1 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
6 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
9 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
10 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
11 INTX_FTYPE_FX, FX_FTYPE_INTX.
12 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
14 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
17 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
19 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
21 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
22 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
25 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
28 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
29 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
30 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
31 (avrxmega6): Increase max flash segments from 5 to 6.
32 * config/avr/t-multilib: Regenerate.
33 * config/avr/avr-tables.opt: Regenerate.
34 * doc/avr-mmcu.texi: Regenerate.
36 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
38 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
39 (avr_device_to_arch): Rename to avr_device_to_ld.
40 (avr_device_to_as): New prototype.
41 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
42 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
43 * config/avr/driver-avr.c (avr_device_to_as): New.
44 (avr_device_to_arch): Rename to avr_device_to_ld.
46 2013-02-27 Jakub Jelinek <jakub@redhat.com>
49 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
50 method on dr_chain and result_chain.
53 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
54 pointer_set_destroy on not_executed_last_iteration.
57 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs
61 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
62 FOR_EACH_DEFINED_FUNCTION when freeing state.
65 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
67 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
71 * ipa-cp.c (decide_whether_version_node): Call vec_free on
72 known_aggs[i].items and release known_aggs vector.
75 * ipa-reference.c (propagate): Free node_info even for alias nodes.
77 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79 * config/microblaze/microblaze.c (microblaze_emit_compare):
80 Use xor for EQ/NE comparisions.
81 * config/microblaze/microblaze.md (cstoresf4): Add constraints
82 (cbranchsf4): Adjust operator to comparison_operator.
84 2013-02-27 Jakub Jelinek <jakub@redhat.com>
87 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
89 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
90 vec_safe_push, always update *slot.
91 (redirect_edge_var_map_clear): Use vec_free.
92 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
93 (free_var_map_entry): Use vec_free.
94 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
95 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
97 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
100 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
101 when the may_trap_p bit of the exprs being merged differs.
102 Reorder tests for speculativeness in the logical and operator.
104 2013-02-27 Jakub Jelinek <jakub@redhat.com>
106 * incpath.c (add_standard_paths): Use reconcat instead of concat
107 where appropriate and avoid leaking memory.
109 * opts.h: Include obstack.h.
110 (opts_concat): New prototype.
111 (opts_obstack): New declaration.
112 * opts.c (opts_concat): New function.
113 (opts_obstack): New variable.
114 (init_options_struct): Call gcc_init_obstack on opts_obstack.
115 (finish_options): Use opts_concat instead of concat
116 and XOBNEWVEC instead of XNEWVEC.
117 * opts-common.c (generate_canonical_option, decode_cmdline_option,
118 generate_option): Likewise.
119 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
120 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
123 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
124 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
126 2013-02-26 Jakub Jelinek <jakub@redhat.com>
129 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
131 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
133 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
134 (arm_block_move_unaligned_straight): Likewise.
135 (arm_adjust_block_mem): Likewise.
137 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
140 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
141 temp, cond and label.
142 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
145 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
146 * config/c6x/c6x.h (dbx_register_map): Update declaration.
149 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
150 of prologue/epilogue functions.
153 * config/tilegx/tilegx.c (tilegx_expand_prologue):
154 Remove unused variable cfa_offset.
155 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
158 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
159 type promotion to unsigned.
162 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
163 for HOST_WIDE_INT of 32 bit / same size as int.
164 (arm_block_move_unaligned_straight): Likewise.
165 (arm_adjust_block_mem): Likewise.
168 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
171 2013-02-26 Marek Polacek <polacek@redhat.com>
173 PR tree-optimization/56426
174 * tree-ssa-loop.c (tree_ssa_loop_init): Always call
177 2013-02-26 Richard Biener <rguenther@suse.de>
180 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
181 unused variable loops.
183 2013-02-26 Jakub Jelinek <jakub@redhat.com>
185 PR tree-optimization/56448
186 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
187 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
188 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
189 later operands of the references, or even first operand for
190 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
192 PR tree-optimization/56443
193 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
194 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
195 to type_for_mode langhook.
197 2013-02-25 Matt Turner <mattst88@gmail.com>
199 * doc/invoke.texi: Document r4700.
201 2013-02-25 Richard Biener <rguenther@suse.de>
203 PR tree-optimization/56175
204 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
206 (simplify_bitwise_binary): ... here. Also guard the conversion
207 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
209 2013-02-25 Catherine Moore <clm@codesourcery.com>
212 2013-02-24 Catherine Moore <clm@codesourcery.com>
213 Maciej W. Rozycki <macro@codesourcery.com>
214 Tom de Vries <tom@codesourcery.com>
215 Nathan Sidwell <nathan@codesourcery.com>
216 Iain Sandoe <iain@codesourcery.com>
217 Nathan Froyd <froydnj@codesourcery.com>
218 Chao-ying Fu <fu@mips.com>
219 * doc/extend.texi: (micromips, nomicromips, nocompression):
220 Document new function attributes.
221 * doc/invoke.texi (minterlink-compressed, mmicromips,
222 m14k, m14ke, m14kec): Document new options.
223 (minterlink-mips16): Update documentation.
224 * doc/md.texi (ZC, ZD): Document new constraints.
225 * configure.ac (gcc_cv_as_micromips): Check if linker
226 supports the .set micromips directive.
227 * configure: Regenerate.
228 * config.in: Regenerate.
229 * config/mips/mips-tables.opt: Regenerate.
230 * config/mips/micromips.md: New file.
231 * constraints.md (ZC, AD): New constraints.
232 * config/mips/predicates.md (movep_src_register): New predicate.
233 (movep_src_operand): New predicate.
234 (non_volatile_mem_operand): New predicate.
235 * config/mips/mips.md (multimem): New type.
236 (length): Differentiate between 17-bit and 18-bit branch offsets.
237 (MOVEP1, MOVEP2): New mode iterator.
238 (mov_<load>l): Use ZC constraint.
239 (mov_<load>r): Likewise.
240 (mov_<store>l): Likewise.
241 (mov_<store>r): Likewise.
242 (*branch_equality<mode>_inverted): Add microMIPS support.
243 (*branch_equality<mode>): Likewise.
244 (*jump_absolute): Likewise.
245 (indirect_jump_<mode>): Likewise.
246 (tablejump_<mode>): Likewise.
247 (<optab>_internal): Likewise.
248 (sibcall_internal): Likewise.
249 (sibcall_value_internal): Likewise.
250 (prefetch): Use constraint ZD.
251 * config/mips/mips.opt (minterlink-compressed): New option.
252 (minterlink-mips16): Now an alias for minterlink-compressed.
253 (mmicromips): New option.
254 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
255 (compare_and_swap_12): Likewise.
256 (sync_add<mode>): Likewise.
257 (sync_<optab>_12): Likewise.
258 (sync_old_<optab>_12): Likewise.
259 (sync_new_<optab>_12): Likewise.
260 (sync_nand_12): Likewise.
261 (sync_old_nand_12): Likewise.
262 (sync_new_nand_12): Likewise.
263 (sync_sub<mode>): Likewise.
264 (sync_old_add<mode>): Likewise.
265 (sync_old_sub<mode>): Likewise.
266 (sync_new_add<mode>): Likewise.
267 (sync_new_sub<mode>): Likewise.
268 (sync_<optab><mode>): Likewise.
269 (sync_old_<optab><mode>): Likewise.
270 (sync_new_<optab><mode>): Likewise.
271 (sync_nand<mode>): Likewise.
272 (sync_old_nand<mode>): Likewise.
273 (sync_new_nand<mode>): Likewise.
274 (sync_lock_test_and_set<mode>): Likewise.
275 (test_and_set_12): Likewise.
276 (atomic_compare_and_swap<mode>): Likewise.
277 (atomic_exchange<mode>_llsc): Likewise.
278 (atomic_fetch_add<mode>_llsc): Likewise.
279 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
280 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
281 (umips_save_restore_pattern_p): Likewise.
282 (umips_load_store_pair_p): Likewise.
283 (umips_output_load_store_pair): Likewise.
284 (umips_movep_target_p): Likewise.
285 (umips_12bit_offset_address_p): Likewise.
286 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
287 (mips_base_mips16): Rename this...
288 (mips_base_compression_flags): ...to this. Update all uses.
289 (mips_attribute_table): Add micromips, nomicromips and nocompression.
290 (mips_mips16_decl_p): Delete.
291 (mips_nomips16_decl_p): Delete.
292 (mips_get_compress_on_flags): New function.
293 (mips_get_compress_off_flags): New function.
294 (mips_get_compress_mode): New function.
295 (mips_get_compress_on_name): New function.
296 (mips_get_compress_off_name): New function.
297 (mips_insert_attributes): Support multiple compression types.
298 (mips_merge_decl_attributes): Likewise.
299 (umips_12bit_offset_address_p): New function.
300 (mips_start_function_definition): Emit .set micromips directive.
301 (mips_call_may_need_jalx_p): New function.
302 (mips_function_ok_for_sibcall): Add microMIPS support.
303 (mips_print_operand_punctuation): Support short delay slots and
305 (umips_swm_mask, umips_swm_encoding): New.
306 (umips_build_save_restore): New function.
307 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
308 (was_mips16_p): Remove.
309 (old_compression_mode): New.
310 (mips_set_compression_mode): New function.
311 (mips_set_current_function): Add microMIPS support.
312 (mips_option_override): Likewise.
313 (umips_save_restore_pattern_p): New function.
314 (umips_output_save_restore): New function.
315 (umips_load_store_pair_p_1): New function.
316 (umips_load_store_pair_p): New function.
317 (umips_output_load_store_pair_1): New function.
318 (umips_output_load_store_pair): New function.
319 (umips_movep_target_p) New function.
320 (mips_prepare_pch_save): Add microMIPS support.
321 * config/mips/mips.h (TARGET_COMPRESSION): New.
322 (TARGET_CPU_CPP_BUILTINS): Update macro
323 to use new compression flags and to support microMIPS.
324 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
325 (MIPS_ARCH_FLOAT_SPEC): Likewise.
326 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
327 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
328 (ASM_SPEC): Support mmicromips and mno-micromips.
329 (M16STORE_REG_P): New macro.
330 (MIPS_CALL): Support TARGET_MICROMIPS.
331 (MICROMIPS_J): New macro.
332 (mips_base_mips16): Rename this...
333 (mips_base_compression_flags): ...to this.
334 (UMIPS_12BIT_OFFSET_P): New macro.
335 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
336 (MULTILIB_DIRNAMES): Likewise.
338 2013-02-25 Tom de Vries <tom@codesourcery.com>
340 PR rtl-optimization/56131
341 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
342 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
343 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
345 2013-02-25 Tobias Burnus <burnus@net-b.de>
347 * doc/invoke.texi (-fsanitize=): Move from optimization
348 to debugging options.
350 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
352 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
354 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
355 Alexander Monakov <amonakov@ispras.ru>
358 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
359 flush pending lists also on non-jumps. Adjust comment.
361 2013-02-24 Catherine Moore <clm@codesourcery.com>
362 Maciej W. Rozycki <macro@codesourcery.com>
363 Tom de Vries <tom@codesourcery.com>
364 Nathan Sidwell <nathan@codesourcery.com>
365 Iain Sandoe <iain@codesourcery.com>
366 Nathan Froyd <froydnj@codesourcery.com>
367 Chao-ying Fu <fu@mips.com>
369 * doc/extend.texi: (micromips, nomicromips, nocompression):
370 Document new function attributes.
371 * doc/invoke.texi (minterlink-compressed, mmicromips,
372 m14k, m14ke, m14kec): Document new options.
373 (minterlink-mips16): Update documentation.
374 * doc/md.texi (ZC, ZD): Document new constraints.
375 * configure.ac (gcc_cv_as_micromips): Check if linker
376 supports the .set micromips directive.
377 * configure: Regenerate.
378 * config.in: Regenerate.
379 * config/mips/mips-tables.opt: Regenerate.
380 * config/mips/micromips.md: New file.
381 * constraints.md (ZC, AD): New constraints.
382 * config/mips/predicates.md (movep_src_register): New predicate.
383 (movep_src_operand): New predicate.
384 (non_volatile_mem_operand): New predicate.
385 * config/mips/mips.md (multimem): New type.
386 (length): Differentiate between 17-bit and 18-bit branch offsets.
387 (MOVEP1, MOVEP2): New mode iterator.
388 (mov_<load>l): Use ZC constraint.
389 (mov_<load>r): Likewise.
390 (mov_<store>l): Likewise.
391 (mov_<store>r): Likewise.
392 (*branch_equality<mode>_inverted): Add microMIPS support.
393 (*branch_equality<mode>): Likewise.
394 (*jump_absolute): Likewise.
395 (indirect_jump_<mode>): Likewise.
396 (tablejump_<mode>): Likewise.
397 (<optab>_internal): Likewise.
398 (sibcall_internal): Likewise.
399 (sibcall_value_internal): Likewise.
400 (prefetch): Use constraint ZD.
401 * config/mips/mips.opt (minterlink-compressed): New option.
402 (minterlink-mips16): Now an alias for minterlink-compressed.
403 (mmicromips): New option.
404 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
405 (compare_and_swap_12): Likewise.
406 (sync_add<mode>): Likewise.
407 (sync_<optab>_12): Likewise.
408 (sync_old_<optab>_12): Likewise.
409 (sync_new_<optab>_12): Likewise.
410 (sync_nand_12): Likewise.
411 (sync_old_nand_12): Likewise.
412 (sync_new_nand_12): Likewise.
413 (sync_sub<mode>): Likewise.
414 (sync_old_add<mode>): Likewise.
415 (sync_old_sub<mode>): Likewise.
416 (sync_new_add<mode>): Likewise.
417 (sync_new_sub<mode>): Likewise.
418 (sync_<optab><mode>): Likewise.
419 (sync_old_<optab><mode>): Likewise.
420 (sync_new_<optab><mode>): Likewise.
421 (sync_nand<mode>): Likewise.
422 (sync_old_nand<mode>): Likewise.
423 (sync_new_nand<mode>): Likewise.
424 (sync_lock_test_and_set<mode>): Likewise.
425 (test_and_set_12): Likewise.
426 (atomic_compare_and_swap<mode>): Likewise.
427 (atomic_exchange<mode>_llsc): Likewise.
428 (atomic_fetch_add<mode>_llsc): Likewise.
429 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
430 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
431 (umips_save_restore_pattern_p): Likewise.
432 (umips_load_store_pair_p): Likewise.
433 (umips_output_load_store_pair): Likewise.
434 (umips_movep_target_p): Likewise.
435 (umips_12bit_offset_address_p): Likewise.
436 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
437 (mips_base_mips16): Rename this...
438 (mips_base_compression_flags): ...to this. Update all uses.
439 (mips_attribute_table): Add micromips, nomicromips and nocompression.
440 (mips_mips16_decl_p): Delete.
441 (mips_nomips16_decl_p): Delete.
442 (mips_get_compress_on_flags): New function.
443 (mips_get_compress_off_flags): New function.
444 (mips_get_compress_mode): New function.
445 (mips_get_compress_on_name): New function.
446 (mips_get_compress_off_name): New function.
447 (mips_insert_attributes): Support multiple compression types.
448 (mips_merge_decl_attributes): Likewise.
449 (umips_12bit_offset_address_p): New function.
450 (mips_start_function_definition): Emit .set micromips directive.
451 (mips_call_may_need_jalx_p): New function.
452 (mips_function_ok_for_sibcall): Add microMIPS support.
453 (mips_print_operand_punctuation): Support short delay slots and
455 (umips_swm_mask, umips_swm_encoding): New.
456 (umips_build_save_restore): New function.
457 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
458 (was_mips16_p): Remove.
459 (old_compression_mode): New.
460 (mips_set_compression_mode): New function.
461 (mips_set_current_function): Add microMIPS support.
462 (mips_option_override): Likewise.
463 (umips_save_restore_pattern_p): New function.
464 (umips_output_save_restore): New function.
465 (umips_load_store_pair_p_1): New function.
466 (umips_load_store_pair_p): New function.
467 (umips_output_load_store_pair_1): New function.
468 (umips_output_load_store_pair): New function.
469 (umips_movep_target_p) New function.
470 (mips_prepare_pch_save): Add microMIPS support.
471 * config/mips/mips.h (TARGET_COMPRESSION): New.
472 (TARGET_CPU_CPP_BUILTINS): Update macro
473 to use new compression flags and to support microMIPS.
474 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
475 (MIPS_ARCH_FLOAT_SPEC): Likewise.
476 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
477 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
478 (ASM_SPEC): Support mmicromips and mno-micromips.
479 (M16STORE_REG_P): New macro.
480 (MIPS_CALL): Support TARGET_MICROMIPS.
481 (MICROMIPS_J): New macro.
482 (mips_base_mips16): Rename this...
483 (mips_base_compression_flags): ...to this.
484 (UMIPS_12BIT_OFFSET_P): New macro.
485 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
486 (MULTILIB_DIRNAMES): Likewise.
488 2013-02-24 Jakub Jelinek <jakub@redhat.com>
491 * target-globals.c (save_target_globals): For init_reg_sets and
492 target_reinit remporarily set this_fn_optabs to this_target_optabs.
494 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
496 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
497 * config/aarch64/t-aarch64
498 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
500 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
503 * lra-constraints.c (process_alt_operands): Reload operand
504 conflicting with earlier clobber only if no more other conflicting
507 2013-02-22 Jakub Jelinek <jakub@redhat.com>
510 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
511 if not linking a shared library.
513 2013-02-22 Seth LaForge <sethml@google.com>
515 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
517 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
519 * config/arm/arm.md (split for extendsidi): Update condition.
520 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
521 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
522 (qhs_zextenddi_cstr): Likewise.
524 2013-02-21 Jakub Jelinek <jakub@redhat.com>
527 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
528 avoid signed wrapping.
529 (expand_mult): Handle properly multiplication by
530 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
531 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
532 in the compiler if coeff is HOST_WIDE_INT_MIN.
533 (expand_divmod): Don't make ext_op1 static, change it's type to
534 uhwi. Avoid undefined behavior in -INTVAL (op1).
536 PR rtl-optimization/50339
537 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
539 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
540 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
541 into splitting_ashiftrt field.
542 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
544 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
547 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
550 * trans-mem.c (execute_tm_mark): Do not expand transactions that
551 are sure to go irrevocable.
553 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
555 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
556 scalars are valid operands.
558 2013-02-21 Martin Jambor <mjambor@suse.cz>
560 PR tree-optimization/56310
561 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
562 only matching indices and non-negative final offsets.
563 (intersect_aggregates_with_edge): Pass src_idx to
564 agg_replacements_to_vector. Pass src_idx insstead of index to
565 intersect_with_agg_replacements.
567 2013-02-21 Martin Jambor <mjambor@suse.cz>
569 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
570 instead of hard-wired defaults.
572 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
574 * doc/invoke.texi (MIPS Options): Update documentation of the
575 floating-point multiply-accumulate instruction restrictions.
577 2013-02-21 Kostya Serebryany <kcc@google.com>
579 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
580 asan_shadow_offset on x86_64 linux.
582 2013-02-21 Richard Biener <rguenther@suse.de>
584 PR tree-optimization/56415
586 2013-02-11 Richard Biener <rguenther@suse.de>
588 PR tree-optimization/56273
589 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
592 2013-02-21 Jakub Jelinek <jakub@redhat.com>
595 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
599 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
600 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
602 2013-02-20 Jan Hubicka <jh@suse.cz>
604 PR tree-optimization/56265
605 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph when target is
606 referenced for firs ttime.
608 2013-02-20 Richard Biener <rguenther@suse.de>
610 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
611 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
612 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
613 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
615 (rename_ssa_copies): Do not remove unused locals.
616 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
617 * tree-ssanames.c (pass_release_ssa_names): Remove unused
619 * passes.c (execute_function_todo): Do not schedule unused locals
620 removal if cleanup_tree_cfg did something.
621 * tree-ssa-live.c (remove_unused_locals): Dump statistics
622 about the number of removed locals.
624 2013-02-20 Richard Biener <rguenther@suse.de>
626 PR tree-optimization/56398
627 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip
630 2013-02-20 Martin Jambor <mjambor@suse.cz>
632 PR tree-optimization/55334
633 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
634 restricted pointers to arrays.
636 2013-02-20 Richard Biener <rguenther@suse.de>
637 Jakub Jelinek <jakub@redhat.com>
639 PR tree-optimization/56396
640 * tree-ssa-ccp.c (n_const_val): New static variable.
641 (get_value): Return NULL for SSA names we don't have a lattice
643 (ccp_initialize): Initialize n_const_val.
644 * tree-ssa-copy.c (n_copy_of): New static variable.
645 (init_copy_prop): Initialize n_copy_of.
646 (get_value): Return NULL_TREE for SSA names we don't have a
649 2013-02-20 Martin Jambor <mjambor@suse.cz>
651 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
653 2013-02-20 Richard Biener <rguenther@suse.de>
655 * genpreds.c (write_lookup_constraint): Do not compare first
656 letter of the constraint again.
658 2013-02-20 Richard Biener <rguenther@suse.de>
660 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
662 (get_use_iv_cost): Terminate hashtable walk when coming across
665 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
667 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
668 reassociation for avx2 targets.
670 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
672 * config/microblaze/microblaze.c: microblaze_has_clz = 0
673 Add version check for v8.10.a to enable microblaze_has_clz
674 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
675 version and TARGET_PATTERN_COMPARE check
676 * config/microblaze/microblaze.md: New clzsi2 instruction
678 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
680 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
681 function before branching.
683 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
685 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
687 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
689 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
693 * sel-sched.c: Include ira.h.
694 (implicit_clobber_conflict_p): New function.
695 (moveup_expr): Use it.
696 * Makefile.in (sel-sched.o): Depend on ira.h.
698 2013-02-19 Richard Biener <rguenther@suse.de>
700 PR tree-optimization/56384
701 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
702 (vn_hash_type): Split out from ...
703 (vn_hash_constant_with_type): ... here.
704 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
705 (vn_phi_eq): Compare types from vn_phi_s structure.
706 (vn_phi_lookup): Populate vn_phi_s type.
707 (vn_phi_insert): Likewise.
709 2013-02-19 Jakub Jelinek <jakub@redhat.com>
711 PR tree-optimization/56350
712 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
713 if haven't found reduction or nested cycle operand, rather than
714 asserting we must find it.
716 PR tree-optimization/56381
717 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
720 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
721 Jakub Jelinek <jakub@redhat.com>
724 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
725 (swap_optab_enable): Same.
726 (init_all_optabs): Use argument instead of global.
727 * tree.h (struct tree_optimization_option): New field
729 * expr.h (init_all_optabs): Add argument to prototype.
730 (TREE_OPTIMIZATION_OPTABS): New.
731 (save_optabs_if_changed): Protoize.
732 * optabs.h: Declare this_fn_optabs.
733 * optabs.c (save_optabs_if_changed): New.
734 Declare this_fn_optabs.
735 (init_optabs): Add argument to init_all_optabs() call.
736 * function.c (invoke_set_current_function_hook): Handle per
738 * function.h (struct function): New field optabs.
739 * config/mips/mips.c (mips_set_mips16_mode): Handle when
740 optimization_current_node has changed.
741 * target-globals.h (save_target_globals_default_opts): Protoize.
742 * target-globals.c (save_target_globals_default_opts): New.
744 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
747 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
748 registers %fr12 and %fr12R as call used.
751 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
752 and HImode, require all displacements to be an integer multiple of their
754 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
755 only allow QImode and HImode when reload is in progress and strict is
756 true. Likewise for symbolic addresses. Use base14_operand to check
757 displacements in REG+BASE addresses.
759 2013-02-18 Richard Biener <rguenther@suse.de>
761 PR tree-optimization/56366
762 * tree-vect-loop.c (get_initial_def_for_induction): Properly
763 handle sign-conversion of outer-loop initial induction value.
765 2013-02-18 Richard Biener <rguenther@suse.de>
768 * cfghooks.c (merge_blocks): If we merge a latch into another
769 block adjust references to it.
770 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
771 (verify_loop_structure): Verify that a recorded latch is in fact
774 2013-02-18 Richard Biener <rguenther@suse.de>
776 PR tree-optimization/56321
777 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
778 order SSA name release and virtual operand unlinking.
780 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
782 * config/microblaze/microblaze.md (save_stack_block): Define.
783 (restore_stack_block): Likewise.
785 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
787 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
788 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
789 * config/microblaze/microblaze.c (microblaze_option_override):
790 Bail out early for PIC modes when target does not support PIC.
792 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
794 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
795 Replace with a microblaze version.
796 (microblaze_trampoline_init): Adapt for microblaze.
797 * gcc/config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
800 2013-02-16 Jakub Jelinek <jakub@redhat.com>
801 Dodji Seketeli <dodji@redhat.com>
804 * asan.c (get_mem_refs_of_builtin_call): White space and style
806 (instrument_mem_region_access): Do not forget to always put
807 instrumentation of the of 'base' and 'base + len' in a "if (len !=
808 0) statement, even for cases where either 'base' or 'base + len'
809 are not instrumented -- because they have been previously
810 instrumented. Simplify the logic by putting all the statements
811 instrument 'base + len' inside a sequence, and then insert that
812 sequence right before the current insertion point. Then, to
813 instrument 'base + len', just get an iterator on that statement.
814 And do not forget to update the pointer to iterator the function
815 received as argument.
817 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
819 PR rtl-optimization/56348
820 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
822 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
824 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
825 (clean_graph_dump_file): Pass base to start_graph_dump.
827 2013-02-14 Richard Henderson <rth@redhat.com>
830 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
832 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
834 * collect2-aix.h: Define F_LOADONLY.
836 2013-02-14 Richard Biener <rguenther@suse.de>
839 * varasm.c (output_constant_def_1): Get the decl representing
840 the constant as argument.
841 (output_constant_def): Wrap output_constant_def_1.
842 (make_decl_rtl): Use output_constant_def_1 with the decl
843 representing the constant.
844 (build_constant_desc): Optionally re-use a decl already
845 representing the constant.
846 (tree_output_constant_def): Adjust.
848 2013-02-14 Dodji Seketeli <dodji@redhat.com>
851 * asan.c (instrument_builtin_call): Really put the length of the
852 second source argument into src1_len.
854 2013-02-13 Jakub Jelinek <jakub@redhat.com>
856 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
857 argument. If it is false, don't create edge from then_bb to
859 (insert_if_then_before_iter): Pass true to it.
860 (build_check_stmt): Pass false to it.
861 (transform_statements): Flush hash table only on extended basic
862 block boundaries, rather than at the beginning of every bb.
863 Don't flush hash table on nonfreeing_call_p calls.
864 * tree-flow.h (nonfreeing_call_p): New prototype.
865 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
867 2013-02-13 David S. Miller <davem@davemloft.net>
869 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
871 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
874 * ira.c (max_regno_before_ira): Move from ...
876 (fix_reg_equiv_init): Use max_regno_before_ira instead of
879 2013-02-13 Jakub Jelinek <jakub@redhat.com>
881 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
883 2013-02-13 Richard Biener <rguenther@suse.de>
886 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
889 2013-02-13 Richard Biener <rguenther@suse.de>
891 * loop-init.c (loop_optimizer_init): Clear loop state when
892 re-initializing preserved loops.
893 * loop-unswitch.c (unswitch_single_loop): Return whether
894 we unswitched the loop. Do not verify loop state here.
895 (unswitch_loops): When we unswitched a loop discover new
898 2013-02-13 Kostya Serebryany <kcc@google.com>
900 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset on x86_64
902 * sanitizer.def: Rename __asan_init to __asan_init_v1.
904 2013-02-12 Dodji Seketeli <dodji@redhat.com>
906 Avoid instrumenting duplicated memory access in the same basic block
907 * Makefile.in (asan.o): Add new dependency on hash-table.h
908 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
909 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
910 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
911 (free_mem_ref_resources, has_mem_ref_been_instrumented)
912 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
913 (get_mem_ref_of_assignment): New functions.
914 (get_mem_refs_of_builtin_call): Extract from
915 instrument_builtin_call and tweak a little bit to make it fit with
917 (instrument_builtin_call): Use the new
918 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
919 of is_gimple_builtin_call.
920 (instrument_derefs, instrument_mem_region_access): Insert the
921 instrumented memory reference into the hash table.
922 (maybe_instrument_assignment): Renamed instrument_assignment into
923 this, and change it to advance the iterator when instrumentation
924 actually happened and return true in that case. This makes it
925 homogeneous with maybe_instrument_assignment, and thus give a
926 chance to callers to be more 'regular'.
927 (transform_statements): Clear the memory reference hash table
928 whenever we enter a new BB, when we cross a function call, or when
929 we are done transforming statements. Use
930 maybe_instrument_assignment instead of instrumentation. No more
931 need to special case maybe_instrument_assignment and advance the
932 iterator after calling it; it's now handled just like
933 maybe_instrument_call. Update comment.
935 2013-02-13 Richard Biener <rguenther@suse.de>
937 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
938 Fix loop discovery code.
940 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
943 * lra-constraints.c (process_alt_operands): Match early clobber
944 operand with itself. Check conflicts with earlyclobber only if
945 the operand is not reloaded. Prefer to reload conflicting operand
946 if earlyclobber and matching operands are the same.
948 2013-02-12 Richard Biener <rguenther@suse.de>
951 * lto-streamer-out.c (write_symbol): Do not output symbols
952 for hard register variables.
954 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
957 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
958 (umulsidi3_insn, mulsidi3_insn): New insns.
960 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
962 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
963 (struct tune_params): Add vec_costs field.
964 * config/arm/arm.c (arm_builtin_vectorization_cost)
965 (arm_add_stmt_cost): New functions.
966 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
967 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
968 (arm_default_vec_cost): New struct of type cpu_vec_costs.
969 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
970 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
971 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
972 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
974 2013-02-12 Richard Biener <rguenther@suse.de>
977 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
978 decls again if possible.
980 2013-02-12 Richard Biener <rguenther@suse.de>
983 * tree-ssa.c (verify_ssa_name): Fix check, move
984 SSA_NAME_IN_FREE_LIST check up.
986 2013-02-12 Jakub Jelinek <jakub@redhat.com>
987 Steven Bosscher <steven@gcc.gnu.org>
989 PR rtl-optimization/56151
990 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
991 equal to op0 or op1, and last_insn pattern is CODE operation
992 with MEM dest and one of the operands matches that MEM.
994 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
996 * doc/extend.texi: Document Function Multiversioning and "default"
997 parameter string to target attribute.
998 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
999 target attribute parameter is "default".
1000 (ix86_compare_version_priority): Remove checks for target attribute.
1001 (ix86_mangle_function_version_assembler_name): Change error to sorry.
1002 Remove check for target attribute equal to NULL. Add assert.
1003 (ix86_generate_version_dispatcher_body): Change error to sorry.
1005 2013-02-11 Iain Sandoe <iain@codesourcery.com>
1006 Jack Howarth <howarth@bromo.med.uc.edu>
1007 Patrick Marlier <patrick.marlier@gmail.com>
1010 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
1011 define ENDFILE_SPEC as TM_DESTRUCTOR.
1012 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
1014 2013-02-11 Alexander Potapenko <glider@google.com>
1015 Jack Howarth <howarth@bromo.med.uc.edu>
1016 Jakub Jelinek <jakub@redhat.com>
1019 * config/darwin.c (cdtor_record): Rename ctor_record.
1020 (sort_cdtor_records): Rename sort_ctor_records.
1021 (finalize_dtors): New routine to sort destructors by
1022 priority before use in assemble_integer.
1023 (machopic_asm_out_destructor): Use finalize_dtors if needed.
1025 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
1027 PR rtl-optimization/56275
1028 * simplify-rtx.c (avoid_constant_pool_reference): Check that
1029 offset is non-negative and less than cmode size before
1030 calling simplify_subreg.
1032 2013-02-11 Richard Biener <rguenther@suse.de>
1034 PR tree-optimization/56264
1035 * cfgloop.h (fix_loop_structure): Adjust prototype.
1036 * loop-init.c (fix_loop_structure): Return the number of
1037 newly discovered loops.
1038 * tree-cfgcleanup.c (repair_loop_structures): When new loops
1039 are discovered, do a full loop-closed SSA rewrite.
1041 2013-02-11 Richard Biener <rguenther@suse.de>
1043 PR tree-optimization/56273
1044 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
1046 (check_array_ref): Fix missing newline in dumps.
1047 (search_for_addr_array): Likewise.
1049 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
1051 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
1053 2013-02-09 Jakub Jelinek <jakub@redhat.com>
1056 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
1058 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
1060 PR rtl-optimization/56246
1061 * lra-constraints.c (simplify_operand_subreg): Try to reuse
1063 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
1064 constraints are satisfied.
1066 2013-02-08 Jeff Law <law@redhat.com>
1069 * emit-rtl.c (reg_is_parm_p): New function.
1070 * regs.h (reg_is_parm_p): New prototype.
1071 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
1072 callee-clobbered registers.
1074 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
1077 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
1078 If there is no implicit builtin declaration, just return NULL.
1080 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
1082 * config/i386/sse.md (FMAMODEM): New mode iterator.
1083 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
1084 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
1086 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
1088 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
1089 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
1090 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
1092 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
1094 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
1095 (microblaze*-*-elf): Likewise.
1096 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
1098 * config/microblaze/microblaze-c.c: Add builtin defines for
1099 _LITTLE_ENDIAN and _BIG_ENDIAN.
1100 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
1101 add to TARGET_DEFAULT flags.
1102 Expand ASM_SPEC and LINK_SPEC.
1103 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
1104 * config/microblaze/microblaze.md: Update extendsidi2 and
1105 movdi_internal instructions to use low-order / high-order reg
1107 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
1108 options and inversemask / mask of LITTLE_ENDIAN.
1109 * config/microblaze/t-microblaze: Expand multilib options to
1110 include mlittle-endian (le) and update exceptions patterns.
1112 2013-02-08 Jakub Jelinek <jakub@redhat.com>
1114 PR rtl-optimization/56195
1115 * lra-constraints.c (get_reload_reg): Don't reuse regs
1116 if they have smaller mode than requested, if they have
1117 wider mode than requested, try to return a SUBREG.
1119 PR tree-optimization/56250
1120 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
1121 if type is unsigned and code isn't MULT_EXPR.
1123 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
1125 PR tree-optimization/56064
1126 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
1127 bits according to mode.
1128 * fixed-value.h (fixed_from_double_int)
1129 (const_fixed_from_double_int): Adjust comments.
1131 2013-02-08 Richard Biener <rguenther@suse.de>
1134 * lto-streamer.h (struct data_in): Remove current_file, current_line
1135 and current_col members.
1136 * lto-streamer-out.c (lto_output_location): Stream changed bits
1137 en-block for efficiency.
1138 * lto-streamer-in.c (clear_line_info): Remove.
1139 (lto_input_location): Cache current file, line and column
1140 globally via local statics. Read changed bits en-block.
1141 (input_function): Do not call clear_line_info.
1142 (lto_read_body): Likewise.
1143 (lto_input_toplevel_asms): Likewise.
1145 2013-02-08 Michael Matz <matz@suse.de>
1147 PR tree-optimization/52448
1148 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
1149 (nt_call_phase): New static.
1150 (add_or_mark_expr): Only mark accesses with newer phase than any
1152 (nonfreeing_call_p): New.
1153 (nt_init_block): Update nt_call_phase, mark blocks as visited.
1154 (nt_fini_block): Keep blocks marked as visited.
1155 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
1157 2013-02-08 Richard Biener <rguenther@suse.de>
1159 * ira.c (ira): Free broken dominator information.
1161 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
1163 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
1165 2013-02-08 Marek Polacek <polacek@redhat.com>
1167 * cfgloop.c (verify_loop_structure): Add more checking of headers.
1169 2013-02-08 Richard Biener <rguenther@suse.de>
1172 * cfgloop.h (flow_loops_find): Adjust.
1173 (bb_loop_header_p): Declare.
1174 * cfgloop.c (bb_loop_header_p): New function split out from ...
1175 (flow_loops_find): ... here. Adjust function signature,
1176 support incremental loop structure update.
1177 (verify_loop_structure): Cleanup. Verify a loop is a loop.
1178 * cfgloopmanip.c (fix_loop_structure): Move ...
1179 * loop-init.c (fix_loop_structure): ... here.
1180 (apply_loop_flags): Split out from ...
1181 (loop_optimizer_init): ... here.
1182 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
1183 in incremental mode, only remove dead loops here.
1185 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
1188 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
1189 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
1190 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
1191 (*round<mode>3.libgcc): New insns for fixed-modes.
1192 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
1193 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
1194 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
1195 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
1196 implementations. Define to __builtin_avr_absFX,
1197 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
1198 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
1199 __builtin_avr_countlsFX, respectively.
1200 * config/avr/avr-c.c (target.h): Include it.
1201 (enum avr_builtin_id): New enum.
1202 (avr_resolve_overloaded_builtin): New static function.
1203 (avr_register_target_pragmas): Use it to set
1204 targetm.resolve_overloaded_builtin.
1205 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
1206 tree nodes used by DEF_BUILTIN.
1207 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
1208 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
1209 <AVR_BUILTIN_xxBITS>: Same.
1211 2013-02-08 Richard Biener <rguenther@suse.de>
1213 * cfgloop.c (verify_loop_structure): Properly handle
1214 a loop exiting to another loop header.
1215 * ira-int.h (ira_loops): Remove.
1216 * ira.c (ira_loops): Remove.
1217 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
1218 (do_reload): Use loop_optimizer_finalize.
1219 * ira-build.c (create_loop_tree_nodes): Use get_loops and
1220 number_of_loops to access the loop tree.
1221 (more_one_region_p): Likewise.
1222 (finish_loop_tree_nodes): Likewise.
1223 (rebuild_regno_allocno_maps): Likewise.
1224 (mark_loops_for_removal): Likewise.
1225 (mark_all_loops_for_removal): Likewise.
1226 (remove_unnecessary_regions): Likewise.
1227 (ira_build): Likewise.
1228 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
1230 2013-02-08 Richard Biener <rguenther@suse.de>
1232 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
1233 * ipa-pure-const.c (analyze_function): Avoid calling
1234 mark_irreducible_loops twice.
1235 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
1237 2013-02-07 David S. Miller <davem@davemloft.net>
1239 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
1241 * var-tracking.c (vt_add_function_parameter): Test the presence of
1242 HAVE_window_save properly and do not remap argument registers when
1243 we have a leaf function.
1245 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
1248 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
1250 * config/i386/i386.c (ix86_print_operand): Ditto.
1252 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
1254 * lra-constraints.c (process_alt_operands): Fix recently added comment.
1256 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
1258 PR rtl-optimization/56225
1259 * lra-constraints.c (process_alt_operands): Check that reload hard
1260 reg can hold value for strict_low_part.
1262 2013-02-07 Jakub Jelinek <jakub@redhat.com>
1265 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
1266 dwarf2out_end_function.
1267 (in_first_function_p, maybe_at_text_label_p,
1268 first_loclabel_num_not_at_text_label): New variables.
1269 (dwarf2out_var_location): In the first function find out
1270 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
1271 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
1274 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
1276 PR rtl-optimization/56178
1277 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
1278 SUBREG of a register. Tidy up related block of code.
1279 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
1280 note if the source is a register or a SUBREG of a register.
1282 2013-02-07 Jakub Jelinek <jakub@redhat.com>
1285 * config/rs6000/rs6000.md (ptrm): New mode attr.
1286 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
1287 call_value_indirect_aix<pttrsize>,
1288 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
1291 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
1293 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
1294 if -bnortl. Convert to strcmp and strncmp.
1296 2013-02-07 Alan Modra <amodra@gmail.com>
1299 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
1300 addresses won't wrap when offsetting.
1301 (rs6000_secondary_reload): Provide secondary reloads needed for
1302 wrapping LO_SUM addresses.
1304 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
1306 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
1307 MACH, just __MACH__.
1309 2013-02-06 Richard Biener <rguenther@suse.de>
1311 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
1312 instead of calling fix_loop_structure.
1314 2013-02-06 Jakub Jelinek <jakub@redhat.com>
1317 * omp-low.c (use_pointer_for_field): Return false if
1318 lower_send_shared_vars doesn't generate any copy-out code.
1320 2013-02-06 Tom de Vries <tom@codesourcery.com>
1322 PR rtl-optimization/56131
1323 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
1324 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
1325 of the label is NULL. Add comment.
1327 2013-02-05 Jakub Jelinek <jakub@redhat.com>
1329 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
1332 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
1333 (STATIC_LIBTSAN_LIBS): Likewise.
1334 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
1335 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
1336 is defined, don't add anything else beyond that.
1337 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
1338 (LINK_COMMAND_SPEC): Use them.
1340 PR tree-optimization/56205
1341 * tree-stdarg.c (check_all_va_list_escapes): Return true if
1342 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
1343 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
1345 2013-02-05 Richard Biener <rguenther@suse.de>
1347 PR tree-optimization/53342
1348 PR tree-optimization/53185
1349 * tree-vectorizer.h (vect_check_strided_load): Remove.
1350 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
1351 not disallow peeling for vectorized strided loads.
1352 (vect_check_strided_load): Make static and simplify.
1353 (vect_analyze_data_refs): Adjust.
1354 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
1355 correctly when vectorizing strided loads.
1357 2013-02-05 Richard Biener <rguenther@suse.de>
1359 * doc/install.texi: Refer to ISL, not PPL.
1361 2013-02-05 Jan Hubicka <jh@suse.cz>
1363 PR tree-optimization/55789
1364 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
1366 2013-02-05 Jan Hubicka <jh@suse.cz>
1368 PR tree-optimization/55789
1369 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
1370 the dead call anyway.
1372 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
1375 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
1377 2013-02-04 Alexander Potapenko <glider@google.com>
1378 Jack Howarth <howarth@bromo.med.uc.edu>
1379 Jakub Jelinek <jakub@redhat.com>
1382 * config/darwin.c (sort_ctor_records): Stabilized qsort
1383 on constructor priority by using original position.
1384 (finalize_ctors): New routine to sort constructors by
1385 priority before use in assemble_integer.
1386 (machopic_asm_out_constructor): Use finalize_ctors if needed.
1388 2013-02-04 Jakub Jelinek <jakub@redhat.com>
1391 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
1392 about visibility on artificial decls.
1393 * config/sol2.c (solaris_assemble_visibility): Likewise.
1395 2013-02-04 Kai Tietz <ktietz@redhat.com>
1398 * config/i386/i386.c (function_value_ms_64): Add additional valtype
1399 argument and improve checking of return-argument types for 16-byte
1401 (ix86_function_value_1): Add additional valtype argument on call
1402 of function_value_64.
1403 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
1404 handling infunction_value_64 function.
1406 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1408 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
1410 2013-02-04 Richard Biener <rguenther@suse.de>
1412 PR tree-optimization/56188
1413 * tree-ssa-structalias.c (label_visit): Consider case with
1414 initially non-empty points-to set.
1415 (perform_var_substitution): Dump node mapping and clean up.
1417 2013-02-04 Richard Guenther <rguenther@suse.de>
1420 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
1421 node prevail as last resort.
1422 (lto_symtab_merge_decls): Remove guard on LTRANS here.
1423 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
1425 2013-02-04 Richard Biener <rguenther@suse.de>
1427 PR tree-optimization/56113
1428 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
1430 (equiv_class_lookup_or_add): ... this.
1431 (label_visit): Adjust and fix error in previous patch.
1432 (perform_var_substitution): Adjust.
1434 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
1436 * config/sh/divtab.c: Fix formatting and comments throughout the file.
1437 * config/sh/sh4-300.md: Likewise.
1438 * config/sh/sh4a.md: Likewise.
1439 * config/sh/constraints.md: Likewise.
1440 * config/sh/sh.md: Likewise.
1441 * config/sh/netbsd-elf.h: Likewise.
1442 * config/sh/predicates.md: Likewise.
1443 * config/sh/sh-protos.h: Likewise.
1444 * config/sh/ushmedia.h: Likewise.
1445 * config/sh/linux.h: Likewise.
1446 * config/sh/sh.c: Likewise.
1447 * config/sh/superh.h: Likewise.
1448 * config/sh/elf.h: Likewise.
1449 * config/sh/sh4.md: Likewise.
1450 * config/sh/sh.h: Likewise.
1452 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
1454 * config/pa/constraints.md: Adjust unused letters. Change "T"
1455 constraint to match_test floating_point_store_memory_operand().
1456 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
1457 (base14_operand): New.
1458 (floating_point_store_memory_operand): New.
1459 (integer_store_memory_operand): Revise to use base14_operand and
1460 reg_plus_base_memory_operand.
1461 (move_dest_operand): Allow symbolic_memory_operands.
1462 (symbolic_memory_operand): Check for LO_SOM.
1463 (symbolic_operand): Change default case to break.
1464 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
1465 CONST_DOUBLE values to be reloaded by putting them into memory when
1466 the destination is a floating point register.
1467 (movdf): Remove code to handle CONST_DOUBLE.
1469 (reload_indf_r1): New.
1470 (reload_insf_r1): New.
1471 Consistently use "Q" and "T" constraints with integer and floating
1472 point move instructions, respectively.
1473 (movdi): Remove FAIL.
1474 Change predicate for source operand unamed DImode move from
1475 general_operand to move_src_operand.
1476 (umulsidi3): Change predicate for destination operand to
1478 Likewise for similar unamed patterns.
1479 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
1480 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
1481 (hppa_legitimize_address): Simplify mask calculation.
1482 (pa_emit_move_sequence): Revised handling of secondary reloads from
1483 REG+D addresses for floating point loads and stores. Directly handle
1484 loading CONST0_RTX (mode) to a floating point register.
1485 (pa_secondary_reload): Handle reloading DF and SFmode constant values
1486 to floating point registers. Don't restrict secondary reloads to
1487 floating point registers to integer modes. Revise some comments and
1489 (TARGET_LEGITIMATE_ADDRESS_P): Define.
1490 (pa_legitimate_address_p): New.
1491 (pa_legitimize_reload_address): New.
1492 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
1493 (STRICT_REG_OK_FOR_BASE_P): New.
1494 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
1495 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
1497 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
1498 Andrew Dixie <andrewd@gentrack.com>
1500 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
1503 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
1505 * expmed.c (extract_bit_field_1): Pass the full width of the
1506 structure to get_best_reg_extraction_insn.
1508 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
1511 * configure.ac (use_cxa_atexit): Add AIX.
1512 * configure: Regenerate.
1514 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
1516 2013-02-01 Jakub Jelinek <jakub@redhat.com>
1519 * final.c (need_profile_function): New variable.
1520 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
1521 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
1522 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
1523 notes, targetm.asm_out.function_prologue doesn't emit anything,
1524 HAVE_prologue and profiler should be emitted before prologue,
1525 set need_profile_function instead of emitting it.
1526 (final_scan_insn): If need_profile_function, emit
1527 profile_function on the first NOTE_INSN_BASIC_BLOCK or
1528 NOTE_INSN_FUNCTION_BEG note.
1530 2013-02-01 Richard Henderson <rth@redhat.com>
1532 * config/rs6000/rs6000.md (smulditi3): New.
1535 * config/alpha/alpha.md (umulditi3): New.
1537 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
1539 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
1540 (ASM_OUTPUT_ALIGNED_LOCAL): New.
1542 2013-02-01 Richard Biener <rguenther@suse.de>
1544 PR tree-optimization/56113
1545 * tree-ssa-structalias.c (label_visit): Reduce work for
1546 single-predecessor nodes.
1548 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
1550 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
1551 range isn't testing for zero.
1553 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
1556 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
1558 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
1559 Nick Clifton <nickc@redhat.com>
1561 * config/v850/constraints.md (Q): Define as a memory constraint.
1562 * config/v850/predicates.md (label_ref_operand): New predicate.
1563 (e3v5_shift_operand): New predicate.
1564 (ior_operator): New predicate.
1565 * config/v850/t-v850: Add e3v5 multilib.
1566 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
1567 (v850_gen_movdi): Prototype.
1568 * config/v850/v850.c: Add support for e3v5 architecture.
1569 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
1571 (construct_save_jarl): Add e3v5 long JARL support.
1572 (v850_adjust_insn_length): New function. Adjust length of call
1573 insns when using e3v5 instructions.
1574 (v850_gen_movdi): New function: Generate instructions to move a
1576 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
1577 (CPP_SPEC): Define __v850e3v5__ as appropriate.
1578 (TARGET_USE_FPU): Enable for e3v5.
1579 (CONST_OK_FOR_W): New macro.
1580 (ADJUST_INSN_LENGTH): Define.
1581 * config/v850/v850.md (UNSPEC_LOOP): Define.
1582 (attr cpu): Add v850e3v5.
1583 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
1584 (movdi): New pattern.
1585 (movdi_internal): New pattern.
1586 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
1587 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
1588 (cstoresf4): Likewise.
1589 (cstoredf4): Likewise.
1590 (insv): New pattern.
1591 (rotlso3_a): New pattern.
1592 (rotlsi3_b): New pattern
1593 (rotlsi3_v850e3v5): New pattern.
1594 (doloop_begin): New pattern.
1595 (fix_loop_counter): New pattern.
1596 (doloop_end): New pattern.
1597 (branch_normal): Add e3v5 long branch support.
1598 (branch_invert): Likewise.
1599 (branch_z_normal): Likewise.
1600 (branch_z_invert): Likewise.
1601 (branch_nz_normal): Likewise.
1602 (branch_nz_invert): Likewise.
1603 (call_internal_short): Add e3v5 register-indirect JARL support.
1604 (call_internal_long): Likewise.
1605 (call_value_internal_short): Likewise.
1606 (call_value_internal_long): Likewise.
1607 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
1608 (mloop): New option.
1609 * config.gcc: Add support for configuring v840e3v5 target.
1610 * doc/invoke.texi: Document new v850 specific command line options.
1612 2013-01-31 Paul Koning <ni1d@arrl.net>
1616 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
1617 children if parent is a class.
1618 (prune_unused_types_prune): Don't add DW_AT_declaration.
1620 2013-01-31 Richard Biener <rguenther@suse.de>
1622 PR tree-optimization/56157
1623 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
1624 match up operand with SLP child.
1626 2013-01-31 Jason Merrill <jason@redhat.com>
1629 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
1630 parameters the first time.
1631 (gen_scheduled_generic_parms_dies): Check completeness here.
1633 2013-01-31 Richard Biener <rguenther@suse.de>
1636 * common.opt (faggressive-loop-optimizations): New flag,
1638 * doc/invoke.texi (faggressive-loop-optimizations): Document.
1639 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
1640 infer_loop_bounds_from_undefined by it.
1642 2013-01-31 Richard Biener <rguenther@suse.de>
1644 PR tree-optimization/56150
1645 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
1646 visit virtual operands.
1647 (find_uses_to_rename_bb): Likewise.
1649 2013-01-31 Richard Biener <rguenther@suse.de>
1651 PR tree-optimization/56150
1652 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
1653 mixed store non-store stmts.
1655 2013-01-30 Jakub Jelinek <jakub@redhat.com>
1658 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
1659 LIBASAN_EARLY_SPEC is defined.
1660 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
1661 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
1663 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
1666 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
1667 invalid args instead of ICEing on it.
1668 (ix86_valid_target_attribute_tree): Return error_mark_node if
1669 ix86_valid_target_attribute_inner_p failed.
1670 (ix86_valid_target_attribute_p): Return false only if
1671 ix86_valid_target_attribute_tree returned error_mark_node. Allow
1672 target("default") attribute.
1673 (sorted_attr_string): Change argument from const char * to tree,
1674 merge in all target attribute arguments rather than just one.
1675 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
1676 instead of free. Avoid using strcat.
1677 (ix86_mangle_function_version_assembler_name): Mangle
1678 target("default") as if no target attribute is present. Adjust
1679 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
1680 instead of xmalloc and XDELETEVEC instead of free.
1681 (ix86_function_versions): Don't return true if one of the decls
1682 doesn't have target attribute. If they don't and one of the decls
1683 is DECL_FUNCTION_VERSIONED, report an error. Adjust
1684 sorted_attr_string caller. Use XDELETEVEC instead of free.
1685 (ix86_supports_function_versions): Remove.
1686 (make_name): Fix up formatting.
1687 (make_dispatcher_decl): Remove resolver_name and its initialization.
1688 Avoid leaking memory.
1689 (is_function_default_version): Return true if there is
1690 target("default") attribute rather than no target attribute at all.
1691 (make_resolver_func): Avoid leaking memory.
1692 (ix86_generate_version_dispatcher_body): Likewise.
1693 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
1694 * target.def (supports_function_versions): Remove.
1695 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
1696 * doc/tm.texi: Regenerated.
1698 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
1700 PR rtl-optimization/56144
1701 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
1702 for values with side effects.
1704 2013-01-30 Richard Biener <rguenther@suse.de>
1706 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
1707 (sparseset_pop): Likewise.
1708 * cfganal.c (compute_idf): Likewise. Increase work-stack size
1709 to be able to use quick_push in the worker loop.
1711 2013-01-30 Marek Polacek <polacek@redhat.com>
1713 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
1715 2013-01-30 Richard Biener <rguenther@suse.de>
1718 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
1720 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
1722 PR tree-optimization/56064
1723 * fixed-value.c (fixed_from_double_int): New function.
1724 * fixed-value.h (fixed_from_double_int): New prototype.
1725 (const_fixed_from_double_int): New static inline function.
1726 * fold-const.c (native_interpret_fixed): New static function.
1727 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
1728 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
1729 (native_encode_fixed): New static function.
1730 (native_encode_expr) <FIXED_CST>: Use it.
1731 (native_interpret_int): Move double_int worker code to...
1732 * double-int.c (double_int::from_buffer): ...this new static method.
1733 * double-int.h (double_int::from_buffer): Prototype it.
1735 2013-01-30 Richard Biener <rguenther@suse.de>
1737 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
1738 New pointer-map and obstack.
1739 (init_alias_vars): Allocate pointer-map and obstack.
1740 (delete_points_to_sets): Free them.
1741 (find_what_var_points_to): Cache result.
1742 (find_what_p_points_to): Adjust for changed interface of
1743 find_what_var_points_to.
1744 (compute_points_to_sets): Likewise.
1745 (ipa_pta_execute): Likewise.
1747 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1749 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
1750 * configure: Regenerate.
1751 * config.in: Regenerate.
1752 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
1753 #nobits/#progbits if supported.
1755 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
1758 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
1759 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
1760 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
1762 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
1764 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
1765 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
1767 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
1769 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
1770 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
1772 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
1774 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
1776 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
1777 * config/arm/cortex-a7.md: New bypasses using
1778 arm_mac_accumulator_is_result.
1780 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
1782 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
1783 (cortex_a7_neon_mla): Likewise.
1784 (cortex_a7_fpfmad): New reservation.
1785 (cortex_a7_fpmacs): Use ffmas and update required units.
1786 (cortex_a7_fpmuld): Update required units and latency.
1787 (cortex_a7_fpmacd): Likewise.
1788 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
1789 (cortex_a7_neon). Likewise.
1790 (bypass) Update participating units.
1792 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
1794 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
1795 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
1797 * config/arm/vfp11.md (vfp_farith): Use ffmas.
1798 (vfp_fmul): Use ffmad.
1799 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
1800 (cortex_r4_fmacd): Use ffmad.
1801 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
1802 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
1803 (cortex_a9_fmacd): Use ffmad.
1804 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
1805 (cortex_a8_vfp_macd): Use ffmad.
1806 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
1807 (cortex_a5_fpmacd): Use ffmad.
1808 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
1809 (cortex_a15_vfp_macd): Use ffmad.
1810 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
1812 2013-01-29 Jason Merrill <jason@redhat.com>
1815 * varasm.c (default_assemble_visibility): Don't warn about
1816 visibility on artificial decls.
1818 2013-01-29 Richard Biener <rguenther@suse.de>
1820 PR tree-optimization/56113
1821 * tree-ssa-structalias.c (equiv_class_lookup): Also return
1823 (label_visit): Free duplicate bitmaps and record the leader instead.
1824 (perform_var_substitution): Adjust.
1826 2013-01-29 Richard Biener <rguenther@suse.de>
1828 PR tree-optimization/55270
1829 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
1830 the CFG, schedule loops for fixup.
1832 2013-01-29 Nick Clifton <nickc@redhat.com>
1834 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
1837 2013-01-28 Leif Ekblad <leif@rdos.net>
1839 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
1840 * config/i386/i386.h (TARGET_RDOS): New macro.
1841 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
1842 * config/i386/i386.c (ix86_option_override_internal): For 64bit
1843 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
1844 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
1845 DEFAULT_LARGE_SECTION_THRESHOLD.
1846 * config/i386/i386.md (R14_REG, R15_REG): New constants.
1847 * config/i386/rdos.h: New file.
1848 * config/i386/rdos64.h: New file.
1850 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
1853 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
1856 2013-01-28 Jakub Jelinek <jakub@redhat.com>
1858 PR rtl-optimization/56117
1859 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
1860 call cselib_lookup_from_insn on the MEM before calling
1861 add_insn_mem_dependence.
1863 2013-01-28 Richard Biener <rguenther@suse.de>
1865 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
1866 to a stmt that didn't have one.
1867 (copy_phis_for_bb): Likewise for PHI arguments.
1868 (copy_debug_stmt): Likewise for debug stmts.
1870 2013-01-28 Richard Biener <rguenther@suse.de>
1872 PR tree-optimization/56034
1873 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
1874 (partition_builtin_p): Adjust.
1875 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
1876 it is the last partition.
1877 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
1878 up the vertex for the definition.
1879 (classify_partition): Classify whether a partition is a
1880 PKIND_REDUCTION, thus has uses outside of the loop.
1881 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
1882 Merge all PKIND_REDUCTION partitions into the last partition.
1883 (tree_loop_distribution): Seed partitions from reductions as well.
1885 2013-01-28 Jakub Jelinek <jakub@redhat.com>
1887 PR tree-optimization/56125
1888 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
1889 pow(x,c) into sqrt(x) * powi(x, n/2) or
1890 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
1891 optimizing for size.
1892 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
1893 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
1896 PR tree-optimization/56094
1897 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
1898 to UNKNOWN_LOCATION while gimplifying expr.
1900 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
1903 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
1904 operand 0 in movabs insn template for -masm=intel asm alternative.
1905 (*movabs<mode>_2): Ditto for operand 1.
1907 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
1910 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
1913 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
1915 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
1918 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
1920 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
1921 Undef to avoid warning.
1923 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
1925 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
1926 * configure: Regenerate.
1928 2013-01-25 Jakub Jelinek <jakub@redhat.com>
1930 PR tree-optimization/56098
1931 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
1932 for stmts with volatile ops.
1933 (cond_store_replacement): Don't optimize if assign has volatile ops.
1934 (cond_if_else_store_replacement_1): Don't optimize if either
1935 then_assign or else_assign have volatile ops.
1936 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
1939 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
1941 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
1943 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
1945 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
1946 missing ':' in asm example.
1948 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
1950 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
1951 entries into lane and laneq entries.
1952 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
1953 Remove AdvSIMD scalar modes.
1954 (aarch64_sq<r>dmulh_laneq<mode>): New.
1955 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
1957 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
1958 builtin implementations to relfect changes in RTL in aarch64-simd.md.
1959 * config/aarch64/iterators.md (VCOND): New.
1962 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
1965 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
1966 Add NULL LIBNAME argument to existing definitions.
1967 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
1968 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
1969 * config/avr/avr.c (DEF_BUILTIN): Same.
1970 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
1971 (avr_expand_builtin): Expand to a vanilla call if a libgcc
1972 implementation is available (DECL_ASSEMBLER_NAME is set).
1973 (avr_fold_absfx): New static function.
1974 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
1975 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
1976 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
1978 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
1979 (abshk, absk, abslk, absllk): Provide as static inline functions.
1981 2013-01-25 Marek Polacek <polacek@redhat.com>
1983 PR tree-optimization/56035
1984 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
1986 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
1988 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
1989 (*movtf_internal_rex64): Add (!o,C) alternative
1990 (*movxf_internal_rex64): Ditto.
1991 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
1993 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
1995 * doc/invoke.texi: fix typo.
1996 * doc/objc.texi: fix typo.
1998 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
2000 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
2001 for the first two alternatives.
2003 2013-01-24 Diego Novillo <dnovillo@google.com>
2005 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
2006 (ggc-zone.o): Remove.
2007 * configure.ac: Remove option --with-gc.
2008 * configure: Re-generate.
2009 * doc/install.texi: Remove documentation for --with-gc.
2010 * gengtype.c (write_enum_defn): Remove. Update all users.
2011 (write_Types_process_field): Remove generation of gt_e_* argument.
2012 (output_type_enum): Remove. Update all users.
2013 (write_enum_defn): Remove. Update all users.
2014 (enum alloc_zone): Remove. Update all users.
2015 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
2016 * ggc-common.c (ggc_splay_alloc): Remove first argument.
2018 (struct ptr_data): Remove field TYPE. Update all users.
2019 (gt_pch_note_object): Remove argument TYPE. Update all users.
2020 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
2022 * ggc-none.c (ggc_alloc_typed_stat): Remove.
2023 (struct alloc_zone): Remove.
2024 (ggc_internal_alloc_zone_stat): Remove.
2025 (ggc_internal_cleared_alloc_zone_stat): Remove.
2026 * ggc-page.c (ggc_alloc_typed_stat): Remove.
2027 (ggc_pch_count_object): Remove last argument. Update all users.
2028 (ggc_pch_alloc_object): Remove last argument. Update all users.
2029 (struct alloc_zone): Remove.
2030 * ggc-zone.c: Remove.
2031 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
2032 (struct alloc_zone): Remove.
2033 (ggc_alloc_typed_stat): Remove.
2034 (ggc_alloc_typed): Remove.
2035 (ggc_splay_alloc): Remove first argument.
2036 (rtl_zone): Remove. Update all users.
2037 (tree_zone): Remove. Update all users.
2038 (tree_id_zone): Remove. Update all users.
2039 (ggc_internal_zone_alloc_stat): Remove. Update all users.
2040 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
2041 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
2042 * tree-ssanames.c: Remove references to zone allocator in comments.
2044 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
2046 * config/avr/avr.c (avr_out_fract): Make register numbers that
2047 might be outside of source operand signed.
2049 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
2051 * config/i386/constraints.md (Yf): New constraint.
2052 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
2053 of f constraint to conditionaly disable x87 register preferences.
2054 (*movdf_internal): Ditto.
2055 (*movsf_internal): Ditto.
2057 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
2060 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
2061 that have operands with impossible constraints.
2062 Add a FIXME for a speed-up opportunity.
2063 * lra-constraints.c (process_alt_operands): Verify that a class
2064 selected from constraints on asms is valid for the operand mode.
2065 (curr_insn_transform): Remove incorrect comment.
2067 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
2069 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
2070 TOC operand is a valid symbol ref in the constant pool.
2072 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
2074 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
2076 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
2079 * config/avr/stdfix.h: New file.
2080 * t-avr (stdfix-gcc.h): New rule to build it.
2081 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
2083 2013-01-23 Kostya Serebryany <kcc@google.com>
2085 * config/darwin.h: remove dependency on
2086 CoreFoundation (asan on Mac OS).
2088 2013-01-23 Jakub Jelinek <jakub@redhat.com>
2091 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
2092 instead of cmpdi_operand for first comparison operand.
2093 Don't assert that comparison operands aren't both constants.
2095 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
2097 * doc/install.texi (Downloading the Source): Update references to
2098 downloading separate components.
2100 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
2102 * doc/extend.texi (__int128): Improve grammar.
2104 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
2107 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
2108 alternative to (o,r).
2109 (*movdi_internal_rex64): Remove (!o,n) alternative.
2110 (DImode immediate->memory splitter): Remove.
2111 (DImode immediate->memory peephole2): Remove.
2112 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
2113 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
2114 alternative to (!o,*r).
2115 (*movtf_internal_sse): New pattern.
2116 (*movxf_internal_rex64): New pattern.
2117 (*movxf_internal): Disable for TARGET_64BIT.
2118 (*movdf_internal_rex64): Remove (!o,F) alternative.
2120 2013-01-22 Jakub Jelinek <jakub@redhat.com>
2123 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
2124 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
2125 * tree-vect-loop-manip.c (find_loop_location): Also ignore
2126 stmt locations where LOCATION_LOCUS of the stmt location is
2127 UNKNOWN_LOCATION or BUILTINS_LOCATION.
2130 * config/i386/i386.md (UNSPEC_STOS): New.
2131 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
2132 *strsetqi_1): Add UNSPEC_STOS.
2134 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
2137 * doc/invoke.texi: Remove left over -Wsynth example.
2139 2013-01-21 Jakub Jelinek <jakub@redhat.com>
2141 PR tree-optimization/56051
2142 * fold-const.c (fold_binary_loc): Don't fold
2143 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
2144 a narrowing conversion, or widening conversion from signed
2147 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
2149 PR rtl-optimization/56023
2150 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
2151 dependent on debug instruction.
2153 2013-01-21 Martin Jambor <mjambor@suse.cz>
2156 * function.c (allocate_struct_function): Call
2157 invoke_set_current_function_hook earlier.
2159 2013-01-21 Jakub Jelinek <jakub@redhat.com>
2161 * reload1.c (init_reload): Only initialize reload_obstack
2162 during the first call.
2164 2013-01-21 Marek Polacek <polacek@redhat.com>
2166 * cfgloop.c (verify_loop_structure): Fix up grammar.
2168 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
2170 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
2171 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
2173 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2176 * config/arm/marvell-pj4.md: Update copyright year.
2177 Fix up use of alu to alu_reg and simple_alu_imm.
2179 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
2181 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
2183 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
2186 * lra-constraints.c (curr_insn_transform): Don't reuse original
2187 insn for secondary memory move when memory mode should be different.
2189 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
2191 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
2192 atomic_storedi_1): New patterns.
2194 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
2196 btver2 pipeline descriptions.
2197 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
2199 * config/i386/i386.md (btver2_decode): New type attributes.
2200 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
2202 * config/i386/btver2.md: New file describing btver2 pipelines.
2204 2013-01-19 Andrew Pinski <apinski@cavium.com>
2206 PR tree-optimization/52631
2207 * tree-ssa-sccvn (visit_use): Before looking up the original
2208 statement, try looking up the simplified expression.
2210 2013-01-19 Anthony Green <green@moxielogic.com>
2212 * config/moxie/moxie.c (moxie_expand_prologue): Set
2213 current_function_static_stack_size.
2215 2013-01-18 Jakub Jelinek <jakub@redhat.com>
2217 PR tree-optimization/56029
2218 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
2219 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
2221 2013-01-18 Sharad Singhai <singhai@google.com>
2223 PR tree-optimization/55995
2224 * dumpfile.c (dump_loc): Print location only if available.
2225 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
2227 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
2230 * lra-constraints.c (curr_insn_transform): Reuse original insn for
2231 secondary memory move.
2232 (inherit_reload_reg): Use rclass instead of cl for
2233 check_secondary_memory_needed_p.
2235 2013-01-18 Jakub Jelinek <jakub@redhat.com>
2238 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
2239 the case where writing real complex part of target modifies op1.
2241 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
2243 * config/aarch64/aarch64-simd.md
2244 (aarch64_vcond_internal<mode>): Handle unordered cases.
2245 * config/aarch64/iterators.md (v_cmp_result): New.
2247 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
2248 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2250 * config/arm/marvell-pj4.md: New file.
2251 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
2252 * config/arm/arm.md (generic_sched): Add marvell_pj4.
2253 (generic_vfp): Likewise.
2254 * config/arm/arm-cores.def: Add marvell-pj4.
2255 * config/arm/arm-tune.md: Regenerate.
2256 * config/arm/arm-tables.opt: Regenerate.
2257 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
2258 * doc/invoke.texi: Document marvell-pj4.
2260 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
2262 * config/aarch64/arm_neon.h: Map scalar types to standard types.
2264 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
2269 * var-tracking.c (negative_power_of_two_p): New.
2270 (global_get_addr_cache, local_get_addr_cache): New.
2271 (get_addr_from_global_cache, get_addr_from_local_cache): New.
2272 (vt_canonicalize_addr): Rewrite using the above. Adjust the
2274 (vt_stack_offset_p): Remove.
2275 (vt_canon_true_dep): Always canonicalize loc's address.
2276 (clobber_overlapping_mems): Make sure we have a MEM.
2277 (local_get_addr_clear_given_value): New.
2278 (val_reset): Clear local cached entries.
2279 (compute_bb_dataflow): Create and release the local cache.
2280 Disable duplicate MEMs clobbering.
2281 (emit_notes_in_bb): Clobber MEMs likewise.
2282 (vt_emit_notes): Create and release the local cache.
2283 (vt_initialize, vt_finalize): Create and release the global
2284 cache, respectively.
2285 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
2287 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
2290 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
2291 not found in the symtab.
2293 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
2296 PR rtl-optimization/55547
2297 PR rtl-optimization/53827
2300 * alias.c (offset_overlap_p): New, factored out of...
2301 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
2302 the conservative special case for symbolic constants. Don't
2303 adjust zero sizes on alignment.
2305 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
2307 PR rtl-optimization/52573
2308 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
2309 REG_UNUSED for the same register.
2311 2013-01-17 Richard Biener <rguenther@suse.de>
2312 Marek Polacek <polacek@redhat.com>
2314 PR rtl-optimization/55833
2315 * loop-unswitch.c (unswitch_loops): Move loop verification...
2316 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
2317 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
2318 Set it to true when we're removing a loop from hierarchy tree in
2319 an irreducible region.
2320 (fix_bb_placements): Adjust caller.
2321 (fix_loop_placements): Likewise.
2323 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
2325 * config/avr/builtins.def (DEF_BUILTIN): Factor out
2326 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
2327 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
2328 Remove ID. Adjust comments.
2329 * config/avr/avr-c.c (avr_builtin_name): Remove.
2330 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
2331 * config/avr/avr.c (avr_tolower): New static function.
2332 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
2333 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
2334 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
2337 2013-01-17 Jan Hubicka <jh@suse.cz>
2339 PR tree-optimization/55273
2340 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
2342 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
2345 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
2346 store through atomic_store<mode>_1.
2347 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
2349 2013-01-17 Martin Jambor <mjambor@suse.cz>
2351 PR tree-optimizations/55264
2352 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
2353 for virtual methods.
2354 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
2355 virtual methods before inlining is over.
2356 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
2358 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
2361 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
2363 PR rtl-optimization/56005
2364 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
2365 pending reads for prefetch.
2367 2013-01-16 Ian Bolton <ian.bolton@arm.com>
2369 * config/aarch64/aarch64.md
2370 (*cstoresi_neg_uxtw): New pattern.
2371 (*cmovsi_insn_uxtw): New pattern.
2372 (*<optab>si3_uxtw): New pattern.
2373 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
2374 (*<optab>si3_insn_uxtw): New pattern.
2375 (*bswapsi2_uxtw): New pattern.
2377 2013-01-16 Richard Biener <rguenther@suse.de>
2379 * tree-inline.c (tree_function_versioning): Remove set but
2380 never used variable.
2382 2013-01-16 Richard Biener <rguenther@suse.de>
2384 PR tree-optimization/55964
2385 * tree-flow.h (rename_variables_in_loop): Remove.
2386 (rename_variables_in_bb): Likewise.
2387 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
2388 (copy_loop_before): Adjust and delete update-ssa status.
2389 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
2390 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2391 (rename_variables_in_loop): Remove.
2392 (slpeel_update_phis_for_duplicate_loop): Likewise.
2393 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
2394 use available cfg machinery instead of duplicating it.
2395 Update PHI nodes and perform poor-mans SSA update here.
2396 (slpeel_tree_peel_loop_to_edge): Adjust.
2398 2013-01-16 Richard Biener <rguenther@suse.de>
2400 PR tree-optimization/54767
2401 PR tree-optimization/53465
2402 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
2403 (vrp_visit_phi_node): For PHI arguments coming via backedges
2404 drop all symbolical range information.
2405 (execute_vrp): Compute backedges.
2407 2013-01-16 Richard Biener <rguenther@suse.de>
2409 * doc/install.texi: Update CLooG and ISL requirements to
2412 2013-01-16 Christian Bruel <christian.bruel@st.com>
2415 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
2416 (broken_move): Handle UNSPECV_SP_SWITCH_B.
2417 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
2419 2013-01-16 DJ Delorie <dj@redhat.com>
2421 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
2422 (UNSPECV_SP_SWITCH_E): New.
2423 (sp_switch_1): Change to an unspec.
2424 (sp_switch_2): Change to an unspec. Don't use post-inc when we
2427 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
2429 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
2430 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
2431 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
2432 (expand_mem_thread_fence): Ditto.
2433 (expand_mem_signal_fence): Ditto.
2434 (expand_atomic_load): Ditto.
2435 (expand_atomic_store): Ditto.
2437 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
2439 PR rtl-optimization/55547
2440 PR rtl-optimization/53827
2443 * alias.c (memrefs_conflict_p): Set sizes to negative after
2446 2013-01-15 Jakub Jelinek <jakub@redhat.com>
2449 * function.c (thread_prologue_and_epilogue_insns): Always
2450 add crtl->drap_reg to set_up_by_prologue.set, even if
2451 stack_realign_drap is false.
2453 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2455 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
2456 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
2457 *call): Fix indention.
2459 2013-01-15 Tom de Vries <tom@codesourcery.com>
2462 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
2465 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
2467 PR rtl-optimization/55153
2468 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
2470 2013-01-15 Martin Jambor <mjambor@suse.cz>
2472 PR tree-optimization/55920
2473 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
2474 accesses as grp_to_be_debug_replaced.
2476 2013-01-15 Jakub Jelinek <jakub@redhat.com>
2478 PR tree-optimization/55920
2479 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
2480 there is non-useless type conversion needed from debug rhs to lhs,
2481 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
2483 2013-01-15 Joseph Myers <joseph@codesourcery.com>
2484 Mikael Pettersson <mikpe@it.uu.se>
2487 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
2489 (ASM_OUTPUT_CASE_LABEL): Remove.
2490 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
2491 * final.c (shorten_branches): Update alignment of labels before
2492 jump tables if CASE_VECTOR_SHORTEN_MODE.
2494 2013-01-15 Richard Biener <rguenther@suse.de>
2497 * system.h: Do not include gmp.h for building host tools.
2499 2013-01-15 Richard Biener <rguenther@suse.de>
2502 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
2503 account for bitpos when computing alignment.
2505 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
2507 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
2508 (ix86_target_macros_internal): Likewise.
2510 * config/i386/i386.c (m_CORE2I7): Removed.
2511 (m_CORE_HASWELL): New macro.
2512 (m_CORE_ALL): Likewise.
2513 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
2514 (initial_ix86_arch_features): Likewise.
2515 (processor_target_table): Initializations for Core avx2.
2516 (cpu_names): New names "core-avx2".
2517 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
2518 PROCESSOR_CORE_HASWELL.
2519 (ix86_issue_rate): New case.
2520 (ia32_multipass_dfa_lookahead): Likewise.
2521 (ix86_sched_init_global): Likewise.
2523 * config/i386/i386.h (TARGET_HASWELL): New macro.
2524 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
2525 (processor_type): New PROCESSOR_HASWELL.
2527 2013-01-15 Jakub Jelinek <jakub@redhat.com>
2529 PR tree-optimization/55955
2530 * tree-vect-loop.c (vectorizable_reduction): Give up early on
2531 *SHIFT_EXPR and *ROTATE_EXPR codes.
2533 PR tree-optimization/48766
2534 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
2535 -ftrapv disable -fwrapv.
2537 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
2540 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
2541 etc. to 1 and not to __flash.
2542 Use LL suffix for __INT24_MAX__ with -mint8.
2543 Use ULL suffix for __UINT24_MAX__ with -mint8.
2545 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
2547 * config/avr/avr-arch.h
2548 (struct base_arch_s): Use typedef avr_arch_t instead.
2549 (struct arch_info_s): Use typedef avr_arch_info_t instead.
2550 (struct mcu_type_s): Use typedef avr_mcu_t instead.
2551 * config/avr/avr.c: Same.
2552 * config/avr/avr-devices.c: Same.
2553 * config/avr/driver-avr.c: Same.
2554 * config/avr/gen-avr-mmcu-texi.c: Same.
2555 * config/avr/avr-mcus.def: Adjust comment.
2557 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
2559 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
2560 * config/aarch64/iterators.md (VALLDI): New.
2562 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
2563 Andi Kleen <ak@linux.intel.com>
2566 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
2567 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
2570 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
2572 * config/avr/avr-stdint.h: Remove trailing blanks.
2573 * config/avr/avr-log.h: Same.
2574 * config/avr/avr-arch.h: Same.
2575 * config/avr/avr-devices.c: Same.
2576 * config/avr/avr-dimode.md: Same.
2577 * config/avr/predicates.md: Same.
2578 * config/avr/avr-c.c: Same. And fix typo.
2580 * config/avr/avr-protos.h: Same. And:
2581 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
2582 (init_cumulative_args): Rename to avr_init_cumulative_args.
2583 (expand_prologue): Rename to avr_expand_prologue.
2584 (expand_epilogue): Rename to avr_expand_epilogue.
2585 (adjust_insn_length): Rename to avr_adjust_insn_length.
2586 (notice_update_cc): Rename to avr_notice_update_cc.
2587 (final_prescan_insn): Rename to avr_final_prescan_insn.
2588 * config/avr/avr.c: Same.
2589 * config/avr/avr.h: Same.
2590 * config/avr/avr.md: Remove trailing blanks.
2591 (prologue): Use avr_expand_prologue.
2592 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
2594 2013-01-14 Richard Biener <rguenther@suse.de>
2596 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
2597 verify_location, collect_subblocks): New functions.
2598 (verify_gimple_in_cfg): Verify that locations only reference
2599 BLOCKs in the functions BLOCK tree.
2601 2013-01-14 Richard Biener <rguenther@suse.de>
2603 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
2605 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
2607 (insert_out_of_ssa_copy_on_edge): Likewise.
2608 (rewrite_close_phi_out_of_ssa): Likewise.
2609 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
2611 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
2612 propagated constants.
2613 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
2616 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
2618 * config/avr/avr-modes.def: Add GPL copyright notice.
2620 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
2622 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
2623 MEMMODEL_MASK to determine memory model.
2624 (atomic_store<mode>): Ditto from operands[2].
2625 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
2627 2013-01-13 Jakub Jelinek <jakub@redhat.com>
2630 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
2631 (fold_gimple_assign): Don't call unshare_expr here.
2632 (fold_ctor_reference): Call unshare_expr.
2634 2013-01-13 Terry Guo <terry.guo@arm.com>
2636 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
2637 * doc/fragments.texi: Document MULTILIB_REUSE.
2638 * gcc.c (multilib_reuse): New internal spec.
2639 (set_multilib_dir): Also search multilib from multilib_reuse.
2640 * genmultilib (tmpmultilib3): Refactor code.
2641 (tmpmultilib4): Ditto.
2642 (multilib_reuse): New multilib argument.
2644 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
2646 * Makefile.in: Update copyright.
2648 2013-01-12 Tom de Vries <tom@codesourcery.com>
2651 * calls.c (expand_call): Check if arg_nr is valid.
2653 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
2655 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
2656 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
2657 documentation. Add missing '__' in front of
2658 __builtin_ia32_packssdw256.
2660 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2663 * config/s390/s390.c (s390_preferred_reload_class): Do not return
2664 NO_REGS for larl operands.
2665 (s390_reload_larl_operand): Use s390_load_address instead of
2668 2013-01-11 Richard Biener <rguenther@suse.de>
2670 * tree-cfg.c (verify_node_sharing_1): Split out from ...
2671 (verify_node_sharing): ... here.
2672 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
2674 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
2676 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
2677 Substitute TREECHECKING.
2678 * configure: Regenerate.
2679 * Makefile.in (TREECHECKING): New.
2681 2013-01-11 Richard Guenther <rguenther@suse.de>
2683 PR tree-optimization/44061
2684 * tree-vrp.c (extract_range_basic): Compute zero as
2685 value-range for __builtin_constant_p of function parameters.
2687 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
2689 Update copyright years.
2691 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
2693 PR rtl-optimization/55672
2694 * lra-eliminations.c (mark_not_eliminable): Permit addition with
2695 const to be eliminable.
2697 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
2699 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
2700 * configure: Regenerate.
2702 2013-01-10 Richard Biener <rguenther@suse.de>
2704 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
2706 2013-01-10 Richard Biener <rguenther@suse.de>
2709 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
2710 locations for virtual PHI arguments.
2711 (rewrite_update_phi_arguments): Likewise.
2713 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
2715 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
2718 2013-01-10 Jakub Jelinek <jakub@redhat.com>
2720 PR tree-optimization/55921
2721 * tree-complex.c (expand_complex_asm): New function.
2722 (expand_complex_operations_1): Call it for GIMPLE_ASM.
2724 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2727 * config/s390/s390.c (s390_symref_operand_p)
2728 (s390_loadrelative_operand_p): Merge the two functions.
2729 (s390_check_qrst_address, print_operand_address): Add parameters
2730 to s390_loadrelative_operand_p invokation.
2731 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
2732 (s390_reload_larl_operand, s390_secondary_reload): Use
2733 s390_loadrelative_operand_p instead of s390_symref_operand_p.
2734 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
2736 2013-01-09 Mike Stump <mikestump@comcast.net>
2738 * dse.c (record_store): Remove unnecessary assert.
2740 2013-01-09 Jan Hubicka <jh@suse.cz>
2742 PR tree-optimization/55569
2743 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
2744 * cfgloop.h (scale_loop_profile): Likewise.
2746 2013-01-09 Jan Hubicka <jh@suse.cz>
2749 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
2751 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
2753 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
2756 * expr.h (maybe_emit_group_store): Declare.
2757 * expr.c (maybe_emit_group_store): New function.
2758 * builtins.c (expand_builtin_int_roundingfn): Call it.
2759 (expand_builtin_int_roundingfn_2): Likewise.
2761 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
2763 PR rtl-optimization/55829
2764 * lra-constraints.c (match_reload): Add code for absent output.
2765 (curr_insn_transform): Add code for reloads of matched inputs
2768 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
2770 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
2771 attribute of movddup insn to DF.
2772 (*vec_interleave_lowv2df): Ditto.
2773 (vec_dupv2df): Ditto.
2775 2013-01-09 Jan Hubicka <jh@suse.cz>
2777 PR tree-optimiation/55875
2778 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
2779 EVERY_ITERATION parameter.
2780 (number_of_iterations_exit): Check if exit is executed every iteration.
2781 (idx_infer_loop_bounds): Similarly here.
2782 (n_of_executions_at_most): Simplify
2783 to only test for cases where statement is dominated by the
2784 particular bound; handle correctly the "postdominance" test.
2785 (scev_probably_wraps_p): Use max loop iterations info
2786 as a global bound first.
2788 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
2789 Nick Clifton <nickc@redhat.com>
2791 * config/v850/v850.md (cbranchsf4): New pattern.
2792 (cstoresf4): New pattern.
2793 (cbranchdf4): New pattern.
2794 (cstoredf4): New pattern.
2795 (movsicc): Disallow floating point comparisons.
2796 (cmpsf_le_insn): Fix order of operators.
2797 (cmpsf_lt_insn): Likewise.
2798 (cmpsf_eq_insn): Likewise.
2799 (cmpdf_le_insn): Likewise.
2800 (cmpdf_lt_insn): Likewise.
2801 (cmpdf_eq_insn): Likewise.
2802 (cmpsf_ge_insn): Use LE comparison.
2803 (cmpdf_ge_insn): Likewise.
2804 (cmpsf_gt_insn): Use LT comparison.
2805 (cmpdf_gt_insn): Likewise.
2806 (cmpsf_ne_insn): Delete pattern.
2807 (cmpdf_ne_insn): Delete pattern.
2808 * config/v850/v850.c (v850_gen_float_compare): Use
2809 gen_cmpdf_eq_insn for NE comparison.
2810 (v850_float_z_comparison_operator)
2811 (v850_float_nz_comparison_operator): Move from here ...
2812 * config/v850/predicates.md: ... to here. Move GT and GE
2813 comparisons into v850_float_z_comparison_operator.
2814 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
2816 (v850_float_nz_comparison_operator): Likewise.
2818 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
2820 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
2821 with calls to gen_insvsi/gen_insvdi.
2823 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
2825 * config/i386/i386.c (initial_ix86_tune_features): Set up
2826 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
2828 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
2829 Jakub Jelinek <jakub@redhat.com>
2831 PR tree-optimization/48189
2832 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
2833 If nitercst is 0, don't predict the exit edge.
2835 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2837 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
2838 in asm_fprintf with reg_names.
2839 (aarch64_print_operand_address): Likewise.
2840 (aarch64_return_addr): Likewise.
2841 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
2843 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
2845 * config/pa/pa.h (VAL_U6_BITS_P): Define.
2846 (INT_U6_BITS): Likewise.
2847 * config/pa/predicates.md (uint6_operand): New predicate.
2848 (shift5_operand, shift6_operand): Likewise.
2849 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
2851 (lshrdi3): Use shift6_operand.
2852 (shrpsi4, shrpdi4): New insn patterns.
2853 (extzv): Delete expander.
2854 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
2855 predicates in unamed zero extract patterns. Tighten common constraint.
2856 (extv): Delete expander.
2857 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
2858 predicates in unamed sign extract patterns. Tighten common constraint.
2859 (insv): Delete expander.
2860 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
2861 predicates in unamed insert patterns. Tighten common constraint.
2862 Change uint32_operand predicate to uint6_operand predicate in unamed
2863 DImode pattern to insert constant values of type 1...1xxxx.
2865 2013-01-04 Jan Hubicka <jh@suse.cz>
2867 PR tree-optimization/55823
2868 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
2871 2013-01-08 Jakub Jelinek <jakub@redhat.com>
2872 Uros Bizjak <ubizjak@gmail.com>
2874 PR rtl-optimization/55845
2875 * df-problems.c (can_move_insns_across): Stop scanning at
2876 volatile_insn_p source instruction or give up if
2877 across_from .. across_to range contains any volatile_insn_p
2880 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
2882 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
2883 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
2885 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
2886 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
2888 2013-01-08 Jakub Jelinek <jakub@redhat.com>
2891 * asan.c (asan_clear_shadow): New function.
2892 (asan_emit_stack_protection): Use it.
2894 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
2896 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
2897 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
2898 with tab instead of space.
2900 2013-01-08 Nick Clifton <nickc@redhat.com>
2902 * config/rl78/rl78.c (rl78_expand_prologue): Always select
2903 register bank 0 at the start of an interrupt handler.
2904 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
2907 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
2909 * config/aarch64/aarch64-simd.md
2910 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
2911 (aarch64_simd_bsl): Likewise.
2912 (aarch64_vcond_internal<mode>): Likewise.
2913 (vcond<mode><mode>): Likewise.
2914 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
2915 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
2917 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
2919 * config/aarch64/aarch64-builtins.c
2920 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
2922 2013-01-08 Martin Jambor <mjambor@suse.cz>
2925 * tree-sra.c (analyze_access_subtree): Return true also after
2926 potentially creating a debug-only replacement.
2928 2013-01-08 Jakub Jelinek <jakub@redhat.com>
2931 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
2933 PR tree-optimization/54120
2934 * tree-vrp.c (range_fits_type_p): Don't allow
2935 src_precision < precision from signed vr to unsigned_p
2936 if vr->min or vr->max is negative.
2937 (simplify_float_conversion_using_ranges): Test can_float_p
2938 against CODE_FOR_nothing.
2940 2013-01-08 Jakub Jelinek <jakub@redhat.com>
2941 Richard Biener <rguenther@suse.de>
2944 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
2945 types instead of just INTEGER_TYPE types.
2947 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
2949 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
2952 2013-01-07 Steve Ellcey <sellcey@mips.com>
2955 * config/mips/mips.opt: Change mad to mmad to match documentation.
2957 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
2960 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
2963 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
2966 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
2967 (avr_addrspace_t): Add .section_name field.
2968 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
2970 (avr_addrspace): Same. Initialize .section_name. Remove last
2971 NULL entry. Put __memx into .progmemx.data.
2972 (progmem_section_prefix): Remove.
2973 (avr_asm_init_sections): No need to initialize progmem_section.
2974 (avr_asm_named_section): Use avr_addrspace[].section_name to get
2975 section name prefix.
2976 (avr_asm_select_section): Ditto. And use get_unnamed_section to
2977 retrieve the progmem section.
2978 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
2979 boundary to run over avr_addrspace[].
2980 (avr_register_target_pragmas): Ditto.
2982 2013-01-06 Jakub Jelinek <jakub@redhat.com>
2984 * varasm.c (output_constant_def_contents): For asan_protect_global
2985 protected strings, adjust DECL_ALIGN if needed, before testing for
2987 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
2988 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
2990 (output_object_block): For asan protected decls, emit asan padding
2991 after their contents.
2992 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
2993 (asan_finish_file): Test it here instead.
2995 2013-01-07 Nick Clifton <nickc@redhat.com>
2996 Matthias Klose <doko@debian.org>
2997 Doug Kwan <dougkwan@google.com>
2998 H.J. Lu <hongjiu.lu@intel.com>
3001 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
3003 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
3005 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
3007 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
3009 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
3011 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
3014 * doc/install.texi (Cross-Compiler-Specific Options): Document
3017 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
3019 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
3020 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
3021 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
3022 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
3023 vqmovun_high_s64): Fix source operand number and update copyright.
3025 2013-01-07 Richard Biener <rguenther@suse.de>
3028 * gimple.h (gimple_call_builtin_p): New overload.
3029 * gimple.c (validate_call): New function.
3030 (gimple_call_builtin_p): Likewise.
3031 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
3032 Use gimple_call_builtin_p.
3033 (find_func_clobbers): Likewise.
3034 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
3035 (strlen_optimize_stmt): Likewise.
3037 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
3039 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
3040 (vld1q_dup_*): Likewise.
3042 (vld1q_*): Likewise.
3043 (vld1_lane_*): Likewise.
3044 (vld1q_lane_*): Likewise.
3046 2013-01-07 Richard Biener <rguenther@suse.de>
3048 * lto-streamer.h (LTO_minor_version): Bump to 2.
3050 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
3052 * config/aarch64/aarch64-protos.h
3053 (aarch64_const_double_zero_rtx_p): Rename to...
3054 (aarch64_float_const_zero_rtx_p): ...this.
3055 (aarch64_float_const_representable_p): New.
3056 (aarch64_output_simd_mov_immediate): Likewise.
3057 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
3058 move immediate case.
3059 * config/aarch64/aarch64.c
3060 (aarch64_const_double_zero_rtx_p): Rename to...
3061 (aarch64_float_const_zero_rtx_p): ...this.
3062 (aarch64_print_operand): Allow printing of new constants.
3063 (aarch64_valid_floating_const): New.
3064 (aarch64_legitimate_constant_p): Check for valid floating-point
3066 (aarch64_simd_valid_immediate): Likewise.
3067 (aarch64_vect_float_const_representable_p): New.
3068 (aarch64_float_const_representable_p): Likewise.
3069 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
3070 (aarch64_output_simd_mov_immediate): New.
3071 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
3072 (*movdf_aarch64): Likewise.
3073 * config/aarch64/constraints.md (Ufc): New.
3074 (Y): call aarch64_float_const_zero_rtx.
3075 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
3077 2013-01-07 Richard Biener <rguenther@suse.de>
3079 PR tree-optimization/55888
3080 PR tree-optimization/55862
3081 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
3082 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
3083 not if it is contained therein.
3085 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
3087 * config/avr/t-avr: Typo.
3089 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
3092 * config/avr/t-avr: Don't automatically rebuild
3093 $(srcdir)/config/avr/t-multilib
3094 $(srcdir)/config/avr/avr-tables.opt
3095 $(srcdir)/doc/avr-mmcu.texi
3096 (avr-mcus): New phony target to build them on request.
3097 (s-avr-mlib, s-avr-mmcu-texi): Remove.
3098 * avr/avr-mcus.def: Adjust comments.
3100 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
3102 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
3104 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
3106 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
3108 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
3110 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
3112 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
3114 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
3115 to generate profiling.
3116 * config/rs6000/aix64.h (LIB_SPEC): Same.
3118 2013-01-04 Andrew Pinski <apinski@cavium.com>
3120 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
3122 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
3124 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
3126 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
3128 (ix86_expand_move): Ditto.
3129 (ix86_zero_extend_to_Pmode): Ditto.
3130 (ix86_expand_call): Ditto.
3131 (ix86_expand_special_args_builtin): Ditto.
3132 (ix86_expand_builtin): Ditto.
3134 2013-01-04 Richard Biener <rguenther@suse.de>
3136 PR tree-optimization/55862
3137 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
3138 translating them through PHI nodes.
3140 2013-01-04 Martin Jambor <mjambor@suse.cz>
3142 PR tree-optimization/55755
3143 * tree-sra.c (sra_modify_assign): Do not check that an access has no
3144 children when trying to avoid producing a VIEW_CONVERT_EXPR.
3146 2013-01-04 Marek Polacek <polacek@redhat.com>
3149 * opts.c (default_options_optimization): Clarify error message.
3151 2013-01-04 Richard Biener <rguenther@suse.de>
3154 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
3157 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
3160 * config/pa/pa.md (movsi): Revert previous change.
3161 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
3164 2013-01-03 Richard Henderson <rth@redhat.com>
3166 * config/i386/i386.c (ix86_expand_move): Always assign to op1
3167 after eliminating TLS symbols.
3169 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
3172 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
3173 * graphite-poly.c (debug_gmp_value): Likewise.
3175 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
3178 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
3179 selected code model, define __code_mode_small__, __code_model_medium__,
3180 __code_model_large__, __code_model_32__ or __code_model_kernel__.
3181 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
3182 xchg temporary register with %k. Declare temporary register as
3184 [__x86_64__]: For medium and large code models, preserve %rbx register.
3186 2013-01-03 Richard Biener <rguenther@suse.de>
3188 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
3189 (dump_subscript): Adjust.
3190 (finalize_ddr_dependent): Do not dump redundant info.
3191 (analyze_siv_subscript): Adjust.
3192 (subscript_dependence_tester): Likewise.
3193 (compute_affine_dependence): Likewise.
3195 2013-01-03 Richard Biener <rguenther@suse.de>
3198 2013-01-03 Richard Biener <rguenther@suse.de>
3200 PR tree-optimization/55857
3201 * tree-vect-stmts.c (vectorizable_load): Do not setup
3202 re-alignment for invariant loads.
3204 2013-01-02 Richard Biener <rguenther@suse.de>
3206 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
3207 invariant load do not generate a vector load from the scalar location.
3209 2013-01-03 Richard Biener <rguenther@suse.de>
3211 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
3212 for not vectorizing.
3213 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
3214 not build INDIRECT_REFs, call get_name once only.
3215 (vect_create_data_ref_ptr): Likewise. Dump base object kind
3216 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
3218 2013-01-03 Richard Biener <rguenther@suse.de>
3220 PR tree-optimization/55857
3221 * tree-vect-stmts.c (vectorizable_load): Do not setup
3222 re-alignment for invariant loads.
3224 2013-01-03 Richard Biener <rguenther@suse.de>
3227 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
3228 prefer a built-in decl.
3230 2013-01-03 Jakub Jelinek <jakub@redhat.com>
3232 * gcc.c (process_command): Update copyright notice dates.
3233 * gcov.c (print_version): Likewise.
3234 * gcov-dump.c (print_version): Likewise.
3236 PR rtl-optimization/55838
3237 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
3238 iv0.step, iv1.step and step.
3240 2013-01-03 Jakub Jelinek <jakub@redhat.com>
3241 Marc Glisse <marc.glisse@inria.fr>
3243 PR tree-optimization/55832
3244 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
3245 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
3246 integer_{one,zero}_node.
3248 2013-01-03 Jakub Jelinek <jakub@redhat.com>
3251 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
3252 * var-tracking.c (reverse_op): Don't add reverse ops to
3253 VALUEs that have already
3254 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
3256 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
3258 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
3260 2013-01-02 Teresa Johnson <tejohnson@google.com>
3262 * dumpfile.c (dump_loc): Print filename with location.
3263 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
3264 new location_t parameter to emit complete unroll message with
3266 (canonicalize_loop_induction_variables): Compute loops location
3267 and pass to try_unroll_loop_completely.
3268 * loop-unroll.c (report_unroll_peel): New function.
3269 (peel_loops_completely): Use new dump format with location
3270 for main dumpfile message, and invoke report_unroll_peel on success.
3271 (decide_unrolling_and_peeling): Ditto.
3272 (decide_peel_once_rolling): Remove old dumpfile message subsumed
3273 by report_unroll_peel.
3274 (decide_peel_completely): Ditto.
3275 (decide_unroll_constant_iterations): Ditto.
3276 (decide_unroll_runtime_iterations): Ditto.
3277 (decide_peel_simple): Ditto.
3278 (decide_unroll_stupid): Ditto.
3279 * cfgloop.c (get_loop_location): New function.
3280 * cfgloop.h (get_loop_location): Declare.
3282 2013-01-02 Sriraman Tallam <tmsriram@google.com>
3284 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
3287 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
3290 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
3291 BLKmode objects when EXPAND_MEMORY is specified.
3293 2013-01-02 Sriraman Tallam <tmsriram@google.com>
3295 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
3297 (fold_builtin_cpu): Do not share cpu model decls across statements.
3299 2013-01-02 Jason Merrill <jason@redhat.com>
3302 * tree.c (build_array_type_1): Revert earlier change.
3304 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
3306 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
3308 * config/aarch64/aarch64-tune.md: Re-generate.
3310 2013-01-02 Richard Biener <rguenther@suse.de>
3312 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
3313 invariant load do not generate a vector load from the scalar location.
3315 2013-01-02 Richard Biener <rguenther@suse.de>
3318 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
3319 * configure: Regenerate.
3321 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3323 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
3324 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
3325 (expand_builtin_int_roundingfn_2): Keep the original target around
3326 for the fallback case.
3328 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3330 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
3331 to be clear for sign changes.
3333 2013-01-01 Jan Hubicka <jh@suse.cz>
3335 * ipa-inline-analysis.c: Fix formatting.
3337 2013-01-01 Jakub Jelinek <jakub@redhat.com>
3339 PR tree-optimization/55831
3340 * tree-vect-loop.c (get_initial_def_for_induction): Use
3341 gsi_after_labels instead of gsi_start_bb.
3343 Copyright (C) 2013 Free Software Foundation, Inc.
3345 Copying and distribution of this file, with or without modification,
3346 are permitted in any medium without royalty provided the copyright
3347 notice and this notice are preserved.