop_num): Likewise.
18534 * tree-chrec.c (chrec_evaluate): Likewise.
18535 (hide_evolution_in_other_loops_than_loop): Likewise.
18536 (chrec_component_in_loop_num): Likewise.
18537 (reset_evolution_in_loop): Likewise.
18538 (evolution_function_is_invariant_rec_p): Likewise.
18539 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18540 * tree-inline.c (copy_loops): Likewise.
18541 (copy_cfg_body): Likewise.
18542 (tree_function_versioning): Likewise.
18543 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18544 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18545 Likewise.
18546 (add_to_evolution_1): Likewise.
18547 (scev_const_prop): Likewise.
18548 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18549 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18550 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18551 (tree_ssa_lim_initialize): Likewise.
18552 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18553 (verify_loop_closed_ssa): Likewise.
18554 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18555 (tree_ssa_loop_im): Likewise.
18556 (tree_ssa_loop_unswitch): Likewise.
18557 (tree_vectorize): Likewise.
18558 (check_data_deps): Likewise.
18559 (tree_ssa_loop_ivcanon): Likewise.
18560 (tree_ssa_loop_bounds): Likewise.
18561 (tree_complete_unroll): Likewise.
18562 (tree_complete_unroll_inner): Likewise.
18563 (tree_parallelize_loops): Likewise.
18564 (tree_ssa_loop_prefetch): Likewise.
18565 (tree_ssa_loop_ivopts): Likewise.
18566 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18567 * tree-vectorizer.c (vectorize_loops): Likewise.
18568
18569 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18570
18571 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18572 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18573 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18574 with EABI_LINK_SPEC.
18575
18576 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18577
18578 PR target/44578
18579 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18580 alternative.
18581
18582 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18583
18584 PR target/57097
18585 * lra-constraints.c (process_alt_operands): Discourage a bit more
18586 using memory for pseudos. Print cost dump for alternatives.
18587 Modify cost values for conflicts with early clobbers.
18588 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18589
18590 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18591
18592 PR target/57098
18593 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18594
18595 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18596
18597 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18598 from/to S register.
18599 (movdi_aarch64): Support LDR/STR from/to D register.
18600
18601 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18602
18603 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18604 or higher by default.
18605
18606 2013-04-29 Richard Biener <rguenther@suse.de>
18607
18608 PR middle-end/57075
18609 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18610 even if not adding abnormal edges for calls that can make
18611 abnormal gotos.
18612
18613 2013-04-29 Richard Biener <rguenther@suse.de>
18614
18615 PR middle-end/57103
18616 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18617 TREE_BLOCK.
18618 (move_stmt_r): Remove redundant checking.
18619
18620 2013-04-29 Teresa Johnson <tejohnson@google.com>
18621
18622 PR bootstrap/57077
18623 * basic-block.h (apply_scale): New function.
18624 (apply_probability): Use apply_scale.
18625 * gimple-streamer-in.c (input_bb): Ditto.
18626 * lto-streamer-in.c (input_cfg): Ditto.
18627 * lto-cgraph.c (merge_profile_summaries): Ditto.
18628 * tree-optimize.c (execute_fixup_cfg): Ditto.
18629 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18630 (copy_edges_for_bb): Ditto.
18631 (copy_cfg_body): Ditto.
18632
18633 2013-04-29 Tom de Vries <tom@codesourcery.com>
18634
18635 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18636 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18637 (tail_merge_optimize): Handle current_loops == NULL.
18638
18639 2013-04-26 Jeff Law <law@redhat.com>
18640
18641 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18642 (simplify_cond_using_ranges): Generalize code to simplify
18643 COND_EXPRs where one argument is a constant and the other
18644 is an SSA_NAME created by an integral type conversion.
18645
18646 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18647
18648 * config/arm/arm.md (store_minmaxsi): Use only when
18649 optimize_insn_for_size_p.
18650
18651 2013-04-29 Christian Bruel <christian.bruel@st.com>
18652
18653 PR target/57108
18654 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18655
18656 2013-04-29 Richard Biener <rguenther@suse.de>
18657
18658 PR middle-end/57089
18659 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18660 loop tree make sure to schedule a fixup for the child as well.
18661 (expand_omp_for_generic): Properly add loops.
18662 (expand_omp_for_static_nochunk): Likewise.
18663 (expand_omp_for_static_chunk): Likewise.
18664 (expand_omp_for): For the degenerate case fixup loops.
18665 (expand_omp_sections): Fix default bb placement in loops.
18666 (expand_omp_atomic_pipeline): Properly add loops.
18667
18668 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18669
18670 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18671
18672 2013-04-29 Tom de Vries <tom@codesourcery.com>
18673
18674 * tree-ssa-tail-merge.c: Update header comment.
18675
18676 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18677
18678 * config/aarch64/arm_neon.h
18679 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18680 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18681 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18682 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18683 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18684 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18685 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18686 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18687
18688 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18689
18690 * config/aarch64/aarch64-simd.md
18691 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18692 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18693 fix_trunc, fixuns_trunc.
18694 (ftrunc<VDQF:mode>2): New.
18695 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18696 (fix_trunc_optab): New.
18697
18698 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18699
18700 * config/aarch64/aarch64-builtins.c
18701 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18702 iceilf, lround, iroundf.
18703
18704 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18705
18706 PR target/54349
18707 * config/i386/i386.h (enum ix86_tune_indices)
18708 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18709 New, split from X86_TUNE_INTER_UNIT_MOVES.
18710 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18711 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18712 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18713 (TARGET_INTER_UNIT_MOVES): Remove.
18714 * config/i386/i386.c (initial_ix86_tune_features): Update.
18715 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18716 (ix86_expand_convert_uns_didf_sse): Use
18717 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18718 (ix86_expand_vector_init_one_nonzero): Ditto.
18719 (ix86_expand_vector_init_interleave): Ditto.
18720 (inline_secondary_memory_needed): Return true for moves from SSE class
18721 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18722 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18723 * config/i386/constraints.md (Yi, Ym): Depend on
18724 TARGET_INTER_UNIT_MOVES_TO_VEC.
18725 (Yj, Yn): New constraints.
18726 * config/i386/i386.md (*movdi_internal): Change constraints of
18727 operand 1 from Yi to Yj and from Ym to Yn.
18728 (*movsi_internal): Ditto.
18729 (*movdf_internal): Ditto.
18730 (*movsf_internal): Ditto.
18731 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18732 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18733 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18734 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18735 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18736 * config/i386/sse.md (movdi_to_sse): Ditto.
18737 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18738 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18739 TARGET_INTER_UNIT_MOVES.
18740 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18741 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18742 instead of TARGET_INTER_UNIT_MOVES.
18743 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18744 operand 1 from Yi to Yj and from Ym to Yn.
18745
18746 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18747
18748 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18749 (float_truncate_hi_): Likewise.
18750 (float_extend_lo_): Likewise.
18751 (float_truncate_lo_): Likewise.
18752 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18753 (aarch64_float_extend_lo_v2df): Likewise.
18754 (vec_unpacks_hi_v4sf): Likewise.
18755 (aarch64_float_truncate_lo_v2sf): Likewise.
18756 (aarch64_float_truncate_hi_v4sf): Likewise.
18757 (vec_pack_trunc_v2df): Likewise.
18758 (vec_pack_trunc_df): Likewise.
18759
18760 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18761
18762 * config/aarch64/aarch64-builtins.c
18763 (aarch64_fold_builtin): Fold float conversions.
18764 * config/aarch64/aarch64-simd-builtins.def
18765 (floatv2si, floatv4si, floatv2di): New.
18766 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18767 * config/aarch64/aarch64-simd.md
18768 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18769 * config/aarch64/iterators.md (FLOATUORS): New.
18770 (optab): Add float, floatuns.
18771 (su_optab): Likewise.
18772
18773 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18774
18775 * config/aarch64/aarch64-builtins.c
18776 (aarch64_builtin_vectorized_function): Use new names for
18777 fcvt builtins.
18778 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18779 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18780 (fcvtzu): Split as...
18781 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18782 (fcvtas): Split as...
18783 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18784 (fcvtau): Split as...
18785 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18786 (fcvtps): Split as...
18787 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18788 (fcvtpu): Split as...
18789 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18790 (fcvtms): Split as...
18791 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18792 (fcvtmu): Split as...
18793 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18794 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18795 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18796 (lfrintnusf, lfrintnudf): Likewise.
18797 * config/aarch64/aarch64-simd.md
18798 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18799 define_insn.
18800 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18801 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18802 (fcvt_pattern): Likewise.
18803
18804 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18805
18806 * config/aarch64/aarch64-simd.md
18807 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18808 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18809
18810 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18811
18812 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18813 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18814 (vrnd<a,m,n,p>_f32): Implement using builtins.
18815 (vrnd<i,x><q>_f<32, 64>): New.
18816
18817 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18818
18819 * config/aarch64/aarch64-builtins.c
18820 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18821 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18822 (frintz): Rename to...
18823 (btrunc): ...this.
18824 (frintp): Rename to...
18825 (ceil): ...this.
18826 (frintm): Rename to...
18827 (floor): ...this.
18828 (frinti): Rename to...
18829 (nearbyint): ...this.
18830 (frintx): Rename to...
18831 (rint): ...this.
18832 (frinta): Rename to...
18833 (round): ...this.
18834 * config/aarch64/aarch64-simd.md
18835 (aarch64_frint<frint_suffix><mode>): Delete.
18836 (<frint_pattern><mode>2): Convert to insn.
18837 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18838 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18839 (frint_pattern): Likewise.
18840 (frint_suffix): Likewise.
18841
18842 2013-04-29 Richard Biener <rguenther@suse.de>
18843
18844 PR tree-optimization/57081
18845 * loop-init.c: Include tree-flow.h.
18846 (loop_optimizer_finalize): Free number of iteration estimates.
18847 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18848
18849 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18850
18851 PR tree-optimization/57083
18852 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18853 non-singleton shift count range, zero extend low_bound for uns case.
18854
18855 * config/i386/predicates.md (general_vector_operand): New predicate.
18856 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18857 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18858 if they aren't nonimmediate operands. If their original values
18859 satisfy const_vector_equal_evenodd_p, don't shift them.
18860 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18861 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18862 if not nonimmediate_operand.
18863 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18864 instead of register_operand.
18865 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18866
18867 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18868
18869 * stor-layout.c (finalize_size_functions): Allocate a structure and
18870 reset cfun before dumping the functions.
18871
18872 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18873
18874 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18875
18876 PR target/56866
18877 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18878 use xop_pmacsdqh if uns_p.
18879 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18880 the immediate rotate count.
18881
18882 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18883
18884 * rtl.h (struct rtx_def): Add comment for field jump.
18885 (LRA_SUBREG_P): New macro.
18886 * recog.c (register_operand): Check LRA_SUBREG_P.
18887 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18888 stack frame.
18889 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18890 (lra_final_code_change): Skip subreg change for operators.
18891 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18892 if there are no operand changes.
18893 * lra-constraints.c (curr_insn_set): New.
18894 (match_reload): Set LRA_SUBREG_P.
18895 (emit_spill_move): Ditto.
18896 (check_and_process_move): Use curr_insn_set. Process only single
18897 set insns. Don't initialize sec_mem_p and change_p.
18898 (simplify_operand_subreg): Use LRA_SUBREG_P.
18899 (reg_in_class_p): New function.
18900 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18901 of #ifdef. Add code to remove cycling.
18902 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18903 non-null disp. Reload inner instead of disp when base and index
18904 are null. Try to put lo_sum into register.
18905 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18906 (check_and_process_move): Move code for move cost check to
18907 simple_move_p. Remove equiv_substitution.
18908 (simple_move_p): New function.
18909 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18910 curr_insn_set. Call check_and_process_move only for single set
18911 insns. Use the new function. Move call of check_and_process_move
18912 after operand equiv substitution and address process.
18913
18914 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18915
18916 PR go/57045
18917 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18918 with nonlocal goto receivers or returns twice calls, ignore
18919 unininitialized values from abnormal edges to nl goto receiver
18920 or returns twice call.
18921
18922 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18923
18924 PR tree-optimization/57051
18925 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18926 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18927 bitsize.
18928
18929 2013-04-26 Richard Biener <rguenther@suse.de>
18930
18931 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18932 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18933 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18934 (fixup_loop_arrays_after_move): New function.
18935 (move_sese_region_to_fn): Properly outline the loop tree parts
18936 of the SESE region.
18937
18938 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18939
18940 * config/i386/i386.md (type, unit): Fix long lines.
18941
18942 2013-04-26 Richard Biener <rguenther@suse.de>
18943
18944 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18945 (lto-streamer-out.o): Likewise.
18946 * cfgloop.c (init_loops_structure): Export, add struct function
18947 argument and adjust.
18948 (flow_loops_find): Adjust.
18949 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18950 (init_loops_structure): Declare.
18951 * lto-streamer-in.c: Include cfgloop.h.
18952 (input_cfg): Input the loop tree.
18953 * lto-streamer-out.c: Include cfgloop.h.
18954 (output_cfg): Output the loop tree.
18955 (output_struct_function_base): Do not drop PROP_loops.
18956
18957 2013-03-26 Richard Biener <rguenther@suse.de>
18958
18959 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18960 (pass_build_cfg): Provide PROP_loops.
18961 (move_sese_region_to_fn): Remove loops that are outlined into fn
18962 for now.
18963 * tree-inline.c: Include cfgloop.h.
18964 (initialize_cfun): Do not drop PROP_loops.
18965 (copy_loops): New function.
18966 (copy_cfg_body): Copy loop structure.
18967 (tree_function_versioning): Initialize destination loop tree.
18968 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18969 (pass_parallelize_loops): Do IL verification.
18970 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18971 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18972 the CFG make sure we fixup loops as well.
18973 * tree-ssa-tail-merge.c: Include cfgloop.h.
18974 (replace_block_by): When merging loop latches mark loops for fixup.
18975 * lto-streamer-out.c (output_struct_function_base): Drop
18976 PROP_loops for now.
18977 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18978 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18979 * ipa-split.c: Include cfgloop.h.
18980 (split_function): Add the new return block to the loop tree root.
18981 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18982 whether we have removed the forwarder block.
18983 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18984 * cfgloop.h (place_new_loop): Declare.
18985 * cfgloopmanip.c (place_new_loop): Export.
18986 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18987 (tree-switch-conversion.o): Likewise.
18988 (tree-complex.o): Likewise.
18989 (tree-inline.o): Likewise.
18990 (tree-ssa-tailmerge.o): Likewise.
18991 (ipa-split.o): Likewise.
18992 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18993 (tree-ssa-copy.o): Likewise.
18994 * tree-switch-conversion.c: Include cfgloop.h
18995 (process_switch): If we emit a bit-test cascade, schedule loops
18996 for fixup.
18997 * tree-complex.c: Include cfgloop.h.
18998 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18999 * asan.c: Include cfgloop.h.
19000 (create_cond_insert_point): Properly add new basic-blocks to
19001 loops, schedule loop fixup.
19002 * cfgloop.c (verify_loop_structure): Check that looks are not
19003 marked for fixup.
19004 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19005 to loops.
19006 (expand_omp_for_generic): Likewise.
19007 (expand_omp_sections): Likewise.
19008 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19009 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19010 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19011 is initialized, not when loops are present.
19012 * tree-parloops.c (parallelize_loops): Remove checking here.
19013 * passes.c (init_optimization_passes): Schedule a copy-propagation
19014 pass before complete unrolling of inner loops.
19015
19016 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19017
19018 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19019 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19020 (colorize_init): Add argument to _WIN32 version.
19021 * toplev.c: Include diagnostic-color.h.
19022 (process_options): Default to -fdiagnostics-color=auto if
19023 GCC_COLORS env var is in the environment.
19024 * common.opt (fdiagnostics-color=): Add Var and Init.
19025 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19026 env var is in the environment, the default is auto rather than never.
19027
19028 * diagnostic.h (file_name_as_prefix): Add context argument.
19029 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19030 the string as locus.
19031 * langhooks.c (lhd_print_error_function): Adjust caller.
19032
19033 2013-04-25 Lawrence Crowl <crowl@google.com>
19034
19035 * var-tracking.c (shared_hash_def::htab):
19036 Change type to hash_table. Update dependent calls and types.
19037
19038 2013-04-25 Lawrence Crowl <crowl@google.com>
19039
19040 * Makefile.in: Update as needed below.
19041
19042 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19043 Move declaration to after the type's method definitons.
19044
19045 * attribs.c (htab_t scoped_attributes::attribute_hash):
19046 Change type to hash_table. Update dependent calls and types.
19047
19048 * bitmap.c (htab_t bitmap_desc_hash):
19049 Change type to hash_table. Update dependent calls and types.
19050
19051 * cselib.c (htab_t cselib_hash_table):
19052 Change type to hash_table. Update dependent calls and types.
19053
19054 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19055 (hash_string_slot_node): Move implementation into lto-streamer.h
19056 struct string_slot_hasher.
19057 (eq_string_slot_node): Likewise.
19058
19059 * data-streamer-out.c: Update output_block::string_hash_table
19060 dependent calls and types.
19061
19062 * dwarf2cfi.c (htab_t trace_index):
19063 Change type to hash_table. Update dependent calls and types.
19064
19065 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19066 Change type to hash_table. Update dependent calls and types.
19067 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19068 (htab_t optimize_external_refs::map): Likewise.
19069 (htab_t output_comp_unit::extern_map): Likewise.
19070 (htab_t output_comdat_type_unit::extern_map): Likewise.
19071 (htab_t output_macinfo::macinfo_htab): Likewise.
19072 (htab_t optimize_location_lists::htab): Likewise.
19073 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19074
19075 * except.c (htab_t ehspec_hash_type):
19076 Change type to hash_table. Update dependent calls and types.
19077 (assign_filter_values::ttypes): Likewise.
19078 (assign_filter_values::ehspec): Likewise.
19079 (sjlj_assign_call_site_values::ar_hash): Likewise.
19080 (convert_to_eh_region_ranges::ar_hash): Likewise.
19081
19082 * gcse.c (htab_t pre_ldst_table):
19083 Change type to hash_table. Update dependent calls and types.
19084
19085 * ggc-common.c (htab_t saving_htab):
19086 Change type to hash_table. Update dependent calls and types.
19087 (htab_t loc_hash): Likewise.
19088 (htab_t ptr_hash): Likewise.
19089 (call_count): Rename ggc_call_count.
19090 (call_alloc): Rename ggc_call_alloc.
19091 (loc_descriptor): Rename make_loc_descriptor.
19092 (add_statistics): Rename ggc_add_statistics.
19093
19094 * ggc-common.c (saving_htab):
19095 Change type to hash_table. Update dependent calls and types.
19096
19097 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19098 (push_gimplify_context): Likewise.
19099 (pop_gimplify_context): Likewise.
19100 (struct gimple_temp_hash_elt): Added.
19101 (struct gimplify_hasher): Likewise.
19102 (struct gimplify_ctx.temp_htab):
19103 Change type to hash_table. Update dependent calls and types.
19104
19105 * gimple-fold.c: Include gimplify-ctx.h.
19106
19107 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19108 Change type to hash_table. Update dependent calls and types.
19109 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19110 avoid potential global name collision.
19111
19112 * gimplify.c: Include gimplify-ctx.h.
19113 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19114 (htab_t gimplify_ctx::temp_htab):
19115 Update dependent calls and types for new type hash_table.
19116 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19117 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19118
19119 * gimplify-ctx.h: New.
19120 (struct gimple_temp_hash_elt): Move from gimplify.c.
19121 (class gimplify_hasher): New.
19122 (struct gimplify_ctx): Move from gimple.h.
19123 (htab_t gimplify_ctx::temp_htab):
19124 Change type to hash_table. Update dependent calls and types.
19125
19126 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19127 (htab_t ivs_params::newivs_index):
19128 Change type to hash_table. Update dependent calls and types.
19129 (htab_t ivs_params::params_index): Likewise.
19130 (htab_t print_generated_program::params_index): Likewise.
19131 (htab_t gloog::newivs_index): Likewise.
19132 (htab_t gloog::params_index): Likewise.
19133
19134 * graphite.c: Include graphite-htab.h.
19135 4htab_t graphite_transform_loops::bb_pbb_mapping):
19136 Change type to hash_table. Update dependent calls and types.
19137
19138 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19139 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19140 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19141
19142 * graphite-dependences.c: Include graphite-htab.h.
19143 (loop_is_parallel_p): Change hash table type of parameter.
19144
19145 * graphite-htab.h: New.
19146 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19147 (extern find_pbb_via_hash): Move from graphite-poly.h.
19148 (extern loop_is_parallel_p): Move from graphite-poly.h.
19149 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19150
19151 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19152 (extern loop_is_parallel_p): Move to graphite-htab.h.
19153 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19154
19155 * haifa-sched.c (htab_t delay_htab):
19156 Change type to hash_table. Update dependent calls and types.
19157 (htab_t delay_htab_i2): Likewise.
19158
19159 * ira-color.c (htab_t allocno_hard_regs_htab):
19160 Change type to hash_table. Update dependent calls and types.
19161
19162 * ira-costs.c (htab_t cost_classes_htab):
19163 Change type to hash_table. Update dependent calls and types.
19164
19165 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19166 Change type to hash_table. Update dependent calls and types.
19167
19168 * loop-iv.c (htab_t bivs):
19169 Change type to hash_table. Update dependent calls and types.
19170
19171 * loop-unroll.c (htab_t opt_info::insns_to_split):
19172 Change type to hash_table. Update dependent calls and types.
19173 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19174
19175 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19176 (struct string_slot_hasher): New.
19177 (htab_t output_block::string_hash_table):
19178 Change type to hash_table. Update dependent calls and types.
19179
19180 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19181 (htab_t file_name_hash_table):
19182 Change type to hash_table. Update dependent calls and types.
19183
19184 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19185 calls and types.
19186
19187 * lto-streamer.c (htab_t tree_htab):
19188 Change type to hash_table. Update dependent calls and types.
19189
19190 * omp-low.c: Include gimplify-ctx.h.
19191
19192 * passes.c (htab_t name_to_pass_map):
19193 Change type to hash_table. Update dependent calls and types.
19194 (pass_traverse): Rename to passes_pass_traverse.
19195
19196 * plugin.c (htab_t event_tab):
19197 Change type to hash_table. Update dependent calls and types.
19198
19199 * postreload-gcse.c (htab_t expr_table):
19200 Change type to hash_table. Update dependent calls and types.
19201 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19202
19203 * sese.c (debug_rename_map_1): Make extern.
19204 (htab_t copy_bb_and_scalar_dependences::rename_map):
19205 Change type to hash_table. Update dependent calls and types.
19206
19207 * sese.h (extern debug_rename_map): Move to .c file.
19208
19209 * store-motion.c (htab_t store_motion_mems_table):
19210 Change type to hash_table. Update dependent calls and types.
19211
19212 * trans-mem.c (htab_t tm_new_mem_hash):
19213 Change type to hash_table. Update dependent calls and types.
19214
19215 * tree-browser.c (htab_t TB_up_ht):
19216 Change type to hash_table. Update dependent calls and types.
19217
19218 * tree-cfg.c (htab_t discriminator_per_locus):
19219 Change type to hash_table. Update dependent calls and types.
19220
19221 * tree-complex.c: Include tree-hasher.h
19222 (htab_t complex_variable_components):
19223 Change type to hash_table. Update dependent calls and types.
19224
19225 * tree-eh.c (htab_t finally_tree):
19226 Change type to hash_table. Update dependent calls and types.
19227
19228 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19229 struct int_tree_hasher.
19230 (extern int_tree_map_eq): Likewise.
19231 (uid_decl_map_hash): Removed.
19232 (extern decl_tree_map_eq): Likewise.
19233
19234 * tree-hasher.h: New.
19235 (struct int_tree_hasher): New.
19236 (typedef int_tree_htab_type): New.
19237
19238 * tree-inline.c: Include gimplify-ctx.h.
19239
19240 * tree-mudflap.c: Include gimplify-ctx.h.
19241
19242 * tree-parloops.c: Include tree-hasher.h.
19243 (htab_t eliminate_local_variables_stmt::decl_address):
19244 Change type to hash_table. Update dependent calls and types.
19245 (htab_t separate_decls_in_region::decl_copies): Likewise.
19246
19247 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19248 Change type to hash_table. Update dependent calls and types.
19249
19250 * tree-sra.c (candidates):
19251 Change type to hash_table. Update dependent calls and types.
19252
19253 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19254 in tree-flow.h.
19255 (int_tree_map_hash): Likewise.
19256
19257 * tree-ssa-dom.c (htab_t avail_exprs):
19258 Change type to hash_table. Update dependent calls and types.
19259
19260 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19261 Change type to hash_table. Update dependent calls and types.
19262
19263 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19264 Change type to hash_table. Update dependent calls and types.
19265
19266 * tree-ssa-phiopt.c (seen_ssa_names):
19267 Change type to hash_table. Update dependent calls and types.
19268
19269 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19270 Change type to hash_table. Update dependent calls and types.
19271
19272 * tree-ssa-uncprop.c (equiv):
19273 Change type to hash_table. Update dependent calls and types.
19274
19275 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19276
19277 PR rtl-optimization/57003
19278 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19279 call note_stores with kill_clobbered_value callback again after
19280 killing regs_invalidated_by_call.
19281
19282 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19283
19284 * config/aarch64/aarch64-simd.md
19285 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19286 (aarch64_simd_bsl<mode>): Likewise.
19287 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19288
19289 2013-04-25 Marek Polacek <polacek@redhat.com>
19290
19291 PR tree-optimization/57066
19292 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19293
19294 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19295
19296 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19297
19298 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19299
19300 * config/aarch64/aarch64-builtins.c
19301 (aarch64_fold_builtin): New.
19302 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19303 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19304 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19305 * config/aarch64/arm_neon.h
19306 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19307
19308 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19309 Tejas Belagod <tejas.belagod@arm.com>
19310
19311 * config/aarch64/aarch64-builtins.c
19312 (aarch64_gimple_fold_builtin): New.
19313 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19314 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19315 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19316 (addvv4sf): Update.
19317 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19318
19319 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19320
19321 * config/aarch64/aarch64.md
19322 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19323
19324 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19325
19326 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19327 (*ngcsi_uxtw): New pattern.
19328
19329 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19330 Julian Brown <julian@codesourcery.com>
19331
19332 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19333 (TB_DREG): Add T_V4HF.
19334 (v4hf_UP): New macro.
19335 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19336 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19337 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19338 built-ins.
19339 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19340 (arm_vector_mode_supported_p): Handle V4HF.
19341 (arm_mangle_map): Handle V4HFmode.
19342 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19343 * config/arm/arm_neon_builtins.def: Add entries for
19344 vcvtv4hfv4sf, vcvtv4sfv4hf.
19345 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19346 (neon_vcvtv4hfv4sf): Likewise.
19347 * config/arm/neon-gen.ml: Handle half-precision floating point
19348 features.
19349 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19350 * config/arm/arm_neon.h: Regenerate.
19351 * config/arm/neon.ml (type elts): Add F16.
19352 (type vectype): Add T_float16x4, T_floatHF.
19353 (type vecmode): Add V4HF.
19354 (type features): Add Requires_FP_bit feature.
19355 (elt_width): Handle F16.
19356 (elt_class): Likewise.
19357 (elt_of_class_width): Likewise.
19358 (mode_of_elt): Refactor.
19359 (type_for_elt): Handle F16, fix error messages.
19360 (vectype_size): Handle T_float16x4.
19361 (vcvt_sh): New function.
19362 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19363 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19364 (string_of_mode): Handle V4HF.
19365 * doc/arm-neon-intrinsics.texi: Regenerate.
19366
19367 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19368
19369 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19370 format specifier in 'X' case.
19371
19372 2013-04-25 Alan Modra <amodra@gmail.com>
19373
19374 PR target/57052
19375 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19376 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19377 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19378 Repeat for many other rotate/shift and mask patterns using subregs.
19379 Name lshiftrt insns.
19380 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19381 on WORDS_BIG_ENDIAN.
19382
19383 2013-04-25 Alan Modra <amodra@gmail.com>
19384
19385 * config.gcc: Support little-endian powerpc-linux targets.
19386 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19387 (LINK_OS_LINUX_SPEC): Define.
19388 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19389 Preserve MASK_LITTLE_ENDIAN.
19390 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19391 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19392 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19393 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19394 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19395 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19396 Correct fp word order for little-endian. Don't shift toc entries
19397 smaller than a word for little-endian.
19398 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19399 (bswapdi2 splits): Correct low-part subreg for little-endian.
19400 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19401 low/high where such is correct only for be.
19402 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19403 little-endian for -mcall-aixdesc.
19404
19405 2013-04-25 Alan Modra <amodra@gmail.com>
19406
19407 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19408 replace_equiv_address_nv.
19409
19410 2013-04-25 Alan Modra <amodra@gmail.com>
19411
19412 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19413
19414 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19415
19416 Revert:
19417 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19418 * rtl.h (struct rtx_def): ...
19419
19420 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19421
19422 PR rtl-optimizations/57046
19423 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19424 for multi-reg splits.
19425
19426 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19427
19428 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19429
19430 2013-04-24 Sterling Augustine <saugustine@google.com>
19431
19432 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19433 (comp_dir_string, debug_str_dwo_section): New.
19434 (DEBUG_STR_DWO_SECTION): Rename to ...
19435 (DEBUG_DWO_STR_SECTION): ... this.
19436 (DEBUG_NORM_STR_SECTION): Delete.
19437 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19438 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19439 (find_AT_string): Move most logic to ...
19440 (find_AT_string_in_table): ... here. New.
19441 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19442 add_skeleton_AT_string. Delete logic.
19443 (output_skeleton_debug_sections): Remove call to
19444 add_top_level_skeleton_die_attrs.
19445 (add_comp_dir_attribute): Move logic to comp_dir_string.
19446 (dwarf2out_init): Initialize debug_str_dwo_section.
19447 (output_indirect_string): Call find_string_form.
19448 (output_indirect_strings): Rewrite.
19449 (prune_unused_types): Empty skeleton_debug_str_hash.
19450 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19451 (dwarf2out_finish): Call output_indirect_strings.
19452
19453 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19454
19455 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19456
19457 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19458
19459 * rtl.h (struct rtx_def): Add comment for field jump.
19460 (LRA_SUBREG_P): New macro.
19461 * recog.c (register_operand): Check LRA_SUBREG_P.
19462 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19463 stack frame.
19464 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19465 (lra_final_code_change): Skip subreg change for operators.
19466 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19467 if there are no operand changes.
19468 * lra-constraints.c (curr_insn_set): New.
19469 (match_reload): Set LRA_SUBREG_P.
19470 (emit_spill_move): Ditto.
19471 (check_and_process_move): Use curr_insn_set. Process only single
19472 set insns. Don't initialize sec_mem_p and change_p.
19473 (simplify_operand_subreg): Use LRA_SUBREG_P.
19474 (reg_in_class_p): New function.
19475 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19476 of #ifdef. Add code to remove cycling.
19477 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19478 non-null disp. Reload inner instead of disp when base and index
19479 are null. Try to put lo_sum into register.
19480 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19481 (check_and_process_move): Move code for move cost check to
19482 simple_move_p. Remove equiv_substitution.
19483 (simple_move_p): New function.
19484 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19485 curr_insn_set. Call check_and_process_move only for single set
19486 insns. Use the new function. Move call of check_and_process_move
19487 after operand equiv substitution and address process.
19488
19489 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19490
19491 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19492 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19493 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19494
19495 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19496
19497 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19498
19499 2013-04-24 Marek Polacek <polacek@redhat.com>
19500
19501 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19502 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19503 (select_loops_exit_conditions): Likewise.
19504 (number_of_iterations_for_all_loops): Likewise.
19505 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19506 (scev_analysis): Likewise.
19507
19508 2013-04-02 Catherine Moore <clm@codesourcery.com>
19509 Chao-ying Fu <fu@mips.com>
19510
19511 * config/mips/micromips.md (jraddiusp): New pattern.
19512 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19513 instruction if possible.
19514
19515 2013-04-24 Alan Modra <amodra@gmail.com>
19516
19517 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19518
19519 2013-04-24 Julian Brown <julian@codesourcery.com>
19520 Chung-Lin Tang <cltang@codesourcery.com>
19521
19522 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19523 dependency behavior in enumeration type DIE generation. Add TODO note
19524 to comments about future DW_FORM_sdata/udata re-work of related code.
19525
19526 2013-04-23 Lawrence Crowl <crowl@google.com>
19527
19528 * Makefile.in: Update as needed below.
19529
19530 * hash-table.h (class hash_table):
19531 Correct many methods with parameter types compare_type to the correct
19532 value_type. (Correct code was unlikely to notice the change.)
19533 (hash_table::elements_with_deleted) New.
19534 (class hashtable::iterator): New.
19535 (hashtable::begin()): New.
19536 (hashtable::end()): New.
19537 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19538
19539 * statistics.c (statistics_hashes):
19540 Change type to hash_table. Update dependent calls and types.
19541
19542 * tree-into-ssa.c (var_infos):
19543 Change type to hash_table. Update dependent calls and types.
19544
19545 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19546 Change type to hash_table. Update dependent calls and types.
19547
19548 * tree-ssa-loop-im.c (struct mem_ref.refs):
19549 Change type to hash_table. Update dependent calls and types.
19550
19551 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19552 Change type to hash_table. Update dependent calls and types.
19553
19554 * tree-ssa-sccvn.c (vn_tables_s::nary):
19555 Change type to hash_table. Update dependent calls and types.
19556 (vn_tables_s::phis): Likewise.
19557 (vn_tables_s::references): Likewise.
19558
19559 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19560 (vn_reference_eq): Update parameter and return types.
19561
19562 * tree-ssa-structalias.c (pointer_equiv_class_table):
19563 Change type to hash_table. Update dependent calls and types.
19564 (location_equiv_class_table): Likewise.
19565
19566 * tree-vect-data-refs.c: Consequential changes for making
19567 peeling a hash_table.
19568
19569 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19570 (destroy_loop_vec_info): Dependent hash_table update.
19571
19572 * tree-vectorizer.h (peeling_htab):
19573 Change type to hash_table. Update dependent calls and types.
19574
19575 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19576
19577 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19578 to check the register content is equal or not.
19579 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19580 to assign register content record.
19581 * lra-eliminations.c (update_reg_eliminate): Use
19582 lra_update_reg_val_offset to update register content offset.
19583 * lra-int.h (struct lra_reg): Add offset member.
19584 (lra_reg_val_equal_p): New static inline function.
19585 (lra_update_reg_val_offset): New static inline function.
19586 (lra_assign_reg_val): New static inline function.
19587 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19588 to assign register content record.
19589 (initialize_lra_reg_info_element): Initial offset to zero.
19590
19591 2013-04-23 Catherine Moore <clm@codesourcery.com>
19592
19593 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19594 operands. Record compression.
19595
19596 2013-04-23 Xinliang David Li <davidxl@google.com>
19597
19598 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19599
19600 2013-04-23 Richard Biener <rguenther@suse.de>
19601
19602 PR middle-end/57036
19603 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19604 parameter, only add abnormal goto edges from the copied body
19605 if the call could perform abnormal gotos.
19606 (copy_cfg_body): Adjust.
19607
19608 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19609
19610 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19611
19612 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19613
19614 * coretypes.h (gimple_stmt_iterator): Add struct to make
19615 compatible with C.
19616
19617 2013-04-23 Richard Biener <rguenther@suse.de>
19618
19619 PR tree-optimization/57026
19620 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19621 from SSA names occuring in abnormal PHI nodes.
19622
19623 2013-04-22 Andi Kleen <ak@linux.intel.com>
19624
19625 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19626
19627 2013-04-22 Andi Kleen <ak@linux.intel.com>
19628
19629 * lto/lto.c (print_lto_report_1): Declare early.
19630 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19631
19632 2013-04-22 Andi Kleen <ak@linux.intel.com>
19633
19634 * common.opt (-flto-report-wpa): Add.
19635 * doc/invoke.texi (-flto-report-wpa): Add.
19636 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19637 (lto_main): dito.
19638
19639 2013-04-22 Xinliang David Li <davidxl@google.com>
19640
19641 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19642 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19643 * Makefile.in: New dependency
19644
19645 David Daney <ddaney.cavm@gmail.com>
19646
19647 * configure.ac (gcc_cv_as_micromips_support): Use the
19648 --fatal-warnings option.
19649 * configure: Regenerate.
19650
19651 2013-04-22 Marek Polacek <polacek@redhat.com>
19652
19653 PR sanitizer/56990
19654 * tsan.c (instrument_expr): Don't instrument expression
19655 in case its size is zero.
19656
19657 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19658
19659 PR target/57032
19660 Revert:
19661 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19662
19663 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19664
19665 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19666
19667 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19668 (gimple_stmt_iterator): New typedef.
19669 * gimple.h (gimple_stmt_iterator): Rename to...
19670 (gimple_stmt_iterator_d): ... This.
19671 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19672 trees be valid for GIMPLE and GENERIC.
19673 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19674 * gimple-fold.c (gimple_fold_call): Call target hook
19675 gimple_fold_builtin.
19676 * hooks.c (hook_bool_gsiptr_false): New.
19677 * hooks.h (hook_bool_gsiptr_false): New.
19678 * target.def (fold_stmt): New.
19679 * doc/tm.texi: Regenerate.
19680
19681 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19682
19683 PR target/57018
19684 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19685 a set sp if no stack realignment.
19686
19687 2013-04-22 Nick Clifton <nickc@redhat.com>
19688
19689 * config.gcc (tilegx-linux): Extend extra_objs rather than
19690 overwriting it.
19691 (tilepro-linux): Likewise.
19692
19693 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19694
19695 * config/aarch64/aarch64-builtins.c
19696 (CF): Remove.
19697 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19698 (VAR<1-12>): Add MAP parameter.
19699 (BUILTIN_*): Likewise.
19700 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19701 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19702 (aarch64_ushl_n<mode>): Likewise.
19703 (aarch64_sshr_n<mode>): Likewise.
19704 (aarch64_ushr_n<mode>): Likewise.
19705 (aarch64_<maxmin><mode>): Likewise.
19706 (aarch64_sqrt<mode>): Likewise.
19707 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19708 (vshr<q>_n_*): Likewise.
19709
19710 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19711
19712 * config/aarch64/aarch64-builtins.c
19713 (aarch64_simd_builtin_type_mode): Handle SF types.
19714 (sf_UP): Define.
19715 (BUILTIN_GPF): Define.
19716 (aarch64_init_simd_builtins): Handle SF types.
19717 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19718 (frecps): Likewise.
19719 (frecpx): Likewise.
19720 * config/aarch64/aarch64-simd.md
19721 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19722 (aarch64_frecpe<mode>): New.
19723 (aarch64_frecps<mode>): Likewise.
19724 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19725 (v8type): Add frecp<esx>.
19726 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19727 (aarch64_frecps<mode>): Likewise.
19728 * config/aarch64/iterators.md (FRECP): New.
19729 (frecp_suffix): Likewise.
19730 * config/aarch64/arm_neon.h
19731 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19732
19733 2013-04-22 Christian Bruel <christian.bruel@st.com>
19734
19735 PR target/56995
19736 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19737 (REG_CLASS_NAMES): Idem.
19738 (REG_CLASS_CONTENTS): Idem.
19739 (REGCLASS_HAS_FP_REG): Idem.
19740 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19741 (sh_conditional_register_usage): Idem.
19742
19743 2013-04-21 Jeff Law <law@redhat.com>
19744
19745 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19746 (ssa_forward_propagate_and_combine): Use it.
19747
19748 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19749
19750 * lra.c: Update the flow chart diagram.
19751
19752 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19753
19754 PR rtl-optimization/56847
19755 * lra-constraints.c (process_alt_operands): Discourage alternative
19756 with non-matche doffsettable memory constraint fro memory with
19757 known offset.
19758
19759 2013-04-19 Richard Biener <rguenther@suse.de>
19760
19761 PR tree-optimization/56982
19762 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19763 function.
19764 * gimplify.c (gimplify_call_expr): Notice special calls.
19765 (gimplify_modify_expr): Likewise.
19766 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19767 abnormal control flow receivers.
19768 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19769 in the same way as cfun->has_nonlocal_labels.
19770 (gimple_purge_dead_abnormal_call_edges): Likewise.
19771 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19772 receivers start a basic-block.
19773
19774 2013-04-19 Richard Biener <rguenther@suse.de>
19775
19776 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19777 member ...
19778 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19779 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19780 (SLP_TREE_LOAD_PERMUTATION): Add.
19781 (vect_transform_slp_perm_load): Adjust prototype.
19782 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19783 (vect_free_slp_instance): Likewise.
19784 (vect_create_new_slp_node): Likewise.
19785 (vect_supported_slp_permutation_p): Remove.
19786 (vect_slp_rearrange_stmts): Adjust.
19787 (vect_supported_load_permutation_p): Likewise. Inline
19788 vect_supported_slp_permutation_p here.
19789 (vect_analyze_slp_instance): Compute load permutations per
19790 slp node instead of per instance.
19791 (vect_get_slp_defs): Adjust.
19792 (vect_transform_slp_perm_load): Likewise.
19793 (vect_schedule_slp_instance): Remove redundant code.
19794 (vect_schedule_slp): Remove hack for PR56270, add it ...
19795 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19796 CSE loads for SLP. Adjust.
19797
19798 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19799
19800 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19801 spelling in two comments.
19802
19803 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19804
19805 PR target/56797
19806 * config/arm/arm.c (load_multiple_sequence): Require SP
19807 as base register for loads if SP is in the register list.
19808
19809 2013-04-19 Martin Jambor <mjambor@suse.cz>
19810
19811 PR tree-optimization/56718
19812 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19813 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19814 and made public. Adjusted all callers.
19815 (ipa_intraprocedural_devirtualization): New function.
19816 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19817 (ipa_intraprocedural_devirtualization): Likewise.
19818 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19819
19820 2013-04-19 Richard Biener <rguenther@suse.de>
19821
19822 PR tree-optimization/57000
19823 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19824
19825 2013-04-19 Terry Guo <terry.guo@arm.com>
19826
19827 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19828 Replace with ...
19829 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19830 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19831 (cortex_m4_fmacs): Use new reservations.
19832 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19833
19834 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19835
19836 PR rtl-optimization/56999
19837 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19838 related code.
19839 (lra_coalesce): Remove split_origin_bitmap and related code.
19840 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19841 ranges if necessary.
19842
19843 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19844
19845 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19846 New array.
19847 (ix86_expand_call): Remove clobbered_registers array and use
19848 x86_64_ms_sysv_extra_clobbered_registers instead.
19849 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19850 Declare here.
19851 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19852 predicate.
19853 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19854 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19855 (*call_value_rex64_ms_sysv): Ditto.
19856
19857 2013-04-18 Cary Coutant <ccoutant@google.com>
19858
19859 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19860 parent instead of die_mark.
19861
19862 2013-04-18 Diego Novillo <dnovillo@google.com>
19863
19864 * gimple.c (create_gimple_tmp): New.
19865 (get_expr_type): New.
19866 (build_assign): New.
19867 (build_type_cast): New.
19868 * gimple.h (enum ssa_mode): Define.
19869 (gimple_seq_set_location): New.
19870 * asan.c (build_check_stmt): Change some gimple_build_* calls
19871 to use build_assign and build_type_cast.
19872
19873 2013-04-18 Richard Biener <rguenther@suse.de>
19874
19875 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19876 handle negative step. Remove redundant checks.
19877 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19878 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19879 for negative step and grouped loads fail to vectorize.
19880
19881 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19882
19883 * emit-rtl.c (reset_insn_used_flags): New function.
19884 (reset_all_used_flags): Use it.
19885 (verify_insn_sharing): New function.
19886 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19887
19888 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19889
19890 PR tree-optimization/56984
19891 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19892 and (x >> M) >= N don't register any assertion if N << M is the
19893 minimum value.
19894
19895 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19896
19897 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19898 do not delete_insn insns that have not yet been emitted, only
19899 unlink them with remove_insn.
19900 * df-scan.c (df_insn_delete): Revert r197492.
19901
19902 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19903
19904 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19905 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19906
19907 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19908
19909 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19910 define_insn_and_split.
19911 (and_scc,ior_scc,negscc): Likewise.
19912 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19913
19914 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19915
19916 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19917 can benefit from using a sequence of LDRD instructions in epilogue
19918 instead of a single LDM instruction.
19919
19920 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19921
19922 PR 45688
19923 * doc/extend.texi: Fix typo.
19924
19925 2013-04-17 Richard Biener <rguenther@suse.de>
19926
19927 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19928 (vect_build_slp_tree): ... here.
19929 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19930 match. Remove special-casing of mismatched complex loads.
19931 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19932 re-try the match with swapped commutative operands.
19933 (vect_supported_load_permutation_p): Remove special-casing of
19934 mismatched complex loads.
19935 (vect_analyze_slp_instance): Adjust.
19936
19937 2013-04-17 Richard Biener <rguenther@suse.de>
19938
19939 PR rtl-optimization/56921
19940 * cfgloop.h (struct loop): Add simple_loop_desc member.
19941 (struct niter_desc): Mark with GTY(()).
19942 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19943 * loop-iv.c (get_simple_loop_desc): Likewise.
19944 (free_simple_loop_desc): Likewise.
19945
19946 Revert
19947 2013-04-16 Richard Biener <rguenther@suse.de>
19948
19949 PR rtl-optimization/56921
19950 * loop-init.c (pass_rtl_move_loop_invariants): Add
19951 TODO_do_not_ggc_collect to todo_flags_finish.
19952 (pass_rtl_unswitch): Same.
19953 (pass_rtl_unroll_and_peel_loops): Same.
19954 (pass_rtl_doloop): Same.
19955
19956 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19957
19958 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19959 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19960 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19961 references.
19962 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19963 * tree-streamer.c (record_common_node): Adjust reference in comment.
19964
19965 2013-04-17 Terry Guo <terry.guo@arm.com>
19966
19967 * config/arm/cortex-m4.md: Add a new bypass.
19968
19969 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19970
19971 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19972 New pattern.
19973 (*subs_<optab><mode>_multp2): New pattern.
19974 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19975 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19976
19977 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19978
19979 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19980 (*subs_mul_imm_<mode>): New pattern.
19981
19982 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19983
19984 PR target/56948
19985 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19986 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19987 (vsx_movti_32bit): Change j->wa to O->wa.
19988
19989 2013-04-16 Richard Biener <rguenther@suse.de>
19990
19991 PR rtl-optimization/56921
19992 * loop-init.c (pass_rtl_move_loop_invariants): Add
19993 TODO_do_not_ggc_collect to todo_flags_finish.
19994 (pass_rtl_unswitch): Same.
19995 (pass_rtl_unroll_and_peel_loops): Same.
19996 (pass_rtl_doloop): Same.
19997
19998 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19999
20000 * config/arm/arm.c (emit_multi_reg_push): New declaration
20001 for an existing function.
20002 (arm_emit_strd_push): New function.
20003 (arm_expand_prologue): Used here.
20004 (arm_emit_ldrd_pop): New function.
20005 (arm_expand_epilogue): Used here.
20006 (arm_get_frame_offsets): Update condition.
20007 (arm_emit_multi_reg_pop): Add a special case for load of a single
20008 register with writeback.
20009
20010 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20011
20012 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20013 description.
20014
20015 2013-04-16 Richard Biener <rguenther@suse.de>
20016
20017 PR tree-optimization/56756
20018 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20019 (first_mem_ref_loc): New.
20020 (execute_sm): Place the load temporarily before a previous
20021 access instead of in the latch edge to ensure its SSA dependencies
20022 are defined at points dominating the load.
20023
20024 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20025
20026 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20027 correct fix by moving header and footer insn to the footer of
20028 the merged basic block. Clear BB_END of the merged-away block.
20029
20030 PR middle-end/43631
20031 * emit-rtl.c (make_note_raw): New function.
20032 (link_insn_into_chain): New static inline function.
20033 (add_insn): Use it.
20034 (add_insn_before, add_insn_after): Factor insn chain linking code...
20035 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20036 using link_insn_into_chain.
20037 (note_outside_basic_block_p): New helper function for emit_note_after
20038 and emit_note_before.
20039 (emit_note_after): Use nobb variant of add_insn_after if the note
20040 should not be contained in a basic block.
20041 (emit_note_before): Use nobb variant of add_insn_before if the note
20042 should not be contained in a basic block.
20043 (emit_note_copy): Use make_note_raw.
20044 (emit_note): Likewise.
20045 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20046 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20047 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20048 the moved barrier the tail of the basic block it follows.
20049 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20050
20051 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20052
20053 PR tree-optimization/56962
20054 * gimple-ssa-strength-reduction.c (record_increment): Only set
20055 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20056 either rhs1 or rhs2 is equal to c->base_expr.
20057
20058 2013-04-15 Richard Biener <rguenther@suse.de>
20059
20060 PR tree-optimization/56933
20061 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20062 member.
20063 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20064 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20065 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20066 dependence check ...
20067 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20068 ... here.
20069 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20070 GROUP_READ_WRITE_DEPENDENCE.
20071
20072 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20073
20074 * emit-rtl.c (reset_all_used_flags): New function.
20075 (verify_rtl_sharing): Call reset_all_used_flags before and after
20076 performing the checks.
20077
20078 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20079
20080 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20081 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20082 * config/arm/constraints.md (De): New constraint.
20083 * config/arm/neon.md (anddi3_neon): Delete.
20084 (neon_vand<mode>): Expand to standard anddi3 pattern.
20085 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20086 Move earlier in the file.
20087 (neon_inv_logic_op2): Likewise.
20088 (arm_anddi_operand_neon): New predicate.
20089
20090 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20091
20092 * configure.ac (gcc_cv_ld_as_needed): Set
20093 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20094 Use -z ignore, -z record on *-*-solaris2*.
20095 (HAVE_LD_AS_NEEDED): Update comment.
20096 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20097 * configure: Regenerate.
20098 * config.in: Regenerate.
20099 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20100 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20101 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20102 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20103 equivalents. Fix markup.
20104 * doc/tm.texi: Regenerate.
20105
20106 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20107
20108 * config/i386/i386.opt: New option mstack-protector-guard=.
20109 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20110 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20111 TARGET_SSP_TLS_GUARD.
20112 * config/i386/i386.c (ix86_option_override_internal): Set
20113 ix86_stack_protector_guard.
20114 * config/i386/i386.md (stack_protect_set): Enable for
20115 TARGET_SSP_TLS_GUARD only.
20116 (stack_protect_set_<mode>): Ditto.
20117 (stack_protect_test): Ditto.
20118 (stack_protect_test_<mode>): Ditto.
20119 * doc/invoke.texi (i386 Option): Document.
20120
20121 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20122
20123 PR target/56890
20124 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20125 (S_MODES): Set H_MODE bit.
20126 (SF_MODES): Set only S_MODE and SF_MODE bits.
20127 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20128 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20129 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20130 <MODE_FLOAT>: Likewise.
20131
20132 2013-04-15 Joey Ye <joey.ye@arm.com>
20133
20134 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20135
20136 2013-04-15 Joey Ye <joey.ye@arm.com>
20137
20138 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20139 for real far jump.
20140 (thumb_far_jump_used_p): Count instruction size and set
20141 far_jump_used.
20142
20143 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20144
20145 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20146 * resource.c (mark_target_live_regs): Reformat conditional block.
20147
20148 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20149
20150 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20151 notes, they are emitted only just before final.
20152 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20153
20154 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20155
20156 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20157 * cfgrtl.c (delete_insn): Call it here instead.
20158 * lra-spills.c (lra_final_code_change): Use delete_insn.
20159 * haifa-sched.c (sched_remove_insn): Likewise.
20160 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20161 returning to the nop pool.
20162 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20163 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20164
20165 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20166
20167 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20168 * doc/tm.texi: Regenerated.
20169
20170 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20171
20172 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20173 QImode checks.
20174
20175 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20176
20177 * df-core.c (df_find_def): Compare register numbers.
20178 (df_find_use): Likewise.
20179
20180 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20181
20182 PR target/56903
20183 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20184 lra_in_progress for return.
20185
20186 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20187
20188 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20189 define_insn into define_insn_and_split and emit movsicc patterns.
20190
20191 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20192
20193 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20194
20195 2013-04-12 Richard Biener <rguenther@suse.de>
20196
20197 * tree-pass.h (TODO_do_not_ggc_collect): New.
20198 * passes.c (execute_one_ipa_transform_pass): Honor
20199 TODO_do_not_ggc_collect.
20200 (execute_one_pass): Likewise.
20201
20202 Revert
20203 2013-04-10 Richard Biener <rguenther@suse.de>
20204
20205 * passes.c (init_optimization_passes): Remove reload pass.
20206 * ira.c (do_reload): Merge into ...
20207 (ira): ... this.
20208 (rest_of_handle_reload): Remove.
20209 (pass_reload): Likewise.
20210 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20211 of reload for vzeroupper pass placement.
20212
20213 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20214
20215 PR tree-optimization/56918
20216 PR tree-optimization/56920
20217 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20218 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20219 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20220 use wide_mul_with_sign method.
20221
20222 2013-04-12 Richard Biener <rguenther@suse.de>
20223
20224 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20225 not be considered a gimple constant.
20226
20227 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20228
20229 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20230 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20231 operations.
20232
20233 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20234 Jakub Jelinek <jakub@redhat.com>
20235
20236 * opts.c: Include diagnostic-color.h.
20237 (common_handle_option): Handle OPT_fdiagnostics_color_.
20238 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20239 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20240 (diagnostic-color.o): New.
20241 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20242 (diagnostic_color_rule): New enum.
20243 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20244 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20245 the location string.
20246 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20247 either NULL, or color kind.
20248 * diagnostic-color.c: New file.
20249 * diagnostic-color.h: New file.
20250 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20251 arguments.
20252 * doc/invoke.texi (-fdiagnostics-color): Document.
20253 * pretty-print.h (pp_show_color): Define.
20254 (struct pretty_print_info): Add show_color field.
20255 * diagnostic.c: Include diagnostic-color.h.
20256 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20257 macros. Colorize error:, warning: etc. strings and also the location
20258 string.
20259 (diagnostic_show_locus): Colorize the caret line.
20260 * pretty-print.c: Include diagnostic-color.h.
20261 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20262 inside of %< %> quotes or quoted through q format modifier.
20263
20264 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20265
20266 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20267
20268 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20269
20270 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20271 code in CC_NZ mode.
20272 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20273 pattern.
20274
20275 2013-04-11 Marek Polacek <polacek@redhat.com>
20276
20277 PR tree-optimization/48184
20278 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20279
20280 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20281
20282 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20283 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20284 (skip_simple_arithmetic): Tidy up.
20285 * tree.h (skip_simple_constant_arithmetic): Declare.
20286
20287 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20288
20289 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20290
20291 2013-04-11 Richard Biener <rguenther@suse.de>
20292
20293 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20294 generate vector constants.
20295
20296 2013-04-11 Richard Biener <rguenther@suse.de>
20297
20298 PR tree-optimization/56878
20299 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20300 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20301 New function.
20302 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20303 Prefer to align the DR with the most invariant base address.
20304
20305 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20306
20307 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20308 comment.
20309
20310 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20311
20312 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20313 floating-point vector comparisons against 0.
20314
20315 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20316
20317 PR tree-optimization/56899
20318 * fold-const.c (extract_muldiv_1): Apply distributive law
20319 only if TYPE_OVERFLOW_WRAPS (ctype).
20320
20321 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20322
20323 PR target/56124
20324 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20325 loading has side effect.
20326
20327 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20328
20329 * config/sparc/sparc.c: Include tree-pass.h.
20330 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20331 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20332 head of file. Change return type. Split off gate function.
20333 (sparc_gate_work_around_errata): New function.
20334 (pass_work_around_errata): New pass definition.
20335 (insert_pass_work_around_errata) New pass insert definition to
20336 insert pass_work_around_errata just after delayed-branch scheduling.
20337 (sparc_option_override): Insert the pass.
20338 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20339
20340 2013-04-10 David S. Miller <davem@davemloft.net>
20341
20342 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20343 or -mcpu=hypersparc.
20344
20345 * target.def (cstore_mode): New hook.
20346 * target.h: Include insn-codes.h
20347 * targhooks.c: Likewise.
20348 (default_cstore_mode): New function.
20349 * targhooks.h: Declare it.
20350 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20351 * doc/tm.texi: Rebuild.
20352 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20353 target hook, rather than inspecting the insn_data.
20354 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20355 (TARGET_CSTORE_MODE): Redefine.
20356 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20357 result patterns.
20358 * config/sparc/predicates.md (cstore_result_operand): New special
20359 predicate.
20360 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20361 Use it for operand 0.
20362 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20363 (*snesi_special): Likewise.
20364 (*snesi_zero): Likewise.
20365 (*seqsi_zero): Likewise.
20366 (*sltu_insn): Likewise.
20367 (*sgeu_insn): Likewise.
20368 (*seqdi_special): Make operand 0 and comparison operation be of
20369 DImode.
20370 (*snedi_special): Likewise.
20371 (*snedi_special_vis3): Likewise.
20372 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20373 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20374 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20375 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20376 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20377 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20378 (*sltu_extend_sp64): Likewise.
20379 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20380 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20381 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20382 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20383 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20384
20385 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20386
20387 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20388 (aarch64_start_file): Use the new function.
20389
20390 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20391 Jason Merrill <jason@redhat.com>
20392
20393 * common.opt: Add -gdwarf.
20394 * opts.c (common_handle_option): Handle it.
20395 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20396
20397 2013-04-10 Richard Biener <rguenther@suse.de>
20398
20399 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20400 (execute_one_ipa_transform_pass): But unconditionally here.
20401 (execute_one_pass): And here.
20402 (init_optimization_passes): Remove reload pass.
20403 * tree-pass.h (TODO_ggc_collect): Remove.
20404 (pass_reload): Likewise.
20405 * ira.c (do_reload): Merge into ...
20406 (ira): ... this.
20407 (rest_of_handle_reload): Remove.
20408 (pass_reload): Likewise.
20409 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20410 of reload for vzeroupper pass placement.
20411 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20412 and todo_flags_finish of all passes.
20413
20414 2013-04-10 Richard Biener <rguenther@suse.de>
20415
20416 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20417 first_const_oprnd field, rename first_def_type to first_op_type.
20418 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20419 (vect_get_and_check_slp_defs): Always use the type of the
20420 operand. Allow mixed vect_external_def, vect_constant_def types.
20421 (vect_get_constant_vectors): Handle mixed vect_external_def,
20422 vect_constant_def types.
20423
20424 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20425
20426 PR tree-optimization/55524
20427 * tree-ssa-math-opts.c
20428 (convert_mult_to_fma): Don't use an fms construct
20429 when we don't have an fms operation, but fnma, and it looks
20430 likely that we'll be able to use the latter.
20431
20432 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20433
20434 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20435 function.
20436 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20437 inline fail caused by overwritable functions.
20438
20439 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20440
20441 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20442 unnecessary bits in the constant power of two case.
20443
20444 2013-04-10 Richard Biener <rguenther@suse.de>
20445
20446 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20447 broken code swapping operands.
20448 (vect_build_slp_tree): Do not compute load permutations here.
20449 (vect_analyze_slp_instance): Compute load permutations here,
20450 after building the SLP tree.
20451
20452 2013-04-09 Christian Bruel <christian.bruel@st.com>
20453
20454 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20455 of next/prev_real_insn.
20456
20457 2013-04-09 Jan Hubicka <jh@suse.cz>
20458
20459 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20460 Drop aliased parameter.
20461 (function_and_variable_visibility): Do not handle alias pairs.
20462 * cgraph.c (varpool_externally_visible_p): Update prototype.
20463 * varpool.c (varpool_add_new_variable): Update.
20464
20465 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20466
20467 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20468
20469 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20470
20471 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20472
20473 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20474
20475 2013-04-09 Marek Polacek <polacek@redhat.com>
20476
20477 PR tree-optimization/48762
20478 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20479
20480 2013-04-09 Richard Biener <rguenther@suse.de>
20481
20482 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20483 dealing with cost.
20484 (vect_build_slp_tree): Likewise.
20485 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20486 calculating the cost of a SLP instance.
20487 (vect_analyze_slp_instance): Use it from here, after building
20488 the SLP tree.
20489
20490 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20491
20492 PR middle-end/56883
20493 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20494 expand_omp_for_static_chunk): Use simple_p = true in
20495 force_gimple_operand_gsi calls when assigning to addressable decls.
20496
20497 2013-04-09 Jeff Law <law@redhat.com>
20498
20499 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20500 when the boolean was created by converting a wider object which
20501 had a boolean range.
20502
20503 2013-04-09 Richard Biener <rguenther@suse.de>
20504
20505 * tree-vectorizer.h (slp_void_p): Remove.
20506 (slp_tree): Typedef before _slp_tree declaration.
20507 (struct _slp_tree): Use a vector of slp_tree as children.
20508 (vect_get_place_in_interleaving_chain): Remove.
20509 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20510 Move ...
20511 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20512 and make static.
20513 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20514 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20515 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20516 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20517 Use slp_node instead of slp_void_p and adjust.
20518
20519 2013-04-09 Richard Biener <rguenther@suse.de>
20520
20521 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20522 work that is not necessary.
20523
20524 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20525
20526 PR tree-optimization/56854
20527 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20528 forward into clobber stmts if it would change MEM_REF lhs into
20529 non-MEM_REF.
20530
20531 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20532
20533 * tree.c (type_hash_lookup, type_hash_add): Make static.
20534 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20535
20536 2013-04-09 Richard Biener <rguenther@suse.de>
20537
20538 * tree.h (unsave_expr_now): Remove.
20539 * tree-inline.c (mark_local_for_remap_r): Remove.
20540 (unsave_expr_1): Likewise.
20541 (unsave_r): Likewise.
20542 (unsave_expr_now): Likewise.
20543 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20544 (propagate_tree_value): Likewise.
20545
20546 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20547
20548 * doc/rtl.texi (sequence): Rewrite documentation to match the
20549 current use of SEQUENCE rtl objects.
20550 * rtl.def (SEQUENCE): Likewise.
20551
20552 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20553 Update documentation.
20554 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20555 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20556
20557 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20558
20559 2013-04-08 Teresa Johnson <tejohnson@google.com>
20560
20561 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20562 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20563 methods.
20564 (estimate_edge_size_and_time): Add comment to suggest using rounding
20565 methods.
20566 (estimate_node_size_and_time): Ditto.
20567 (remap_edge_change_prob): Use helper rounding divide methods.
20568 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20569 (gimple_mod_pow2_value_transform): Ditto.
20570 (gimple_mod_subtract_transform): Ditto.
20571 (gimple_ic_transform): Ditto.
20572 (gimple_stringops_transform): Ditto.
20573 * stmt.c (conditional_probability): Ditto.
20574 (emit_case_dispatch_table): Ditto.
20575 * lto-cgraph.c (merge_profile_summaries): Ditto.
20576 * tree-optimize.c (execute_fixup_cfg): Ditto.
20577 * cfgcleanup.c (try_forward_edges): Ditto.
20578 * cfgloopmanip.c (scale_loop_profile): Ditto.
20579 (loopify): Ditto.
20580 (duplicate_loop_to_header_edge): Ditto.
20581 (lv_adjust_loop_entry_edge): Ditto.
20582 * tree-vect-loop.c (vect_transform_loop): Ditto.
20583 * profile.c (compute_branch_probabilities): Ditto.
20584 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20585 * lto-streamer-in.c (input_cfg): Ditto.
20586 * gimple-streamer-in.c (input_bb): Ditto.
20587 * ipa-cp.c (update_profiling_info): Ditto.
20588 (update_specialized_profile): Ditto.
20589 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20590 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20591 rounding methods.
20592 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20593 (compute_trg_info): Ditto.
20594 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20595 (purge_dead_edges): Ditto.
20596 * loop-unswitch.c (unswitch_loop): Ditto.
20597 * cgraphclones.c (cgraph_clone_edge): Ditto.
20598 (cgraph_clone_node): Ditto.
20599 * tree-inline.c (copy_bb): Ditto.
20600 (copy_edges_for_bb): Ditto.
20601 (initialize_cfun): Ditto.
20602 (copy_cfg_body): Ditto.
20603 (expand_call_inline): Ditto.
20604
20605 2013-04-08 Kai Tietz <ktietz@redhat.com>
20606
20607 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20608 TARGET_CYGWIN64 by TARGET_64BIT.
20609
20610 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20611
20612 * config/epiphany/epiphany.md (GPR_1): New constant.
20613 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20614 * config/epiphany/epiphany.c (gen_compare_reg):
20615 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20616 is already in place.
20617 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20618 Don't require being called during rtl expansion; If y operlaps r0,
20619 return 0.
20620 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20621 (epiphany_expand_epilogue): Likewise.
20622
20623 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20624 Don't use CC_FPmode for ORDERED / UNORDERED.
20625 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20626
20627 * config/epiphany/constraints.md (CnL): New constraint.
20628 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20629 * config/epiphany/predicates.md (add_operand): Allow 1024.
20630
20631 * config/epiphany/epiphany.md (logical_op): New code iterator.
20632 (op_mnc): New code attribute.
20633 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20634 (mov_f+1, mov_f+2): New peephole2 patterns.
20635
20636 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20637 (cstoresi4): Also allow re-use of zero result when doing a NE
20638 comparison to a non-zero operand.
20639 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20640
20641 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20642 Use gen_addsi3_i / gen_subsi3_i.
20643
20644 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20645
20646 PR c++/34949
20647 PR c++/50243
20648 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20649 contain anything but clobbers, at most one __builtin_stack_restore,
20650 optionally debug stmts and final resx, and if it has at least one
20651 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20652 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20653 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20654 which isn't defaut definition, remove them.
20655 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20656 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20657 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20658 with MEM_REF LHS with SSA_NAME address.
20659
20660 2013-04-08 Jeff Law <law@redhat.com>
20661
20662 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20663
20664 2013-04-08 Richard Biener <rguenther@suse.de>
20665
20666 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20667 extra newline.
20668 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20669 determined vector type.
20670 (vect_analyze_data_refs): Likewise.
20671 (vect_get_new_vect_var): Adjust.
20672 (vect_create_destination_var): Preserve SSA name versions.
20673 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20674 not dump anything here.
20675
20676 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20677
20678 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20679 Add member lr_slot_known.
20680 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20681 if necessary.
20682 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20683 Remove code that sets lr_slot_offset according to what a previous
20684 version of epiphany_emit_save_restore used to do.
20685 (epiphany_emit_save_restore): When doing an lr save or restore,
20686 set/verify lr_slot_known and lr_slot_offset.
20687
20688 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20689
20690 PR target/54338
20691 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20692 in ALL_REGS.
20693
20694 2013-04-08 Richard Biener <rguenther@suse.de>
20695
20696 * alias.c (find_base_term): Fix thinko in previous change.
20697
20698 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20699
20700 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20701 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20702 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20703 if possible to compute val.
20704 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20705 For QImode integers don't require anything about precision. Use
20706 const_with_all_bytes_same to find out if the constant doesn't have
20707 repeated bytes in it.
20708
20709 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20710
20711 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20712 within mode size.
20713
20714 2013-04-08 Marek Polacek <polacek@redhat.com>
20715
20716 PR rtl-optimization/48182
20717 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20718 value to 1.
20719
20720 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20721
20722 PR target/55487
20723 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20724 nuses, make sure we have a label.
20725
20726 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20727
20728 PR target/56843
20729 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20730 (rs6000_emit_swdiv_low_precision): Remove.
20731 (rs6000_emit_swdiv): Rewrite to handle between one and four
20732 iterations of Newton-Raphson generally; modify required number of
20733 iterations for some cases.
20734 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20735
20736 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20737
20738 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20739 set-but-unused variable.
20740
20741 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20742 basic blocks of released function bodies garbage-collectable.
20743
20744 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20745 (struct rtl_opt_pass): Add TODO_df_finish.
20746
20747 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20748
20749 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20750
20751 * config/arm/constraints.md (q): New constraint.
20752 * config/arm/ldrdstrd.md: New file.
20753 * config/arm/arm.md (ldrdstrd.md) New include.
20754 (arm_movdi): Use "q" instead of "r" constraint
20755 for double-word memory access.
20756 (movdf_soft_insn): Likewise.
20757 * config/arm/vfp.md (movdi_vfp): Likewise.
20758 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20759 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20760 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20761 (mem_ok_for_ldrd_strd): Likewise.
20762 (output_move_double): Update assertion.
20763
20764 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20765
20766 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20767
20768 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20769
20770 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20771 define_insn_and_split.
20772 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20773
20774 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20775
20776 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20777 define_insn_and_split.
20778 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20779 (shiftsi3_compare): New pattern.
20780 (rrx): New pattern.
20781 * config/arm/unspecs.md (UNSPEC_RRX): New.
20782
20783 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20784
20785 * config/arm/arm.md (negdi_extendsidi): New pattern.
20786 (negdi_zero_extendsidi): Likewise.
20787
20788 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20789
20790 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20791 define_insn_and_split.
20792 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20793 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20794
20795 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20796
20797 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20798 define_insn_and_split.
20799 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20800 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20801
20802 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20803
20804 * config/arm/arm.md (subsi3_carryin): New pattern.
20805 (subsi3_carryin_const): Likewise.
20806 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20807 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20808
20809 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20810
20811 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20812
20813 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20814
20815 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20816 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20817
20818 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20819
20820 * config/arm/arm.c (arm_expand_builtin): Change fcode
20821 type to unsigned int.
20822
20823 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20824
20825 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20826
20827 2013-04-04 Ian Lance Taylor <iant@google.com>
20828
20829 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20830 language standard.
20831
20832 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20833
20834 PR middle-end/56729
20835 * df-scan.c (df_insn_delete): Disable failing assert.
20836
20837 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20838
20839 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20840 New function prototype.
20841 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20842 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20843 (arm_builtin_vectorized_function): New function.
20844
20845 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20846
20847 * config/arm/arm_neon_builtins.def: New file.
20848 * config/arm/arm.c (neon_builtin_data): Move contents to
20849 arm_neon_builtins.def.
20850 (enum arm_builtins): Include neon builtin definitions.
20851 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20852 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20853
20854 2013-04-04 Marek Polacek <polacek@redhat.com>
20855
20856 PR tree-optimization/48186
20857 * predict.c (maybe_hot_frequency_p): Return false if
20858 HOT_BB_FREQUENCY_FRACTION is 0.
20859 (cgraph_maybe_hot_edge_p): Likewise.
20860
20861 2013-04-04 Richard Biener <rguenther@suse.de>
20862
20863 PR tree-optimization/56826
20864 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20865 more accurately.
20866
20867 2013-04-04 Richard Biener <rguenther@suse.de>
20868
20869 PR tree-optimization/56213
20870 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20871 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20872
20873 2013-04-04 Richard Biener <rguenther@suse.de>
20874
20875 PR tree-optimization/56837
20876 * tree-loop-distribution.c (classify_partition): For non-zero
20877 values require that the value has the same precision as its
20878 mode to be useful as memset value.
20879
20880 2013-04-03 Nick Clifton <nickc@redhat.com>
20881
20882 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20883 (fmssf4): Use fmsf.s on E3V5 architectures.
20884 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20885 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20886
20887 2013-04-03 Jeff Law <law@redhat.com>
20888
20889 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20890 (lra-eliminations.o): Likewise.
20891
20892 2013-04-03 Teresa Johnson <tejohnson@google.com>
20893
20894 * gcov-io.c (compute_working_sets): Moved most of body of old
20895 compute_working_sets here from profile.c.
20896 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20897 (gcov_working_set_t): Moved typedef here from basic-block.h
20898 (compute_working_set): Declare.
20899 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20900 (get_working_sets): Renamed from compute_working_set,
20901 replace most of body with call to new compute_working_sets.
20902 (get_exec_counts): Replace call to compute_working_sets
20903 to get_working_sets.
20904 * profile.h (get_working_sets): Renamed from compute_working_set.
20905 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20906 to get_working_sets.
20907 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20908 * gcov-dump.c (dump_working_sets): New function.
20909
20910 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20911
20912 * hwint.c (sext_hwi, zext_hwi): New functions.
20913 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20914 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20915 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20916 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20917 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20918 (sext_hwi, zext_hwi): New functions.
20919
20920 2013-04-03 Jeff Law <law@redhat.com>
20921
20922 PR tree-optimization/56799
20923 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20924 back test for widening conversion erroneously dropped in prior change.
20925
20926 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20927
20928 PR target/56809
20929 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20930 instead of next_real_insn.
20931
20932 2013-04-03 Marek Polacek <polacek@redhat.com>
20933
20934 PR sanitizer/55702
20935 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20936
20937 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20938
20939 PR target/56809
20940 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20941 next_real_insn.
20942 (thumb1_output_casesi): Likewise.
20943 (thumb2_output_casesi): Likewise.
20944
20945 2013-04-03 Richard Biener <rguenther@suse.de>
20946
20947 PR tree-optimization/56817
20948 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20949 Split out ...
20950 (tree_unroll_loops_completely_1): ... new function to manually
20951 walk the loop tree, properly defering outer loops of unrolled
20952 loops to later iterations.
20953
20954 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20955
20956 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20957 (vectorizable_load): Likewise.
20958 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20959 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20960
20961 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20962
20963 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20964 BIT_FIELD_REF.
20965
20966 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20967
20968 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20969
20970 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20971
20972 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20973
20974 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20975
20976 PR tree-optimization/56790
20977 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20978 folding.
20979
20980 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20981
20982 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20983 Handle VEC_MERGE.
20984 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20985 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20986 equal arguments.
20987
20988 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20989
20990 PR c/19449
20991 * tree.h (force_folding_builtin_constant_p): New decl.
20992 * builtins.c (force_folding_builtin_constant_p): New variable.
20993 (fold_builtin_constant_p): Fold immediately also if
20994 force_folding_builtin_constant_p.
20995
20996 2013-04-03 Richard Biener <rguenther@suse.de>
20997
20998 PR tree-optimization/56812
20999 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21000 DRs of the same interleaving chain are independent.
21001
21002 2013-04-02 Jason Merrill <jason@redhat.com>
21003
21004 * gdbinit.in (pbb): Use debug fn.
21005
21006 2013-04-02 Lawrence Crowl <crowl@google.com>
21007
21008 * sese.h (struct ivtype_map_elt_s): Remove unused.
21009 (extern debug_ivtype_map): Remove unused.
21010 (extern eq_ivtype_map_elts): Remove unused.
21011 * sese.c (debug_ivtype_map): Removed unused.
21012 (debug_ivtype_map_1): Removed unused.
21013 (debug_ivtype_elt): Remove unused.
21014 (eq_ivtype_map_elts): Remove unused.
21015
21016
21017 2013-04-02 Kai Tietz <ktietz@redhat.com>
21018
21019 PR target/52790
21020 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21021 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21022 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21023 function.
21024 (legitimize_pe_coff_symbol): Likewise.
21025 (is_imported_p): New helper-function.
21026 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21027 for Windows x64 targets.
21028 (ix86_expand_prologue): Optimize for pe-coff targets.
21029 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21030 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21031 medium/large code-model.
21032 (legitimize_pic_address): Likewise.
21033 (legitimize_tls_address): Likewise.
21034 (ix86_expand_call): Likewise.
21035 (x86_output_mi_thunk): Likewise.
21036 (get_dllimport_decl): Add new beimport argument.
21037 (construct_plt_address): Don't assert for x64 pe-coff targets.
21038 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21039 targets.
21040 (SYMBOL_FLAG_STUBVAR): New macro.
21041 (SYMBOL_REF_STUBVAR_P): Likewise.
21042 * config/i386/winnt.c (stub_list): New structure.
21043 (stub_head): New local variable.
21044 (i386_pe_record_stub): New function.
21045 (i386_pe_file_end): Emit refptr-stubs.
21046
21047 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21048
21049 PR rtl-optimization/56745
21050 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21051 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21052
21053 PR c++/34949
21054 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21055 and both of them are MEM_REFs, just compare first argument for
21056 equality and attempt to deal even with differing offsets.
21057
21058 PR c++/34949
21059 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21060 of gimple_clobber_p to be MEM_REF.
21061 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21062 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21063 after gimplification.
21064 * asan.c (get_mem_ref_of_assignment): Don't instrument
21065 gimple_clobber_p stmts.
21066 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21067 gimple_clobber_p stmt if they have MEM_REF lhs and
21068 are dead because of another gimple_clobber_p stmt.
21069 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21070 gimple_clobber_p stmts like debug stmts.
21071 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21072 that refer to unused VAR_DECLs or uninitialized values.
21073 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21074 gimple_clobber_p stmts if they refer to removed parameters.
21075 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21076 formatting.
21077
21078 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21079
21080 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21081 using SWI48 mode attribute.
21082
21083 2013-04-02 Wei Mi <wmi@google.com>
21084
21085 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21086 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21087 *<rotate_insn><mode>3_mask in i386.md.
21088
21089 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21090
21091 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21092
21093 2013-04-02 Richard Biener <rguenther@suse.de>
21094
21095 PR tree-optimization/56778
21096 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21097 Runtime alias tests are not supported for gather loads.
21098 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21099 stmts referenced from SSA operands before updating SSA form.
21100
21101 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21102 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21103
21104 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21105 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21106 * config/arm/cortex-a53.md: New file.
21107 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21108 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21109 * config/arm/arm.c (arm_issue_rate): Likewise.
21110 * config/arm/arm-tune.md: Regenerate
21111 * config/arm/arm-tables.opt: Regenerate.
21112 * config/arm/arm-cores.def: Add cortex-a53.
21113
21114 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21115
21116 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21117 non-static link.
21118
21119 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21120
21121 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21122 scalar load/store operations using B/H registers.
21123 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21124
21125 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21126
21127 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21128 scalar move.
21129 * config/aarch64/aarch64.c
21130 (aarch64_simd_scalar_immediate_valid_for_move): New.
21131 * config/aarch64/aarch64-protos.h
21132 (aarch64_simd_scalar_immediate_valid_for_move): New.
21133 * config/aarch64/constraints.md (Dh, Dq): New.
21134 * config/aarch64/iterators.md (hq): New.
21135
21136 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21137
21138 * reorg.c (get_branch_condition): Deal with conditional returns.
21139 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21140
21141 2013-04-01 Wei Mi <wmi@google.com>
21142
21143 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21144 Truncate operand 2 using %b asm operand modifier.
21145 (*<shift_insn><mode>3_mask): Ditto.
21146 (*<rotate_insn><mode>3_mask): Ditto.
21147
21148 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21149
21150 PR middle-end/56798
21151 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21152
21153 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21154
21155 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21156 of next_real_insn.
21157 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21158
21159 2013-03-30 Lawrence Crowl <crowl@google.com>
21160
21161 * dse.c (clear_alias_sets): Remove never set.
21162 (disqualified_clear_alias_sets): Remove never set.
21163 (clear_alias_mode_pool): Remove never set.
21164 (dse_step0): Remove condition that is never true.
21165 (canon_address): Remove condition that is never true.
21166 (dse_step7): Remove condition that is never true.
21167 (rest_of_handle_dse): Remove condition that is never true.
21168 (rest_of_handle_dse::did_global): Remove never read from above.
21169 (dse_step2_spill): Remove never called from above.
21170 (dse_step5_spill): Remove never called from above.
21171
21172 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21173
21174 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21175 JUMP_TABLE_DATA changes.
21176 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21177 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21178 (Insns) <jump_table_data>: New entry.
21179 * doc/tm.texi: Regenerate.
21180
21181 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21182
21183 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21184 for table jump at the end of a basic block using tablejump_p.
21185 * targhooks.c (default_invalid_within_doloop): Likewise.
21186 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21187 target hook implementation that is identical to the default hook.
21188 (rs6000_invalid_within_doloop): Remove.
21189
21190 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21191 unused variable from tablejump_p call.
21192
21193 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21194 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21195 (INSN_DELETED_P): Likewise.
21196 (emit_jump_table_data): New prototype.
21197 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21198 after 4th as unused.
21199 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21200 * sched-vis.c (print_insn): Likewise.
21201 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21202 insn for compatibility with back ends that use next_active_insn to
21203 identify jump table data.
21204 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21205 (remove_insn): Likewise.
21206 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21207 to be emitted.
21208 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21209 (emit_jump_table_data): New function.
21210
21211 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21212 basic block, a JUMP_TABLE_DATA never is.
21213 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21214 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21215 off from code handling real insns.
21216 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21217 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21218 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21219 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21220 is not a NONDEBUG_INSN_P.
21221 * ira-costs.c (scan_one_insn): Likewise.
21222 * jump.c (mark_all_labels): Likewise.
21223 (mark_jump_label_1): Likewise.
21224 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21225 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21226 a basic block.
21227 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21228 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21229 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21230 JUMP_TABLE_DATA_P insns.
21231 (calculate_elim_costs_all_insns): Likewise.
21232 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21233 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21234 (delete_output_reload): Code style fixups.
21235 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21236 insn flags on this non-insn.
21237 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21238 as scheduling barriers, for pre-change compatibility.
21239 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21240 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21241
21242 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21243 redundant JUMP_TABLE_DATA_P test.
21244 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21245 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21246 (frv_for_each_packet): Likewise.
21247 * config/i386/i386.c (min_insn_size): Likewise.
21248 (ix86_avoid_jump_mispredicts): Likewise.
21249 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21250 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21251 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21252 (mips16_insn_length): Robustify.
21253 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21254 (mips16_split_long_branches): Likewise.
21255 * config/pa/pa.c (pa_combine_instructions): Likewise.
21256 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21257 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21258 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21259 as contributing to pool range lengths.
21260 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21261 Remove redundant JUMP_TABLE_DATA_P test.
21262 (sh_loop_align): Likewise.
21263 (split_branches): Likewise.
21264 (sh_insn_length_adjustment): Likewise.
21265 * config/spu/spu.c (get_branch_target): Likewise.
21266
21267 2013-03-29 Jan Hubicka <jh@suse.cz>
21268
21269 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21270 gcov streaming; stream hot bb threshold to ltrans.
21271 * predict.c (get_hot_bb_threshold): Break out from ....
21272 (maybe_hot_count_p): ... here.
21273 (set_hot_bb_threshold): New function.
21274 * lto-section-in.c (lto_section_name): Add profile.
21275 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21276 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21277 and data-streamer.h
21278 (histogram_entry): New structure.
21279 (histogram, histogram_pool): New global vars.
21280 (histogram_hash): New structure.
21281 (histogram_hash::hash): New method.
21282 (histogram_hash::equal): Likewise.
21283 (account_time_size): New function.
21284 (cmp_counts): New function.
21285 (dump_histogram): New function.
21286 (ipa_profile_generate_summary): New function.
21287 (ipa_profile_write_summary): New function.
21288 (ipa_profile_read_summary): New function.
21289 (ipa_profile): Decide on threshold.
21290 (pass_ipa_profile): Add ipa_profile_write_summary and
21291 ipa_profile_read_summary.
21292 * Makefile.in (ipa.o): Update dependencies.
21293 * lto-streamer.h (LTO_section_ipa_profile): New section.
21294
21295 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21296
21297 * tree.h (VAR_P): New.
21298
21299 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21300
21301 PR lto/56777
21302 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21303
21304 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21305
21306 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21307 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21308 (control_flow_insn_p): Likewise.
21309 * cfgrtl.c (duplicate_insn_chain): Likewise.
21310 * final.c (get_attr_length_1): Likewise.
21311 (shorten_branches): Likewise.
21312 (final_scan_insn): Likewise.
21313 * function.c (instantiate_virtual_regs): Likewise.
21314 * gcse.c (insert_insn_end_basic_block): Likewise.
21315 * ira-costs.c (scan_one_insn): Likewise.
21316 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21317 * lra.c (check_rtl): Likewise.
21318 * reload1.c (elimination_costs_in_insn): Likewise.
21319 * reorg.c (follow_jumps): Likewise.
21320
21321 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21322 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21323 (thumb_far_jump_used_p): Likewise.
21324 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21325 (workaround_speculation): Likewise.
21326 (add_sched_insns_for_speculation): Likewise.
21327 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21328 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21329 (frv_for_each_packet): Likewise.
21330 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21331 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21332 (final_emit_insn_group_barriers): Likewise.
21333 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21334 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21335 (mips16_insn_length): Likewise.
21336 * config/pa/pa.c (pa_reorg): Likewise.
21337 (pa_combine_instructions): Likewise.
21338 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21339 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21340 (sh_reorg): Likewise.
21341 (split_branches): Likewise.
21342 * config/spu/spu.c (get_branch_target): Likewise.
21343
21344 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21345 JUMP_TABLE_DATA_P.
21346
21347 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21348
21349 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21350 Fix declaration name.
21351
21352 2013-03-28 Lawrence Crowl <crowl@google.com>
21353
21354 * graphds.h (struct graph.indicies): Remove unused.
21355 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21356 (SCOP_ORIGINAL_PDDRS): Remove unused.
21357 * sese.h (extern insert_loop_close_phis): Removed unused.
21358 (extern insert_guard_phis): Removed unused.
21359 (extern ivtype_map_elt_info): Removed unused.
21360 (new_ivtype_map_elt): Removed unused.
21361 * sese.c (ivtype_map_elt_info): Removed unused.
21362
21363 2013-03-28 Lawrence Crowl <crowl@google.com>
21364
21365 * Makefile.in: Add several missing include dependences.
21366 (DUMPFILE_H): New.
21367 (test-dump.o): New. This object is not added to any executable,
21368 but is present for ad-hoc testing.
21369 * bitmap.c
21370 (debug (const bitmap_head_def &)): New.
21371 (debug (const bitmap_head_def *)): New.
21372 * bitmap.h
21373 (extern debug (const bitmap_head_def &)): New.
21374 (extern debug (const bitmap_head_def *)): New.
21375 * cfg.c
21376 (debug (edge_def &)): New.
21377 (debug (edge_def *)): New.
21378 * cfghooks.c
21379 (debug (basic_block_def &)): New.
21380 (debug (basic_block_def *)): New.
21381 * dumpfile.h
21382 (dump_node (const_tree, int, FILE *)): Correct source file.
21383 * dwarf2out.c
21384 (debug (die_struct &)): New.
21385 (debug (die_struct *)): New.
21386 * dwarf2out.h
21387 (extern debug (die_struct &)): New.
21388 (extern debug (die_struct *)): New.
21389 * gimple-pretty-print.c
21390 (debug (gimple_statement_d &)): New.
21391 (debug (gimple_statement_d *)): New.
21392 * gimple-pretty-print.h
21393 (extern debug (gimple_statement_d &)): New.
21394 (extern debug (gimple_statement_d *)): New.
21395 * ira-build.c
21396 (debug (ira_allocno_copy &)): New.
21397 (debug (ira_allocno_copy *)): New.
21398 (debug (ira_allocno &)): New.
21399 (debug (ira_allocno *)): New.
21400 * ira-int.h
21401 (extern debug (ira_allocno_copy &)): New.
21402 (extern debug (ira_allocno_copy *)): New.
21403 (extern debug (ira_allocno &)): New.
21404 (extern debug (ira_allocno *)): New.
21405 * ira-lives.c
21406 (debug (live_range &)): New.
21407 (debug (live_range *)): New.
21408 * lra-int.h
21409 (debug (lra_live_range &)): New.
21410 (debug (lra_live_range *)): New.
21411 * lra-lives.c
21412 (debug (lra_live_range &)): New.
21413 (debug (lra_live_range *)): New.
21414 * omega.c
21415 (debug (omega_pb_d &)): New.
21416 (debug (omega_pb_d *)): New.
21417 * omega.h
21418 (extern debug (omega_pb_d &)): New.
21419 (extern debug (omega_pb_d *)): New.
21420 * print-rtl.c
21421 (debug (const rtx_def &)): New.
21422 (debug (const rtx_def *)): New.
21423 * print-tree.c
21424 (debug_tree (tree): Move within file.
21425 (debug_raw (const tree_node &)): New.
21426 (debug_raw (const tree_node *)): New.
21427 (dump_tree_via_hooks (const tree_node *, int)): New.
21428 (debug (const tree_node &)): New.
21429 (debug (const tree_node *)): New.
21430 (debug_verbose (const tree_node &)): New.
21431 (debug_verbose (const tree_node *)): New.
21432 (debug_head (const tree_node &)): New.
21433 (debug_head (const tree_node *)): New.
21434 (debug_body (const tree_node &)): New.
21435 (debug_body (const tree_node *)): New.
21436 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21437 (debug (vec<tree, va_gc> &)): New.
21438 (debug (vec<tree, va_gc> *)): New.
21439 * rtl.h
21440 (extern debug (const rtx_def &)): New.
21441 (extern debug (const rtx_def *)): New.
21442 * sbitmap.c
21443 (debug_raw (simple_bitmap_def &)): New.
21444 (debug_raw (simple_bitmap_def *)): New.
21445 (debug (simple_bitmap_def &)): New.
21446 (debug (simple_bitmap_def *)): New.
21447 * sbitmap.h
21448 (extern debug (simple_bitmap_def &)): New.
21449 (extern debug (simple_bitmap_def *)): New.
21450 (extern debug_raw (simple_bitmap_def &)): New.
21451 (extern debug_raw (simple_bitmap_def *)): New.
21452 * sel-sched-dump.c
21453 (debug (vinsn_def &)): New.
21454 (debug (vinsn_def *)): New.
21455 (debug_verbose (vinsn_def &)): New.
21456 (debug_verbose (vinsn_def *)): New.
21457 (debug (expr_def &)): New.
21458 (debug (expr_def *)): New.
21459 (debug_verbose (expr_def &)): New.
21460 (debug_verbose (expr_def *)): New.
21461 (debug (vec<rtx> &)): New.
21462 (debug (vec<rtx> *)): New.
21463 * sel-sched-dump.h
21464 (extern debug (vinsn_def &)): New.
21465 (extern debug (vinsn_def *)): New.
21466 (extern debug_verbose (vinsn_def &)): New.
21467 (extern debug_verbose (vinsn_def *)): New.
21468 (extern debug (expr_def &)): New.
21469 (extern debug (expr_def *)): New.
21470 (extern debug_verbose (expr_def &)): New.
21471 (extern debug_verbose (expr_def *)): New.
21472 (extern debug (vec<rtx> &)): New.
21473 (extern debug (vec<rtx> *)): New.
21474 * sel-sched-ir.h
21475 (_list_iter_cond_expr): Make inline instead of static.
21476 * sreal.c
21477 (debug (sreal &)): New.
21478 (debug (sreal *)): New.
21479 * sreal.h
21480 (extern debug (sreal &)): New.
21481 (extern debug (sreal *)): New.
21482 * tree.h
21483 (extern debug_raw (const tree_node &)): New.
21484 (extern debug_raw (const tree_node *)): New.
21485 (extern debug (const tree_node &)): New.
21486 (extern debug (const tree_node *)): New.
21487 (extern debug_verbose (const tree_node &)): New.
21488 (extern debug_verbose (const tree_node *)): New.
21489 (extern debug_head (const tree_node &)): New.
21490 (extern debug_head (const tree_node *)): New.
21491 (extern debug_body (const tree_node &)): New.
21492 (extern debug_body (const tree_node *)): New.
21493 (extern debug (vec<tree, va_gc> &)): New.
21494 (extern debug (vec<tree, va_gc> *)): New.
21495 * tree-cfg.c
21496 (debug (struct loop &)): New.
21497 (debug (struct loop *)): New.
21498 (debug_verbose (struct loop &)): New.
21499 (debug_verbose (struct loop *)): New.
21500 * tree-dump.c: Add header dependence.
21501 * tree-flow.h
21502 (extern debug (struct loop &)): New.
21503 (extern debug (struct loop *)): New.
21504 (extern debug_verbose (struct loop &)): New.
21505 (extern debug_verbose (struct loop *)): New.
21506 * tree-data-ref.c
21507 (debug (data_reference &)): New.
21508 (debug (data_reference *)): New.
21509 (debug (vec<data_reference_p> &)): New.
21510 (debug (vec<data_reference_p> *)): New.
21511 (debug (vec<ddr_p> &)): New.
21512 (debug (vec<ddr_p> *)): New.
21513 * tree-data-ref.h
21514 (extern debug (data_reference &)): New.
21515 (extern debug (data_reference *)): New.
21516 (extern debug (vec<data_reference_p> &)): New.
21517 (extern debug (vec<data_reference_p> *)): New.
21518 (extern debug (vec<ddr_p> &)): New.
21519 (extern debug (vec<ddr_p> *)): New.
21520 * tree-ssa-alias.c
21521 (debug (pt_solution &)): New.
21522 (debug (pt_solution *)): New.
21523 * tree-ssa-alias.h
21524 (extern debug (pt_solution &)): New.
21525 (extern debug (pt_solution *)): New.
21526 * tree-ssa-alias.c
21527 (debug (_var_map &)): New.
21528 (debug (_var_map *)): New.
21529 (debug (tree_live_info_d &)): New.
21530 (debug (tree_live_info_d *)): New.
21531 * tree-ssa-alias.h
21532 (extern debug (_var_map &)): New.
21533 (extern debug (_var_map *)): New.
21534 (extern debug (tree_live_info_d &)): New.
21535 (extern debug (tree_live_info_d *)): New.
21536
21537 2013-03-28 Jan Hubicka <jh@suse.cz>
21538
21539 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21540
21541 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21542
21543 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21544 record only when desired or required.
21545
21546 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21547
21548 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21549 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21550
21551 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21552
21553 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21554 (*andsi3_compare0_uxtw): New pattern.
21555 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21556 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21557
21558 2013-03-28 Jan Hubicka <jh@suse.cz>
21559
21560 * data-streamer-in.c (streamer_read_gcov_count): New function.
21561 * gimple-streamer-out.c: Include value-prof.h.
21562 (output_gimple_stmt): Output histogram.
21563 (output_bb): Use streamer_write_gcov_count.
21564 * value-prof.c: Include data-streamer.h
21565 (dump_histogram_value): Add HIST_TYPE_MAX.
21566 (stream_out_histogram_value): New function.
21567 (stream_in_histogram_value): New function.
21568 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21569 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21570 * data-streamer-out.c (streamer_write_gcov_count): New function.
21571 (streamer_write_gcov_count_stream): New function.
21572 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21573 (lto_output_node): Likewise.
21574 (input_node, input_edge): Likewise.
21575 * lto-streamer-out.c (output_cfg): Update streaming.
21576 * lto-streamer-in.c (input_cfg): Likewise.
21577 * data-streamer.h (streamer_write_gcov_count,
21578 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21579 * gimple-streamer-in.c: Include value-prof.h
21580 (input_gimple_stmt): Input histograms.
21581 (input_bb): Update profile streaming.
21582
21583 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21584
21585 * genmodes.c (emit_max_int): New function.
21586 (emit_insn_modes_h): Added call to emit_max_function.
21587 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21588 Added doc.
21589 * machmode.def: Fixed comment.
21590
21591 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21592
21593 * combine.c (try_combine): Removed useless assert.
21594 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21595
21596 2013-03-28 Marek Polacek <polacek@redhat.com>
21597 Richard Biener <rguenther@suse.de>
21598
21599 PR tree-optimization/56695
21600 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21601 build signed result of a vector comparison.
21602 * tree-cfg.c (verify_gimple_comparison): Check that a result
21603 of a vector comparison has signed type.
21604
21605 2013-03-28 Richard Biener <rguenther@suse.de>
21606
21607 PR tree-optimization/37021
21608 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21609 do not restrict gaps between groups.
21610 * tree-vect-stmts.c (vectorizable_load): Properly account for
21611 a gap between groups.
21612
21613 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21614
21615 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21616 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21617 is not enabled.
21618
21619 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21620
21621 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21622 * doc/extend.texi (Named Address Spaces): Ditto.
21623 (Variable Attributes): Ditto.
21624
21625 2013-03-27 Kai Tietz <ktietz@redhat.com>
21626
21627 * config.build: Add support for cygwin x64 target.
21628 * config.gcc: Likewise.
21629 * config.host: Likewise.
21630 * configure.ac: Likewise
21631 * configure: Regenerated.
21632
21633 2013-03-27 Kai Tietz <ktietz@redhat.com>
21634
21635 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21636 * config/i386/t-cygwin-w64: New file.
21637 * config/i386/cygwin-w64.h: New file.
21638 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21639 and add support for x64-cygwin target.
21640 (CPP_SPEC): Likewise.
21641 (CXX_WRAP_SPEC_LIST): Undefine before define.
21642 (LIBGCJ_SONAME): Use 15 as version.
21643
21644 2013-03-27 Richard Biener <rguenther@suse.de>
21645
21646 PR tree-optimization/56716
21647 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21648 dumping for ref nodes.
21649
21650 2013-03-27 Martin Jambor <mjambor@suse.cz>
21651
21652 PR tree-optimization/55334
21653 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21654 restricted pointers to arrays.
21655
21656 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21657
21658 * Makefile.in (.SUFFIXES): Add .cc.
21659 (.c.o): Apply same recipe for implicit rule .cc.o.
21660
21661 2013-03-27 Richard Biener <rguenther@suse.de>
21662
21663 PR tree-optimization/37021
21664 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21665 REALPART/IMAGPART_EXPRs around the supported refs.
21666 * tree-ssa-structalias.c (find_func_aliases): Assume that
21667 floating-point values are not used to transfer pointers.
21668
21669 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21670
21671 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21672 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21673 * doc/tm.texi: Regenerate.
21674 * targhooks.h (default_has_ifunc_p): New.
21675 * targhooks.c (default_has_ifunc_p): Ditto.
21676 * config/linux-protos.h: New file.
21677 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21678 hook for linux which disables support of indirect functions in android.
21679 * config/linux-android.c: New file.
21680 * config/t-linux-android.c: Ditto.
21681 * config.gcc: Added new object file linux-android.o.
21682 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21683 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21684 * varasm.c (do_assemble_alias): Likewise.
21685 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21686 doesn't support indirect functions.
21687 * configure: Regenerate.
21688
21689 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21690
21691 PR target/56102
21692 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21693 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21694 mult-word mode.
21695
21696 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21697
21698 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21699
21700 2013-03-27 Terry Guo <terry.guo@arm.com>
21701
21702 * config/arm/arm-cores.def: Added core cortex-r7.
21703 * config/arm/arm-tune.md: Regenerated.
21704 * config/arm/arm-tables.opt: Regenerated.
21705 * doc/invoke.texi: Added entry for core cortex-r7.
21706
21707 2013-03-27 Walter Lee <walt@tilera.com>
21708
21709 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21710 double-decrement of next_scratch_regno.
21711
21712 2013-03-27 Walter Lee <walt@tilera.com>
21713
21714 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21715 input operands.
21716 (insn_v1mulus): Ditto.
21717 (insn_v2muls): Ditto.
21718
21719 2013-03-27 Walter Lee <walt@tilera.com>
21720
21721 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21722 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21723
21724 2013-03-27 Walter Lee <walt@tilera.com>
21725
21726 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21727 (*sibcall_value): Ditto.
21728
21729 2013-03-27 Walter Lee <walt@tilera.com>
21730
21731 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21732 (insn_mnz_v8qi): ... this ...
21733 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21734 vector equivalent.
21735 (insn_v<n>mnz): Replaced by ...
21736 (insn_v1mnz): ... this ...
21737 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21738 equivalent.
21739 (insn_mz_<mode>): Replaced by ...
21740 (insn_mz_v8qi): ... this ...
21741 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21742 vector equivalent.
21743 (insn_v<n>mz): Replaced by ...
21744 (insn_v1mz): ... this ...
21745 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21746 equivalent.
21747
21748 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21749
21750 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21751
21752 2013-03-26 Roland McGrath <mcgrathr@google.com>
21753
21754 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21755 than fprintf with a non-constant, non-format string.
21756
21757 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21758
21759 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21760 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21761 operand 0 predicate.
21762 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21763 attribute. Use general_x64nomem_operand as operand 1 predicate.
21764 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21765 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21766 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21767 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21768 (mov<mode>_insv_1): Remove expander. Merge insn with
21769 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21770 Use general_x64nomem_operand as operand 1 predicate.
21771 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21772 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21773 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21774 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21775 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21776 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21777 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21778 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21779 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21780 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21781 (general_x64nomem_operand): Ditto.
21782
21783 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21784
21785 * config/rtems.opt: Add -pthread option.
21786
21787 2013-03-26 Richard Biener <rguenther@suse.de>
21788
21789 * alias.c (find_base_term): Avoid redundant and not used recursion.
21790 (base_alias_check): Get the initial base term from the caller.
21791 (true_dependence_1): Compute and pass base terms to base_alias_check.
21792 (write_dependence_p): Likewise.
21793 (may_alias_p): Likewise.
21794
21795 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21796
21797 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21798 PC-relative load in SI modes and above only.
21799
21800 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21801
21802 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21803 * config/arm/iwmmxt.md (WCGR0): Update.
21804 (WCGR1, WCGR2, WCGR3): Likewise.
21805
21806 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21807
21808 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21809 Use x64 and nox64 isa attributes.
21810
21811 2013-03-26 Richard Biener <rguenther@suse.de>
21812
21813 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21814 alignment computations and rely on get_object_alignment_1
21815 for the !TYPE_P case.
21816 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21817
21818 2013-03-26 Walter Lee <walt@tilera.com>
21819
21820 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21821 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21822
21823 2013-03-25 Jeff Law <law@redhat.com>
21824
21825 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21826 check for INTEGRAL_TYPE_P that was missing due to checking in
21827 wrong version of prior patch.
21828
21829 2013-03-25 Walter Lee <walt@tilera.com>
21830
21831 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21832 TILEGX_INSN_SHUFFLEBYTES1.
21833 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21834 shufflebytes1.
21835 (tilegx_builtins): Ditto.
21836 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21837
21838 2013-03-25 Walter Lee <walt@tilera.com>
21839
21840 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21841 (floatunssisf2): New pattern.
21842 (floatsidf2): New pattern.
21843 (floatunssidf2): New pattern.
21844
21845 2013-03-25 Walter Lee <walt@tilera.com>
21846
21847 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21848 tests for constraint J, K, N, P.
21849
21850 2013-03-25 Walter Lee <walt@tilera.com>
21851
21852 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21853 Use indirect/pcrel encoding.
21854 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21855 Ditto.
21856
21857 2013-03-25 Steve Ellcey <sellcey@mips.com>
21858
21859 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21860 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21861 * config/mips/mips.c (mips_option_override): Set IMADD default.
21862 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21863 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21864 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21865 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21866 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21867
21868 2013-03-25 Jeff Law <law@redhat.com>
21869
21870 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21871 slightly to avoid creating and folding useless trees. Simplify
21872 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21873
21874 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21875
21876 * config/i386/i386.md (*zero_extendsidi2): Merge with
21877 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21878 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21879 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21880 targets. Clarify comment.
21881
21882 2013-03-25 Martin Jambor <mjambor@suse.cz>
21883
21884 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21885 pass-through jump functions differently.
21886 (ipa_read_jump_function): Likewise. Also use setter functions to set
21887 up jump functions.
21888
21889 2013-03-25 Martin Jambor <mjambor@suse.cz>
21890
21891 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21892 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21893 process it.
21894 (ipa_get_indirect_edge_target): New function.
21895 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21896 ipa_get_indirect_edge_target. Update all callers.
21897 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21898 ipa_get_indirect_edge_target_1 instead of calling
21899 ipa_get_indirect_edge_target.
21900 (create_specialized_node): Pass aggvlas to
21901 ipcp_discover_new_direct_edges.
21902
21903 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21904
21905 * config/arm/arm.md (f_sels, f_seld): New types.
21906 (*cmov<mode>): New pattern.
21907 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21908 predicate.
21909
21910 2013-03-25 Kai Tietz <ktietz@redhat.com>
21911
21912 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21913 POSIX-printf for mingw-hosted builds.
21914
21915 2013-03-25 Richard Biener <rguenther@suse.de>
21916
21917 PR middle-end/56694
21918 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21919 must-not-throw stmt location.
21920
21921 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21922
21923 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21924 Emit load-acquire versions when acq is true.
21925 (arm_emit_store_exclusive): Add rel parameter.
21926 Emit store-release versions when rel is true.
21927 (arm_split_compare_and_swap): Use acquire-release instructions
21928 instead.
21929 of barriers when appropriate.
21930 (arm_split_atomic_op): Likewise.
21931 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21932 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21933 (VUNSPEC_SLX): Likewise.
21934 (VUNSPEC_LDA): Likewise.
21935 (VUNSPEC_STL): Likewise.
21936 * config/arm/sync.md (atomic_load<mode>): New pattern.
21937 (atomic_store<mode>): Likewise.
21938 (arm_load_acquire_exclusive<mode>): Likewise.
21939 (arm_load_acquire_exclusivesi): Likewise.
21940 (arm_load_acquire_exclusivedi): Likewise.
21941 (arm_store_release_exclusive<mode>): Likewise.
21942
21943 2013-03-25 Catherine Moore <clm@codesourcery.com>
21944
21945 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21946 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21947 * config/mip/predicates.md (lwsp_swsp_operand,
21948 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21949 sb16_operand, db4_operand, db7_operand, ib3_operand,
21950 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21951 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21952 andi16_operand): New predicates.
21953 * config/mips/mips.md (compression): New attribute.
21954 (enabled): New attribute.
21955 (length): Consider compression in computing length.
21956 (shift_compression): New code attribute.
21957 (*add<mode>3): New operands. Record compression.
21958 (sub<mode>3): Likewise.
21959 (one_cmpl<mode>2): Likewise.
21960 (*and<mode>3): Likewise.
21961 (*ior<mode>3): Likewise.
21962 (unnamed pattern for xor): Likewise.
21963 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21964 (*<optab><mode>3): Likewise.
21965 (*mov<mode>_internal: Likewise.
21966 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21967 (mips_unsigned_immediate_p): New.
21968 (umips_lwsp_swsp_address_p): New.
21969 (m16_based_address_p): New.
21970 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21971 (mips_unsigned_immediate_p): New prototype.
21972 (lwsp_swsp_address_p): New prototype.
21973 (m16_based_address_p): New prototype.
21974 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21975 (mips_signed_immediate_p): New function.
21976 (m16_based_address_p): New function.
21977 (lwsp_swsp_address_p): New function.
21978 (mips_print_operand_punctuation): Recognize short delay slot insns
21979 for microMIPS.add<mode>3"
21980
21981 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21982
21983 PR target/56720
21984 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21985 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21986
21987 2013-03-25 Richard Biener <rguenther@suse.de>
21988
21989 PR tree-optimization/56689
21990 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21991 any edge.
21992
21993 2013-03-25 Richard Biener <rguenther@suse.de>
21994
21995 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21996 of bitmap.
21997 (memory_references): Likewise.
21998 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21999 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22000 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22001 (gather_mem_refs_in_loops): Fold into ...
22002 (analyze_memory_references): ... this. Move initialization
22003 to tree_ssa_lim_initialize.
22004 (fill_always_executed_in): Rename to ...
22005 (fill_always_executed_in_1): ... this.
22006 (fill_always_executed_in): Move contains_call computation to
22007 this new function from ...
22008 (tree_ssa_lim_initialize): ... here.
22009 (tree_ssa_lim): Call fill_always_executed_in.
22010
22011 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22012
22013 * postreload.c (reload_combine): Fix code detecting returns.
22014
22015 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22016
22017 * function.c (emit_use_return_register_into_block): On cc0 targets,
22018 do not emit the sequence between cc0 setter and user.
22019
22020 2013-03-25 Kai Tietz <ktietz@redhat.com>
22021
22022 * config/i386/predicates.md (local_symbolic_operand): Interpret
22023 dll-imported symbols as none-local.
22024
22025 2013-03-25 Richard Biener <rguenther@suse.de>
22026
22027 * tree-ssa-loop-im.c (struct depend): Remove.
22028 (struct lim_aux_data): Make depends a vec of gimples.
22029 (free_lim_aux_data): Adjust.
22030 (add_dependency): Likewise.
22031 (set_level): Likewise.
22032
22033 2013-03-25 Richard Biener <rguenther@suse.de>
22034
22035 PR middle-end/56434
22036 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22037 the pointer returned by calls with ECF_MALLOC set.
22038
22039 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22040
22041 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22042
22043 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22044
22045 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22046 using MMXMODE mode iterator.
22047 (*move<mode>_internal): Merge with *movv2sf_internal and
22048 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22049
22050 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22051
22052 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22053 (record_last_mem_set_info): Likewise.
22054
22055 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22056 of XNEWVEC followed by memset.
22057 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22058
22059 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22060
22061 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22062 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22063 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22064 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22065 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22066 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22067 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22068 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22069 BARRIER_P instead of GET_CODE.
22070
22071 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22072
22073 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22074 inaccuracy in the probing code.
22075
22076 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22077 (ctrapdi4): Likewise.
22078
22079 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22080
22081 * calls.c (expand_call): Add missing guard to code handling return
22082 of non-BLKmode structures in MSB.
22083 * function.c (expand_function_end): Likewise.
22084
22085 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22086
22087 * combine.c (try_combine): Adjust comment. Do not add the set of
22088 insn #0 if the destination indirectly is set or dies in insn #2.
22089 Tidy up code to distribute a new note.
22090
22091 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22092
22093 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22094 also for alternatives 16 and 17.
22095
22096 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22097
22098 * config/i386/sse.md (*mov<mode>_internal): Merge with
22099 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22100 Emit insn template depending on type attribute. Use
22101 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22102 movd instead of movq mnemonic for interunit moves. Rewrite mode
22103 attribute calculation. Remove unit attribute calculation.
22104 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22105 Set prefix_data16 attribute for DImode ssemov types.
22106 Use Ym instead of y for SSE-MMX conversion alternatives.
22107 Reorder operand constraints.
22108
22109 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22110
22111 * df.h (df_insn_delete): Adjust prototype.
22112 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22113 and let it decide whether mark the basic block dirty.
22114 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22115 * df-scan.c (df_insn_info_delete): New helper function, split
22116 off from df_insn_delete.
22117 (df_scan_free_bb_info): Use it.
22118 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22119 Likewise.
22120 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22121 that the insn is actually an insn and it has a non-NULL basic block.
22122 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22123
22124 2013-03-22 Richard Biener <rguenther@suse.de>
22125
22126 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22127 dep_ref members.
22128 (mem_ref_alloc): Do not allocate them.
22129 (refs_independent_p): Do not query or maintain a cache.
22130
22131 2013-03-22 Richard Biener <rguenther@suse.de>
22132
22133 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22134 (gather_mem_refs_in_loops): Do not compute it.
22135 (analyze_memory_references): Do not allocate it.
22136 (tree_ssa_lim_finalize): Do not free it.
22137 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22138
22139 2013-03-22 Richard Biener <rguenther@suse.de>
22140
22141 * is-a.h (as_a): Use gcc_checking_assert.
22142
22143 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22144
22145 * config/aarch64/aarch64.c (aarch64_print_operand): New
22146 format specifier for printing a constant in hex.
22147 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22148 format specifier for printing second operand.
22149
22150 2013-03-22 Richard Biener <rguenther@suse.de>
22151
22152 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22153 bitmaps.
22154 (gather_mem_refs_in_loops): Perform store accumulation here.
22155 (create_vop_ref_mapping_loop): Remove.
22156 (create_vop_ref_mapping): Likewise.
22157 (analyze_memory_references): Initialize refs_stored_in_loop.
22158 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22159 (record_indep_loop): Remove.
22160 (record_dep_loop): New function.
22161 (ref_indep_loop_p_1): Adjust to only walk over references
22162 in the loop, not its subloops.
22163 (ref_indep_loop_p): Rename to ...
22164 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22165 maintaining a more fine-grained cache.
22166 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22167 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22168
22169 2013-03-22 Richard Biener <rguenther@suse.de>
22170
22171 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22172 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22173 aggregate mem_ref_loc.
22174 (free_mem_ref_locs): Inline into ...
22175 (memref_free): ... this and adjust.
22176 (mem_ref_alloc): Adjust.
22177 (mem_ref_locs_alloc): Remove.
22178 (record_mem_ref_loc): Adjust.
22179 (get_all_locs_in_loop): Rewrite into ...
22180 (for_all_locs_in_loop): ... this iterator.
22181 (rewrite_mem_ref_loc): New functor.
22182 (rewrite_mem_refs): Use for_all_locs_in_loop.
22183 (sm_set_flag_if_changed): New functor.
22184 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22185 (ref_always_accessed): New functor.
22186 (ref_always_accessed_p): Use for_all_locs_in_loop.
22187
22188 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22189
22190 * tree-pass.h (PROP_gimple_lvec): New.
22191 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22192 (init_optimization_passes): Move pass_lower_vector.
22193 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22194 PROP_gimple_lvec.
22195 (pass_lower_vector): Provide PROP_gimple_lvec.
22196 (pass_lower_vector_ssa): Likewise.
22197 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22198
22199 2013-03-21 Mark Wielaard <mjw@redhat.com>
22200
22201 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22202
22203 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22204
22205 * config/i386/i386.md (*movdi_internal): Disparage slightly
22206 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22207 conversion alternatives.
22208
22209 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22210
22211 PR middle-end/48087
22212 * diagnostic.def (DK_WERROR): New kind.
22213 * diagnostic.h (werrorcount): Define.
22214 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22215 promoted to DK_ERROR, increment DK_WERROR counter instead of
22216 DK_ERROR counter.
22217 * toplev.c (toplev_main): Call print_ignored_options even if
22218 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22219 even if just werrorcount is non-zero.
22220
22221 PR debug/55608
22222 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22223 on failure.
22224 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22225 (string_cst_pool_decl): New function.
22226 (optimize_one_addr_into_implicit_ptr): New function.
22227 (resolve_addr_in_expr): Optimize DWARF location expression
22228 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22229 which doesn't live in memory, but has DW_AT_location or
22230 DW_AT_const_value, or refers to a string literal, into
22231 DW_OP_GNU_implicit_pointer.
22232 (optimize_location_into_implicit_ptr): New function.
22233 (resolve_addr): If removing DW_AT_location of a variable because
22234 it was DW_OP_addr of address of the variable, but the variable doesn't
22235 live in memory, try to emit const value attribute for the initializer.
22236
22237 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22238
22239 * tree.h (VECTOR_TYPE_P): New macro.
22240 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22241 TYPE_MODE): Use it.
22242 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22243 VEC_COND_EXPR cannot be lvalues.
22244 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22245
22246 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22247
22248 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22249 Restrict the transformation to equal modes.
22250
22251 2013-03-21 Richard Biener <rguenther@suse.de>
22252
22253 PR tree-optimization/39326
22254 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22255 (MEM_ANALYZABLE): Adjust.
22256 (record_mem_ref_loc): Move bitmap ops ...
22257 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22258 unanalyzable refs, do not record locations for it.
22259 (analyze_memory_references): Allocate ref zero as shared
22260 unanalyzable ref.
22261 (refs_independent_p): Do not test for unanalyzed mems here.
22262 (ref_indep_loop_p_1): Special-case disambiguation against
22263 the unanalyzed ref.
22264 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22265
22266 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22267
22268 * config/arm/arm-protos.h (tune_params): Add
22269 prefer_neon_for_64bits field.
22270 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22271 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22272 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22273 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22274 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22275 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22276 (arm_option_override): Handle -mneon-for-64bits new option.
22277 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22278 (prefer_neon_for_64bits): Declare new variable.
22279 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22280 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22281 (arch_enabled): Handle new arch types. Remove support for onlya8
22282 and nota8.
22283 (one_cmpldi2): Use new arch names.
22284 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22285 * config/arm/arm.opt (mneon-for-64bits): Add option.
22286 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22287 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22288 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22289 of onlya8.
22290 * doc/invoke.texi (-mneon-for-64bits): Document.
22291
22292 2013-03-21 Richard Biener <rguenther@suse.de>
22293
22294 PR tree-optimization/39326
22295 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22296 (sort_bbs_in_loop_postorder_cmp): New function.
22297 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22298
22299 2013-03-21 Richard Biener <rguenther@suse.de>
22300
22301 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22302 (vect_insert_into_interleaving_chain): Likewise.
22303 (vect_drs_dependent_in_basic_block): Inline ...
22304 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22305 split out from ...
22306 (vect_analyze_data_ref_dependence): ... here. Simplify.
22307 (vect_check_interleaving): Simplify.
22308 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22309 (vect_slp_analyze_data_ref_dependences): ... this new function.
22310 (dr_group_sort_cmp): New function.
22311 (vect_analyze_data_ref_accesses): Compute data-reference groups
22312 here instead of in vect_analyze_data_ref_dependence. Use
22313 a more efficient algorithm.
22314 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22315 vect_slp_analyze_data_ref_dependences. Call
22316 vect_analyze_data_ref_accesses earlier.
22317 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22318 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22319 (vect_slp_analyze_data_ref_dependences): New prototype.
22320
22321 2013-03-21 Richard Biener <rguenther@suse.de>
22322
22323 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22324 ref is stored in the loop.
22325 (find_refs_for_sm): Walk only over all stores.
22326 (store_motion_loop): Allocate from lim_bitmap_obstack.
22327 (store_motion): Likewise.
22328
22329 2013-03-21 Richard Biener <rguenther@suse.de>
22330
22331 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22332 Update virtual SSA form.
22333
22334 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22335
22336 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22337 * configure: Regenerate.
22338 * config.in: Regenerate.
22339 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22340 if !HAVE_LD_EH_FRAME_CIEV3.
22341
22342 2013-03-21 Richard Biener <rguenther@suse.de>
22343
22344 * tree-cfg.c (verify_expr_no_block): New function.
22345 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22346 nor DECL_VALUE_EXPR have locations with associated blocks.
22347 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22348 (clear_unused_block_pointer): Remove code dealing with
22349 blocks in DECL_DEBUG_EXPR locations.
22350
22351 2013-03-21 Richard Biener <rguenther@suse.de>
22352
22353 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22354 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22355 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22356 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22357 instead of DECL_DEBUG_EXPR_IS_FROM.
22358 * gimplify.c (gimplify_modify_expr): Likewise.
22359 * tree-cfg.c (verify_expr_location_1): Likewise.
22360 * tree-complex.c (create_one_component_var): Likewise.
22361 * tree-sra.c (create_access_replacement): Likewise.
22362 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22363 (clear_unused_block_pointer): Likewise.
22364 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22365 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22366 * var-tracking.c (var_debug_decl): Likewise.
22367 (track_expr_p): Likewise.
22368 * tree-inline.c (add_local_variables): Likewise. Set
22369 DECL_HAS_DEBUG_EXPR_P after copying it.
22370 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22371 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22372
22373 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22374
22375 PR bootstrap/56656
22376 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22377 * configure: Regenerate.
22378 * config.in: Regenerate.
22379 * config/i386/i386.md (*movdf_internal): Use
22380 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22381 movd instead of movq mnemonic for interunit moves.
22382 (*movdi_internal): Ditto.
22383
22384 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22385
22386 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22387 (abd<mode>_3): New pattern.
22388 (aba<mode>_3): New pattern.
22389 (fabd<mode>_3): New pattern.
22390
22391 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22392
22393 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22394 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22395 occurrence of REGISTER_PREFIX as its empty string.
22396
22397 2013-03-20 Jeff Law <law@redhat.com>
22398
22399 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22400 addititional equivalences for equality comparisons between an SSA_NAME
22401 and a constant where the SSA_NAME was set from a widening conversion.
22402
22403 2013-03-20 Walter Lee <walt@tilera.com>
22404
22405 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22406
22407 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22408
22409 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22410 depending on type attribute.
22411 (*movti_internal): Ditto.
22412 (*movtf_internal): Ditto.
22413 (*movxf_internal): Ditto.
22414 (*movdf_internal): Ditto.
22415 (*movsf_internal): Ditto.
22416
22417 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22418
22419 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22420 maybe_vex for sselog1 and ssemov types.
22421 (*movdi_internal): Reorder operand constraints.
22422 (*movsi_internal): Ditto. Set prefix attribute to
22423 maybe_vex for sselog1 and ssemov types.
22424 (*movtf_internal): Set prefix attribute to maybe_vex
22425 for sselog1 and ssemov types.
22426 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22427 DImode ssemov types. Reorder operand constraints.
22428 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22429 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22430 attribute for SImode ssemov types. Reorder operand constraints.
22431
22432 2013-03-20 Martin Jambor <mjambor@suse.cz>
22433
22434 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22435 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22436
22437 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22438
22439 * config/rs6000/predicates.md (indexed_address, update_address_mem
22440 update_indexed_address_mem): New predicates.
22441 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22442 attribute for load/store instructions.
22443 * config/rs6000/dfp.md (movsd_store): Likewise.
22444 (movsd_load): Likewise.
22445 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22446 (unnamed HI->DI extend define_insn): Likewise.
22447 (unnamed SI->DI extend define_insn): Likewise.
22448 (unnamed QI->SI extend define_insn): Likewise.
22449 (unnamed QI->HI extend define_insn): Likewise.
22450 (unnamed HI->SI extend define_insn): Likewise.
22451 (unnamed HI->SI extend define_insn): Likewise.
22452 (extendsfdf2_fpr): Likewise.
22453 (movsi_internal1): Likewise.
22454 (movsi_internal1_single): Likewise.
22455 (movhi_internal): Likewise.
22456 (movqi_internal): Likewise.
22457 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22458 attribute for load/store instructions.
22459 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22460 instructions.
22461 (mov<mode>_softfloat): Likewise.
22462 (mov<mode>_hardfloat32): Likewise.
22463 (mov<mode>_hardfloat64): Likewise.
22464 (mov<mode>_softfloat64): Likewise.
22465 (movdi_internal32): Likewise.
22466 (movdi_internal64): Likewise.
22467 (probe_stack_<mode>): Likewise.
22468
22469 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22470
22471 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22472 floating point, and decimal floating point to reload iterator.
22473
22474 * config/rs6000/constraints.md (wl constraint): New constraints to
22475 return FLOAT_REGS if certain options are used to reduce the number
22476 of separate patterns that exist in the file.
22477 (wx constraint): Likewise.
22478 (wz constraint): Likewise.
22479
22480 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22481 -mdebug=reg, print wg, wl, wx, and wz constraints.
22482 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22483 Initialize the reload functions for 64-bit binary/decimal floating
22484 point types.
22485 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22486 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22487 create the buffer on the stack to overcome not having a 32-bit
22488 load and store.
22489 (rs6000_emit_move): Likewise.
22490 (rs6000_secondary_memory_needed_rtx): Likewise.
22491 (rs6000_alloc_sdmode_stack_slot): Likewise.
22492 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22493 via xxlxor, just like DFmode 0.0.
22494
22495 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22496 define as 1 if we are running on a power7 or newer.
22497 (enum r6000_reg_class_enum): Add new constraints.
22498
22499 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22500 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22501 with other moves by using conditional constraits (wg). Use LFIWZX
22502 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22503 (movsd splitter): Likewise.
22504 (movsd_hardfloat): Likewise.
22505 (movsd_softfloat): Likewise.
22506
22507 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22508 binary and decimal floating point moves.
22509 (fmove_ok): New attributes to combine binary and decimal floating
22510 point moves, and to combine power6x (mfpgpr) moves along normal
22511 floating moves.
22512 (real_value_to_target): Likewise.
22513 (f32_lr): Likewise.
22514 (f32_lm): Likewise.
22515 (f32_li): Likewise.
22516 (f32_sr): Likewise.
22517 (f32_sm): Likewise.
22518 (f32_si): Likewise.
22519 (movsf): Combine binary and decimal floating point moves. Combine
22520 power6x (mfpgpr) moves with other moves by using conditional
22521 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22522 (mov<mode> for SFmode/SDmode); Likewise.
22523 (SFmode/SDmode splitters): Likewise.
22524 (movsf_hardfloat): Likewise.
22525 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22526 (movsf_softfloat): Likewise.
22527 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22528
22529 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22530 wx and wz constraints.
22531
22532 * config/rs6000/constraints.md (wg constraint): New constraint to
22533 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22534
22535 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22536 constraint.
22537
22538 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22539 -mdebug=reg, print wg, wl, wx, and wz constraints.
22540 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22541 Initialize the reload functions for 64-bit binary/decimal floating
22542 point types.
22543 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22544 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22545 create the buffer on the stack to overcome not having a 32-bit
22546 load and store.
22547 (rs6000_emit_move): Likewise.
22548 (rs6000_secondary_memory_needed_rtx): Likewise.
22549 (rs6000_alloc_sdmode_stack_slot): Likewise.
22550 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22551 via xxlxor, just like DFmode 0.0.
22552
22553 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22554 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22555 with other moves by using conditional constraits (wg). Use LFIWZX
22556 and STFIWX for loading SDmode on power7.
22557 (movdd splitters): Likewise.
22558 (movdd_hardfloat32): Likewise.
22559 (movdd_softfloat32): Likewise.
22560 (movdd_hardfloat64_mfpgpr): Likewise.
22561 (movdd_hardfloat64): Likewise.
22562 (movdd_softfloat64): Likewise.
22563
22564 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22565 64-bit binary and decimal floating point moves.
22566 (FMOVE64X): Likewise.
22567 (movdf): Combine 64-bit binary and decimal floating point moves.
22568 Combine power6x (mfpgpr) moves with other moves by using
22569 conditional constraits (wg).
22570 (mov<mode> for DFmode/DDmode): Likewise.
22571 (DFmode/DDmode splitters): Likewise.
22572 (movdf_hardfloat32): Likewise.
22573 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22574 (movdf_softfloat32): Likewise.
22575 (movdf_hardfloat64_mfpgpr): Likewise.
22576 (movdf_hardfloat64): Likewise.
22577 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22578 (movdf_softfloat64): Likewise.
22579 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22580 (reload_<mode>_load): Move to later in the file so they aren't in
22581 the middle of the floating point move insns.
22582 (reload_<mode>_store): Likewise.
22583
22584 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22585 constraint.
22586
22587 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22588 constraint if -mdebug=reg.
22589 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22590 Enable using dd reload support if needed.
22591
22592 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22593 binary and decimal floating point moves in rs6000.md.
22594 (movtd_internal): Likewise.
22595
22596 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22597 decimal floating point moves.
22598 (movtf): Likewise.
22599 (movtf_internal): Likewise.
22600 (mov<mode>_internal, TDmode/TFmode): Likewise.
22601 (movtf_softfloat): Likewise.
22602 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22603
22604 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22605 movdi_internal64, using wg constraint for move direct operations.
22606 (movdi_internal64): Likewise.
22607
22608 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22609 MODES_TIEABLE_P for selected modes. Print the numerical value of
22610 the various virtual registers. Use GPR/FPR first/last values,
22611 instead of hard coding the register numbers. Print which modes
22612 have reload functions registered.
22613 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22614 settings before/after setting cpu, target and subtarget settings.
22615 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22616 and for secondary reload failures in rs6000_secondary_reload_inner.
22617 (rs6000_secondary_reload_fail): Likewise.
22618 (rs6000_secondary_reload_inner): Likewise.
22619
22620 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22621 macros for first/last GPR and FPR registers.
22622 (LAST_GPR_REGNO): Likewise.
22623 (FIRST_FPR_REGNO): Likewise.
22624 (LAST_FPR_REGNO): Likewise.
22625
22626 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22627 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22628 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22629 (vcond<mode><mode>): Likewise.
22630 (vcondu<mode><mode>): Likewise.
22631 (vector_gtu<mode>): Likewise.
22632 (vector_gte<mode>): Likewise.
22633 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22634 to prevent the compiler from converting DImode operations to TImode.
22635 (ior<mode>3): Likewise.
22636 (and<mode>3): Likewise.
22637 (one_cmpl<mode>2): Likewise.
22638 (nor<mode>3): Likewise.
22639 (andc<mode>3): Likewise.
22640
22641 * config/rs6000/constraints.md (wt constraint): New constraint
22642 that returns VSX_REGS if TImode is allowed in VSX registers.
22643
22644 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22645 constant under VSX.
22646
22647 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22648 similar to TImode, but it is restricted to being in the GPRs.
22649
22650 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22651 TImode to occupy a single VSX register.
22652
22653 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22654 -mvsx-timode for power7/power8.
22655 (power7 cpu): Likewise.
22656 (power8 cpu): Likewise.
22657
22658 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22659 sure that TFmode/TDmode take up two registers if they are ever
22660 allowed in the upper VSX registers.
22661 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22662 registers.
22663 (rs6000_init_hard_regno_mode_ok): Likewise.
22664 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22665 constraint. Print if LRA is turned on.
22666 (rs6000_option_override_internal): Give an error if -mvsx-timode
22667 and VSX is not enabled.
22668 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22669 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22670 to reg+offset addressing. Use PTImode when checking offset
22671 addresses for validity.
22672 (reg_offset_addressing_ok_p): Likewise.
22673 (rs6000_legitimate_offset_address_p): Likewise.
22674 (rs6000_legitimize_address): Likewise.
22675 (rs6000_legitimize_reload_address): Likewise.
22676 (rs6000_legitimate_address_p): Likewise.
22677 (rs6000_eliminate_indexed_memrefs): Likewise.
22678 (rs6000_emit_move): Likewise.
22679 (rs6000_secondary_reload): Likewise.
22680 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22681 reloads to fpr registers to continue to use reg+offset addressing,
22682 but 64-bit reloads to altivec registers need reg+reg addressing.
22683 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22684 it. Treat LO_SUM like a PLUS operation.
22685 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22686 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22687 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22688 registers to share a register with a smaller sized type, since VSX
22689 puts scalars in the upper 64-bits.
22690 (print_operand): Add support for PTImode.
22691 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22692 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22693 registers, but don't have arithmetic support.
22694 (rs6000_memory_move_cost): Add test for VSX.
22695 (rs6000_opt_masks): Add -mvsx-timode.
22696
22697 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22698 for TImode.
22699 (VSs): Likewise.
22700 (VSr): Use wt constraint for TImode.
22701 (VSv): Drop TImode support.
22702 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22703 (vsx_movti_64bit): Likewise.
22704 (vsx_movti_32bit): Likewise.
22705 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22706 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22707 one '?' on the appropriate output constraint. Do not allow TImode
22708 logical operations on 32-bit systems.
22709 (vsx_ior<mode>3): Likewise.
22710 (vsx_xor<mode>3): Likewise.
22711 (vsx_one_cmpl<mode>2): Likewise.
22712 (vsx_nor<mode>3): Likewise.
22713 (vsx_andc<mode>3): Likewise.
22714 (vsx_concat_<mode>): Likewise.
22715 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22716
22717 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22718 OPTION_MASK_VSX_TIMODE.
22719 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22720 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22721
22722 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22723 (TI2 iterator): New iterator for TImode, PTImode.
22724 (wd mode attribute): Add values for vector types.
22725 (movti_string): Replace TI move operations with operations for TImode
22726 and PTImode. Add support for TImode being allowed in VSX registers.
22727 (mov<mode>_string, TImode/PTImode): Likewise.
22728 (movti_ppc64): Likewise.
22729 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22730 (TI mode splitters): Likewise.
22731
22732 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22733 constraint.
22734
22735 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22736
22737 PR tree-optimization/56355
22738 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22739 Also handle integers with undefined overflow.
22740
22741 2013-03-20 Catherine Moore <clm@codesourcery.com>
22742 Maciej W. Rozycki <macro@codesourcery.com>
22743 Tom de Vries <tom@codesourcery.com>
22744 Nathan Sidwell <nathan@codesourcery.com>
22745 Iain Sandoe <iain@codesourcery.com>
22746 Nathan Froyd <froydnj@codesourcery.com>
22747 Chao-ying Fu <fu@mips.com>
22748
22749 * doc/extend.texi: (micromips, nomicromips, nocompression):
22750 Document new function attributes.
22751 * doc/invoke.texi (minterlink-compressed, mmicromips,
22752 m14k, m14ke, m14kec): Document new options.
22753 (minterlink-mips16): Update documentation.
22754 * doc/md.texi (ZC, ZD): Document new constraints.
22755 * configure.ac (gcc_cv_as_micromips): Check if linker
22756 supports the .set micromips directive.
22757 * configure: Regenerate.
22758 * config.in: Regenerate.
22759 * config/mips/mips-tables.opt: Regenerate.
22760 * config/mips/micromips.md: New file.
22761 * constraints.md (ZC, ZD): New constraints.
22762 * config/mips/predicates.md (movep_src_register): New predicate.
22763 (movep_src_operand): New predicate.
22764 (non_volatile_mem_operand): New predicate.
22765 * config/mips/mips.md (multimem): New type.
22766 (length): Differentiate between 17-bit and 18-bit branch offsets.
22767 (MOVEP1, MOVEP2): New mode iterator.
22768 (mov_<load>l): Use ZC constraint.
22769 (mov_<load>r): Likewise.
22770 (mov_<store>l): Likewise.
22771 (mov_<store>r): Likewise.
22772 (*branch_equality<mode>_inverted): Add microMIPS support.
22773 (*branch_equality<mode>): Likewise.
22774 (*jump_absolute): Likewise.
22775 (indirect_jump_<mode>): Likewise.
22776 (tablejump_<mode>): Likewise.
22777 (<optab>_internal): Likewise.
22778 (sibcall_internal): Likewise.
22779 (sibcall_value_internal): Likewise.
22780 (prefetch): Use constraint ZD.
22781 * config/mips/mips.opt (minterlink-compressed): New option.
22782 (minterlink-mips16): Now an alias for minterlink-compressed.
22783 (mmicromips): New option.
22784 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22785 (compare_and_swap_12): Likewise.
22786 (sync_add<mode>): Likewise.
22787 (sync_<optab>_12): Likewise.
22788 (sync_old_<optab>_12): Likewise.
22789 (sync_new_<optab>_12): Likewise.
22790 (sync_nand_12): Likewise.
22791 (sync_old_nand_12): Likewise.
22792 (sync_new_nand_12): Likewise.
22793 (sync_sub<mode>): Likewise.
22794 (sync_old_add<mode>): Likewise.
22795 (sync_old_sub<mode>): Likewise.
22796 (sync_new_add<mode>): Likewise.
22797 (sync_new_sub<mode>): Likewise.
22798 (sync_<optab><mode>): Likewise.
22799 (sync_old_<optab><mode>): Likewise.
22800 (sync_new_<optab><mode>): Likewise.
22801 (sync_nand<mode>): Likewise.
22802 (sync_old_nand<mode>): Likewise.
22803 (sync_new_nand<mode>): Likewise.
22804 (sync_lock_test_and_set<mode>): Likewise.
22805 (test_and_set_12): Likewise.
22806 (atomic_compare_and_swap<mode>): Likewise.
22807 (atomic_exchange<mode>_llsc): Likewise.
22808 (atomic_fetch_add<mode>_llsc): Likewise.
22809 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22810 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22811 (umips_save_restore_pattern_p): Likewise.
22812 (umips_load_store_pair_p): Likewise.
22813 (umips_output_load_store_pair): Likewise.
22814 (umips_movep_target_p): Likewise.
22815 (umips_12bit_offset_address_p): Likewise.
22816 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22817 (mips_base_mips16): Rename this...
22818 (mips_base_compression_flags): ...to this. Update all uses.
22819 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22820 (mips_mips16_decl_p): Delete.
22821 (mips_nomips16_decl_p): Delete.
22822 (mips_get_compress_on_flags): New function.
22823 (mips_get_compress_off_flags): New function.
22824 (mips_get_compress_mode): New function.
22825 (mips_get_compress_on_name): New function.
22826 (mips_get_compress_off_name): New function.
22827 (mips_insert_attributes): Support multiple compression types.
22828 (mips_merge_decl_attributes): Likewise.
22829 (umips_12bit_offset_address_p): New function.
22830 (mips_start_function_definition): Emit .set micromips directive.
22831 (mips_call_may_need_jalx_p): New function.
22832 (mips_function_ok_for_sibcall): Add microMIPS support.
22833 (mips_print_operand_punctuation): Support short delay slots and
22834 compact jumps.
22835 (umips_swm_mask, umips_swm_encoding): New.
22836 (umips_build_save_restore): New function.
22837 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22838 (was_mips16_p): Remove.
22839 (old_compression_mode): New.
22840 (mips_set_compression_mode): New function.
22841 (mips_set_current_function): Add microMIPS support.
22842 (mips_option_override): Likewise.
22843 (umips_save_restore_pattern_p): New function.
22844 (umips_output_save_restore): New function.
22845 (umips_load_store_pair_p_1): New function.
22846 (umips_load_store_pair_p): New function.
22847 (umips_output_load_store_pair_1): New function.
22848 (umips_output_load_store_pair): New function.
22849 (umips_movep_target_p) New function.
22850 (mips_prepare_pch_save): Add microMIPS support.
22851 * config/mips/mips.h (TARGET_COMPRESSION): New.
22852 (TARGET_CPU_CPP_BUILTINS): Update macro
22853 to use new compression flags and to support microMIPS.
22854 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22855 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22856 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22857 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22858 (ASM_SPEC): Support mmicromips and mno-micromips.
22859 (M16STORE_REG_P): New macro.
22860 (MIPS_CALL): Support TARGET_MICROMIPS.
22861 (MICROMIPS_J): New macro.
22862 (mips_base_mips16): Rename this...
22863 (mips_base_compression_flags): ...to this.
22864 (UMIPS_12BIT_OFFSET_P): New macro.
22865 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22866 (MULTILIB_DIRNAMES): Likewise.
22867 2013-03-20 Richard Biener <rguenther@suse.de>
22868
22869 PR tree-optimization/56661
22870 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22871 the result does not have to be distinct.
22872
22873 2013-03-20 Richard Biener <rguenther@suse.de>
22874
22875 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22876 remap_gimple_op_r.
22877
22878 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22879 Steven Bosscher <steven@gcc.gnu.org>
22880
22881 PR rtl-optimization/56605
22882 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22883
22884 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22885
22886 PR bootstrap/56656
22887 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22888 that require movd instead of movq.
22889
22890 2013-03-20 Richard Biener <rguenther@suse.de>
22891
22892 * tree-ssa-structalias.c (struct variable_info): Add pointer
22893 to the first field of an aggregate with sub-vars. Make
22894 this and the pointer to the next subfield its ID.
22895 (vi_next): New function.
22896 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22897 storedanything_id, integer_id): Increment by one.
22898 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22899 get_call_clobber_vi): Adjust.
22900 (solution_set_expand): Simplify and speedup.
22901 (solution_set_add): Inline into ...
22902 (set_union_with_increment): ... this. Adjust accordingly.
22903 (do_sd_constraint): Likewise.
22904 (do_ds_constraint): Likewise.
22905 (do_complex_constraint): Simplify.
22906 (build_pred_graph): Adjust.
22907 (solve_graph): Likewise. Simplify and speedup.
22908 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22909 get_constraint_for_component_ref, get_constraint_for_1,
22910 first_vi_for_offset, first_or_preceding_vi_for_offset,
22911 create_function_info_for, create_variable_info_for_1,
22912 create_variable_info_for, intra_create_variable_infos): Adjust.
22913 (init_base_vars): Push NULL for ID zero.
22914 (compute_points_to_sets): Adjust.
22915
22916 2013-03-20 Richard Biener <rguenther@suse.de>
22917
22918 * cfgloop.c (verify_loop_structure): Streamline and avoid
22919 ICEing on corrupt loop tree.
22920 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22921 loop tree.
22922
22923 2013-03-20 Richard Biener <rguenther@suse.de>
22924
22925 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22926 check whether an SSA update is needed.
22927
22928 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22929
22930 * config/mips/constraints.md (T): Rename to...
22931 (Yf): ...this.
22932 (U): Rename to...
22933 (Yd): ...this.
22934 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22935 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22936
22937 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22938
22939 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22940 (*subsi3_carryin_uxtw): Likewise.
22941
22942 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22943
22944 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22945 (*rorsi3_insn_uxtw): Likewise.
22946
22947 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22948
22949 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22950 (*extrsi5_insn_uxtw): Likewise.
22951
22952 2013-03-19 Richard Biener <rguenther@suse.de>
22953
22954 PR tree-optimization/56273
22955 * passes.c (init_optimization_passes): Move second VRP after DOM.
22956
22957 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22958
22959 * config/i386/i386.md (*movti_internal): Merge from
22960 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22961 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22962 nox64 isa attributes.
22963
22964 2013-03-18 Richard Biener <rguenther@suse.de>
22965
22966 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22967 (unite): Likewise.
22968 (merge_node_constraints): Likewise.
22969 (build_succ_graph): Likewise.
22970 (valid_graph_edge): Inline into single caller.
22971 (unify_nodes): Likewise. Use bitmap_set_bit return value
22972 and cache varinfo.
22973 (scc_visit): Fix formatting and variable use.
22974 (do_sd_constraint): Use gcc_checking_assert.
22975 (do_ds_constraint): Likewise.
22976 (do_complex_constraint): Likewise.
22977 (condense_visit): Likewise. Cleanup.
22978 (dump_pred_graph): New function.
22979 (perform_var_substitution): Dump the pred-graph before
22980 variable substitution.
22981 (find_equivalent_node): Use gcc_checking_assert.
22982 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22983
22984 2013-03-18 Richard Biener <rguenther@suse.de>
22985
22986 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22987 Remove cond_expr_stmt_list argument and do not gimplify the
22988 built expression.
22989 (vect_loop_versioning): Adjust.
22990 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22991 Cleanup to use less temporaries.
22992 (vect_create_data_ref_ptr): Cleanup.
22993
22994 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22995
22996 PR tree-optimization/56635
22997 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22998 require types_compatible_p types.
22999
23000 2013-03-18 Nick Clifton <nickc@redhat.com>
23001
23002 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23003 spurious backslash.
23004
23005 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23006 Add missing line to comment describing function.
23007
23008 2013-03-18 Richard Biener <rguenther@suse.de>
23009
23010 PR tree-optimization/56210
23011 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23012 Handle string / character search functions.
23013 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23014
23015 2013-03-18 Richard Biener <rguenther@suse.de>
23016
23017 PR middle-end/56483
23018 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23019 and implement properly.
23020 * gimple.h (gimple_cond_single_var_p): Remove.
23021
23022 2013-03-18 Richard Biener <rguenther@suse.de>
23023
23024 * tree-data-ref.h (find_data_references_in_loop): Declare.
23025 * tree-data-ref.c (get_references_in_stmt): Use a stack
23026 vector pre-allocated in the callers.
23027 (find_data_references_in_stmt): Adjust.
23028 (graphite_find_data_references_in_stmt): Likewise.
23029 (create_rdg_vertices): Likewise.
23030 (find_data_references_in_loop): Export.
23031 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23032 Compute dependences here...
23033 (vect_analyze_data_refs): ...not here. When we encounter
23034 a non-vectorizable data reference in basic-block vectorization
23035 truncate the data reference vector. Do not bother to
23036 fixup data-dependence information for gather loads.
23037 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23038 of data references, as reported.
23039
23040 2013-03-18 Richard Biener <rguenther@suse.de>
23041
23042 PR tree-optimization/3713
23043 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23044 has_constants and expr.
23045 (stmt_has_constants): Properly valueize SSA names when deciding
23046 whether the stmt has constants.
23047
23048 2013-03-18 Richard Biener <rguenther@suse.de>
23049
23050 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23051 whole function when there is nothing to do.
23052 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23053 * tree-vectorizer.c (vectorize_loops): Update virtual and
23054 loop-closed SSA once.
23055 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23056
23057 2013-03-18 Richard Biener <rguenther@suse.de>
23058
23059 PR middle-end/56113
23060 * domwalk.c (bb_postorder): New global static.
23061 (cmp_bb_postorder): New function.
23062 (walk_dominator_tree): Replace scheme imposing an order for
23063 visiting dominator sons by one sorting them at the time they
23064 are pushed on the stack.
23065
23066 2013-03-18 Richard Biener <rguenther@suse.de>
23067
23068 PR tree-optimization/39326
23069 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23070 (struct mem_ref): Replace mem member with ao_ref typed member.
23071 (MEM_ANALYZABLE): Adjust.
23072 (memref_eq): Likewise.
23073 (mem_ref_alloc): Likewise.
23074 (gather_mem_refs_stmt): Likewise.
23075 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23076 (execute_sm_if_changed_flag_set): Adjust.
23077 (execute_sm): Likewise.
23078 (ref_always_accessed_p): Likewise.
23079 (refs_independent_p): Likewise.
23080 (can_sm_ref_p): Likewise.
23081
23082 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23083
23084 PR c/56566
23085 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23086 return 1 even for !unsignedp.
23087
23088 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23089
23090 * config/i386/i386.md (isa): Add x64 and nox64.
23091 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23092 (*pushtf): Enable *roF alternative for x64 isa only.
23093 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23094 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23095 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23096 (*movtf_internal): Merge from *movtf_internal_rex64 and
23097 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23098 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23099 nox64 isa attributes.
23100 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23101 nox64 isa attributes.
23102 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23103
23104 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23105
23106 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23107
23108 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23109
23110 PR target/56640
23111 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23112 class names. Remove trailing comma after "ALL_REGS".
23113
23114 2013-03-16 Jan Hubicka <jh@suse.cz>
23115
23116 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23117 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23118 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23119 of cgraph_get_create_node.
23120 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23121
23122 2013-03-16 Jason Merrill <jason@redhat.com>
23123
23124 PR debug/49090
23125 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23126 with DW_AT_default_value.
23127
23128 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23129
23130 * BASE-VER: Set to 4.9.0.
23131
23132 2013-03-14 Andi Kleen <ak@linux.intel.com>
23133
23134 PR target/56619
23135 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23136 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23137 Document _x* TSX intrinsics.
23138
23139 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23140 David Holsgrove <david.holsgrove@xilinx.com>
23141
23142 * configure.ac: Add MicroBlaze TLS support detection.
23143 * configure: Regenerate.
23144 * config/microblaze/microblaze-protos.h
23145 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23146 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23147 * config/microblaze/microblaze.c (microblaze_address_type): Add
23148 ADDRESS_TLS and tls_reloc address types.
23149 (microblaze_address_info): Add tls_reloc.
23150 (TARGET_HAVE_TLS): Define.
23151 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23152 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23153 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23154 load_tls_operand, microblaze_call_tls_get_addr,
23155 microblaze_legitimize_tls_address): New functions.
23156 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23157 (get_base_reg): Use microblaze_tls_symbol_p.
23158 (microblaze_classify_address): Handle TLS.
23159 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23160 label_mentioned_p and microblaze_tls_referenced_p.
23161 (microblaze_legitimize_address): Handle TLS.
23162 (microblaze_address_insns): Handle ADDRESS_TLS.
23163 (pic_address_needs_scratch): Handle TLS.
23164 (print_operand_address): Handle TLS.
23165 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23166 (microblaze_expand_move): Handle TLS.
23167 (microblaze_legitimate_constant_p): Check
23168 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23169 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23170 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23171 (PIC_OFFSET_TABLE_REGNUM): Set.
23172 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23173 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23174 (addsi3, movsi_internal2, movdf_internal): Update constraints
23175 * config/microblaze/predicates.md (arith_plus_operand): Define
23176 (move_operand): Redefine as move_src_operand,
23177 check microblaze_tls_referenced_p.
23178
23179 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23180
23181 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23182 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23183
23184 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23185
23186 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23187 CC mode for AND.
23188
23189 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23190
23191 PR tree-optimization/53265
23192 * common.opt (Waggressive-loop-optimizations): New option.
23193 * tree-ssa-loop-niter.c: Include tree-pass.h.
23194 (do_warn_aggressive_loop_optimizations): New function.
23195 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23196 if number_of_latch_executions returned constant.
23197 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23198 early. If number_of_latch_executions returned constant, set
23199 nb_iterations_upper_bound back to it.
23200 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23201 field.
23202 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23203 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23204
23205 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23206 (MULTILIB_OSDIRNAMES): Set.
23207 * genmultilib: If defaultosdirname doesn't start with :: , set
23208 defaultosdirname2 instead, clear it and emit two . multilib_raw
23209 entries instead of just one.
23210
23211 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23212
23213 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23214 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23215 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23216 (SUBTARGET_OVERRIDE_OPTIONS): New.
23217
23218 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23219
23220 PR target/49880
23221 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23222 (musermode): Convert to Var(TARGET_USERMODE).
23223 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23224 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23225 * config/sh/sh.c (sh_option_override): Use
23226 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23227 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23228 condition.
23229 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23230 TARGET_SH4.
23231 (udivsi3_i4_single, divsi3_i4_single): Use
23232 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23233
23234 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23235
23236 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23237 default setting.
23238
23239 2013-03-13 Richard Biener <rguenther@suse.de>
23240
23241 PR tree-optimization/56608
23242 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23243 calls when vectorizing basic-blocks.
23244
23245 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23246
23247 PR plugins/45078
23248 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23249 tm_file.
23250
23251 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23252
23253 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23254
23255 2013-03-11 Jan Hubicka <jh@suse.cz>
23256
23257 PR lto/56557
23258 * lto-streamer-out.c (output_symbol_p): Skip references from
23259 constructors of external variables.
23260
23261 2013-03-11 Jan Hubicka <jh@suse.cz>
23262
23263 PR middle-end/56571
23264 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23265 from pseudos.
23266 * emit-rtl.c (verify_rtx_sharing): Likewise.
23267 (copy_insn_1): Likewise.
23268 * rtl.c (copy_rtx): Likewise.
23269
23270 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23271
23272 PR target/56591
23273 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23274 output_operand_lossage message.
23275
23276 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23277
23278 PR target/56470
23279 * arm.c (shift_op): Validate RTL pattern on the fly.
23280 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23281 the RTL.
23282
23283 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23284
23285 PR target/56347
23286 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23287 new call patterns that clobber %fr12.
23288 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23289 split and postreload patterns.
23290 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23291 registers %fr12 and %fr12R as call used.
23292
23293 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23294
23295 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23296 (canon_address, record_store, replace_read, check_mem_read_rtx,
23297 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23298 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23299 rest_of_handle_dse): Likewise.
23300
23301 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23302
23303 PR middle-end/56524
23304 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23305 Add base_optabs.
23306 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23307 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23308 (save_optabs_if_changed): Replace with...
23309 (init_tree_optimization_optabs): ...this.
23310 * optabs.c (save_optabs_if_changed): Rename to...
23311 (init_tree_optimization_optabs): ...this. Take the optimization node
23312 as argument. Do nothing if the base optabs are already correct.
23313 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23314 to recompute optabs.
23315 * function.h (function): Remove optabs field.
23316 * function.c (invoke_set_current_function_hook): Call
23317 init_tree_optimization_optabs. Use the result to initialize
23318 this_fn_optabs.
23319
23320 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23321
23322 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23323 if GTMA_HAS_NO_INSTRUMENTATION.
23324 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23325 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23326 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23327 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23328 GTMA_HAS_NO_INSTRUMENTATION.
23329
23330 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23331
23332 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23333 libasan_preinit.o.
23334
23335 2013-03-08 Marek Polacek <polacek@redhat.com>
23336 Jakub Jelinek <jakub@redhat.com>
23337
23338 PR tree-optimization/56478
23339 * predict.c (is_comparison_with_loop_invariant_p): Change the
23340 type of loop_step to tree.
23341 (predict_loops): Adjust.
23342 (predict_iv_comparison): Perform the computations on double_ints.
23343
23344 2013-03-08 Richard Biener <rguenther@suse.de>
23345
23346 PR tree-optimization/56570
23347 * tree-cfg.c (verify_expr_location_1): Verify locations for
23348 DECL_DEBUG_EXPR.
23349 * tree-sra.c (create_access_replacement): Strip locations
23350 from DECL_DEBUG_EXPRs.
23351
23352 2013-03-08 Richard Biener <rguenther@suse.de>
23353
23354 * tree-inline.c (expand_call_inline): Do not associate
23355 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23356 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23357
23358 2013-03-08 Richard Biener <rguenther@suse.de>
23359
23360 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23361 or block changes with -Og. Fix for location / block encoding
23362 changes and PHI arguments with locations.
23363
23364 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23365
23366 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23367 for all counters.
23368 (struct output_info): Likewise.
23369 (register_overhead): Remove bad gcc_assert.
23370 (bitmap_find_bit): If there is only a single bitmap element, do not
23371 count a miss as a search.
23372 (print_statistics): Update for counter type changes.
23373 (dump_bitmap_statistics): Likewise. Print headers such that they
23374 are properly lined up with the printed counters.
23375
23376 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23377
23378 PR tree-optimization/56559
23379 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23380 check that it has only a single use.
23381
23382 2013-03-07 Richard Biener <rguenther@suse.de>
23383
23384 * doc/invoke.texi (fwhole-program): Discourage use in combination
23385 with -flto.
23386
23387 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23388
23389 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23390
23391 PR tree-optimization/56539
23392 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23393 instead of GSI_CONTINUE_LINKING as last argument to
23394 force_gimple_operand_gsi. Adjust function comment.
23395
23396 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23397 aarch64-cores.def.
23398
23399 PR middle-end/56548
23400 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23401 promoted mode, convert the result back to the original mode.
23402
23403 2013-03-06 Richard Biener <rguenther@suse.de>
23404
23405 PR middle-end/56294
23406 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23407 (insert_updated_phi_nodes_compare_uids): New function.
23408 (update_ssa): Sort symbols_to_rename after UID before
23409 traversing it to insert PHI nodes.
23410
23411 2013-03-06 Richard Biener <rguenther@suse.de>
23412
23413 PR middle-end/50494
23414 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23415 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23416
23417 Revert
23418 2013-02-13 Richard Biener <rguenther@suse.de>
23419
23420 PR lto/50494
23421 * varasm.c (output_constant_def_1): Get the decl representing
23422 the constant as argument.
23423 (output_constant_def): Wrap output_constant_def_1.
23424 (make_decl_rtl): Use output_constant_def_1 with the decl
23425 representing the constant.
23426 (build_constant_desc): Optionally re-use a decl already
23427 representing the constant.
23428 (tree_output_constant_def): Adjust.
23429
23430 2013-03-06 Joey Ye <joey.ye@arm.com>
23431
23432 PR lto/50293
23433 * gcc.c (convert_white_space): New function.
23434 (main): Handles white space in function name.
23435
23436 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23437
23438 PR target/56529
23439 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23440 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23441 to SH_DIV_CALL_TABLE for TARGET_SH2.
23442 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23443 list.
23444 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23445 call-table options.
23446
23447 2013-03-05 Sterling Augustine <saugustine@google.com>
23448 Cary Coutant <ccoutant@google.com>
23449
23450 PR debug/55364
23451 * dwarf2out.c (resolve_addr): Don't call
23452 remove_loc_list_addr_table_entries a second time for the same
23453 expression.
23454
23455 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23456
23457 PR debug/56510
23458 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23459 (avoid_complex_debug_insns): New function.
23460 (expand_debug_locations): Call it.
23461
23462 PR rtl-optimization/56484
23463 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23464 lifetimes of hard registers on small register class machines.
23465
23466 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23467
23468 * config/microblaze/microblaze-protos.h: Rename
23469 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23470 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23471 fast_interrupt.
23472 (microblaze_fast_interrupt_function_p): New function.
23473 (microblaze_is_interrupt_handler): Rename to
23474 microblaze_is_interrupt_variant and add fast_interrupt check.
23475 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23476 (save_restore_insns): Likewise.
23477 (compute_frame_size): Likewise.
23478 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23479 (microblaze_globalize_label): Likewise.
23480 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23481 * config/microblaze/microblaze.md: Use wrapper
23482 microblaze_is_interrupt_variant.
23483
23484 2013-03-05 Kai Tietz <ktietz@redhat.com>
23485
23486 * sdbout.c (sdbout_one_type): Switch to current function's section
23487 supporting cold/hot.
23488
23489 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23490
23491 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23492 -mxl-reorder.
23493
23494 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23495
23496 PR middle-end/56461
23497 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23498 if VALGRIND_GET_VBITS is defined, temporarily make object
23499 memory all defined, and restore previous valgrind addressability
23500 and definability afterwards. Free this_object at the end.
23501
23502 PR middle-end/56461
23503 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23504 right before calling lra_create_live_ranges, also call it
23505 when clearing live_p. Only call lra_clear_live_ranges
23506 at the end if live_p.
23507
23508 PR middle-end/56461
23509 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23510
23511 2013-03-05 Richard Biener <rguenther@suse.de>
23512
23513 PR tree-optimization/56521
23514 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23515 value-id.
23516
23517 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23518
23519 PR c++/55135
23520 * except.h (remove_unreachable_eh_regions): New prototype.
23521 * except.c (remove_eh_handler_splicer): New function, split out
23522 of remove_eh_handler.
23523 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23524 warning about running it on many EH regions one at a time.
23525 (remove_unreachable_eh_regions_worker): New function, walk the
23526 EH tree in depth-first order and remove non-marked regions.
23527 (remove_unreachable_eh_regions): New function.
23528 * tree-eh.c (mark_reachable_handlers): New function, split out
23529 from remove_unreachable_handlers.
23530 (remove_unreachable_handlers): Use mark_reachable_handlers and
23531 remove_unreachable_eh_regions.
23532 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23533 and remove_unreachable_eh_regions.
23534
23535 2013-03-05 Richard Biener <rguenther@suse.de>
23536
23537 PR middle-end/56525
23538 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23539 not freeing them until the end.
23540
23541 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23542
23543 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23544
23545 2013-03-05 Richard Biener <rguenther@suse.de>
23546
23547 PR tree-optimization/56270
23548 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23549 of loads after scheduling an SLP instance.
23550
23551 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23552
23553 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23554 tic6x.exp.
23555 (check_gcc_parallelize): Run guality.exp as a separate job from
23556 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23557 struct-layout-1.exp with stackalign.exp.
23558
23559 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23560
23561 PR middle-end/56461
23562 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23563 load_index sbitmap even if some bit in it isn't set.
23564
23565 PR middle-end/56461
23566 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23567 (discover_iteration_bound_by_body_walk): Change queues to
23568 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23569 spelling in comment. Call safe_push on queues[bound_index] directly.
23570 Release queues[queue_index] in every iteration unconditionally.
23571 Release bounds vector.
23572
23573 PR middle-end/56461
23574 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23575 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23576 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23577 inner_phis vector.
23578
23579 2013-03-05 Richard Biener <rguenther@suse.de>
23580
23581 PR lto/56515
23582 * tree-inline.c (remap_blocks_to_null): New function.
23583 (expand_call_inline): When expanding a call stmt without
23584 an associated block inline remap all callee blocks to NULL.
23585
23586 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23587
23588 PR rtl-optimization/56494
23589 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23590 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23591 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23592
23593 PR middle-end/56461
23594 * sel-sched-ir.c (free_sched_pools): Release
23595 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23596 if succs_info_pool.max_top isn't -1.
23597
23598 PR bootstrap/56509
23599 * opts.c (opts_obstack, opts_concat): Moved to...
23600 * opts-common.c (opts_obstack, opts_concat): ... here.
23601
23602 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23603
23604 PR middle-end/56461
23605 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23606
23607 2013-03-04 Martin Jambor <mjambor@suse.cz>
23608
23609 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23610 all appropriate places.
23611
23612 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23613
23614 PR tree-optimization/56424
23615 * ipa-split.c (split_function): Do not set the RSO flag if result is
23616 not by reference and its type is a register type.
23617
23618 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23619
23620 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23621 (microblaze_legitimate_pic_operand): Likewise
23622 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23623 new function microblaze_legitimate_pic_operand
23624 * config/microblaze/microblaze-protos.h
23625 (microblaze_legitimate_pic_operand): Declare.
23626
23627 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23628
23629 * config/microblaze/predicates.md (call_insn_simple_operand):
23630 New predicate for supported rtx code types.
23631 * config/microblaze/microblaze.md (call_internal1): Use
23632 call_insn_simple_operand predicate.
23633
23634 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23635
23636 PR middle-end/56461
23637 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23638 partitions.ordered_remove.
23639
23640 PR middle-end/56461
23641 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23642 vec_oprnds0.create (1) for modifier == NONE.
23643
23644 PR middle-end/56461
23645 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23646 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23647 vec_oprnds1 right before pushing anything to it for
23648 scalar_shift_arg.
23649
23650 PR middle-end/56461
23651 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23652 set nbbs to 0 instead of having separate code path.
23653 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23654 instead of false as last argument if returning NULL.
23655
23656 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23657
23658 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23659 the attribute is now called "target" instead of "option".
23660 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23661 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23662 attribute/pragma name for TARGET_OPTION_VALID_P and
23663 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23664 * doc/tm.texi: Regenerated.
23665
23666 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23667
23668 * config/microblaze/microblaze.c:
23669 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23670 * config/microblaze/microblaze.h: Add -mxl-reorder to
23671 DRIVER_SELF_SPECS.
23672 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23673 instructions emitted if TARGET_REORDER.
23674 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23675 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23676 separately.
23677
23678 2013-03-01 Xinliang David Li <davidxl@google.com>
23679
23680 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23681 walk length.
23682
23683 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23684
23685 PR middle-end/56461
23686 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23687 vector even when returning true. Fix up function comment formatting.
23688
23689 PR middle-end/56461
23690 * ira-build.c (ira_loop_nodes_count): New variable.
23691 (create_loop_tree_nodes): Initialize it.
23692 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23693
23694 PR middle-end/56461
23695 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23696 method on dr_chain and result_chain.
23697 * tree-vect-stmts.c (vectorizable_store): Only call
23698 result_chain.create if j == 0.
23699
23700 PR middle-end/56461
23701 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23702 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23703 before overwriting it.
23704
23705 2013-03-01 Tobias Burnus <burnus@net-b.de>
23706
23707 * doc/extended.texi (C Extensions): Change order in @menu
23708 to match @node.
23709 (Other MIPS Built-in Functions): Move last MIPS entry before
23710 "picoChip Built-in Functions".
23711 (SH Built-in Functions): Move after RX Built-in Functions.
23712 * doc/gcc.texi (Introduction): Change order in @menu
23713 to match @node.
23714 * doc/md.texi (Constraints): Ditto.
23715 * gty.texi (Type Information): Ditto.
23716 (User-provided marking routines for template types): Make
23717 subsection.
23718 * doc/invoke.texi (AArch64 Options): Move before
23719 "Adapteva Epiphany Options".
23720
23721 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23722 Jakub Jelinek <jakub@redhat.com>
23723
23724 PR sanitizer/56454
23725 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23726 no_address_safety_analysis attribute.
23727 * doc/extend.texi (no_address_safety_attribute): Rename to
23728 no_sanitize_address attribute, mention no_address_safety_analysis
23729 attribute as deprecated alias.
23730
23731 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23732
23733 PR middle-end/56461
23734 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23735 type to vec<vec<tree> > *.
23736 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23737 to be vec<tree> instead of vec<tree> *, set vec_defs
23738 to vNULL and call vec_defs.create (number_of_vects), adjust other
23739 uses of vec_defs.
23740 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23741 vectorizable_condition): Adjust vect_get_slp_defs callers.
23742
23743 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23744
23745 * config/aarch64/aarch64.c
23746 (aarch64_float_const_representable): Remove unused variable.
23747
23748 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23749
23750 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23751
23752 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23753
23754 * config/aarch64/aarch64-builtins.c
23755 (aarch64_init_simd_builtins): Make static.
23756
23757 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23758
23759 * config/aarch64/aarch64.c
23760 (aarch64_simd_make_constant): Make static.
23761
23762 2013-02-28 Martin Jambor <mjambor@suse.cz>
23763
23764 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23765 with no initialization to the RHS of debug statements.
23766
23767 2013-02-28 Martin Jambor <mjambor@suse.cz>
23768
23769 PR tree-optimization/56294
23770 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23771 Adjust dumping.
23772 (get_access_replacement): Do not call create_access_replacement.
23773 Assert a replacement exists.
23774 (get_repl_default_def_ssa_name): Create the replacement declaration
23775 itself.
23776
23777 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23778
23779 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23780 final_end_function.
23781
23782 2013-02-28 Marek Polacek <polacek@redhat.com>
23783
23784 PR rtl-optimization/56466
23785 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23786 if we're changing a loop.
23787 (peel_loops_completely): Likewise.
23788
23789 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23790
23791 PR c++/55813
23792 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23793
23794 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23795
23796 PR target/56445
23797 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23798 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23799 INTX_FTYPE_FX, FX_FTYPE_INTX.
23800 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23801
23802 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23803
23804 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23805 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23806 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23807 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23808 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23809 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23810 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23811 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23812 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23813 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23814 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23815 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23816 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23817 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23818 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23819 (avrxmega6): Increase max flash segments from 5 to 6.
23820 * config/avr/t-multilib: Regenerate.
23821 * config/avr/avr-tables.opt: Regenerate.
23822 * doc/avr-mmcu.texi: Regenerate.
23823
23824 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23825
23826 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23827 (avr_device_to_arch): Rename to avr_device_to_ld.
23828 (avr_device_to_as): New prototype.
23829 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23830 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23831 * config/avr/driver-avr.c (avr_device_to_as): New.
23832 (avr_device_to_arch): Rename to avr_device_to_ld.
23833
23834 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23835
23836 PR middle-end/56461
23837 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23838 method on dr_chain and result_chain.
23839
23840 PR middle-end/56461
23841 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23842 pointer_set_destroy on not_executed_last_iteration.
23843
23844 PR middle-end/56461
23845 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23846
23847 PR middle-end/56461
23848 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23849 FOR_EACH_DEFINED_FUNCTION when freeing state.
23850
23851 PR middle-end/56461
23852 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23853 pool_free.
23854 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23855 overwriting it.
23856
23857 PR middle-end/56461
23858 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23859 known_aggs[i].items and release known_aggs vector.
23860
23861 PR middle-end/56461
23862 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23863
23864 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23865
23866 * config/microblaze/microblaze.c (microblaze_emit_compare):
23867 Use xor for EQ/NE comparisions.
23868 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23869 (cbranchsf4): Adjust operator to comparison_operator.
23870
23871 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23872
23873 PR middle-end/56461
23874 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23875 vector.
23876 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23877 vec_safe_push, always update *slot.
23878 (redirect_edge_var_map_clear): Use vec_free.
23879 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23880 (free_var_map_entry): Use vec_free.
23881 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23882 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23883
23884 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23885
23886 PR middle-end/45472
23887 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23888 when the may_trap_p bit of the exprs being merged differs.
23889 Reorder tests for speculativeness in the logical and operator.
23890
23891 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23892
23893 * incpath.c (add_standard_paths): Use reconcat instead of concat
23894 where appropriate and avoid leaking memory.
23895
23896 * opts.h: Include obstack.h.
23897 (opts_concat): New prototype.
23898 (opts_obstack): New declaration.
23899 * opts.c (opts_concat): New function.
23900 (opts_obstack): New variable.
23901 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23902 (finish_options): Use opts_concat instead of concat
23903 and XOBNEWVEC instead of XNEWVEC.
23904 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23905 generate_option): Likewise.
23906 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23907 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23908
23909 PR target/56455
23910 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23911 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23912
23913 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23914
23915 PR middle-end/56461
23916 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23917
23918 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23919
23920 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23921 (arm_block_move_unaligned_straight): Likewise.
23922 (arm_adjust_block_mem): Likewise.
23923
23924 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23925
23926 PR target/48901
23927 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23928 temp, cond and label.
23929 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23930
23931 PR target/52500
23932 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23933 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23934
23935 PR target/52501
23936 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23937 of prologue/epilogue functions.
23938
23939 PR target/52550
23940 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23941 Remove unused variable cfa_offset.
23942 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23943
23944 PR target/54639
23945 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23946 type promotion to unsigned.
23947
23948 PR target/54640
23949 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23950 for HOST_WIDE_INT of 32 bit / same size as int.
23951 (arm_block_move_unaligned_straight): Likewise.
23952 (arm_adjust_block_mem): Likewise.
23953
23954 PR target/54662
23955 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23956 ALL_CFLAGS.
23957
23958 2013-02-26 Marek Polacek <polacek@redhat.com>
23959
23960 PR tree-optimization/56426
23961 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23962
23963 2013-02-26 Richard Biener <rguenther@suse.de>
23964
23965 PR target/56444
23966 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23967 unused variable loops.
23968
23969 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23970
23971 PR tree-optimization/56448
23972 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23973 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23974 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23975 later operands of the references, or even first operand for
23976 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23977
23978 PR tree-optimization/56443
23979 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23980 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23981 to type_for_mode langhook.
23982
23983 2013-02-25 Matt Turner <mattst88@gmail.com>
23984
23985 * doc/invoke.texi: Document r4700.
23986
23987 2013-02-25 Richard Biener <rguenther@suse.de>
23988
23989 PR tree-optimization/56175
23990 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23991 split out from ...
23992 (simplify_bitwise_binary): ... here. Also guard the conversion
23993 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23994
23995 2013-02-25 Catherine Moore <clm@codesourcery.com>
23996
23997 Revert:
23998 2013-02-24 Catherine Moore <clm@codesourcery.com>
23999 Maciej W. Rozycki <macro@codesourcery.com>
24000 Tom de Vries <tom@codesourcery.com>
24001 Nathan Sidwell <nathan@codesourcery.com>
24002 Iain Sandoe <iain@codesourcery.com>
24003 Nathan Froyd <froydnj@codesourcery.com>
24004 Chao-ying Fu <fu@mips.com>
24005
24006 * doc/extend.texi: (micromips, nomicromips, nocompression):
24007 Document new function attributes.
24008 * doc/invoke.texi (minterlink-compressed, mmicromips,
24009 m14k, m14ke, m14kec): Document new options.
24010 (minterlink-mips16): Update documentation.
24011 * doc/md.texi (ZC, ZD): Document new constraints.
24012 * configure.ac (gcc_cv_as_micromips): Check if linker
24013 supports the .set micromips directive.
24014 * configure: Regenerate.
24015 * config.in: Regenerate.
24016 * config/mips/mips-tables.opt: Regenerate.
24017 * config/mips/micromips.md: New file.
24018 * constraints.md (ZC, AD): New constraints.
24019 * config/mips/predicates.md (movep_src_register): New predicate.
24020 (movep_src_operand): New predicate.
24021 (non_volatile_mem_operand): New predicate.
24022 * config/mips/mips.md (multimem): New type.
24023 (length): Differentiate between 17-bit and 18-bit branch offsets.
24024 (MOVEP1, MOVEP2): New mode iterator.
24025 (mov_<load>l): Use ZC constraint.
24026 (mov_<load>r): Likewise.
24027 (mov_<store>l): Likewise.
24028 (mov_<store>r): Likewise.
24029 (*branch_equality<mode>_inverted): Add microMIPS support.
24030 (*branch_equality<mode>): Likewise.
24031 (*jump_absolute): Likewise.
24032 (indirect_jump_<mode>): Likewise.
24033 (tablejump_<mode>): Likewise.
24034 (<optab>_internal): Likewise.
24035 (sibcall_internal): Likewise.
24036 (sibcall_value_internal): Likewise.
24037 (prefetch): Use constraint ZD.
24038 * config/mips/mips.opt (minterlink-compressed): New option.
24039 (minterlink-mips16): Now an alias for minterlink-compressed.
24040 (mmicromips): New option.
24041 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24042 (compare_and_swap_12): Likewise.
24043 (sync_add<mode>): Likewise.
24044 (sync_<optab>_12): Likewise.
24045 (sync_old_<optab>_12): Likewise.
24046 (sync_new_<optab>_12): Likewise.
24047 (sync_nand_12): Likewise.
24048 (sync_old_nand_12): Likewise.
24049 (sync_new_nand_12): Likewise.
24050 (sync_sub<mode>): Likewise.
24051 (sync_old_add<mode>): Likewise.
24052 (sync_old_sub<mode>): Likewise.
24053 (sync_new_add<mode>): Likewise.
24054 (sync_new_sub<mode>): Likewise.
24055 (sync_<optab><mode>): Likewise.
24056 (sync_old_<optab><mode>): Likewise.
24057 (sync_new_<optab><mode>): Likewise.
24058 (sync_nand<mode>): Likewise.
24059 (sync_old_nand<mode>): Likewise.
24060 (sync_new_nand<mode>): Likewise.
24061 (sync_lock_test_and_set<mode>): Likewise.
24062 (test_and_set_12): Likewise.
24063 (atomic_compare_and_swap<mode>): Likewise.
24064 (atomic_exchange<mode>_llsc): Likewise.
24065 (atomic_fetch_add<mode>_llsc): Likewise.
24066 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24067 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24068 (umips_save_restore_pattern_p): Likewise.
24069 (umips_load_store_pair_p): Likewise.
24070 (umips_output_load_store_pair): Likewise.
24071 (umips_movep_target_p): Likewise.
24072 (umips_12bit_offset_address_p): Likewise.
24073 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24074 (mips_base_mips16): Rename this...
24075 (mips_base_compression_flags): ...to this. Update all uses.
24076 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24077 (mips_mips16_decl_p): Delete.
24078 (mips_nomips16_decl_p): Delete.
24079 (mips_get_compress_on_flags): New function.
24080 (mips_get_compress_off_flags): New function.
24081 (mips_get_compress_mode): New function.
24082 (mips_get_compress_on_name): New function.
24083 (mips_get_compress_off_name): New function.
24084 (mips_insert_attributes): Support multiple compression types.
24085 (mips_merge_decl_attributes): Likewise.
24086 (umips_12bit_offset_address_p): New function.
24087 (mips_start_function_definition): Emit .set micromips directive.
24088 (mips_call_may_need_jalx_p): New function.
24089 (mips_function_ok_for_sibcall): Add microMIPS support.
24090 (mips_print_operand_punctuation): Support short delay slots and
24091 compact jumps.
24092 (umips_swm_mask, umips_swm_encoding): New.
24093 (umips_build_save_restore): New function.
24094 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24095 (was_mips16_p): Remove.
24096 (old_compression_mode): New.
24097 (mips_set_compression_mode): New function.
24098 (mips_set_current_function): Add microMIPS support.
24099 (mips_option_override): Likewise.
24100 (umips_save_restore_pattern_p): New function.
24101 (umips_output_save_restore): New function.
24102 (umips_load_store_pair_p_1): New function.
24103 (umips_load_store_pair_p): New function.
24104 (umips_output_load_store_pair_1): New function.
24105 (umips_output_load_store_pair): New function.
24106 (umips_movep_target_p) New function.
24107 (mips_prepare_pch_save): Add microMIPS support.
24108 * config/mips/mips.h (TARGET_COMPRESSION): New.
24109 (TARGET_CPU_CPP_BUILTINS): Update macro
24110 to use new compression flags and to support microMIPS.
24111 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24112 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24113 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24114 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24115 (ASM_SPEC): Support mmicromips and mno-micromips.
24116 (M16STORE_REG_P): New macro.
24117 (MIPS_CALL): Support TARGET_MICROMIPS.
24118 (MICROMIPS_J): New macro.
24119 (mips_base_mips16): Rename this...
24120 (mips_base_compression_flags): ...to this.
24121 (UMIPS_12BIT_OFFSET_P): New macro.
24122 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24123 (MULTILIB_DIRNAMES): Likewise.
24124
24125 2013-02-25 Tom de Vries <tom@codesourcery.com>
24126
24127 PR rtl-optimization/56131
24128 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24129 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24130 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24131
24132 2013-02-25 Tobias Burnus <burnus@net-b.de>
24133
24134 * doc/invoke.texi (-fsanitize=): Move from optimization
24135 to debugging options.
24136
24137 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24138
24139 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24140
24141 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24142 Alexander Monakov <amonakov@ispras.ru>
24143
24144 PR middle-end/56077
24145 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24146 flush pending lists also on non-jumps. Adjust comment.
24147
24148 2013-02-24 Catherine Moore <clm@codesourcery.com>
24149 Maciej W. Rozycki <macro@codesourcery.com>
24150 Tom de Vries <tom@codesourcery.com>
24151 Nathan Sidwell <nathan@codesourcery.com>
24152 Iain Sandoe <iain@codesourcery.com>
24153 Nathan Froyd <froydnj@codesourcery.com>
24154 Chao-ying Fu <fu@mips.com>
24155
24156 * doc/extend.texi: (micromips, nomicromips, nocompression):
24157 Document new function attributes.
24158 * doc/invoke.texi (minterlink-compressed, mmicromips,
24159 m14k, m14ke, m14kec): Document new options.
24160 (minterlink-mips16): Update documentation.
24161 * doc/md.texi (ZC, ZD): Document new constraints.
24162 * configure.ac (gcc_cv_as_micromips): Check if linker
24163 supports the .set micromips directive.
24164 * configure: Regenerate.
24165 * config.in: Regenerate.
24166 * config/mips/mips-tables.opt: Regenerate.
24167 * config/mips/micromips.md: New file.
24168 * constraints.md (ZC, AD): New constraints.
24169 * config/mips/predicates.md (movep_src_register): New predicate.
24170 (movep_src_operand): New predicate.
24171 (non_volatile_mem_operand): New predicate.
24172 * config/mips/mips.md (multimem): New type.
24173 (length): Differentiate between 17-bit and 18-bit branch offsets.
24174 (MOVEP1, MOVEP2): New mode iterator.
24175 (mov_<load>l): Use ZC constraint.
24176 (mov_<load>r): Likewise.
24177 (mov_<store>l): Likewise.
24178 (mov_<store>r): Likewise.
24179 (*branch_equality<mode>_inverted): Add microMIPS support.
24180 (*branch_equality<mode>): Likewise.
24181 (*jump_absolute): Likewise.
24182 (indirect_jump_<mode>): Likewise.
24183 (tablejump_<mode>): Likewise.
24184 (<optab>_internal): Likewise.
24185 (sibcall_internal): Likewise.
24186 (sibcall_value_internal): Likewise.
24187 (prefetch): Use constraint ZD.
24188 * config/mips/mips.opt (minterlink-compressed): New option.
24189 (minterlink-mips16): Now an alias for minterlink-compressed.
24190 (mmicromips): New option.
24191 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24192 (compare_and_swap_12): Likewise.
24193 (sync_add<mode>): Likewise.
24194 (sync_<optab>_12): Likewise.
24195 (sync_old_<optab>_12): Likewise.
24196 (sync_new_<optab>_12): Likewise.
24197 (sync_nand_12): Likewise.
24198 (sync_old_nand_12): Likewise.
24199 (sync_new_nand_12): Likewise.
24200 (sync_sub<mode>): Likewise.
24201 (sync_old_add<mode>): Likewise.
24202 (sync_old_sub<mode>): Likewise.
24203 (sync_new_add<mode>): Likewise.
24204 (sync_new_sub<mode>): Likewise.
24205 (sync_<optab><mode>): Likewise.
24206 (sync_old_<optab><mode>): Likewise.
24207 (sync_new_<optab><mode>): Likewise.
24208 (sync_nand<mode>): Likewise.
24209 (sync_old_nand<mode>): Likewise.
24210 (sync_new_nand<mode>): Likewise.
24211 (sync_lock_test_and_set<mode>): Likewise.
24212 (test_and_set_12): Likewise.
24213 (atomic_compare_and_swap<mode>): Likewise.
24214 (atomic_exchange<mode>_llsc): Likewise.
24215 (atomic_fetch_add<mode>_llsc): Likewise.
24216 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24217 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24218 (umips_save_restore_pattern_p): Likewise.
24219 (umips_load_store_pair_p): Likewise.
24220 (umips_output_load_store_pair): Likewise.
24221 (umips_movep_target_p): Likewise.
24222 (umips_12bit_offset_address_p): Likewise.
24223 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24224 (mips_base_mips16): Rename this...
24225 (mips_base_compression_flags): ...to this. Update all uses.
24226 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24227 (mips_mips16_decl_p): Delete.
24228 (mips_nomips16_decl_p): Delete.
24229 (mips_get_compress_on_flags): New function.
24230 (mips_get_compress_off_flags): New function.
24231 (mips_get_compress_mode): New function.
24232 (mips_get_compress_on_name): New function.
24233 (mips_get_compress_off_name): New function.
24234 (mips_insert_attributes): Support multiple compression types.
24235 (mips_merge_decl_attributes): Likewise.
24236 (umips_12bit_offset_address_p): New function.
24237 (mips_start_function_definition): Emit .set micromips directive.
24238 (mips_call_may_need_jalx_p): New function.
24239 (mips_function_ok_for_sibcall): Add microMIPS support.
24240 (mips_print_operand_punctuation): Support short delay slots and
24241 compact jumps.
24242 (umips_swm_mask, umips_swm_encoding): New.
24243 (umips_build_save_restore): New function.
24244 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24245 (was_mips16_p): Remove.
24246 (old_compression_mode): New.
24247 (mips_set_compression_mode): New function.
24248 (mips_set_current_function): Add microMIPS support.
24249 (mips_option_override): Likewise.
24250 (umips_save_restore_pattern_p): New function.
24251 (umips_output_save_restore): New function.
24252 (umips_load_store_pair_p_1): New function.
24253 (umips_load_store_pair_p): New function.
24254 (umips_output_load_store_pair_1): New function.
24255 (umips_output_load_store_pair): New function.
24256 (umips_movep_target_p) New function.
24257 (mips_prepare_pch_save): Add microMIPS support.
24258 * config/mips/mips.h (TARGET_COMPRESSION): New.
24259 (TARGET_CPU_CPP_BUILTINS): Update macro
24260 to use new compression flags and to support microMIPS.
24261 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24262 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24263 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24264 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24265 (ASM_SPEC): Support mmicromips and mno-micromips.
24266 (M16STORE_REG_P): New macro.
24267 (MIPS_CALL): Support TARGET_MICROMIPS.
24268 (MICROMIPS_J): New macro.
24269 (mips_base_mips16): Rename this...
24270 (mips_base_compression_flags): ...to this.
24271 (UMIPS_12BIT_OFFSET_P): New macro.
24272 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24273 (MULTILIB_DIRNAMES): Likewise.
24274
24275 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24276
24277 PR target/52555
24278 * target-globals.c (save_target_globals): For init_reg_sets and
24279 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24280
24281 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24282
24283 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24284 * config/aarch64/t-aarch64
24285 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24286
24287 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24288
24289 PR inline-asm/56148
24290 * lra-constraints.c (process_alt_operands): Reload operand
24291 conflicting with earlier clobber only if no more other conflicting
24292 operands.
24293
24294 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24295
24296 PR sanitizer/56393
24297 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24298 if not linking a shared library.
24299
24300 2013-02-22 Seth LaForge <sethml@google.com>
24301
24302 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24303
24304 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24305
24306 * config/arm/arm.md (split for extendsidi): Update condition.
24307 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24308 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24309 (qhs_zextenddi_cstr): Likewise.
24310
24311 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24312
24313 PR middle-end/56420
24314 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24315 avoid signed wrapping.
24316 (expand_mult): Handle properly multiplication by
24317 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24318 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24319 in the compiler if coeff is HOST_WIDE_INT_MIN.
24320 (expand_divmod): Don't make ext_op1 static, change it's type to
24321 uhwi. Avoid undefined behavior in -INTVAL (op1).
24322
24323 PR rtl-optimization/50339
24324 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24325 field.
24326 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24327 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24328 into splitting_ashiftrt field.
24329 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24330 ASHIFTRT.
24331 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24332 choices.
24333
24334 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24335
24336 PR middle-end/56108
24337 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24338 are sure to go irrevocable.
24339
24340 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24341
24342 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24343 scalars are valid operands.
24344
24345 2013-02-21 Martin Jambor <mjambor@suse.cz>
24346
24347 PR tree-optimization/56310
24348 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24349 only matching indices and non-negative final offsets.
24350 (intersect_aggregates_with_edge): Pass src_idx to
24351 agg_replacements_to_vector. Pass src_idx insstead of index to
24352 intersect_with_agg_replacements.
24353
24354 2013-02-21 Martin Jambor <mjambor@suse.cz>
24355
24356 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24357 instead of hard-wired defaults.
24358
24359 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24360
24361 * doc/invoke.texi (MIPS Options): Update documentation of the
24362 floating-point multiply-accumulate instruction restrictions.
24363
24364 2013-02-21 Kostya Serebryany <kcc@google.com>
24365
24366 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24367 asan_shadow_offset on x86_64 linux.
24368
24369 2013-02-21 Richard Biener <rguenther@suse.de>
24370
24371 PR tree-optimization/56415
24372 Revert
24373 2013-02-11 Richard Biener <rguenther@suse.de>
24374
24375 PR tree-optimization/56273
24376 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24377 first VRP run.
24378
24379 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24380
24381 PR bootstrap/56258
24382 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24383 instead of @itemx.
24384
24385 PR inline-asm/56405
24386 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24387 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24388
24389 2013-02-20 Jan Hubicka <jh@suse.cz>
24390
24391 PR tree-optimization/56265
24392 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24393 when target is referenced for first time.
24394
24395 2013-02-20 Richard Biener <rguenther@suse.de>
24396
24397 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24398 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24399 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24400 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24401 not return anything.
24402 (rename_ssa_copies): Do not remove unused locals.
24403 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24404 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24405 * passes.c (execute_function_todo): Do not schedule unused locals
24406 removal if cleanup_tree_cfg did something.
24407 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24408 about the number of removed locals.
24409
24410 2013-02-20 Richard Biener <rguenther@suse.de>
24411
24412 PR tree-optimization/56398
24413 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24414
24415 2013-02-20 Martin Jambor <mjambor@suse.cz>
24416
24417 PR tree-optimization/55334
24418 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24419 restricted pointers to arrays.
24420
24421 2013-02-20 Richard Biener <rguenther@suse.de>
24422 Jakub Jelinek <jakub@redhat.com>
24423
24424 PR tree-optimization/56396
24425 * tree-ssa-ccp.c (n_const_val): New static variable.
24426 (get_value): Return NULL for SSA names we don't have a lattice
24427 entry for.
24428 (ccp_initialize): Initialize n_const_val.
24429 * tree-ssa-copy.c (n_copy_of): New static variable.
24430 (init_copy_prop): Initialize n_copy_of.
24431 (get_value): Return NULL_TREE for SSA names we don't have a
24432 lattice entry for.
24433
24434 2013-02-20 Martin Jambor <mjambor@suse.cz>
24435
24436 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24437
24438 2013-02-20 Richard Biener <rguenther@suse.de>
24439
24440 * genpreds.c (write_lookup_constraint): Do not compare first
24441 letter of the constraint again.
24442
24443 2013-02-20 Richard Biener <rguenther@suse.de>
24444
24445 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24446 and ceil_log2.
24447 (get_use_iv_cost): Terminate hashtable walk when coming across
24448 an empty entry.
24449
24450 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24451
24452 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24453 reassociation for avx2 targets.
24454
24455 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24456
24457 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24458 Add version check for v8.10.a to enable microblaze_has_clz
24459 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24460 version and TARGET_PATTERN_COMPARE check
24461 * config/microblaze/microblaze.md: New clzsi2 instruction
24462
24463 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24464
24465 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24466 function before branching.
24467
24468 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24469
24470 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24471 DUMP_INSN_RTX_UID.
24472 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24473
24474 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24475
24476 PR middle-end/55889
24477 * sel-sched.c: Include ira.h.
24478 (implicit_clobber_conflict_p): New function.
24479 (moveup_expr): Use it.
24480 * Makefile.in (sel-sched.o): Depend on ira.h.
24481
24482 2013-02-19 Richard Biener <rguenther@suse.de>
24483
24484 PR tree-optimization/56384
24485 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24486 (vn_hash_type): Split out from ...
24487 (vn_hash_constant_with_type): ... here.
24488 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24489 (vn_phi_eq): Compare types from vn_phi_s structure.
24490 (vn_phi_lookup): Populate vn_phi_s type.
24491 (vn_phi_insert): Likewise.
24492
24493 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24494
24495 PR tree-optimization/56350
24496 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24497 if haven't found reduction or nested cycle operand, rather than
24498 asserting we must find it.
24499
24500 PR tree-optimization/56381
24501 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24502 to fold_build3.
24503
24504 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24505 Jakub Jelinek <jakub@redhat.com>
24506
24507 PR target/52555
24508 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24509 (swap_optab_enable): Same.
24510 (init_all_optabs): Use argument instead of global.
24511 * tree.h (struct tree_optimization_option): New field target_optabs.
24512 * expr.h (init_all_optabs): Add argument to prototype.
24513 (TREE_OPTIMIZATION_OPTABS): New.
24514 (save_optabs_if_changed): Protoize.
24515 * optabs.h: Declare this_fn_optabs.
24516 * optabs.c (save_optabs_if_changed): New.
24517 Declare this_fn_optabs.
24518 (init_optabs): Add argument to init_all_optabs() call.
24519 * function.c (invoke_set_current_function_hook): Handle per
24520 function optabs.
24521 * function.h (struct function): New field optabs.
24522 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24523 optimization_current_node has changed.
24524 * target-globals.h (save_target_globals_default_opts): Protoize.
24525 * target-globals.c (save_target_globals_default_opts): New.
24526
24527 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24528
24529 PR target/56347
24530 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24531 registers %fr12 and %fr12R as call used.
24532
24533 PR target/56214
24534 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24535 and HImode, require all displacements to be an integer multiple of
24536 their mode size.
24537 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24538 only allow QImode and HImode when reload is in progress and strict is
24539 true. Likewise for symbolic addresses. Use base14_operand to check
24540 displacements in REG+BASE addresses.
24541
24542 2013-02-18 Richard Biener <rguenther@suse.de>
24543
24544 PR tree-optimization/56366
24545 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24546 handle sign-conversion of outer-loop initial induction value.
24547
24548 2013-02-18 Richard Biener <rguenther@suse.de>
24549
24550 PR middle-end/56349
24551 * cfghooks.c (merge_blocks): If we merge a latch into another
24552 block adjust references to it.
24553 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24554 (verify_loop_structure): Verify that a recorded latch is in fact
24555 a latch.
24556
24557 2013-02-18 Richard Biener <rguenther@suse.de>
24558
24559 PR tree-optimization/56321
24560 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24561 order SSA name release and virtual operand unlinking.
24562
24563 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24564
24565 * config/microblaze/microblaze.md (save_stack_block): Define.
24566 (restore_stack_block): Likewise.
24567
24568 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24569
24570 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24571 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24572 * config/microblaze/microblaze.c (microblaze_option_override):
24573 Bail out early for PIC modes when target does not support PIC.
24574
24575 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24576
24577 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24578 Replace with a microblaze version.
24579 (microblaze_trampoline_init): Adapt for microblaze.
24580 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24581 microblaze.
24582
24583 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24584 Dodji Seketeli <dodji@redhat.com>
24585
24586 PR asan/56330
24587 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24588 (instrument_mem_region_access): Do not forget to always put
24589 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24590 0) statement, even for cases where either 'base' or 'base + len'
24591 are not instrumented -- because they have been previously
24592 instrumented. Simplify the logic by putting all the statements
24593 instrument 'base + len' inside a sequence, and then insert that
24594 sequence right before the current insertion point. Then, to
24595 instrument 'base + len', just get an iterator on that statement.
24596 And do not forget to update the pointer to iterator the function
24597 received as argument.
24598
24599 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24600
24601 PR rtl-optimization/56348
24602 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24603
24604 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24605
24606 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24607 (clean_graph_dump_file): Pass base to start_graph_dump.
24608
24609 2013-02-14 Richard Henderson <rth@redhat.com>
24610
24611 PR target/55941
24612 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24613
24614 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24615
24616 * collect2-aix.h: Define F_LOADONLY.
24617
24618 2013-02-14 Richard Biener <rguenther@suse.de>
24619
24620 PR lto/50494
24621 * varasm.c (output_constant_def_1): Get the decl representing
24622 the constant as argument.
24623 (output_constant_def): Wrap output_constant_def_1.
24624 (make_decl_rtl): Use output_constant_def_1 with the decl
24625 representing the constant.
24626 (build_constant_desc): Optionally re-use a decl already
24627 representing the constant.
24628 (tree_output_constant_def): Adjust.
24629
24630 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24631
24632 Fix an asan crash
24633 * asan.c (instrument_builtin_call): Really put the length of the
24634 second source argument into src1_len.
24635
24636 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24637
24638 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24639 argument. If it is false, don't create edge from then_bb to
24640 fallthru_bb.
24641 (insert_if_then_before_iter): Pass true to it.
24642 (build_check_stmt): Pass false to it.
24643 (transform_statements): Flush hash table only on extended basic
24644 block boundaries, rather than at the beginning of every bb.
24645 Don't flush hash table on nonfreeing_call_p calls.
24646 * tree-flow.h (nonfreeing_call_p): New prototype.
24647 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24648
24649 2013-02-13 David S. Miller <davem@davemloft.net>
24650
24651 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24652
24653 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24654
24655 PR target/56184
24656 * ira.c (max_regno_before_ira): Move from ...
24657 (ira): ... here.
24658 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24659 vec_safe_length.
24660
24661 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24662
24663 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24664
24665 2013-02-13 Richard Biener <rguenther@suse.de>
24666
24667 PR lto/56295
24668 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24669 globals in MEM_REFs.
24670
24671 2013-02-13 Richard Biener <rguenther@suse.de>
24672
24673 * loop-init.c (loop_optimizer_init): Clear loop state when
24674 re-initializing preserved loops.
24675 * loop-unswitch.c (unswitch_single_loop): Return whether
24676 we unswitched the loop. Do not verify loop state here.
24677 (unswitch_loops): When we unswitched a loop discover new loops.
24678
24679 2013-02-13 Kostya Serebryany <kcc@google.com>
24680
24681 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24682 on x86_64 linux.
24683 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24684
24685 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24686
24687 Avoid instrumenting duplicated memory access in the same basic block
24688 * Makefile.in (asan.o): Add new dependency on hash-table.h
24689 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24690 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24691 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24692 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24693 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24694 (get_mem_ref_of_assignment): New functions.
24695 (get_mem_refs_of_builtin_call): Extract from
24696 instrument_builtin_call and tweak a little bit to make it fit with
24697 the new signature.
24698 (instrument_builtin_call): Use the new
24699 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24700 of is_gimple_builtin_call.
24701 (instrument_derefs, instrument_mem_region_access): Insert the
24702 instrumented memory reference into the hash table.
24703 (maybe_instrument_assignment): Renamed instrument_assignment into
24704 this, and change it to advance the iterator when instrumentation
24705 actually happened and return true in that case. This makes it
24706 homogeneous with maybe_instrument_assignment, and thus give a
24707 chance to callers to be more 'regular'.
24708 (transform_statements): Clear the memory reference hash table
24709 whenever we enter a new BB, when we cross a function call, or when
24710 we are done transforming statements. Use
24711 maybe_instrument_assignment instead of instrumentation. No more
24712 need to special case maybe_instrument_assignment and advance the
24713 iterator after calling it; it's now handled just like
24714 maybe_instrument_call. Update comment.
24715
24716 2013-02-13 Richard Biener <rguenther@suse.de>
24717
24718 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24719 Fix loop discovery code.
24720
24721 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24722
24723 PR inline-asm/56148
24724 * lra-constraints.c (process_alt_operands): Match early clobber
24725 operand with itself. Check conflicts with earlyclobber only if
24726 the operand is not reloaded. Prefer to reload conflicting operand
24727 if earlyclobber and matching operands are the same.
24728
24729 2013-02-12 Richard Biener <rguenther@suse.de>
24730
24731 PR lto/56297
24732 * lto-streamer-out.c (write_symbol): Do not output symbols
24733 for hard register variables.
24734
24735 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24736
24737 PR target/54222
24738 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24739 (umulsidi3_insn, mulsidi3_insn): New insns.
24740
24741 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24742
24743 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24744 (struct tune_params): Add vec_costs field.
24745 * config/arm/arm.c (arm_builtin_vectorization_cost)
24746 (arm_add_stmt_cost): New functions.
24747 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24748 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24749 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24750 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24751 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24752 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24753 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24754
24755 2013-02-12 Richard Biener <rguenther@suse.de>
24756
24757 PR lto/56295
24758 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24759 decls again if possible.
24760
24761 2013-02-12 Richard Biener <rguenther@suse.de>
24762
24763 PR middle-end/56288
24764 * tree-ssa.c (verify_ssa_name): Fix check, move
24765 SSA_NAME_IN_FREE_LIST check up.
24766
24767 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24768 Steven Bosscher <steven@gcc.gnu.org>
24769
24770 PR rtl-optimization/56151
24771 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24772 equal to op0 or op1, and last_insn pattern is CODE operation
24773 with MEM dest and one of the operands matches that MEM.
24774
24775 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24776
24777 * doc/extend.texi: Document Function Multiversioning and "default"
24778 parameter string to target attribute.
24779 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24780 target attribute parameter is "default".
24781 (ix86_compare_version_priority): Remove checks for target attribute.
24782 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24783 Remove check for target attribute equal to NULL. Add assert.
24784 (ix86_generate_version_dispatcher_body): Change error to sorry.
24785
24786 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24787 Jack Howarth <howarth@bromo.med.uc.edu>
24788 Patrick Marlier <patrick.marlier@gmail.com>
24789
24790 PR libitm/55693
24791 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24792 define ENDFILE_SPEC as TM_DESTRUCTOR.
24793 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24794
24795 2013-02-11 Alexander Potapenko <glider@google.com>
24796 Jack Howarth <howarth@bromo.med.uc.edu>
24797 Jakub Jelinek <jakub@redhat.com>
24798
24799 PR sanitizer/55617
24800 * config/darwin.c (cdtor_record): Rename ctor_record.
24801 (sort_cdtor_records): Rename sort_ctor_records.
24802 (finalize_dtors): New routine to sort destructors by
24803 priority before use in assemble_integer.
24804 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24805
24806 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24807
24808 PR rtl-optimization/56275
24809 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24810 offset is non-negative and less than cmode size before
24811 calling simplify_subreg.
24812
24813 2013-02-11 Richard Biener <rguenther@suse.de>
24814
24815 PR tree-optimization/56264
24816 * cfgloop.h (fix_loop_structure): Adjust prototype.
24817 * loop-init.c (fix_loop_structure): Return the number of
24818 newly discovered loops.
24819 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24820 are discovered, do a full loop-closed SSA rewrite.
24821
24822 2013-02-11 Richard Biener <rguenther@suse.de>
24823
24824 PR tree-optimization/56273
24825 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24826 first VRP run.
24827 (check_array_ref): Fix missing newline in dumps.
24828 (search_for_addr_array): Likewise.
24829
24830 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24831
24832 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24833
24834 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24835
24836 PR target/56256
24837 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24838
24839 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24840
24841 PR rtl-optimization/56246
24842 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24843 reload pseudo.
24844 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24845 constraints are satisfied.
24846
24847 2013-02-08 Jeff Law <law@redhat.com>
24848
24849 PR debug/53948
24850 * emit-rtl.c (reg_is_parm_p): New function.
24851 * regs.h (reg_is_parm_p): New prototype.
24852 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24853 callee-clobbered registers.
24854
24855 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24856
24857 PR target/56043
24858 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24859 If there is no implicit builtin declaration, just return NULL.
24860
24861 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24862
24863 * config/i386/sse.md (FMAMODEM): New mode iterator.
24864 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24865 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24866
24867 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24868
24869 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24870 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24871 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24872
24873 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24874
24875 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24876 (microblaze*-*-elf): Likewise.
24877 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24878 LINK_SPEC.
24879 * config/microblaze/microblaze-c.c: Add builtin defines for
24880 _LITTLE_ENDIAN and _BIG_ENDIAN.
24881 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24882 add to TARGET_DEFAULT flags.
24883 Expand ASM_SPEC and LINK_SPEC.
24884 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24885 * config/microblaze/microblaze.md: Update extendsidi2 and
24886 movdi_internal instructions to use low-order / high-order reg
24887 print_operands.
24888 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24889 options and inversemask / mask of LITTLE_ENDIAN.
24890 * config/microblaze/t-microblaze: Expand multilib options to
24891 include mlittle-endian (le) and update exceptions patterns.
24892
24893 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24894
24895 PR rtl-optimization/56195
24896 * lra-constraints.c (get_reload_reg): Don't reuse regs
24897 if they have smaller mode than requested, if they have
24898 wider mode than requested, try to return a SUBREG.
24899
24900 PR tree-optimization/56250
24901 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24902 if type is unsigned and code isn't MULT_EXPR.
24903
24904 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24905
24906 PR tree-optimization/56064
24907 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24908 bits according to mode.
24909 * fixed-value.h (fixed_from_double_int)
24910 (const_fixed_from_double_int): Adjust comments.
24911
24912 2013-02-08 Richard Biener <rguenther@suse.de>
24913
24914 PR lto/56231
24915 * lto-streamer.h (struct data_in): Remove current_file, current_line
24916 and current_col members.
24917 * lto-streamer-out.c (lto_output_location): Stream changed bits
24918 en-block for efficiency.
24919 * lto-streamer-in.c (clear_line_info): Remove.
24920 (lto_input_location): Cache current file, line and column
24921 globally via local statics. Read changed bits en-block.
24922 (input_function): Do not call clear_line_info.
24923 (lto_read_body): Likewise.
24924 (lto_input_toplevel_asms): Likewise.
24925
24926 2013-02-08 Michael Matz <matz@suse.de>
24927
24928 PR tree-optimization/52448
24929 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24930 (nt_call_phase): New static.
24931 (add_or_mark_expr): Only mark accesses with newer phase than any
24932 call seen.
24933 (nonfreeing_call_p): New.
24934 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24935 (nt_fini_block): Keep blocks marked as visited.
24936 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24937
24938 2013-02-08 Richard Biener <rguenther@suse.de>
24939
24940 * ira.c (ira): Free broken dominator information.
24941
24942 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24943
24944 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24945
24946 2013-02-08 Marek Polacek <polacek@redhat.com>
24947
24948 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24949
24950 2013-02-08 Richard Biener <rguenther@suse.de>
24951
24952 PR middle-end/56181
24953 * cfgloop.h (flow_loops_find): Adjust.
24954 (bb_loop_header_p): Declare.
24955 * cfgloop.c (bb_loop_header_p): New function split out from ...
24956 (flow_loops_find): ... here. Adjust function signature,
24957 support incremental loop structure update.
24958 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24959 * cfgloopmanip.c (fix_loop_structure): Move ...
24960 * loop-init.c (fix_loop_structure): ... here.
24961 (apply_loop_flags): Split out from ...
24962 (loop_optimizer_init): ... here.
24963 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24964 in incremental mode, only remove dead loops here.
24965
24966 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24967
24968 PR target/54222
24969 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24970 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24971 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24972 (*round<mode>3.libgcc): New insns for fixed-modes.
24973 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24974 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24975 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24976 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24977 implementations. Define to __builtin_avr_absFX,
24978 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24979 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24980 __builtin_avr_countlsFX, respectively.
24981 * config/avr/avr-c.c (target.h): Include it.
24982 (enum avr_builtin_id): New enum.
24983 (avr_resolve_overloaded_builtin): New static function.
24984 (avr_register_target_pragmas): Use it to set
24985 targetm.resolve_overloaded_builtin.
24986 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24987 tree nodes used by DEF_BUILTIN.
24988 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24989 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24990 <AVR_BUILTIN_xxBITS>: Same.
24991
24992 2013-02-08 Richard Biener <rguenther@suse.de>
24993
24994 * cfgloop.c (verify_loop_structure): Properly handle
24995 a loop exiting to another loop header.
24996 * ira-int.h (ira_loops): Remove.
24997 * ira.c (ira_loops): Remove.
24998 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24999 (do_reload): Use loop_optimizer_finalize.
25000 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25001 number_of_loops to access the loop tree.
25002 (more_one_region_p): Likewise.
25003 (finish_loop_tree_nodes): Likewise.
25004 (rebuild_regno_allocno_maps): Likewise.
25005 (mark_loops_for_removal): Likewise.
25006 (mark_all_loops_for_removal): Likewise.
25007 (remove_unnecessary_regions): Likewise.
25008 (ira_build): Likewise.
25009 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25010
25011 2013-02-08 Richard Biener <rguenther@suse.de>
25012
25013 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25014 * ipa-pure-const.c (analyze_function): Avoid calling
25015 mark_irreducible_loops twice.
25016 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25017
25018 2013-02-07 David S. Miller <davem@davemloft.net>
25019
25020 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25021 on 'reg'.
25022 * var-tracking.c (vt_add_function_parameter): Test the presence of
25023 HAVE_window_save properly and do not remap argument registers when
25024 we have a leaf function.
25025
25026 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25027
25028 PR bootstrap/56227
25029 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25030 instead of "ll".
25031 * config/i386/i386.c (ix86_print_operand): Ditto.
25032
25033 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25034
25035 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25036
25037 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25038
25039 PR rtl-optimization/56225
25040 * lra-constraints.c (process_alt_operands): Check that reload hard
25041 reg can hold value for strict_low_part.
25042
25043 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25044
25045 PR debug/56154
25046 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25047 dwarf2out_end_function.
25048 (in_first_function_p, maybe_at_text_label_p,
25049 first_loclabel_num_not_at_text_label): New variables.
25050 (dwarf2out_var_location): In the first function find out
25051 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25052 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25053 functions.
25054
25055 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25056
25057 PR rtl-optimization/56178
25058 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25059 SUBREG of a register. Tidy up related block of code.
25060 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25061 note if the source is a register or a SUBREG of a register.
25062
25063 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25064
25065 PR target/56228
25066 * config/rs6000/rs6000.md (ptrm): New mode attr.
25067 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25068 call_value_indirect_aix<pttrsize>,
25069 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25070 m in constraints.
25071
25072 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25073
25074 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25075 if -bnortl. Convert to strcmp and strncmp.
25076
25077 2013-02-07 Alan Modra <amodra@gmail.com>
25078
25079 PR target/54009
25080 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25081 addresses won't wrap when offsetting.
25082 (rs6000_secondary_reload): Provide secondary reloads needed for
25083 wrapping LO_SUM addresses.
25084
25085 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25086
25087 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25088 MACH, just __MACH__.
25089
25090 2013-02-06 Richard Biener <rguenther@suse.de>
25091
25092 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25093 instead of calling fix_loop_structure.
25094
25095 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25096
25097 PR middle-end/56217
25098 * omp-low.c (use_pointer_for_field): Return false if
25099 lower_send_shared_vars doesn't generate any copy-out code.
25100
25101 2013-02-06 Tom de Vries <tom@codesourcery.com>
25102
25103 PR rtl-optimization/56131
25104 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25105 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25106 of the label is NULL. Add comment.
25107
25108 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25109
25110 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25111
25112 PR sanitizer/55374
25113 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25114 (STATIC_LIBTSAN_LIBS): Likewise.
25115 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25116 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25117 is defined, don't add anything else beyond that.
25118 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25119 (LINK_COMMAND_SPEC): Use them.
25120
25121 PR tree-optimization/56205
25122 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25123 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25124 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25125
25126 2013-02-05 Richard Biener <rguenther@suse.de>
25127
25128 PR tree-optimization/53342
25129 PR tree-optimization/53185
25130 * tree-vectorizer.h (vect_check_strided_load): Remove.
25131 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25132 not disallow peeling for vectorized strided loads.
25133 (vect_check_strided_load): Make static and simplify.
25134 (vect_analyze_data_refs): Adjust.
25135 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25136 correctly when vectorizing strided loads.
25137
25138 2013-02-05 Richard Biener <rguenther@suse.de>
25139
25140 * doc/install.texi: Refer to ISL, not PPL.
25141
25142 2013-02-05 Jan Hubicka <jh@suse.cz>
25143
25144 PR tree-optimization/55789
25145 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25146
25147 2013-02-05 Jan Hubicka <jh@suse.cz>
25148
25149 PR tree-optimization/55789
25150 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25151 the dead call anyway.
25152
25153 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25154
25155 PR sanitizer/55374
25156 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25157
25158 2013-02-04 Alexander Potapenko <glider@google.com>
25159 Jack Howarth <howarth@bromo.med.uc.edu>
25160 Jakub Jelinek <jakub@redhat.com>
25161
25162 PR sanitizer/55617
25163 * config/darwin.c (sort_ctor_records): Stabilized qsort
25164 on constructor priority by using original position.
25165 (finalize_ctors): New routine to sort constructors by
25166 priority before use in assemble_integer.
25167 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25168
25169 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25170
25171 PR libstdc++/54314
25172 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25173 about visibility on artificial decls.
25174 * config/sol2.c (solaris_assemble_visibility): Likewise.
25175
25176 2013-02-04 Kai Tietz <ktietz@redhat.com>
25177
25178 PR target/56186
25179 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25180 argument and improve checking of return-argument types for 16-byte
25181 modes.
25182 (ix86_function_value_1): Add additional valtype argument on call
25183 of function_value_64.
25184 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25185 handling infunction_value_64 function.
25186
25187 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25188
25189 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25190
25191 2013-02-04 Richard Biener <rguenther@suse.de>
25192
25193 PR tree-optimization/56188
25194 * tree-ssa-structalias.c (label_visit): Consider case with
25195 initially non-empty points-to set.
25196 (perform_var_substitution): Dump node mapping and clean up.
25197
25198 2013-02-04 Richard Guenther <rguenther@suse.de>
25199
25200 PR lto/56168
25201 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25202 node prevail as last resort.
25203 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25204 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25205
25206 2013-02-04 Richard Biener <rguenther@suse.de>
25207
25208 PR tree-optimization/56113
25209 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25210 Merge into ...
25211 (equiv_class_lookup_or_add): ... this.
25212 (label_visit): Adjust and fix error in previous patch.
25213 (perform_var_substitution): Adjust.
25214
25215 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25216
25217 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25218 * config/sh/sh4-300.md: Likewise.
25219 * config/sh/sh4a.md: Likewise.
25220 * config/sh/constraints.md: Likewise.
25221 * config/sh/sh.md: Likewise.
25222 * config/sh/netbsd-elf.h: Likewise.
25223 * config/sh/predicates.md: Likewise.
25224 * config/sh/sh-protos.h: Likewise.
25225 * config/sh/ushmedia.h: Likewise.
25226 * config/sh/linux.h: Likewise.
25227 * config/sh/sh.c: Likewise.
25228 * config/sh/superh.h: Likewise.
25229 * config/sh/elf.h: Likewise.
25230 * config/sh/sh4.md: Likewise.
25231 * config/sh/sh.h: Likewise.
25232
25233 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25234
25235 * config/pa/constraints.md: Adjust unused letters. Change "T"
25236 constraint to match_test floating_point_store_memory_operand().
25237 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25238 (base14_operand): New.
25239 (floating_point_store_memory_operand): New.
25240 (integer_store_memory_operand): Revise to use base14_operand and
25241 reg_plus_base_memory_operand.
25242 (move_dest_operand): Allow symbolic_memory_operands.
25243 (symbolic_memory_operand): Check for LO_SOM.
25244 (symbolic_operand): Change default case to break.
25245 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25246 CONST_DOUBLE values to be reloaded by putting them into memory when
25247 the destination is a floating point register.
25248 (movdf): Remove code to handle CONST_DOUBLE.
25249 (movsf): Likewise.
25250 (reload_indf_r1): New.
25251 (reload_insf_r1): New.
25252 Consistently use "Q" and "T" constraints with integer and floating
25253 point move instructions, respectively.
25254 (movdi): Remove FAIL.
25255 Change predicate for source operand unamed DImode move from
25256 general_operand to move_src_operand.
25257 (umulsidi3): Change predicate for destination operand to
25258 register_operand.
25259 Likewise for similar unamed patterns.
25260 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25261 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25262 (hppa_legitimize_address): Simplify mask calculation.
25263 (pa_emit_move_sequence): Revised handling of secondary reloads from
25264 REG+D addresses for floating point loads and stores. Directly handle
25265 loading CONST0_RTX (mode) to a floating point register.
25266 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25267 to floating point registers. Don't restrict secondary reloads to
25268 floating point registers to integer modes. Revise some comments and
25269 cleanup some code.
25270 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25271 (pa_legitimate_address_p): New.
25272 (pa_legitimize_reload_address): New.
25273 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25274 (STRICT_REG_OK_FOR_BASE_P): New.
25275 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25276 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25277
25278 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25279 Andrew Dixie <andrewd@gentrack.com>
25280
25281 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25282 flag set.
25283
25284 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25285
25286 * expmed.c (extract_bit_field_1): Pass the full width of the
25287 structure to get_best_reg_extraction_insn.
25288
25289 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25290
25291 PR target/54601
25292 * configure.ac (use_cxa_atexit): Add AIX.
25293 * configure: Regenerate.
25294
25295 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25296
25297 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25298
25299 PR debug/54793
25300 * final.c (need_profile_function): New variable.
25301 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25302 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25303 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25304 notes, targetm.asm_out.function_prologue doesn't emit anything,
25305 HAVE_prologue and profiler should be emitted before prologue,
25306 set need_profile_function instead of emitting it.
25307 (final_scan_insn): If need_profile_function, emit
25308 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25309 NOTE_INSN_FUNCTION_BEG note.
25310
25311 2013-02-01 Richard Henderson <rth@redhat.com>
25312
25313 * config/rs6000/rs6000.md (smulditi3): New.
25314 (umulditi3): New.
25315
25316 * config/alpha/alpha.md (umulditi3): New.
25317
25318 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25319
25320 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25321 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25322
25323 2013-02-01 Richard Biener <rguenther@suse.de>
25324
25325 PR tree-optimization/56113
25326 * tree-ssa-structalias.c (label_visit): Reduce work for
25327 single-predecessor nodes.
25328
25329 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25330
25331 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25332 range isn't testing for zero.
25333
25334 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25335
25336 PR middle-end/56113
25337 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25338
25339 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25340 Nick Clifton <nickc@redhat.com>
25341
25342 * config/v850/constraints.md (Q): Define as a memory constraint.
25343 * config/v850/predicates.md (label_ref_operand): New predicate.
25344 (e3v5_shift_operand): New predicate.
25345 (ior_operator): New predicate.
25346 * config/v850/t-v850: Add e3v5 multilib.
25347 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25348 (v850_gen_movdi): Prototype.
25349 * config/v850/v850.c: Add support for e3v5 architecture.
25350 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25351 TARGET_V850E_UP.
25352 (construct_save_jarl): Add e3v5 long JARL support.
25353 (v850_adjust_insn_length): New function. Adjust length of call
25354 insns when using e3v5 instructions.
25355 (v850_gen_movdi): New function: Generate instructions to move a
25356 DImode value.
25357 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25358 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25359 (TARGET_USE_FPU): Enable for e3v5.
25360 (CONST_OK_FOR_W): New macro.
25361 (ADJUST_INSN_LENGTH): Define.
25362 * config/v850/v850.md (UNSPEC_LOOP): Define.
25363 (attr cpu): Add v850e3v5.
25364 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25365 (movdi): New pattern.
25366 (movdi_internal): New pattern.
25367 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25368 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25369 (cstoresf4): Likewise.
25370 (cstoredf4): Likewise.
25371 (insv): New pattern.
25372 (rotlso3_a): New pattern.
25373 (rotlsi3_b): New pattern
25374 (rotlsi3_v850e3v5): New pattern.
25375 (doloop_begin): New pattern.
25376 (fix_loop_counter): New pattern.
25377 (doloop_end): New pattern.
25378 (branch_normal): Add e3v5 long branch support.
25379 (branch_invert): Likewise.
25380 (branch_z_normal): Likewise.
25381 (branch_z_invert): Likewise.
25382 (branch_nz_normal): Likewise.
25383 (branch_nz_invert): Likewise.
25384 (call_internal_short): Add e3v5 register-indirect JARL support.
25385 (call_internal_long): Likewise.
25386 (call_value_internal_short): Likewise.
25387 (call_value_internal_long): Likewise.
25388 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25389 (mloop): New option.
25390 * config.gcc: Add support for configuring v840e3v5 target.
25391 * doc/invoke.texi: Document new v850 specific command line options.
25392
25393 2013-01-31 Paul Koning <ni1d@arrl.net>
25394
25395 PR debug/55059
25396 PR debug/54508
25397 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25398 children if parent is a class.
25399 (prune_unused_types_prune): Don't add DW_AT_declaration.
25400
25401 2013-01-31 Richard Biener <rguenther@suse.de>
25402
25403 PR tree-optimization/56157
25404 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25405 match up operand with SLP child.
25406
25407 2013-01-31 Jason Merrill <jason@redhat.com>
25408
25409 PR debug/54410
25410 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25411 parameters the first time.
25412 (gen_scheduled_generic_parms_dies): Check completeness here.
25413
25414 2013-01-31 Richard Biener <rguenther@suse.de>
25415
25416 PR middle-end/53073
25417 * common.opt (faggressive-loop-optimizations): New flag,
25418 enabled by default.
25419 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25420 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25421 infer_loop_bounds_from_undefined by it.
25422
25423 2013-01-31 Richard Biener <rguenther@suse.de>
25424
25425 PR tree-optimization/56150
25426 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25427 visit virtual operands.
25428 (find_uses_to_rename_bb): Likewise.
25429
25430 2013-01-31 Richard Biener <rguenther@suse.de>
25431
25432 PR tree-optimization/56150
25433 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25434 mixed store non-store stmts.
25435
25436 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25437
25438 PR sanitizer/55374
25439 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25440 LIBASAN_EARLY_SPEC is defined.
25441 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25442 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25443 before %o.
25444 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25445
25446 PR c++/55742
25447 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25448 invalid args instead of ICEing on it.
25449 (ix86_valid_target_attribute_tree): Return error_mark_node if
25450 ix86_valid_target_attribute_inner_p failed.
25451 (ix86_valid_target_attribute_p): Return false only if
25452 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25453 target("default") attribute.
25454 (sorted_attr_string): Change argument from const char * to tree,
25455 merge in all target attribute arguments rather than just one.
25456 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25457 instead of free. Avoid using strcat.
25458 (ix86_mangle_function_version_assembler_name): Mangle
25459 target("default") as if no target attribute is present. Adjust
25460 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25461 instead of xmalloc and XDELETEVEC instead of free.
25462 (ix86_function_versions): Don't return true if one of the decls
25463 doesn't have target attribute. If they don't and one of the decls
25464 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25465 sorted_attr_string caller. Use XDELETEVEC instead of free.
25466 (ix86_supports_function_versions): Remove.
25467 (make_name): Fix up formatting.
25468 (make_dispatcher_decl): Remove resolver_name and its initialization.
25469 Avoid leaking memory.
25470 (is_function_default_version): Return true if there is
25471 target("default") attribute rather than no target attribute at all.
25472 (make_resolver_func): Avoid leaking memory.
25473 (ix86_generate_version_dispatcher_body): Likewise.
25474 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25475 * target.def (supports_function_versions): Remove.
25476 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25477 * doc/tm.texi: Regenerated.
25478
25479 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25480
25481 PR rtl-optimization/56144
25482 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25483 for values with side effects.
25484
25485 2013-01-30 Richard Biener <rguenther@suse.de>
25486
25487 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25488 (sparseset_pop): Likewise.
25489 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25490 to be able to use quick_push in the worker loop.
25491
25492 2013-01-30 Marek Polacek <polacek@redhat.com>
25493
25494 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25495
25496 2013-01-30 Richard Biener <rguenther@suse.de>
25497
25498 PR lto/56147
25499 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25500
25501 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25502
25503 PR tree-optimization/56064
25504 * fixed-value.c (fixed_from_double_int): New function.
25505 * fixed-value.h (fixed_from_double_int): New prototype.
25506 (const_fixed_from_double_int): New static inline function.
25507 * fold-const.c (native_interpret_fixed): New static function.
25508 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25509 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25510 (native_encode_fixed): New static function.
25511 (native_encode_expr) <FIXED_CST>: Use it.
25512 (native_interpret_int): Move double_int worker code to...
25513 * double-int.c (double_int::from_buffer): ...this new static method.
25514 * double-int.h (double_int::from_buffer): Prototype it.
25515
25516 2013-01-30 Richard Biener <rguenther@suse.de>
25517
25518 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25519 New pointer-map and obstack.
25520 (init_alias_vars): Allocate pointer-map and obstack.
25521 (delete_points_to_sets): Free them.
25522 (find_what_var_points_to): Cache result.
25523 (find_what_p_points_to): Adjust for changed interface of
25524 find_what_var_points_to.
25525 (compute_points_to_sets): Likewise.
25526 (ipa_pta_execute): Likewise.
25527
25528 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25529
25530 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25531 * configure: Regenerate.
25532 * config.in: Regenerate.
25533 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25534 #nobits/#progbits if supported.
25535
25536 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25537
25538 PR target/56121
25539 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25540 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25541 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25542
25543 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25544
25545 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25546 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25547
25548 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25549
25550 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25551 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25552
25553 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25554
25555 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25556 declaration.
25557 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25558 * config/arm/cortex-a7.md: New bypasses using
25559 arm_mac_accumulator_is_result.
25560
25561 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25562
25563 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25564 (cortex_a7_neon_mla): Likewise.
25565 (cortex_a7_fpfmad): New reservation.
25566 (cortex_a7_fpmacs): Use ffmas and update required units.
25567 (cortex_a7_fpmuld): Update required units and latency.
25568 (cortex_a7_fpmacd): Likewise.
25569 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25570 (cortex_a7_neon). Likewise.
25571 (bypass) Update participating units.
25572
25573 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25574
25575 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25576 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25577 from fmac to ffma.
25578 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25579 (vfp_fmul): Use ffmad.
25580 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25581 (cortex_r4_fmacd): Use ffmad.
25582 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25583 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25584 (cortex_a9_fmacd): Use ffmad.
25585 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25586 (cortex_a8_vfp_macd): Use ffmad.
25587 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25588 (cortex_a5_fpmacd): Use ffmad.
25589 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25590 (cortex_a15_vfp_macd): Use ffmad.
25591 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25592
25593 2013-01-29 Jason Merrill <jason@redhat.com>
25594
25595 PR libstdc++/54314
25596 * varasm.c (default_assemble_visibility): Don't warn about
25597 visibility on artificial decls.
25598
25599 2013-01-29 Richard Biener <rguenther@suse.de>
25600
25601 PR tree-optimization/56113
25602 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25603 the bitmap leader.
25604 (label_visit): Free duplicate bitmaps and record the leader instead.
25605 (perform_var_substitution): Adjust.
25606
25607 2013-01-29 Richard Biener <rguenther@suse.de>
25608
25609 PR tree-optimization/55270
25610 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25611 the CFG, schedule loops for fixup.
25612
25613 2013-01-29 Nick Clifton <nickc@redhat.com>
25614
25615 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25616 SP_REG.
25617
25618 2013-01-28 Leif Ekblad <leif@rdos.net>
25619
25620 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25621 * config/i386/i386.h (TARGET_RDOS): New macro.
25622 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25623 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25624 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25625 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25626 DEFAULT_LARGE_SECTION_THRESHOLD.
25627 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25628 * config/i386/rdos.h: New file.
25629 * config/i386/rdos64.h: New file.
25630
25631 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25632
25633 PR other/54814
25634 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25635 TEST_HARD_REG_BIT.
25636
25637 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25638
25639 PR rtl-optimization/56117
25640 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25641 call cselib_lookup_from_insn on the MEM before calling
25642 add_insn_mem_dependence.
25643
25644 2013-01-28 Richard Biener <rguenther@suse.de>
25645
25646 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25647 to a stmt that didn't have one.
25648 (copy_phis_for_bb): Likewise for PHI arguments.
25649 (copy_debug_stmt): Likewise for debug stmts.
25650
25651 2013-01-28 Richard Biener <rguenther@suse.de>
25652
25653 PR tree-optimization/56034
25654 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25655 (partition_builtin_p): Adjust.
25656 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25657 it is the last partition.
25658 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25659 up the vertex for the definition.
25660 (classify_partition): Classify whether a partition is a
25661 PKIND_REDUCTION, thus has uses outside of the loop.
25662 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25663 Merge all PKIND_REDUCTION partitions into the last partition.
25664 (tree_loop_distribution): Seed partitions from reductions as well.
25665
25666 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25667
25668 PR tree-optimization/56125
25669 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25670 pow(x,c) into sqrt(x) * powi(x, n/2) or
25671 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25672 optimizing for size.
25673 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25674 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25675 integer.
25676
25677 PR tree-optimization/56094
25678 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25679 to UNKNOWN_LOCATION while gimplifying expr.
25680
25681 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25682
25683 PR target/56114
25684 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25685 operand 0 in movabs insn template for -masm=intel asm alternative.
25686 (*movabs<mode>_2): Ditto for operand 1.
25687
25688 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25689
25690 PR target/54663
25691 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25692 of microblaze-c.o
25693
25694 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25695
25696 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25697 tm_file.
25698
25699 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25700
25701 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25702 Undef to avoid warning.
25703
25704 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25705
25706 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25707 * configure: Regenerate.
25708
25709 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25710
25711 PR tree-optimization/56098
25712 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25713 for stmts with volatile ops.
25714 (cond_store_replacement): Don't optimize if assign has volatile ops.
25715 (cond_if_else_store_replacement_1): Don't optimize if either
25716 then_assign or else_assign have volatile ops.
25717 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25718 volatile ops.
25719
25720 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25721
25722 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25723
25724 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25725
25726 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25727 missing ':' in asm example.
25728
25729 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25730
25731 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25732 entries into lane and laneq entries.
25733 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25734 Remove AdvSIMD scalar modes.
25735 (aarch64_sq<r>dmulh_laneq<mode>): New.
25736 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25737 modes.
25738 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25739 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25740 * config/aarch64/iterators.md (VCOND): New.
25741 (VCONQ): New.
25742
25743 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25744
25745 PR target/54222
25746 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25747 Add NULL LIBNAME argument to existing definitions.
25748 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25749 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25750 * config/avr/avr.c (DEF_BUILTIN): Same.
25751 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25752 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25753 implementation is available (DECL_ASSEMBLER_NAME is set).
25754 (avr_fold_absfx): New static function.
25755 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25756 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25757 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25758 AVR_BUILTIN_ABSLLK.
25759 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25760 (abshk, absk, abslk, absllk): Provide as static inline functions.
25761
25762 2013-01-25 Marek Polacek <polacek@redhat.com>
25763
25764 PR tree-optimization/56035
25765 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25766
25767 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25768
25769 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25770 (*movtf_internal_rex64): Add (!o,C) alternative
25771 (*movxf_internal_rex64): Ditto.
25772 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25773
25774 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25775
25776 * doc/invoke.texi: fix typo.
25777 * doc/objc.texi: fix typo.
25778
25779 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25780
25781 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25782 for the first two alternatives.
25783
25784 2013-01-24 Diego Novillo <dnovillo@google.com>
25785
25786 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25787 (ggc-zone.o): Remove.
25788 * configure.ac: Remove option --with-gc.
25789 * configure: Re-generate.
25790 * doc/install.texi: Remove documentation for --with-gc.
25791 * gengtype.c (write_enum_defn): Remove. Update all users.
25792 (write_Types_process_field): Remove generation of gt_e_* argument.
25793 (output_type_enum): Remove. Update all users.
25794 (write_enum_defn): Remove. Update all users.
25795 (enum alloc_zone): Remove. Update all users.
25796 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25797 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25798 Update all callers.
25799 (struct ptr_data): Remove field TYPE. Update all users.
25800 (gt_pch_note_object): Remove argument TYPE. Update all users.
25801 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25802 Update all users.
25803 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25804 (struct alloc_zone): Remove.
25805 (ggc_internal_alloc_zone_stat): Remove.
25806 (ggc_internal_cleared_alloc_zone_stat): Remove.
25807 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25808 (ggc_pch_count_object): Remove last argument. Update all users.
25809 (ggc_pch_alloc_object): Remove last argument. Update all users.
25810 (struct alloc_zone): Remove.
25811 * ggc-zone.c: Remove.
25812 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25813 (struct alloc_zone): Remove.
25814 (ggc_alloc_typed_stat): Remove.
25815 (ggc_alloc_typed): Remove.
25816 (ggc_splay_alloc): Remove first argument.
25817 (rtl_zone): Remove. Update all users.
25818 (tree_zone): Remove. Update all users.
25819 (tree_id_zone): Remove. Update all users.
25820 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25821 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25822 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25823 * tree-ssanames.c: Remove references to zone allocator in comments.
25824
25825 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25826
25827 * config/avr/avr.c (avr_out_fract): Make register numbers that
25828 might be outside of source operand signed.
25829
25830 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25831
25832 * config/i386/constraints.md (Yf): New constraint.
25833 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25834 of f constraint to conditionaly disable x87 register preferences.
25835 (*movdf_internal): Ditto.
25836 (*movsf_internal): Ditto.
25837
25838 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25839
25840 PR inline-asm/55934
25841 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25842 that have operands with impossible constraints.
25843 Add a FIXME for a speed-up opportunity.
25844 * lra-constraints.c (process_alt_operands): Verify that a class
25845 selected from constraints on asms is valid for the operand mode.
25846 (curr_insn_transform): Remove incorrect comment.
25847
25848 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25849
25850 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25851 TOC operand is a valid symbol ref in the constant pool.
25852
25853 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25854
25855 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25856
25857 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25858
25859 PR target/54222
25860 * config/avr/stdfix.h: New file.
25861 * t-avr (stdfix-gcc.h): New rule to build it.
25862 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25863
25864 2013-01-23 Kostya Serebryany <kcc@google.com>
25865
25866 * config/darwin.h: remove dependency on
25867 CoreFoundation (asan on Mac OS).
25868
25869 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25870
25871 PR target/49069
25872 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25873 instead of cmpdi_operand for first comparison operand.
25874 Don't assert that comparison operands aren't both constants.
25875
25876 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25877
25878 * doc/install.texi (Downloading the Source): Update references to
25879 downloading separate components.
25880
25881 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25882
25883 * doc/extend.texi (__int128): Improve grammar.
25884
25885 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25886
25887 PR target/56028
25888 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25889 alternative to (o,r).
25890 (*movdi_internal_rex64): Remove (!o,n) alternative.
25891 (DImode immediate->memory splitter): Remove.
25892 (DImode immediate->memory peephole2): Remove.
25893 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25894 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25895 alternative to (!o,*r).
25896 (*movtf_internal_sse): New pattern.
25897 (*movxf_internal_rex64): New pattern.
25898 (*movxf_internal): Disable for TARGET_64BIT.
25899 (*movdf_internal_rex64): Remove (!o,F) alternative.
25900
25901 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25902
25903 PR middle-end/56074
25904 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25905 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25906 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25907 stmt locations where LOCATION_LOCUS of the stmt location is
25908 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25909
25910 PR target/55686
25911 * config/i386/i386.md (UNSPEC_STOS): New.
25912 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25913 *strsetqi_1): Add UNSPEC_STOS.
25914
25915 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25916
25917 PR c++/56067
25918 * doc/invoke.texi: Remove left over -Wsynth example.
25919
25920 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25921
25922 PR tree-optimization/56051
25923 * fold-const.c (fold_binary_loc): Don't fold
25924 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25925 a narrowing conversion, or widening conversion from signed
25926 to unsigned.
25927
25928 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25929
25930 PR rtl-optimization/56023
25931 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25932 dependent on debug instruction.
25933
25934 2013-01-21 Martin Jambor <mjambor@suse.cz>
25935
25936 PR middle-end/56022
25937 * function.c (allocate_struct_function): Call
25938 invoke_set_current_function_hook earlier.
25939
25940 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25941
25942 * reload1.c (init_reload): Only initialize reload_obstack
25943 during the first call.
25944
25945 2013-01-21 Marek Polacek <polacek@redhat.com>
25946
25947 * cfgloop.c (verify_loop_structure): Fix up grammar.
25948
25949 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25950
25951 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25952 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25953
25954 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25955
25956 PR target/56058
25957 * config/arm/marvell-pj4.md: Update copyright year.
25958 Fix up use of alu to alu_reg and simple_alu_imm.
25959
25960 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25961
25962 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25963
25964 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25965
25966 PR target/55433
25967 * lra-constraints.c (curr_insn_transform): Don't reuse original
25968 insn for secondary memory move when memory mode should be different.
25969
25970 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25971
25972 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25973 atomic_storedi_1): New patterns.
25974
25975 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25976
25977 btver2 pipeline descriptions.
25978 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25979 descriptions.
25980 * config/i386/i386.md (btver2_decode): New type attributes.
25981 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25982 type attributes.
25983 * config/i386/btver2.md: New file describing btver2 pipelines.
25984
25985 2013-01-19 Andrew Pinski <apinski@cavium.com>
25986
25987 PR tree-optimization/52631
25988 * tree-ssa-sccvn (visit_use): Before looking up the original
25989 statement, try looking up the simplified expression.
25990
25991 2013-01-19 Anthony Green <green@moxielogic.com>
25992
25993 * config/moxie/moxie.c (moxie_expand_prologue): Set
25994 current_function_static_stack_size.
25995
25996 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25997
25998 PR tree-optimization/56029
25999 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26000 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26001
26002 2013-01-18 Sharad Singhai <singhai@google.com>
26003
26004 PR tree-optimization/55995
26005 * dumpfile.c (dump_loc): Print location only if available.
26006 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26007
26008 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26009
26010 PR target/55433
26011 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26012 secondary memory move.
26013 (inherit_reload_reg): Use rclass instead of cl for
26014 check_secondary_memory_needed_p.
26015
26016 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26017
26018 PR middle-end/56015
26019 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26020 the case where writing real complex part of target modifies op1.
26021
26022 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26023
26024 * config/aarch64/aarch64-simd.md
26025 (aarch64_vcond_internal<mode>): Handle unordered cases.
26026 * config/aarch64/iterators.md (v_cmp_result): New.
26027
26028 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26029 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26030
26031 * config/arm/marvell-pj4.md: New file.
26032 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26033 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26034 (generic_vfp): Likewise.
26035 * config/arm/arm-cores.def: Add marvell-pj4.
26036 * config/arm/arm-tune.md: Regenerate.
26037 * config/arm/arm-tables.opt: Regenerate.
26038 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26039 * doc/invoke.texi: Document marvell-pj4.
26040
26041 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26042
26043 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26044
26045 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26046
26047 PR debug/54114
26048 PR debug/54402
26049 PR debug/49888
26050 * var-tracking.c (negative_power_of_two_p): New.
26051 (global_get_addr_cache, local_get_addr_cache): New.
26052 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26053 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26054 heading comment.
26055 (vt_stack_offset_p): Remove.
26056 (vt_canon_true_dep): Always canonicalize loc's address.
26057 (clobber_overlapping_mems): Make sure we have a MEM.
26058 (local_get_addr_clear_given_value): New.
26059 (val_reset): Clear local cached entries.
26060 (compute_bb_dataflow): Create and release the local cache.
26061 Disable duplicate MEMs clobbering.
26062 (emit_notes_in_bb): Clobber MEMs likewise.
26063 (vt_emit_notes): Create and release the local cache.
26064 (vt_initialize, vt_finalize): Create and release the global
26065 cache, respectively.
26066 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26067
26068 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26069
26070 PR libmudflap/53359
26071 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26072 not found in the symtab.
26073
26074 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26075
26076 PR debug/56006
26077 PR rtl-optimization/55547
26078 PR rtl-optimization/53827
26079 PR debug/53671
26080 PR debug/49888
26081 * alias.c (offset_overlap_p): New, factored out of...
26082 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26083 the conservative special case for symbolic constants. Don't
26084 adjust zero sizes on alignment.
26085
26086 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26087
26088 PR rtl-optimization/52573
26089 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26090 REG_UNUSED for the same register.
26091
26092 2013-01-17 Richard Biener <rguenther@suse.de>
26093 Marek Polacek <polacek@redhat.com>
26094
26095 PR rtl-optimization/55833
26096 * loop-unswitch.c (unswitch_loops): Move loop verification...
26097 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26098 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26099 Set it to true when we're removing a loop from hierarchy tree in
26100 an irreducible region.
26101 (fix_bb_placements): Adjust caller.
26102 (fix_loop_placements): Likewise.
26103
26104 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26105
26106 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26107 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26108 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26109 Remove ID. Adjust comments.
26110 * config/avr/avr-c.c (avr_builtin_name): Remove.
26111 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26112 * config/avr/avr.c (avr_tolower): New static function.
26113 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26114 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26115 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26116 default expansion.
26117
26118 2013-01-17 Jan Hubicka <jh@suse.cz>
26119
26120 PR tree-optimization/55273
26121 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26122
26123 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26124
26125 PR target/55981
26126 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26127 store through atomic_store<mode>_1.
26128 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26129
26130 2013-01-17 Martin Jambor <mjambor@suse.cz>
26131
26132 PR tree-optimizations/55264
26133 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26134 for virtual methods.
26135 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26136 virtual methods before inlining is over.
26137 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26138 virtual functions.
26139 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26140 non-virtual.
26141
26142 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26143
26144 PR rtl-optimization/56005
26145 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26146 pending reads for prefetch.
26147
26148 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26149
26150 * config/aarch64/aarch64.md
26151 (*cstoresi_neg_uxtw): New pattern.
26152 (*cmovsi_insn_uxtw): New pattern.
26153 (*<optab>si3_uxtw): New pattern.
26154 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26155 (*<optab>si3_insn_uxtw): New pattern.
26156 (*bswapsi2_uxtw): New pattern.
26157
26158 2013-01-16 Richard Biener <rguenther@suse.de>
26159
26160 * tree-inline.c (tree_function_versioning): Remove set but
26161 never used variable.
26162
26163 2013-01-16 Richard Biener <rguenther@suse.de>
26164
26165 PR tree-optimization/55964
26166 * tree-flow.h (rename_variables_in_loop): Remove.
26167 (rename_variables_in_bb): Likewise.
26168 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26169 (copy_loop_before): Adjust and delete update-ssa status.
26170 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26171 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26172 (rename_variables_in_loop): Remove.
26173 (slpeel_update_phis_for_duplicate_loop): Likewise.
26174 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26175 use available cfg machinery instead of duplicating it.
26176 Update PHI nodes and perform poor-mans SSA update here.
26177 (slpeel_tree_peel_loop_to_edge): Adjust.
26178
26179 2013-01-16 Richard Biener <rguenther@suse.de>
26180
26181 PR tree-optimization/54767
26182 PR tree-optimization/53465
26183 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26184 (vrp_visit_phi_node): For PHI arguments coming via backedges
26185 drop all symbolical range information.
26186 (execute_vrp): Compute backedges.
26187
26188 2013-01-16 Richard Biener <rguenther@suse.de>
26189
26190 * doc/install.texi: Update CLooG and ISL requirements to
26191 0.18.0 and 0.11.1.
26192
26193 2013-01-16 Christian Bruel <christian.bruel@st.com>
26194
26195 PR target/55301
26196 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26197 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26198 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26199
26200 2013-01-16 DJ Delorie <dj@redhat.com>
26201
26202 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26203 (UNSPECV_SP_SWITCH_E): New.
26204 (sp_switch_1): Change to an unspec.
26205 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26206 replace $r15.
26207
26208 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26209
26210 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26211 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26212 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26213 (expand_mem_thread_fence): Ditto.
26214 (expand_mem_signal_fence): Ditto.
26215 (expand_atomic_load): Ditto.
26216 (expand_atomic_store): Ditto.
26217
26218 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26219
26220 PR rtl-optimization/55547
26221 PR rtl-optimization/53827
26222 PR debug/53671
26223 PR debug/49888
26224 * alias.c (memrefs_conflict_p): Set sizes to negative after
26225 AND adjustments.
26226
26227 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26228
26229 PR target/55940
26230 * function.c (thread_prologue_and_epilogue_insns): Always
26231 add crtl->drap_reg to set_up_by_prologue.set, even if
26232 stack_realign_drap is false.
26233
26234 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26235
26236 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26237 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26238 *call): Fix indention.
26239
26240 2013-01-15 Tom de Vries <tom@codesourcery.com>
26241
26242 PR target/55876
26243 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26244 Update comment.
26245
26246 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26247
26248 PR rtl-optimization/55153
26249 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26250
26251 2013-01-15 Martin Jambor <mjambor@suse.cz>
26252
26253 PR tree-optimization/55920
26254 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26255 accesses as grp_to_be_debug_replaced.
26256
26257 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26258
26259 PR tree-optimization/55920
26260 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26261 there is non-useless type conversion needed from debug rhs to lhs,
26262 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26263
26264 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26265 Mikael Pettersson <mikpe@it.uu.se>
26266
26267 PR target/43961
26268 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26269 Thumb.
26270 (ASM_OUTPUT_CASE_LABEL): Remove.
26271 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26272 * final.c (shorten_branches): Update alignment of labels before
26273 jump tables if CASE_VECTOR_SHORTEN_MODE.
26274
26275 2013-01-15 Richard Biener <rguenther@suse.de>
26276
26277 PR bootstrap/55961
26278 * system.h: Do not include gmp.h for building host tools.
26279
26280 2013-01-15 Richard Biener <rguenther@suse.de>
26281
26282 PR middle-end/55882
26283 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26284 account for bitpos when computing alignment.
26285
26286 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26287
26288 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26289 (ix86_target_macros_internal): Likewise.
26290
26291 * config/i386/i386.c (m_CORE2I7): Removed.
26292 (m_CORE_HASWELL): New macro.
26293 (m_CORE_ALL): Likewise.
26294 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26295 (initial_ix86_arch_features): Likewise.
26296 (processor_target_table): Initializations for Core avx2.
26297 (cpu_names): New names "core-avx2".
26298 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26299 PROCESSOR_CORE_HASWELL.
26300 (ix86_issue_rate): New case.
26301 (ia32_multipass_dfa_lookahead): Likewise.
26302 (ix86_sched_init_global): Likewise.
26303
26304 * config/i386/i386.h (TARGET_HASWELL): New macro.
26305 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26306 (processor_type): New PROCESSOR_HASWELL.
26307
26308 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26309
26310 PR tree-optimization/55955
26311 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26312 *SHIFT_EXPR and *ROTATE_EXPR codes.
26313
26314 PR tree-optimization/48766
26315 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26316 -ftrapv disable -fwrapv.
26317
26318 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26319
26320 PR target/55974
26321 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26322 etc. to 1 and not to __flash.
26323 Use LL suffix for __INT24_MAX__ with -mint8.
26324 Use ULL suffix for __UINT24_MAX__ with -mint8.
26325
26326 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26327
26328 * config/avr/avr-arch.h
26329 (struct base_arch_s): Use typedef avr_arch_t instead.
26330 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26331 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26332 * config/avr/avr.c: Same.
26333 * config/avr/avr-devices.c: Same.
26334 * config/avr/driver-avr.c: Same.
26335 * config/avr/gen-avr-mmcu-texi.c: Same.
26336 * config/avr/avr-mcus.def: Adjust comment.
26337
26338 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26339
26340 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26341 * config/aarch64/iterators.md (VALLDI): New.
26342
26343 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26344 Andi Kleen <ak@linux.intel.com>
26345
26346 PR target/55948
26347 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26348 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26349 memmodel flag.
26350
26351 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26352
26353 * config/avr/avr-stdint.h: Remove trailing blanks.
26354 * config/avr/avr-log.h: Same.
26355 * config/avr/avr-arch.h: Same.
26356 * config/avr/avr-devices.c: Same.
26357 * config/avr/avr-dimode.md: Same.
26358 * config/avr/predicates.md: Same.
26359 * config/avr/avr-c.c: Same. And fix typo.
26360
26361 * config/avr/avr-protos.h: Same. And:
26362 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26363 (init_cumulative_args): Rename to avr_init_cumulative_args.
26364 (expand_prologue): Rename to avr_expand_prologue.
26365 (expand_epilogue): Rename to avr_expand_epilogue.
26366 (adjust_insn_length): Rename to avr_adjust_insn_length.
26367 (notice_update_cc): Rename to avr_notice_update_cc.
26368 (final_prescan_insn): Rename to avr_final_prescan_insn.
26369 * config/avr/avr.c: Same.
26370 * config/avr/avr.h: Same.
26371 * config/avr/avr.md: Remove trailing blanks.
26372 (prologue): Use avr_expand_prologue.
26373 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26374
26375 2013-01-14 Richard Biener <rguenther@suse.de>
26376
26377 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26378 verify_location, collect_subblocks): New functions.
26379 (verify_gimple_in_cfg): Verify that locations only reference
26380 BLOCKs in the functions BLOCK tree.
26381
26382 2013-01-14 Richard Biener <rguenther@suse.de>
26383
26384 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26385 PHI argument.
26386 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26387 unshare reference.
26388 (insert_out_of_ssa_copy_on_edge): Likewise.
26389 (rewrite_close_phi_out_of_ssa): Likewise.
26390 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26391 debug expressions.
26392 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26393 propagated constants.
26394 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26395 can not be shared.
26396
26397 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26398
26399 * config/avr/avr-modes.def: Add GPL copyright notice.
26400
26401 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26402
26403 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26404 MEMMODEL_MASK to determine memory model.
26405 (atomic_store<mode>): Ditto from operands[2].
26406 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26407
26408 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26409
26410 PR fortran/55935
26411 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26412 (fold_gimple_assign): Don't call unshare_expr here.
26413 (fold_ctor_reference): Call unshare_expr.
26414
26415 2013-01-13 Terry Guo <terry.guo@arm.com>
26416
26417 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26418 * doc/fragments.texi: Document MULTILIB_REUSE.
26419 * gcc.c (multilib_reuse): New internal spec.
26420 (set_multilib_dir): Also search multilib from multilib_reuse.
26421 * genmultilib (tmpmultilib3): Refactor code.
26422 (tmpmultilib4): Ditto.
26423 (multilib_reuse): New multilib argument.
26424
26425 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26426
26427 * Makefile.in: Update copyright.
26428
26429 2013-01-12 Tom de Vries <tom@codesourcery.com>
26430
26431 PR middle-end/55890
26432 * calls.c (expand_call): Check if arg_nr is valid.
26433
26434 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26435
26436 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26437 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26438 documentation. Add missing '__' in front of
26439 __builtin_ia32_packssdw256.
26440
26441 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26442
26443 PR target/55719
26444 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26445 NO_REGS for larl operands.
26446 (s390_reload_larl_operand): Use s390_load_address instead of
26447 emit_move_insn.
26448
26449 2013-01-11 Richard Biener <rguenther@suse.de>
26450
26451 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26452 (verify_node_sharing): ... here.
26453 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26454
26455 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26456
26457 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26458 Substitute TREECHECKING.
26459 * configure: Regenerate.
26460 * Makefile.in (TREECHECKING): New.
26461
26462 2013-01-11 Richard Guenther <rguenther@suse.de>
26463
26464 PR tree-optimization/44061
26465 * tree-vrp.c (extract_range_basic): Compute zero as
26466 value-range for __builtin_constant_p of function parameters.
26467
26468 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26469
26470 Update copyright years.
26471
26472 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26473
26474 PR rtl-optimization/55672
26475 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26476 const to be eliminable.
26477
26478 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26479
26480 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26481 * configure: Regenerate.
26482
26483 2013-01-10 Richard Biener <rguenther@suse.de>
26484
26485 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26486
26487 2013-01-10 Richard Biener <rguenther@suse.de>
26488
26489 PR bootstrap/55792
26490 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26491 locations for virtual PHI arguments.
26492 (rewrite_update_phi_arguments): Likewise.
26493
26494 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26495
26496 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26497 on to assembler.
26498
26499 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26500
26501 PR tree-optimization/55921
26502 * tree-complex.c (expand_complex_asm): New function.
26503 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26504
26505 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26506
26507 PR target/55718
26508 * config/s390/s390.c (s390_symref_operand_p)
26509 (s390_loadrelative_operand_p): Merge the two functions.
26510 (s390_check_qrst_address, print_operand_address): Add parameters
26511 to s390_loadrelative_operand_p invokation.
26512 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26513 (s390_reload_larl_operand, s390_secondary_reload): Use
26514 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26515 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26516
26517 2013-01-09 Mike Stump <mikestump@comcast.net>
26518
26519 * dse.c (record_store): Remove unnecessary assert.
26520
26521 2013-01-09 Jan Hubicka <jh@suse.cz>
26522
26523 PR tree-optimization/55569
26524 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26525 * cfgloop.h (scale_loop_profile): Likewise.
26526
26527 2013-01-09 Jan Hubicka <jh@suse.cz>
26528
26529 PR lto/45375
26530 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26531 functions.
26532 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26533
26534 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26535
26536 PR middle-end/55114
26537 * expr.h (maybe_emit_group_store): Declare.
26538 * expr.c (maybe_emit_group_store): New function.
26539 * builtins.c (expand_builtin_int_roundingfn): Call it.
26540 (expand_builtin_int_roundingfn_2): Likewise.
26541
26542 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26543
26544 PR rtl-optimization/55829
26545 * lra-constraints.c (match_reload): Add code for absent output.
26546 (curr_insn_transform): Add code for reloads of matched inputs
26547 without output.
26548
26549 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26550
26551 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26552 attribute of movddup insn to DF.
26553 (*vec_interleave_lowv2df): Ditto.
26554 (vec_dupv2df): Ditto.
26555
26556 2013-01-09 Jan Hubicka <jh@suse.cz>
26557
26558 PR tree-optimiation/55875
26559 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26560 EVERY_ITERATION parameter.
26561 (number_of_iterations_exit): Check if exit is executed every iteration.
26562 (idx_infer_loop_bounds): Similarly here.
26563 (n_of_executions_at_most): Simplify
26564 to only test for cases where statement is dominated by the
26565 particular bound; handle correctly the "postdominance" test.
26566 (scev_probably_wraps_p): Use max loop iterations info
26567 as a global bound first.
26568
26569 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26570 Nick Clifton <nickc@redhat.com>
26571
26572 * config/v850/v850.md (cbranchsf4): New pattern.
26573 (cstoresf4): New pattern.
26574 (cbranchdf4): New pattern.
26575 (cstoredf4): New pattern.
26576 (movsicc): Disallow floating point comparisons.
26577 (cmpsf_le_insn): Fix order of operators.
26578 (cmpsf_lt_insn): Likewise.
26579 (cmpsf_eq_insn): Likewise.
26580 (cmpdf_le_insn): Likewise.
26581 (cmpdf_lt_insn): Likewise.
26582 (cmpdf_eq_insn): Likewise.
26583 (cmpsf_ge_insn): Use LE comparison.
26584 (cmpdf_ge_insn): Likewise.
26585 (cmpsf_gt_insn): Use LT comparison.
26586 (cmpdf_gt_insn): Likewise.
26587 (cmpsf_ne_insn): Delete pattern.
26588 (cmpdf_ne_insn): Delete pattern.
26589 * config/v850/v850.c (v850_gen_float_compare): Use
26590 gen_cmpdf_eq_insn for NE comparison.
26591 (v850_float_z_comparison_operator)
26592 (v850_float_nz_comparison_operator): Move from here ...
26593 * config/v850/predicates.md: ... to here. Move GT and GE
26594 comparisons into v850_float_z_comparison_operator.
26595 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26596 Delete prototype.
26597 (v850_float_nz_comparison_operator): Likewise.
26598
26599 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26600
26601 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26602 with calls to gen_insvsi/gen_insvdi.
26603
26604 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26605
26606 * config/i386/i386.c (initial_ix86_tune_features): Set up
26607 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26608
26609 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26610 Jakub Jelinek <jakub@redhat.com>
26611
26612 PR tree-optimization/48189
26613 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26614 If nitercst is 0, don't predict the exit edge.
26615
26616 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26617
26618 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26619 in asm_fprintf with reg_names.
26620 (aarch64_print_operand_address): Likewise.
26621 (aarch64_return_addr): Likewise.
26622 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26623
26624 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26625
26626 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26627 (INT_U6_BITS): Likewise.
26628 * config/pa/predicates.md (uint6_operand): New predicate.
26629 (shift5_operand, shift6_operand): Likewise.
26630 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26631 arith32_operand.
26632 (lshrdi3): Use shift6_operand.
26633 (shrpsi4, shrpdi4): New insn patterns.
26634 (extzv): Delete expander.
26635 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26636 predicates in unamed zero extract patterns. Tighten common constraint.
26637 (extv): Delete expander.
26638 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26639 predicates in unamed sign extract patterns. Tighten common constraint.
26640 (insv): Delete expander.
26641 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26642 predicates in unamed insert patterns. Tighten common constraint.
26643 Change uint32_operand predicate to uint6_operand predicate in unamed
26644 DImode pattern to insert constant values of type 1...1xxxx.
26645
26646 2013-01-04 Jan Hubicka <jh@suse.cz>
26647
26648 PR tree-optimization/55823
26649 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26650 issue.
26651
26652 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26653 Uros Bizjak <ubizjak@gmail.com>
26654
26655 PR rtl-optimization/55845
26656 * df-problems.c (can_move_insns_across): Stop scanning at
26657 volatile_insn_p source instruction or give up if
26658 across_from .. across_to range contains any volatile_insn_p
26659 instructions.
26660
26661 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26662
26663 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26664 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26665 Declare.
26666 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26667 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26668
26669 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26670
26671 PR fortran/55341
26672 * asan.c (asan_clear_shadow): New function.
26673 (asan_emit_stack_protection): Use it.
26674
26675 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26676
26677 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26678 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26679 with tab instead of space.
26680
26681 2013-01-08 Nick Clifton <nickc@redhat.com>
26682
26683 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26684 register bank 0 at the start of an interrupt handler.
26685 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26686 MDBH registers.
26687
26688 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26689
26690 * config/aarch64/aarch64-simd.md
26691 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26692 (aarch64_simd_bsl): Likewise.
26693 (aarch64_vcond_internal<mode>): Likewise.
26694 (vcond<mode><mode>): Likewise.
26695 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26696 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26697
26698 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26699
26700 * config/aarch64/aarch64-builtins.c
26701 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26702
26703 2013-01-08 Martin Jambor <mjambor@suse.cz>
26704
26705 PR debug/55579
26706 * tree-sra.c (analyze_access_subtree): Return true also after
26707 potentially creating a debug-only replacement.
26708
26709 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26710
26711 PR middle-end/55890
26712 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26713
26714 PR tree-optimization/54120
26715 * tree-vrp.c (range_fits_type_p): Don't allow
26716 src_precision < precision from signed vr to unsigned_p
26717 if vr->min or vr->max is negative.
26718 (simplify_float_conversion_using_ranges): Test can_float_p
26719 against CODE_FOR_nothing.
26720
26721 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26722 Richard Biener <rguenther@suse.de>
26723
26724 PR middle-end/55851
26725 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26726 types instead of just INTEGER_TYPE types.
26727
26728 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26729
26730 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26731 TF_SIZE): Define.
26732
26733 2013-01-07 Steve Ellcey <sellcey@mips.com>
26734
26735 PR target/42661
26736 * config/mips/mips.opt: Change mad to mmad to match documentation.
26737
26738 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26739
26740 PR target/55897
26741 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26742 .progmemx.data now.
26743
26744 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26745
26746 PR target/55897
26747 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26748 (avr_addrspace_t): Add .section_name field.
26749 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26750 array size.
26751 (avr_addrspace): Same. Initialize .section_name. Remove last
26752 NULL entry. Put __memx into .progmemx.data.
26753 (progmem_section_prefix): Remove.
26754 (avr_asm_init_sections): No need to initialize progmem_section.
26755 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26756 section name prefix.
26757 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26758 retrieve the progmem section.
26759 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26760 boundary to run over avr_addrspace[].
26761 (avr_register_target_pragmas): Ditto.
26762
26763 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26764
26765 * varasm.c (output_constant_def_contents): For asan_protect_global
26766 protected strings, adjust DECL_ALIGN if needed, before testing for
26767 anchored symbols.
26768 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26769 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26770 normal decls.
26771 (output_object_block): For asan protected decls, emit asan padding
26772 after their contents.
26773 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26774 (asan_finish_file): Test it here instead.
26775
26776 2013-01-07 Nick Clifton <nickc@redhat.com>
26777 Matthias Klose <doko@debian.org>
26778 Doug Kwan <dougkwan@google.com>
26779 H.J. Lu <hongjiu.lu@intel.com>
26780
26781 PR driver/55470
26782 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26783
26784 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26785
26786 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26787
26788 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26789
26790 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26791
26792 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26793
26794 PR target/54461
26795 * doc/install.texi (Cross-Compiler-Specific Options): Document
26796 --with-avrlibc.
26797
26798 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26799
26800 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26801 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26802 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26803 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26804 vqmovun_high_s64): Fix source operand number and update copyright.
26805
26806 2013-01-07 Richard Biener <rguenther@suse.de>
26807
26808 PR middle-end/55890
26809 * gimple.h (gimple_call_builtin_p): New overload.
26810 * gimple.c (validate_call): New function.
26811 (gimple_call_builtin_p): Likewise.
26812 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26813 Use gimple_call_builtin_p.
26814 (find_func_clobbers): Likewise.
26815 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26816 (strlen_optimize_stmt): Likewise.
26817
26818 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26819
26820 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26821 (vld1q_dup_*): Likewise.
26822 (vld1_*): Likewise.
26823 (vld1q_*): Likewise.
26824 (vld1_lane_*): Likewise.
26825 (vld1q_lane_*): Likewise.
26826
26827 2013-01-07 Richard Biener <rguenther@suse.de>
26828
26829 * lto-streamer.h (LTO_minor_version): Bump to 2.
26830
26831 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26832
26833 * config/aarch64/aarch64-protos.h
26834 (aarch64_const_double_zero_rtx_p): Rename to...
26835 (aarch64_float_const_zero_rtx_p): ...this.
26836 (aarch64_float_const_representable_p): New.
26837 (aarch64_output_simd_mov_immediate): Likewise.
26838 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26839 move immediate case.
26840 * config/aarch64/aarch64.c
26841 (aarch64_const_double_zero_rtx_p): Rename to...
26842 (aarch64_float_const_zero_rtx_p): ...this.
26843 (aarch64_print_operand): Allow printing of new constants.
26844 (aarch64_valid_floating_const): New.
26845 (aarch64_legitimate_constant_p): Check for valid floating-point
26846 constants.
26847 (aarch64_simd_valid_immediate): Likewise.
26848 (aarch64_vect_float_const_representable_p): New.
26849 (aarch64_float_const_representable_p): Likewise.
26850 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26851 (aarch64_output_simd_mov_immediate): New.
26852 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26853 (*movdf_aarch64): Likewise.
26854 * config/aarch64/constraints.md (Ufc): New.
26855 (Y): call aarch64_float_const_zero_rtx.
26856 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26857
26858 2013-01-07 Richard Biener <rguenther@suse.de>
26859
26860 PR tree-optimization/55888
26861 PR tree-optimization/55862
26862 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26863 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26864 not if it is contained therein.
26865
26866 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26867
26868 * config/avr/t-avr: Typo.
26869
26870 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26871
26872 PR55243
26873 * config/avr/t-avr: Don't automatically rebuild
26874 $(srcdir)/config/avr/t-multilib
26875 $(srcdir)/config/avr/avr-tables.opt
26876 $(srcdir)/doc/avr-mmcu.texi
26877 (avr-mcus): New phony target to build them on request.
26878 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26879 * avr/avr-mcus.def: Adjust comments.
26880
26881 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26882
26883 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26884
26885 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26886
26887 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26888
26889 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26890
26891 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26892
26893 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26894
26895 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26896 to generate profiling.
26897 * config/rs6000/aix64.h (LIB_SPEC): Same.
26898
26899 2013-01-04 Andrew Pinski <apinski@cavium.com>
26900
26901 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26902 New function.
26903 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26904
26905 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26906
26907 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26908 unconditionally.
26909 (ix86_expand_move): Ditto.
26910 (ix86_zero_extend_to_Pmode): Ditto.
26911 (ix86_expand_call): Ditto.
26912 (ix86_expand_special_args_builtin): Ditto.
26913 (ix86_expand_builtin): Ditto.
26914
26915 2013-01-04 Richard Biener <rguenther@suse.de>
26916
26917 PR tree-optimization/55862
26918 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26919 translating them through PHI nodes.
26920
26921 2013-01-04 Martin Jambor <mjambor@suse.cz>
26922
26923 PR tree-optimization/55755
26924 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26925 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26926
26927 2013-01-04 Marek Polacek <polacek@redhat.com>
26928
26929 PR middle-end/55859
26930 * opts.c (default_options_optimization): Clarify error message.
26931
26932 2013-01-04 Richard Biener <rguenther@suse.de>
26933
26934 PR middle-end/55863
26935 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26936 reassociation.
26937
26938 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26939
26940 PR target/53789
26941 * config/pa/pa.md (movsi): Revert previous change.
26942 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26943 references.
26944
26945 2013-01-03 Richard Henderson <rth@redhat.com>
26946
26947 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26948 after eliminating TLS symbols.
26949
26950 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26951
26952 PR bootstrap/50167
26953 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26954 * graphite-poly.c (debug_gmp_value): Likewise.
26955
26956 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26957
26958 PR target/55712
26959 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26960 selected code model, define __code_mode_small__, __code_model_medium__,
26961 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26962 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26963 xchg temporary register with %k. Declare temporary register as
26964 early clobbered.
26965 [__x86_64__]: For medium and large code models, preserve %rbx register.
26966
26967 2013-01-03 Richard Biener <rguenther@suse.de>
26968
26969 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26970 (dump_subscript): Adjust.
26971 (finalize_ddr_dependent): Do not dump redundant info.
26972 (analyze_siv_subscript): Adjust.
26973 (subscript_dependence_tester): Likewise.
26974 (compute_affine_dependence): Likewise.
26975
26976 2013-01-03 Richard Biener <rguenther@suse.de>
26977
26978 Revert
26979 2013-01-03 Richard Biener <rguenther@suse.de>
26980
26981 PR tree-optimization/55857
26982 * tree-vect-stmts.c (vectorizable_load): Do not setup
26983 re-alignment for invariant loads.
26984
26985 2013-01-02 Richard Biener <rguenther@suse.de>
26986
26987 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26988 invariant load do not generate a vector load from the scalar location.
26989
26990 2013-01-03 Richard Biener <rguenther@suse.de>
26991
26992 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26993 for not vectorizing.
26994 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26995 not build INDIRECT_REFs, call get_name once only.
26996 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26997 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26998
26999 2013-01-03 Richard Biener <rguenther@suse.de>
27000
27001 PR tree-optimization/55857
27002 * tree-vect-stmts.c (vectorizable_load): Do not setup
27003 re-alignment for invariant loads.
27004
27005 2013-01-03 Richard Biener <rguenther@suse.de>
27006
27007 PR lto/55848
27008 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27009 prefer a built-in decl.
27010
27011 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27012
27013 * gcc.c (process_command): Update copyright notice dates.
27014 * gcov.c (print_version): Likewise.
27015 * gcov-dump.c (print_version): Likewise.
27016
27017 PR rtl-optimization/55838
27018 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27019 iv0.step, iv1.step and step.
27020
27021 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27022 Marc Glisse <marc.glisse@inria.fr>
27023
27024 PR tree-optimization/55832
27025 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27026 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27027 integer_{one,zero}_node.
27028
27029 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27030
27031 PR debug/54402
27032 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27033 * var-tracking.c (reverse_op): Don't add reverse ops to
27034 VALUEs that have already
27035 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27036
27037 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27038
27039 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27040
27041 2013-01-02 Teresa Johnson <tejohnson@google.com>
27042
27043 * dumpfile.c (dump_loc): Print filename with location.
27044 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27045 new location_t parameter to emit complete unroll message with
27046 new dump framework.
27047 (canonicalize_loop_induction_variables): Compute loops location
27048 and pass to try_unroll_loop_completely.
27049 * loop-unroll.c (report_unroll_peel): New function.
27050 (peel_loops_completely): Use new dump format with location
27051 for main dumpfile message, and invoke report_unroll_peel on success.
27052 (decide_unrolling_and_peeling): Ditto.
27053 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27054 by report_unroll_peel.
27055 (decide_peel_completely): Ditto.
27056 (decide_unroll_constant_iterations): Ditto.
27057 (decide_unroll_runtime_iterations): Ditto.
27058 (decide_peel_simple): Ditto.
27059 (decide_unroll_stupid): Ditto.
27060 * cfgloop.c (get_loop_location): New function.
27061 * cfgloop.h (get_loop_location): Declare.
27062
27063 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27064
27065 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27066 NULL.
27067
27068 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27069
27070 PR middle-end/55198
27071 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27072 BLKmode objects when EXPAND_MEMORY is specified.
27073
27074 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27075
27076 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27077 in loop predicate.
27078 (fold_builtin_cpu): Do not share cpu model decls across statements.
27079
27080 2013-01-02 Jason Merrill <jason@redhat.com>
27081
27082 PR c++/55804
27083 * tree.c (build_array_type_1): Revert earlier change.
27084
27085 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27086
27087 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27088 "cortex-a57".
27089 * config/aarch64/aarch64-tune.md: Re-generate.
27090
27091 2013-01-02 Richard Biener <rguenther@suse.de>
27092
27093 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27094 invariant load do not generate a vector load from the scalar location.
27095
27096 2013-01-02 Richard Biener <rguenther@suse.de>
27097
27098 PR bootstrap/55784
27099 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27100 * configure: Regenerate.
27101
27102 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27103
27104 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27105 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27106 (expand_builtin_int_roundingfn_2): Keep the original target around
27107 for the fallback case.
27108
27109 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27110
27111 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27112 to be clear for sign changes.
27113
27114 2013-01-01 Jan Hubicka <jh@suse.cz>
27115
27116 * ipa-inline-analysis.c: Fix formatting.
27117
27118 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27119
27120 PR tree-optimization/55831
27121 * tree-vect-loop.c (get_initial_def_for_induction): Use
27122 gsi_after_labels instead of gsi_start_bb.
27123 \f
27124 Copyright (C) 2013 Free Software Foundation, Inc.
27125
27126 Copying and distribution of this file, with or without modification,
27127 are permitted in any medium without royalty provided the copyright
27128 notice and this notice are preserved.