1 2013-10-22 Jan Hubicka <jh@suse.cz>
3 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
4 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
5 * i386.c (expand_small_movmem_or_setmem): New function.
6 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
7 (alg_usable_p): Add support for value ranges; cleanup.
8 (ix86_expand_set_or_movmem): Add support for misaligned moves.
10 2013-10-22 Sterling Augustine <saugustine@google.com>
12 * doc/invoke.texi: Document -ggnu-pubnames.
13 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
15 * dwarf2out.c: Include gdb/gdb-index.h.
16 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
17 debug_generate_pub_sections.
18 (is_java, output_pubtables, output_pubname): New functions.
19 (include_pubname_in_output): Handle debug_generate_pub_sections at
21 (size_of_pubnames): Use new local space_for_flags based on
22 debug_generate_pub_sections.
23 (output_pubnames): Unify pubnames and pubtypes output logic.
24 Genericize comments. Call output_pubname.
25 (dwarf2out_finish): Move logic to output_pubnames and call it.
27 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
30 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
31 Remove CCCmode handling.
32 <case LTU>: Return 'c' suffix for CCCmode.
33 <case GEU>: Return 'nc' suffix for CCCmode.
34 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
35 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
36 (*sub<mode>3_cc_overflow): Ditto.
37 (*subsi3_zext_cc_overflow): Ditto.
39 2013-10-22 Steve Ellcey <sellcey@mips.com>
41 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
42 (AND (NOT OP1) (NOT OP2)).
44 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
46 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
47 meaning of merge-high and merge-low masks for little endian; avoid
48 use of vector-pack masks for little endian for mismatched modes.
50 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
52 * config/tilepro/tilepro.c: Include "tree.h".
54 2013-10-22 Andreas Schwab <schwab@suse.de>
56 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
59 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
61 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
62 [Ed Smith-Rowland]): New entries.
63 ([Stephen M. Webb]): Update.
65 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
67 * tree-ssa-ter.h: Remove duplicate copy of file contents.
69 2013-10-21 Marek Polacek <polacek@redhat.com>
72 * fold-const.c (fold_range_test): Return 0 if the type is not
75 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
77 * system.h: Move hwint.h include further down.
78 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
80 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
82 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
84 Fix volatile issues in optimize_bit_field_compare.
85 * fold-const.c (optimize_bit_field_compare): Bail out if
86 lvolatilep or rvolatilep.
88 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
90 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
91 and get_inner_reference returning different pmode for non-volatile
92 bit-field members dependent on flag_strict_volatile_bitfields.
93 * stor-layout.c (layout_decl): Remove special handling of
94 flag_strict_volatile_bitfields.
95 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
96 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
98 2013-10-21 Paulo Matos <pmatos@broadcom.com>
100 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
103 2013-10-21 Jeff Law <law@redhat.com>
105 * tree-ssa-threadedge.c (thread_through_normal_block): New
106 argument VISITED. Remove VISISTED as a local variable. When we
107 have a threadable jump, verify the destination of the jump has not
109 (thread_across_edge): Allocate VISITED bitmap once at function
110 scope and use it throughout. Make sure to set appropriate bits in
111 VISITED for E (start of jump thread path).
112 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
113 through a joiner if any edge on the path has a recorded jump thread.
115 2013-10-21 Ian Lance Taylor <iant@google.com>
117 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
118 don't imply that attributes can solve all problems.
119 (Directory Options): Fix typo.
121 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
123 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
124 extend and extend_add.
126 2013-10-21 Richard Biener <rguenther@suse.de>
128 PR tree-optimization/58794
129 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
130 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
132 2013-10-21 Richard Biener <rguenther@suse.de>
135 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
136 to (T) X for sign-changing conversions (or no conversion).
138 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
140 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
142 2013-10-20 Jan Hubicka <jh@suse.cz>
144 * config/i386/i386-tune.def: Add comment; organize into categories
146 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
148 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
149 argument. Update function comment.
150 (expand_set_or_movmem_via_rep): New function combining
151 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
152 (expand_movmem_via_rep_mov): Remove.
153 expand_setmem_via_rep_stos): Remove.
154 (expand_movmem_epilogue): Update calls correspondingly.
155 (expand_setmem_epilogue_via_loop): Likewise.
157 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
158 (expand_set_or_movmem_prologue): New function combining
159 expand_movmem_prologue and expand_setmem_prologue.
160 (expand_movmem_prologue): Remove.
161 (expand_setmem_prologue): Remove.
162 (expand_set_or_movmem_constant_prologue): New function combining
163 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
164 (expand_constant_movmem_prologue): Remove.
165 (expand_constant_setmem_prologue): Remove.
166 (promote_duplicated_reg): Allow vector-const0 value.
167 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
168 and ix86_expand_setmem.
169 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
170 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
172 2013-10-21 Diego Novillo <dnovillo@google.com>
174 * asan.c: Include tree.h
175 * bb-reorder.c: Likewise.
176 * cfgcleanup.c: Likewise.
177 * cfgloopmanip.c: Likewise.
178 * data-streamer-in.c: Likewise.
179 * data-streamer-out.c: Likewise.
180 * data-streamer.c: Likewise.
181 * dwarf2cfi.c: Likewise.
182 * graphite-blocking.c: Likewise.
183 * graphite-clast-to-gimple.c: Likewise.
184 * graphite-dependences.c: Likewise.
185 * graphite-interchange.c: Likewise.
186 * graphite-optimize-isl.c: Likewise.
187 * graphite-poly.c: Likewise.
188 * graphite-scop-detection.c: Likewise.
189 * graphite-sese-to-poly.c: Likewise.
190 * graphite.c: Likewise.
191 * ipa-devirt.c: Likewise.
192 * ipa-profile.c: Likewise.
195 * loop-init.c: Likewise.
196 * loop-unroll.c: Likewise.
197 * lower-subreg.c: Likewise.
198 * lto/lto-object.c: Likewise.
200 * reginfo.c: Likewise.
201 * tree-loop-distribution.c: Likewise.
202 * tree-parloops.c: Likewise.
203 * tree-ssa-strlen.c: Likewise.
204 * tree-streamer.c: Likewise.
205 * value-prof.c: Likewise.
206 * target-globals.c: Likewise.
207 * expr.h: Include tree-core.h instead of tree.h.
208 * gimple.h: Likewise.
209 * ipa-prop.h: Likewise.
210 * ipa-utils.h: Likewise.
211 * lto-streamer.h: Likewise.
212 * streamer-hooks.h: Likewise.
213 * ipa-reference.h: Include cgraph.h instead of tree.h.
214 * cgraph.h: Include basic-block.h instead of tree.h.
215 * tree-streamer.h: Do not include tree.h.
216 * genattrtab.c (write_header): Generate inclusion of
218 * genautomata.c (main): Likewise.
219 * genemit.c: Likewise.
220 * genopinit.c: Likewise.
221 * genoutput.c (output_prologue): Likewise.
222 * genpeep.c: Likewise.
224 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
226 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
228 (vec_unpacku_hi_v8hi): Likewise.
229 (vec_unpacku_lo_v16qi): Likewise.
230 (vec_unpacku_lo_v8hi): Likewise.
232 2013-10-20 Jan Hubicka <jh@suse.cz>
234 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
235 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
236 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
238 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
240 * config/mips/mips.h (ISA_HAS_WSBH): Define.
241 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
243 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
245 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
248 * system.h: Undef m_slot.
250 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
252 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
253 all elements for both endian flavors.
255 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
258 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
259 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
260 and SI_REG for 64bit SYSV ABI targets.
262 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
264 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
265 to multi_reg_return. Clarify that we are skipping USEs of multiple
266 return registers. Use bool type where appropriate.
268 2013-10-18 Jan Hubicka <jh@suse.cz>
270 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
272 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
273 (X86_TUNE_PUSH_MEMORY): Likewise.
274 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
275 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
276 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
278 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
279 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
281 (ix86_option_override_internal): Update to use tune features instead
284 2013-10-18 Cong Hou <congh@google.com>
286 PR tree-optimization/58508
287 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
288 statement that contains data refs with zero-step.
290 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
292 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
293 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
294 * gimple-low.c (gimple_check_call_arg,
295 gimple_check_call_matching_types): Move to cgraph.c.
296 * gimple-low.h: Remove prototype.
297 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
298 Relocate from gimple-low.c.
299 * cgraph.h: Add prototype. Don't include basic-block.h.
300 * gimplify.c: Add gimple-low to include list.
301 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
302 * tree-eh.c: Add gimple-low to include list.
303 * tree-nested.c: Likewise.
304 * cfgexpand.c: Add tree-ssa-address.h to include list.
306 * gimple-fold.c: Likewise.
307 * gimple-ssa-strength-reduction.c: Likewise.
308 * trans-mem.c: Likewise.
309 * tree-mudflap.c: Likewise.
310 * tree-ssa-loop-ivopts.c: Likewise.
311 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
312 (degenerate_phi_result): Move to tree-phinodes.c.
313 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
314 * tree-ssa-threadedge.c: Likewise.
315 * tree-vrp.c: Likewise.
316 * tree-phinodes.c (degenerate_phi_result): Relocate here.
317 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
318 * tree-phinodes.h (degenerate_phi_result): Add prototype.
319 * tree-ssa-copy.c: Include tree-ssa-dom.h.
320 * tree-ssa-forwprop.c: Likewise.
321 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
322 pass_data_cleanup_cfg_post_optimizing,
323 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
324 * tree-optimize.c: Delete File.
325 * graphite.c: Include tree-cfgcleanup.h.
326 * passes.c: Likewise.
327 * tree-cfg.c: Likewise.
328 * tree-profile.c: Likewise.
329 * tree-ssa-dse.c: Likewise.
330 * tree-ssa-loop-ivcanon.c: Likewise.
331 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
332 * tree-outof-ssa.c: Include sbitmap.h.
333 * tree-ssa-live.c: Likewise.
334 * tree-ssa-propagate.c: Likewise.
335 * tree-ssa-structalias.c: Likewise.
336 * tree-stdarg.c: Likewise.
337 * Makefile.in (OBJS): Delete tree-optimize.o.
338 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
339 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
340 * varasm.c: Include basic-block.h.
341 * cfgloop.h: Include function.h instead of basic-block.h
342 (bb_loop_depth): Move to cfgloop.c.
343 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
345 2013-10-18 Teresa Johnson <tejohnson@google.com>
347 * predict.c (probably_never_executed): Compare frequency-based
348 count to number of training runs.
349 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
351 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
353 * config/arm/arm.c (cortexa9_extra_costs): New table.
354 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
356 2013-10-18 Jeff Law <law@redhat.com>
358 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
360 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
361 (register_jump_thread): Add "registered_jump_thread" debug
363 * dbgcnt.def (registered_jump_thread): New debug counter.
365 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
367 * config/rs6000/rs6000.c: Include cgraph.h.
369 2013-10-18 Teresa Johnson <tejohnson@google.com>
371 * tree-ssa-tail-merge.c (replace_block_by): Update edge
372 weights during merging.
374 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
376 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
377 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
378 * tree-ssa-operands.h: Remove prototype.
379 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
380 * gimple.c (virtual_operand_p): Relocate from gimple.c.
381 * gimple.h: Add prototype.
382 * gimple-ssa.h: Include tree-ssa-operands.h.
383 * tree-dump.c: Add tree-cfg.h to include list.
384 * tree-ssa-alias.c: Add ipa-reference.h to include list.
385 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
386 * config/i386/i386.c: Don't include tree-flow.h.
387 * config/rs6000/rs6000.c: Likewise.
389 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
391 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
393 2013-10-18 Richard Biener <rguenther@suse.de>
395 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
396 or TYPE_UNSIGNED of integral types.
397 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
398 NULL_TREE for zero-precision integral types.
400 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
402 * config/aarch64/arm_neon.h
403 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
405 2013-10-17 Sriraman Tallam <tmsriram@google.com>
407 * opth-gen.awk: Define target_flags_explicit.
409 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
411 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
412 fields to the reg_addr array that describes the valid addressing
413 mode for any register, general purpose registers, floating point
414 registers, and Altivec registers.
415 (FIRST_RELOAD_REG_CLASS): Likewise.
416 (LAST_RELOAD_REG_CLASS): Likewise.
417 (struct reload_reg_map_type): Likewise.
418 (reload_reg_map_type): Likewise.
419 (RELOAD_REG_VALID): Likewise.
420 (RELOAD_REG_MULTIPLE): Likewise.
421 (RELOAD_REG_INDEXED): Likewise.
422 (RELOAD_REG_OFFSET): Likewise.
423 (RELOAD_REG_PRE_INCDEC): Likewise.
424 (RELOAD_REG_PRE_MODIFY): Likewise.
425 (reg_addr): Likewise.
426 (mode_supports_pre_incdec_p): New helper functions to say whether
427 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
428 (mode_supports_pre_modify_p): Likewise.
429 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
430 print the valid address mode bits for each mode.
431 (rs6000_debug_print_mode): Likewise.
432 (rs6000_debug_reg_global): Likewise.
433 (rs6000_setup_reg_addr_masks): New function to set up the address
434 mask bits for each type.
435 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
436 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
437 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
438 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
439 PRE_MODIFY are supported.
440 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
441 registers, instead of {src,dest}_av_p.
442 (rs6000_print_options_internal): Tweak the debug output slightly.
444 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
446 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
449 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
451 * tree-flow.h (struct omp_region): Move to omp-low.c.
452 Remove omp_ prototypes and variables.
453 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
454 (copy_var_decl): Relocate prototype from tree-flow.h.
455 * gimple.c (copy_var_decl): Relocate from omp-low.c.
456 * tree.h: Move prototype to omp-low.h.
457 * omp-low.h: New File. Relocate prototypes here.
458 * omp-low.c (struct omp_region): Make local here.
459 (root_omp_region): Make static.
460 (copy_var_decl) Move to gimple.c.
461 (new_omp_region): Make static.
462 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
463 * tree-cfg.c: Include omp-low.h.
464 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
465 * gimplify.c: Include omp-low.h.
466 * tree-parloops.c: Likewise.
468 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
470 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
471 operands are in memory, prefer to force non-matched operand 1 to
474 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
477 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
478 restrict TImode addresses to single indirect registers if both
479 -mquad-memory and -mvsx-timode are used.
480 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
481 we should emit load/store quad. Remove using %y for quad memory
484 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
485 constraints to allow load/store quad on machines where TImode is
486 not allowed in VSX registers. Use 'n' instead of 'F' constraint
487 for TImode to load integer constants.
489 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
491 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
493 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
495 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
496 handling of STACK_REG.
498 2013-10-17 Richard Biener <rguenther@suse.de>
500 PR tree-optimization/58143
501 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
503 (rewrite_to_defined_overflow): Likewise.
504 (move_computations_dom_walker::before_dom): Rewrite stmts
505 with undefined signed overflow that are not always executed
506 into unsigned arithmetic.
508 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
511 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
512 explicit isa flag to be an options variable, instead of using
513 global_options_set. Remove define from rs6000.h.
514 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
516 * config/rs6000/rs6000.c (rs6000_option_override_internal):
517 Initialize rs6000_isa_flags_explicit.
518 (rs6000_function_specific_save): Add gcc_options* parameter, so
519 that the powerpc builds after the 2013-10-15 changes.
520 (rs6000_function_specific_restore): Likewise.
522 2013-10-16 DJ Delorie <dj@redhat.com>
524 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
525 op is a REG before checking REGNO.
526 (rl78_alloc_physical_registers): Verify pattern is a SET before
529 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
531 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
533 (vec_unpacks_lo_v4sf): Likewise.
534 (vec_unpacks_float_hi_v4si): Likewise.
535 (vec_unpacks_float_lo_v4si): Likewise.
536 (vec_unpacku_float_hi_v4si): Likewise.
537 (vec_unpacku_float_lo_v4si): Likewise.
539 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
541 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
542 (vsx_concat_v2sf): Likewise.
544 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
546 * config/aarch64/aarch64.md
547 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
549 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
551 PR tree-optimization/58697
552 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
553 estimated_loop_iterations_int.
554 (max_stmt_executions_int): Call get_max_loop_iterations_int.
555 (get_max_loop_iterations_int): New. HWINT version of
556 get_max_loop_iterations.
557 * cfgloop.h: Add prototypes.
558 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
559 * loop-unroll.c (decide_peel_once_rolling): Call
560 get_estimated_loop_iterations_int.
561 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
562 * tree-ssa-loop-niter.h: Tweak prototypes.
564 2013-10-16 David Malcolm <dmalcolm@redhat.com>
566 * gengtype-parse.c (struct_field_seq): Ignore access-control
567 keywords ("public:" etc).
569 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
571 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
572 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
574 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
576 * config/arm/arm.opt (mlra): New option.
577 * config/arm/arm.c (arm_lra_p): New function.
578 (TARGET_LRA_P): Define.
580 2013-10-16 Paulo Matos <pmatos@broadcom.com>
582 * tree-core.h (tree_code_name): Remove.
583 * tree.h (get_tree_code_name): New prototype.
584 * tree.c (tree_code_name): Make static.
585 (get_tree_code_name): New function.
586 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
587 tree_class_check_failed, tree_range_check_failed,
588 tree_not_class_check_failed, omp_clause_check_failed,
589 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
590 wrapper get_tree_code_name instead of calling tree_code_name directly.
591 * tree-vrp.c (dump_asserts_for): Likewise.
592 * tree-dump.c (dequeue_and_dump): Likewise.
593 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
594 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
595 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
596 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
597 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
598 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
599 dump_gimple_omp_for): Likewise.
600 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
601 * tree-ssa-pre.c (print_pre_expr): Likewise.
602 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
603 * print-tree.c (print_node_brief, print_node): Likewise.
604 * gimple.c (gimple_check_failed): Likewise.
605 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
606 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
607 * config/mep/mep.c (mep_validate_vliw): Likewise.
608 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
609 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
611 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
613 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
616 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
618 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
619 (MULTILIB_MATCHES): Add multilib for -march=v8.
621 2013-10-15 Sriraman Tallam <tmsriram@google.com>
624 * optc-save-gen.awk: Add extra parameter to the save and restore
626 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
627 * tree.c (build_optimization_node): New parameter. Add extra parameter
628 to call to cl_optimization_save.
629 (build_target_option_node): New parameter. Add extra parameter
630 to call to cl_target_option_save.
631 * tree.h (build_optimization_node): New parameter.
632 (build_target_option_node): New parameter.
633 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
634 build_optimization_node and build_target_option_node.
635 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
636 (handle_pragma_push_options): Ditto.
637 * toplev.c (process_options): Ditto.
638 * opts.c (init_options_struct): Check for opts_set non-null.
639 * target.def (target_option.save): New parameter.
640 (target_option.restore): New parameter.
642 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
643 (ix86_pragma_target_parse): Ditto.
644 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
646 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
647 to build_optimization_node and build_target_option_node.
648 (rs6000_valid_attribute_p): Ditto.
649 (rs6000_pragma_target_parse): Ditto.
650 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
652 * config/i386/i386.h:
653 TARGET_64BIT_P: New Macro
654 TARGET_MMX_P: New Macro.
655 TARGET_3DNOW_P: New Macro.
656 TARGET_3DNOW_A_P: New Macro.
657 TARGET_SSE_P: New Macro.
658 TARGET_SSE2_P: New Macro.
659 TARGET_SSE3_P: New Macro.
660 TARGET_SSSE3_P: New Macro.
661 TARGET_SSE4_1_P: New Macro.
662 TARGET_SSE4_2_P: New Macro.
663 TARGET_AVX_P: New Macro.
664 TARGET_AVX2_P: New Macro.
665 TARGET_AVX512F_P: New Macro.
666 TARGET_AVX512PF_P: New Macro.
667 TARGET_AVX512ER_P: New Macro.
668 TARGET_AVX512CD_P: New Macro.
669 TARGET_FMA_P: New Macro.
670 TARGET_SSE4A_P: New Macro.
671 TARGET_FMA4_P: New Macro.
672 TARGET_XOP_P: New Macro.
673 TARGET_LWP_P: New Macro.
674 TARGET_ABM_P: New Macro.
675 TARGET_BMI_P: New Macro.
676 TARGET_BMI2_P: New Macro.
677 TARGET_LZCNT_P: New Macro.
678 TARGET_TBM_P: New Macro.
679 TARGET_POPCNT_P: New Macro.
680 TARGET_SAHF_P: New Macro.
681 TARGET_MOVBE_P: New Macro.
682 TARGET_CRC32_P: New Macro.
683 TARGET_AES_P: New Macro.
684 TARGET_PCLMUL_P: New Macro.
685 TARGET_CMPXCHG16B_P: New Macro.
686 TARGET_FSGSBASE_P: New Macro.
687 TARGET_RDRND_P: New Macro.
688 TARGET_F16C_P: New Macro.
689 TARGET_RTM_P: New Macro.
690 TARGET_HLE_P: New Macro.
691 TARGET_RDSEED_P: New Macro.
692 TARGET_PRFCHW_P: New Macro.
693 TARGET_ADX_P: New Macro.
694 TARGET_FXSR_P: New Macro.
695 TARGET_XSAVE_P: New Macro.
696 TARGET_XSAVEOPT_P: New Macro.
697 TARGET_LP64_P: New Macro.
698 TARGET_X32_P: New Macro.
699 TARGET_FPMATH_DEFAULT_P: New Macro.
700 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
701 * config/i386/i386.c (ix86_option_override_internal): New parameters.
703 Change ix86_tune_string to access opts->x_ix86_tune_string.
704 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
705 Change ix86_arch_string to access opts->x_ix86_arch_string.
706 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
707 Change ix86_pmode to access opts->x_ix86_pmode.
708 Change ix86_abi to access opts->x_ix86_abi.
709 Change ix86_cmodel to access opts->x_ix86_cmodel.
710 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
711 Change ix86_isa_flags_explicit to access
712 opts->x_ix86_isa_flags_explicit.
713 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
714 Change ix86_regparm to access opts->x_ix86_regparm.
715 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
716 Change ix86_preferred_stack_boundary_arg to access
717 opts->x_ix86_preferred_stack_boundary_arg.
718 Change ix86_force_align_arg_pointer to access
719 opts->x_ix86_force_align_arg_pointer.
720 Change ix86_incoming_stack_boundar_arg to access
721 opts->x_ix86_incoming_stack_boundar_arg.
722 Change ix86_fpmath to access opts->x_ix86_fpmath.
723 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
724 Change ix86_recip_name to access opts->x_ix86_recip_name.
725 Change ix86_stack_protector_guard to access
726 opts->x_ix86_stack_protector_guard.
727 Change ix86_tune_memcpy_strategy to access
728 opts->x_ix86_tune_memcpy_strategy.
729 Change ix86_tune_memset_strategy to access
730 opts->x_ix86_tune_memset_strategy.
731 Change global_options to access opts.
732 Change global_options_set to access opts_set.
733 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
734 Change TARGET_MMX to TARGET_MMX_P (opts->...).
735 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
736 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
737 Change TARGET_SSE to TARGET_SSE_P (opts->...).
738 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
739 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
740 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
741 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
742 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
743 Change TARGET_AVX to TARGET_AVX_P (opts->...).
744 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
745 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
746 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
747 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
748 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
749 Change TARGET_FMA to TARGET_FMA_P (opts->...).
750 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
751 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
752 Change TARGET_XOP to TARGET_XOP_P (opts->...).
753 Change TARGET_LWP to TARGET_LWP_P (opts->...).
754 Change TARGET_ABM to TARGET_ABM_P (opts->...).
755 Change TARGET_BMI to TARGET_BMI_P (opts->...).
756 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
757 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
758 Change TARGET_TBM to TARGET_TBM_P (opts->...).
759 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
760 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
761 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
762 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
763 Change TARGET_AES to TARGET_AES_P (opts->...).
764 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
765 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
766 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
767 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
768 Change TARGET_F16C to TARGET_F16C_P (opts->...).
769 Change TARGET_RTM to TARGET_RTM_P (opts->...).
770 Change TARGET_HLE to TARGET_HLE_P (opts->...).
771 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
772 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
773 Change TARGET_ADX to TARGET_ADX_P (opts->...).
774 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
775 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
776 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
777 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
778 Change TARGET_X32 to TARGET_X32_P (opts->...).
779 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
780 Change TARGET_FLOAT_RETURNS_IN_80387 to
781 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
782 (ix86_function_specific_save): New parameter. Use opts-> fields
783 to replace global fields.
784 (ix86_function_specific_restore): Ditto.
785 (ix86_valid_target_attribute_inner_p): New parameters.
787 Fix call to ix86_handle_option and set_option.
788 (ix86_valid_target_attribute_tree): New parameters.
789 Change global_options to access opts.
790 Change global_options_set to access opts_set.
791 Fix call to ix86_valid_target_attribute_inner_p.
792 Change ix86_tune_string to access opts->x_ix86_tune_string.
793 Change ix86_arch_string to access opts->x_ix86_arch_string.
794 Change ix86_fpmath to access opts->x_ix86_fpmath
795 Fix call to ix86_option_override_internal.
796 Fix call to ix86_add_new_builtins.
797 Fix calls to build_optimization_node and build_target_option_node.
798 (ix86_valid_target_attribute_p): Remove access to global_options.
799 Use new gcc_options structure func_options.
800 Fix call to ix86_valid_target_attribute_tree.
801 Fix call to build_optimization_node.
802 (get_builtin_code_for_version): Fix call to
803 ix86_valid_target_attribute_tree.
805 2013-10-15 David Malcolm <dmalcolm@redhat.com>
807 * Makefile.in (PICFLAG): New.
808 (enable_host_shared): New.
809 (INTERNAL_CFLAGS): Use PICFLAG.
810 (LIBIBERTY): Use pic build of libiberty.a if configured with
811 --enable-host-shared.
812 * configure.ac: Add --enable-host-shared, setting up new
814 * configure: Regenerate.
815 * doc/install.texi (--enable-shared): Add note contrasting it with ...
816 (--enable-host-shared): New option.
818 2013-10-15 Richard Biener <rguenther@suse.de>
820 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
821 for built-in functions.
823 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
825 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
826 (optimize_range_tests_1): New function,
827 extracted from optimize_range_tests.
828 (optimize_range_tests_xor): Similarly.
829 (optimize_range_tests_diff): New function.
830 (optimize_range_tests): Use optimize_range_tests_1.
832 2013-10-15 Cong Hou <congh@google.com>
834 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
835 requirement of the reduction pattern so that one operand of the
836 reduction operation can come from outside of the loop.
838 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
840 * config/arm/neon-schedgen.ml: Remove.
841 * config/arm/cortex-a9-neon.md: Remove comment regarding
844 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
846 * config/arm/types: Remove old neon types.
848 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
850 * config/arm/cortex-a7.md
851 (cortex_a7_neon_type): New.
852 (cortex_a7_neon_mul): Update for new types.
853 (cortex_a7_neon_mla): Likewise.
854 (cortex_a7_neon): Likewise.
856 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
858 * config/arm/cortex-a15-neon.md
859 (cortex_a15_neon_type): New,
861 (cortex_a15_neon_int_1): Remove.
862 (cortex_a15_neon_int_2): Likewise.
863 (cortex_a15_neon_int_3): Likewise.
864 (cortex_a15_neon_int_4): Likewise.
865 (cortex_a15_neon_int_5): Likewise.
866 (cortex_a15_neon_vqneg_vqabs): Likewise.
867 (cortex_a15_neon_vmov): Likewise.
868 (cortex_a15_neon_vaba): Likewise.
869 (cortex_a15_neon_vaba_qqq): Likewise.
870 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
871 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
872 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
874 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
875 (cortex_a15_neon_mla_qqq_8_16): Likewise.
876 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
877 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
878 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
879 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
880 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
881 (cortex_a15_neon_shift_1): Likewise.
882 (cortex_a15_neon_shift_2): Likewise.
883 (cortex_a15_neon_shift_3): Likewise.
884 (cortex_a15_neon_vshl_ddd): Likewise.
885 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
886 (cortex_a15_neon_vsra_vrsra): Likewise.
887 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
888 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
889 (cortex_a15_neon_bp_3cycle): Likewise.
890 (cortex_a15_neon_ldm_2): Likewise.
891 (cortex_a15_neon_stm_2): Likewise.
892 (cortex_a15_neon_mcr): Likewise.
893 (cortex_a15_neon_mrc): Likewise.
894 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
895 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
896 (cortex_a15_neon_fp_vmul_ddd): Likewise.
897 (cortex_a15_neon_fp_vmul_qqd): Likewise.
898 (cortex_a15_neon_fp_vmla_ddd): Likewise.
899 (cortex_a15_neon_fp_vmla_qqq): Likewise.
900 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
901 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
902 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
903 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
904 (cortex_a15_neon_bp_simple): Likewise.
905 (cortex_a15_neon_bp_2cycle): Likewise.
906 (cortex_a15_neon_bp_3cycle): Likewise.
907 (cortex_a15_neon_vld1_1_2_regs): Likewise.
908 (cortex_a15_neon_vld1_3_4_regs): Likewise.
909 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
910 (cortex_a15_neon_vld2_4_regs): Likewise.
911 (cortex_a15_neon_vld3_vld4): Likewise.
912 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
913 (cortex_a15_neon_vst1_3_4_regs): Likewise.
914 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
915 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
916 (cortex_a15_neon_vst3_vst4): Rename to...
917 (cortex_a15_neon_vst4): This, update for new attributes.
918 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
919 (cortex_a15_neon_vld3_vld4_lane): Likewise.
920 (cortex_a15_neon_vst1_vst2_lane): Likewise.
921 (cortex_a15_neon_vst3_vst4_lane): Likewise.
922 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
923 (cortex_a15_neon_ldm_2): Likewise.
924 (cortex_a15_neon_stm_2): Likewise.
925 (cortex_a15_neon_mcr): Likewise.
926 (cortex_a15_neon_mcr_2_mcrr): Likewise.
927 (cortex_a15_neon_mrc): Likewise.
928 (cortex_a15_neon_mrrc): Likewise.
930 (cortex_a15_neon_abd): New.
931 (cortex_a15_neon_abd_q): Likewise.
932 (cortex_a15_neon_aba): Likewise.
933 (cortex_a15_neon_aba_q): Likewise.
934 (cortex_a15_neon_acc): Likewise.
935 (cortex_a15_neon_acc_q): Likewise.
936 (cortex_a15_neon_arith_basic): Likewise.
937 (cortex_a15_neon_arith_complex): Likewise.
938 (cortex_a15_neon_multiply): Likewise.
939 (cortex_a15_neon_multiply_q): Likewise.
940 (cortex_a15_neon_mla): Likewise.
941 (cortex_a15_neon_mla_q): Likewise.
942 (cortex_a15_neon_sat_mla_long): Likewise.
943 (cortex_a15_neon_shift_acc): Likewise.
944 (cortex_a15_neon_shift_imm_basic): Likewise.
945 (cortex_a15_neon_shift_imm_complex): Likewise.
946 (cortex_a15_neon_shift_reg_basic): Likewise.
947 (cortex_a15_neon_shift_reg_basic_q): Likewise.
948 (cortex_a15_neon_shift_reg_complex): Likewise.
949 (cortex_a15_neon_shift_reg_complex_q): Likewise.
950 (cortex_a15_neon_fp_negabs): Likewise
951 (cortex_a15_neon_fp_arith): Likewise
952 (cortex_a15_neon_fp_arith_q): Likewise
953 (cortex_a15_neon_fp_cvt_int): Likewise
954 (cortex_a15_neon_fp_cvt_int_q): Likewise
955 (cortex_a15_neon_fp_cvt_16): Likewise
956 (cortex_a15_neon_fp_mul): Likewise
957 (cortex_a15_neon_fp_mul_q): Likewise
958 (cortex_a15_neon_fp_mla): Likewise
959 (cortex_a15_neon_fp_mla_q): Likewise
960 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
961 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
962 (cortex_a15_neon_bitops): Likewise.
963 (cortex_a15_neon_bitops_q): Likewise.
964 (cortex_a15_neon_from_gp): Likewise.
965 (cortex_a15_neon_from_gp_q): Likewise.
966 (cortex_a15_neon_tbl3_tbl4): Likewise.
967 (cortex_a15_neon_zip_q): Likewise.
968 (cortex_a15_neon_to_gp): Likewise.
969 (cortex_a15_neon_load_a): Likewise.
970 (cortex_a15_neon_load_b): Likewise.
971 (cortex_a15_neon_load_c): Likewise.
972 (cortex_a15_neon_load_d): Likewise.
973 (cortex_a15_neon_load_e): Likewise.
974 (cortex_a15_neon_load_f): Likewise.
975 (cortex_a15_neon_store_a): Likewise.
976 (cortex_a15_neon_store_b): Likewise.
977 (cortex_a15_neon_store_c): Likewise.
978 (cortex_a15_neon_store_d): Likewise.
979 (cortex_a15_neon_store_e): Likewise.
980 (cortex_a15_neon_store_f): Likewise.
981 (cortex_a15_neon_store_g): Likewise.
982 (cortex_a15_neon_store_h): Likewise.
983 (cortex_a15_vfp_to_from_gp): Likewise.
985 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
987 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
989 (cortex_a9_neon_vshl_ddd): Remove.
990 (cortex_a9_neon_vst3_vst4): Likewise.
991 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
993 (cortex_a9_neon_bit_ops_q): New.
995 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
996 (cortex_a9_neon_int_2): Likewise.
997 (cortex_a9_neon_int_3): Likewise.
998 (cortex_a9_neon_int_4): Likewise.
999 (cortex_a9_neon_int_5): Likewise.
1000 (cortex_a9_neon_vqneg_vqabs): Likewise.
1001 (cortex_a9_neon_vmov): Likewise.
1002 (cortex_a9_neon_vaba): Likewise.
1003 (cortex_a9_neon_vaba_qqq): Likewise.
1004 (cortex_a9_neon_shift_1): Likewise.
1005 (cortex_a9_neon_shift_2): Likewise.
1006 (cortex_a9_neon_shift_3): Likewise.
1007 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1008 (cortex_a9_neon_vsra_vrsra): Likewise.
1009 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1010 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1011 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1013 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1014 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1015 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1017 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1018 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1019 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1020 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1021 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1022 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1023 (cortex_a9_neon_fp_vsum): Likewise.
1024 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1025 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1026 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1027 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1028 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1029 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1030 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1031 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1032 (cortex_a9_neon_bp_simple): Likewise.
1033 (cortex_a9_neon_bp_2cycle): Likewise.
1034 (cortex_a9_neon_bp_3cycle): Likewise.
1035 (cortex_a9_neon_ldr): Likewise.
1036 (cortex_a9_neon_str): Likewise.
1037 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1038 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1039 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1040 (cortex_a9_neon_vld2_4_regs): Likewise.
1041 (cortex_a9_neon_vld3_vld4): Likewise.
1042 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1043 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1044 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1045 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1046 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1047 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1048 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1049 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1050 (cortex_a9_neon_mcr): Likewise.
1051 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1052 (cortex_a9_neon_mrc): Likewise.
1053 (cortex_a9_neon_mrrc): Likewise.
1055 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1057 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1059 (cortex_a8_neon_vshl_ddd): Remove.
1060 (cortex_a8_neon_vst3_vst4): Likewise.
1061 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1063 (cortex_a8_neon_bit_ops_q): New.
1065 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1066 (cortex_a8_neon_int_2): Likewise..
1067 (cortex_a8_neon_int_3): Likewise.
1068 (cortex_a8_neon_int_5): Likewise.
1069 (cortex_a8_neon_vqneg_vqabs): Likewise.
1070 (cortex_a8_neon_int_4): Likewise.
1071 (cortex_a8_neon_vaba): Likewise.
1072 (cortex_a8_neon_vaba_qqq): Likewise.
1073 (cortex_a8_neon_shift_1): Likewise.
1074 (cortex_a8_neon_shift_2): Likewise.
1075 (cortex_a8_neon_shift_3): Likewise.
1076 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1077 (cortex_a8_neon_vsra_vrsra): Likewise.
1078 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1079 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1080 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1082 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1083 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1084 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1086 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1087 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1088 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1089 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1090 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1091 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1092 (cortex_a8_neon_fp_vsum): Likewise.
1093 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1094 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1095 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1096 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1097 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1098 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1099 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1100 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1101 (cortex_a8_neon_bp_simple): Likewise.
1102 (cortex_a8_neon_bp_2cycle): Likewise.
1103 (cortex_a8_neon_bp_3cycle): Likewise.
1104 (cortex_a8_neon_ldr): Likewise.
1105 (cortex_a8_neon_str): Likewise.
1106 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1107 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1108 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1109 (cortex_a8_neon_vld2_4_regs): Likewise.
1110 (cortex_a8_neon_vld3_vld4): Likewise.
1111 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1112 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1113 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1114 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1115 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1116 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1117 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1118 (cortex_a8_neon_mcr): Likewise.
1119 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1120 (cortex_a8_neon_mrc): Likewise.
1121 (cortex_a8_neon_mrrc): Likewise.
1123 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1125 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1127 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1128 (aarch64_simd_dup<mode>): Add "type" attribute.
1129 (aarch64_dup_lane<mode>): Likewise.
1130 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1131 (*aarch64_simd_mov<mode>): Likewise.
1132 (aarch64_simd_mov_from_<mode>low): Likewise.
1133 (aarch64_simd_mov_from_<mode>high): Likewise.
1134 (orn<mode>3): Likewise.
1135 (bic<mode>3): Likewise.
1136 (add<mode>3): Likewise.
1137 (sub<mode>3): Likewise.
1138 (mul<mode>3): Likewise.
1139 (*aarch64_mul3_elt<mode>): Likewise.
1140 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1141 (*aarch64_mul3_elt_to_128df): Likewise.
1142 (*aarch64_mul3_elt_to_64v2df): Likewise.
1143 (neg<mode>2): Likewise.
1144 (abs<mode>2): Likewise.
1145 (abd<mode>_3): Likewise.
1146 (aba<mode>_3): Likewise.
1147 (fabd<mode>_3): Likewise.
1148 (*fabd_scalar<mode>3): Likewise.
1149 (and<mode>3): Likewise.
1150 (ior<mode>3): Likewise.
1151 (xor<mode>3): Likewise.
1152 (one_cmpl<mode>2): Likewise.
1153 (aarch64_simd_vec_set<mode>): Likewise.
1154 (aarch64_simd_lshr<mode>): Likewise.
1155 (aarch64_simd_ashr<mode>): Likewise.
1156 (aarch64_simd_imm_shl<mode>): Likewise.
1157 (aarch64_simd_reg_sshl<mode): Likewise.
1158 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1159 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1160 (aarch64_simd_vec_setv2di): Likewise.
1161 (aarch64_simd_vec_set<mode>): Likewise.
1162 (aarch64_mla<mode>): Likewise.
1163 (*aarch64_mla_elt<mode>): Likewise.
1164 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1165 (aarch64_mls<mode>): Likewise.
1166 (*aarch64_mls_elt<mode>): Likewise.
1167 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1168 (<su><maxmin><mode>3): Likewise.
1169 (move_lo_quad_<mode>): Likewise.
1170 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1171 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1172 (vec_pack_trunc_<mode>): Likewise.
1173 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1174 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1175 (*aarch64_<su>mlal_lo<mode>): Likewise.
1176 (*aarch64_<su>mlal_hi<mode>): Likewise.
1177 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1178 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1179 (*aarch64_<su>mlal<mode>): Likewise.
1180 (*aarch64_<su>mlsl<mode>): Likewise.
1181 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1182 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1183 (add<mode>3): Likewise.
1184 (sub<mode>3): Likewise.
1185 (mul<mode>3): Likewise.
1186 (div<mode>3): Likewise.
1187 (neg<mode>2): Likewise.
1188 (abs<mode>2): Likewise.
1189 (fma<mode>4): Likewise.
1190 (*aarch64_fma4_elt<mode>): Likewise.
1191 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1192 (*aarch64_fma4_elt_to_128df): Likewise.
1193 (*aarch64_fma4_elt_to_64v2df): Likewise.
1194 (fnma<mode>4): Likewise.
1195 (*aarch64_fnma4_elt<mode>): Likewise.
1196 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1197 (*aarch64_fnma4_elt_to_128df): Likewise.
1198 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1199 (<frint_pattern><mode>2): Likewise.
1200 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1201 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1202 (vec_unpacks_lo_v4sf): Likewise.
1203 (aarch64_float_extend_lo_v2df): Likewise.
1204 (vec_unpacks_hi_v4sf): Likewise.
1205 (aarch64_float_truncate_lo_v2sf): Likewise.
1206 (aarch64_float_truncate_hi_v4sf): Likewise.
1207 (aarch64_vmls<mode>): Likewise.
1208 (<su><maxmin><mode>3): Likewise.
1209 (<maxmin_uns><mode>3): Likewise.
1210 (reduc_<sur>plus_<mode>): Likewise.
1211 (reduc_<sur>plus_v2di): Likewise.
1212 (reduc_<sur>plus_v2si): Likewise.
1213 (reduc_<sur>plus_<mode>): Likewise.
1214 (aarch64_addpv4sf): Likewise.
1215 (clz<mode>2): Likewise.
1216 (reduc_<maxmin_uns>_<mode>): Likewise.
1217 (reduc_<maxmin_uns>_v2di): Likewise.
1218 (reduc_<maxmin_uns>_v2si): Likewise.
1219 (reduc_<maxmin_uns>_<mode>): Likewise.
1220 (reduc_<maxmin_uns>_v4sf): Likewise.
1221 (aarch64_simd_bsl<mode>_internal): Likewise.
1222 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1223 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1224 (aarch64_get_lane<mode>): Likewise.
1225 (*aarch64_combinez<mode>): Likewise.
1226 (aarch64_combine<mode>): Likewise.
1227 (aarch64_simd_combine<mode>): Likewise.
1228 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1229 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1230 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1231 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1232 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1233 (aarch64_<sur>h<addsub><mode>): Likewise.
1234 (aarch64_<sur><addsub>hn<mode>): Likewise.
1235 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1236 (aarch64_pmul<mode>): Likewise.
1237 (aarch64_<su_optab><optab><mode>): Likewise.
1238 (aarch64_<sur>qadd<mode>): Likewise.
1239 (aarch64_sqmovun<mode>): Likewise.
1240 (aarch64_<sur>qmovn<mode>): Likewise.
1241 (aarch64_s<optab><mode>): Likewise.
1242 (aarch64_sq<r>dmulh<mode>): Likewise.
1243 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1244 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1245 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1246 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1247 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1248 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1249 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1250 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1251 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1252 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1253 (aarch64_sqdmull<mode>): Likewise.
1254 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1255 (aarch64_sqdmull_n<mode>): Likewise.
1256 (aarch64_sqdmull2<mode>_internal): Likewise.
1257 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1258 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1259 (aarch64_<sur>shl<mode>): Likewise.
1260 (aarch64_<sur>q<r>shl<mode>
1261 (aarch64_<sur>shll_n<mode>): Likewise.
1262 (aarch64_<sur>shll2_n<mode>): Likewise.
1263 (aarch64_<sur>shr_n<mode>): Likewise.
1264 (aarch64_<sur>sra_n<mode>): Likewise.
1265 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1266 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1267 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1268 (aarch64_cm<optab><mode>): Likewise.
1269 (aarch64_cm<optab>di): Likewise.
1270 (aarch64_cm<optab><mode>): Likewise.
1271 (aarch64_cm<optab>di): Likewise.
1272 (aarch64_cmtst<mode>): Likewise.
1273 (aarch64_cmtstdi): Likewise.
1274 (aarch64_cm<optab><mode>): Likewise.
1275 (*aarch64_fac<optab><mode>): Likewise.
1276 (aarch64_addp<mode>): Likewise.
1277 (aarch64_addpdi): Likewise.
1278 (sqrt<mode>2): Likewise.
1279 (vec_load_lanesoi<mode>): Likewise.
1280 (vec_store_lanesoi<mode>): Likewise.
1281 (vec_load_lanesci<mode>): Likewise.
1282 (vec_store_lanesci<mode>): Likewise.
1283 (vec_load_lanesxi<mode>): Likewise.
1284 (vec_store_lanesxi<mode>): Likewise.
1285 (*aarch64_mov<mode>): Likewise.
1286 (aarch64_ld2<mode>_dreg): Likewise.
1287 (aarch64_ld2<mode>_dreg): Likewise.
1288 (aarch64_ld3<mode>_dreg): Likewise.
1289 (aarch64_ld3<mode>_dreg): Likewise.
1290 (aarch64_ld4<mode>_dreg): Likewise.
1291 (aarch64_ld4<mode>_dreg): Likewise.
1292 (aarch64_tbl1<mode>): Likewise.
1293 (aarch64_tbl2v16qi): Likewise.
1294 (aarch64_combinev16qi): Likewise.
1295 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1296 (aarch64_st2<mode>_dreg): Likewise.
1297 (aarch64_st2<mode>_dreg): Likewise.
1298 (aarch64_st3<mode>_dreg): Likewise.
1299 (aarch64_st3<mode>_dreg): Likewise.
1300 (aarch64_st4<mode>_dreg): Likewise.
1301 (aarch64_st4<mode>_dreg): Likewise.
1302 (*aarch64_simd_ld1r<mode>): Likewise.
1303 (aarch64_frecpe<mode>): Likewise.
1304 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1305 (aarch64_frecps<mode>): Likewise.
1307 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1309 * config/arm/iterators.md (V_elem_ch): New.
1311 (VQH_type): Likewise.
1312 * config/arm/arm.md (is_neon_type): New.
1313 (conds): Use is_neon_type.
1314 (anddi3_insn): Update type attribute.
1315 (xordi3_insn): Likewise.
1316 (one_cmpldi2): Likewise.
1317 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1318 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1319 (*movmisalign<mode>_neon_store): Likewise.
1320 (*movmisalign<mode>_neon_load): Likewise.
1321 (vec_set<mode>_internal): Likewise.
1322 (vec_set<mode>_internal): Likewise.
1323 (vec_setv2di_internal): Likewise.
1324 (vec_extract<mode>): Likewise.
1325 (vec_extract<mode>): Likewise.
1326 (vec_extractv2di): Likewise.
1327 (*add<mode>3_neon): Likewise.
1328 (adddi3_neon): Likewise.
1329 (*sub<mode>3_neon): Likewise.
1330 (subdi3_neon): Likewise.
1331 (fma<VCVTF:mode>4): Likewise.
1332 (fma<VCVTF:mode>4_intrinsic): Likewise.
1333 (*fmsub<VCVTF:mode>4): Likewise.
1334 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1335 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1336 (ior<mode>3): Likewise.
1337 (and<mode>3): Likewise.
1338 (orn<mode>3_neon): Likewise.
1339 (orndi3_neon): Likewise.
1340 (bic<mode>3_neon): Likewise.
1341 (bicdi3_neon): Likewise.
1342 (xor<mode>3): Likewise.
1343 (one_cmpl<mode>2): Likewise.
1344 (abs<mode>2): Likewise.
1345 (neg<mode>2): Likewise.
1346 (negdi2_neon): Likewise.
1347 (*umin<mode>3_neon): Likewise.
1348 (*umax<mode>3_neon): Likewise.
1349 (*smin<mode>3_neon): Likewise.
1350 (*smax<mode>3_neon): Likewise.
1351 (vashl<mode>3): Likewise.
1352 (vashr<mode>3_imm): Likewise.
1353 (vlshr<mode>3_imm): Likewise.
1354 (ashl<mode>3_signed): Likewise.
1355 (ashl<mode>3_unsigned): Likewise.
1356 (neon_load_count): Likewise.
1357 (ashldi3_neon_noclobber): Likewise.
1358 (ashldi3_neon): Likewise.
1359 (signed_shift_di3_neon): Likewise.
1360 (unsigned_shift_di3_neon): Likewise.
1361 (ashrdi3_neon_imm_noclobber): Likewise.
1362 (lshrdi3_neon_imm_noclobber): Likewise.
1363 (<shift>di3_neon): Likewise.
1364 (widen_ssum<mode>3): Likewise.
1365 (widen_usum<mode>3): Likewise.
1366 (quad_halves_<code>v4si): Likewise.
1367 (quad_halves_<code>v4sf): Likewise.
1368 (quad_halves_<code>v8hi): Likewise.
1369 (quad_halves_<code>v16qi): Likewise.
1370 (reduc_splus_v2di): Likewise.
1371 (neon_vpadd_internal<mode>): Likewise.
1372 (neon_vpsmin<mode>): Likewise.
1373 (neon_vpsmax<mode>): Likewise.
1374 (neon_vpumin<mode>): Likewise.
1375 (neon_vpumax<mode>): Likewise.
1376 (*ss_add<mode>_neon): Likewise.
1377 (*us_add<mode>_neon): Likewise.
1378 (*ss_sub<mode>_neon): Likewise.
1379 (*us_sub<mode>_neon): Likewise.
1380 (neon_vadd<mode>_unspec): Likewise.
1381 (neon_vaddl<mode>): Likewise.
1382 (neon_vaddw<mode>): Likewise.
1383 (neon_vhadd<mode>): Likewise.
1384 (neon_vqadd<mode>): Likewise.
1385 (neon_vaddhn<mode>): Likewise.
1386 (neon_vmul<mode>): Likewise.
1387 (neon_vfms<VCVTF:mode>): Likewise.
1388 (neon_vmlal<mode>): Likewise.
1389 (neon_vmls<mode>): Likewise.
1390 (neon_vmlsl<mode>): Likewise.
1391 (neon_vqdmulh<mode>): Likewise.
1392 (neon_vqdmlal<mode>): Likewise.
1393 (neon_vqdmlsl<mode>): Likewise.
1394 (neon_vmull<mode>): Likewise.
1395 (neon_vqdmull<mode>): Likewise.
1396 (neon_vsub<mode>_unspec): Likewise.
1397 (neon_vsubl<mode>): Likewise.
1398 (neon_vsubw<mode>): Likewise.
1399 (neon_vqsub<mode>): Likewise.
1400 (neon_vhsub<mode>): Likewise.
1401 (neon_vsubhn<mode>): Likewise.
1402 (neon_vceq<mode>): Likewise.
1403 (neon_vcge<mode>): Likewise.
1404 (neon_vcgeu<mode>): Likewise.
1405 (neon_vcgt<mode>): Likewise.
1406 (neon_vcgtu<mode>): Likewise.
1407 (neon_vcle<mode>): Likewise.
1408 (neon_vclt<mode>): Likewise.
1409 (neon_vcage<mode>): Likewise.
1410 (neon_vcagt<mode>): Likewise.
1411 (neon_vtst<mode>): Likewise.
1412 (neon_vabd<mode>): Likewise.
1413 (neon_vabdl<mode>): Likewise.
1414 (neon_vaba<mode>): Likewise.
1415 (neon_vabal<mode>): Likewise.
1416 (neon_vmax<mode>): Likewise.
1417 (neon_vmin<mode>): Likewise.
1418 (neon_vpaddl<mode>): Likewise.
1419 (neon_vpadal<mode>): Likewise.
1420 (neon_vpmax<mode>): Likewise.
1421 (neon_vpmin<mode>): Likewise.
1422 (neon_vrecps<mode>): Likewise.
1423 (neon_vrsqrts<mode>): Likewise.
1424 (neon_vqabs<mode>): Likewise.
1425 (neon_vqneg<mode>): Likewise.
1426 (neon_vcls<mode>): Likewise.
1427 (clz<mode>2): Likewise.
1428 (popcount<mode>2): Likewise.
1429 (neon_vrecpe<mode>): Likewise.
1430 (neon_vrsqrte<mode>): Likewise.
1431 (neon_vget_lane<mode>_sext_internal): Likewise.
1432 (neon_vget_lane<mode>_zext_internal): Likewise.
1433 (neon_vdup_n<mode>): Likewise.
1434 (neon_vdup_n<mode>): Likewise.
1435 (neon_vdup_nv2di): Likewise.
1436 (neon_vdup_lane<mode>_interal): Likewise.
1437 (*neon_vswp<mode>): Likewise.
1438 (neon_vcombine<mode>): Likewise.
1439 (float<mode><V_cvtto>2): Likewise.
1440 (floatuns<mode><V_cvtto>2): Likewise.
1441 (fix_trunc<mode><V_cvtto>2): Likewise.
1442 (fixuns_trunc<mode><V_cvtto>2
1443 (neon_vcvt<mode>): Likewise.
1444 (neon_vcvt<mode>): Likewise.
1445 (neon_vcvtv4sfv4hf): Likewise.
1446 (neon_vcvtv4hfv4sf): Likewise.
1447 (neon_vcvt_n<mode>): Likewise.
1448 (neon_vcvt_n<mode>): Likewise.
1449 (neon_vmovn<mode>): Likewise.
1450 (neon_vqmovn<mode>): Likewise.
1451 (neon_vqmovun<mode>): Likewise.
1452 (neon_vmovl<mode>): Likewise.
1453 (neon_vmul_lane<mode>): Likewise.
1454 (neon_vmul_lane<mode>): Likewise.
1455 (neon_vmull_lane<mode>): Likewise.
1456 (neon_vqdmull_lane<mode>): Likewise.
1457 (neon_vqdmulh_lane<mode>): Likewise.
1458 (neon_vqdmulh_lane<mode>): Likewise.
1459 (neon_vmla_lane<mode>): Likewise.
1460 (neon_vmla_lane<mode>): Likewise.
1461 (neon_vmlal_lane<mode>): Likewise.
1462 (neon_vqdmlal_lane<mode>): Likewise.
1463 (neon_vmls_lane<mode>): Likewise.
1464 (neon_vmls_lane<mode>): Likewise.
1465 (neon_vmlsl_lane<mode>): Likewise.
1466 (neon_vqdmlsl_lane<mode>): Likewise.
1467 (neon_vext<mode>): Likewise.
1468 (neon_vrev64<mode>): Likewise.
1469 (neon_vrev32<mode>): Likewise.
1470 (neon_vrev16<mode>): Likewise.
1471 (neon_vbsl<mode>_internal): Likewise.
1472 (neon_vshl<mode>): Likewise.
1473 (neon_vqshl<mode>): Likewise.
1474 (neon_vshr_n<mode>): Likewise.
1475 (neon_vshrn_n<mode>): Likewise.
1476 (neon_vqshrn_n<mode>): Likewise.
1477 (neon_vqshrun_n<mode>): Likewise.
1478 (neon_vshl_n<mode>): Likewise.
1479 (neon_vqshl_n<mode>): Likewise.
1480 (neon_vqshlu_n<mode>): Likewise.
1481 (neon_vshll_n<mode>): Likewise.
1482 (neon_vsra_n<mode>): Likewise.
1483 (neon_vsri_n<mode>): Likewise.
1484 (neon_vsli_n<mode>): Likewise.
1485 (neon_vtbl1v8qi): Likewise.
1486 (neon_vtbl2v8qi): Likewise.
1487 (neon_vtbl3v8qi): Likewise.
1488 (neon_vtbl4v8qi): Likewise.
1489 (neon_vtbl1v16qi): Likewise.
1490 (neon_vtbl2v16qi): Likewise.
1491 (neon_vcombinev16qi): Likewise.
1492 (neon_vtbx1v8qi): Likewise.
1493 (neon_vtbx2v8qi): Likewise.
1494 (neon_vtbx3v8qi): Likewise.
1495 (neon_vtbx4v8qi): Likewise.
1496 (*neon_vtrn<mode>_insn): Likewise.
1497 (*neon_vzip<mode>_insn): Likewise.
1498 (*neon_vuzp<mode>_insn): Likewise.
1499 (neon_vld1<mode>): Likewise.
1500 (neon_vld1_lane<mode>): Likewise.
1501 (neon_vld1_lane<mode>): Likewise.
1502 (neon_vld1_dup<mode>): Likewise.
1503 (neon_vld1_dup<mode>): Likewise.
1504 (neon_vld1_dupv2di): Likewise.
1505 (neon_vst1<mode>): Likewise.
1506 (neon_vst1_lane<mode>): Likewise.
1507 (neon_vst1_lane<mode>): Likewise.
1508 (neon_vld2<mode>): Likewise.
1509 (neon_vld2<mode>): Likewise.
1510 (neon_vld2_lane<mode>): Likewise.
1511 (neon_vld2_lane<mode>): Likewise.
1512 (neon_vld2_dup<mode>): Likewise.
1513 (neon_vst2<mode>): Likewise.
1514 (neon_vst2<mode>): Likewise.
1515 (neon_vst2_lane<mode>): Likewise.
1516 (neon_vst2_lane<mode>): Likewise.
1517 (neon_vld3<mode>): Likewise.
1518 (neon_vld3qa<mode>): Likewise.
1519 (neon_vld3qb<mode>): Likewise.
1520 (neon_vld3_lane<mode>): Likewise.
1521 (neon_vld3_lane<mode>): Likewise.
1522 (neon_vld3_dup<mode>): Likewise.
1523 (neon_vst3<mode>): Likewise.
1524 (neon_vst3qa<mode>): Likewise.
1525 (neon_vst3qb<mode>): Likewise.
1526 (neon_vst3_lane<mode>): Likewise.
1527 (neon_vst3_lane<mode>): Likewise.
1528 (neon_vld4<mode>): Likewise.
1529 (neon_vld4qa<mode>): Likewise.
1530 (neon_vld4qb<mode>): Likewise.
1531 (neon_vld4_lane<mode>): Likewise.
1532 (neon_vld4_lane<mode>): Likewise.
1533 (neon_vld4_dup<mode>): Likewise.
1534 (neon_vst4<mode>): Likewise.
1535 (neon_vst4qa<mode>): Likewise.
1536 (neon_vst4qb<mode>): Likewise.
1537 (neon_vst4_lane<mode>): Likewise.
1538 (neon_vst4_lane<mode>): Likewise.
1539 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1540 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1541 (neon_vec_<US>mult_lo_<mode>): Likewise.
1542 (neon_vec_<US>mult_hi_<mode>): Likewise.
1543 (neon_vec_<US>shiftl_<mode>): Likewise.
1544 (neon_unpack<US>_<mode>): Likewise.
1545 (neon_vec_<US>mult_<mode>): Likewise.
1546 (vec_pack_trunc_<mode>): Likewise.
1547 (neon_vec_pack_trunc_<mode>): Likewise.
1548 (neon_vabd<mode>_2): Likewise.
1549 (neon_vabd<mode>_3): Likewise.
1551 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1553 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1554 (load_pair): Update type attribute.
1555 (store_pair): Update type attribute.
1556 * config/aarch64/iterators.md (q): New.
1558 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1560 * config/arm/types.md: Add new types for Neon insns.
1562 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1563 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1564 Sergey Lega <sergey.s.lega@intel.com>
1565 Anna Tikhonova <anna.tikhonova@intel.com>
1566 Ilya Tocar <ilya.tocar@intel.com>
1567 Andrey Turetskiy <andrey.turetskiy@intel.com>
1568 Ilya Verbin <ilya.verbin@intel.com>
1569 Kirill Yukhin <kirill.yukhin@intel.com>
1570 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1572 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1573 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1574 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1576 (srcp14<mode>): Ditto.
1577 (rsqrt14<mode>): Ditto.
1578 (rsqrt14<mode>): Ditto.
1579 (avx512f_vmscalef<mode>): Ditto.
1580 (avx512f_scalef<mode>): Ditto.
1581 (avx512f_getexp<mode>): Ditto.
1582 (avx512f_sgetexp<mode>): Ditto.
1583 (avx512f_fixupimm<mode>): Ditto.
1584 (avx512f_sfixupimm<mode>): Ditto.
1585 (avx512f_rndscale<mode>): Ditto.
1586 (*avx512er_exp2<mode>): Ditto.
1587 (*avx512er_rcp28<mode>): Ditto.
1588 (avx512er_rsqrt28<mode>): Ditto.
1589 (avx512f_getmant<mode>): Ditto.
1590 (avx512f_getmant<mode>): Ditto.
1591 (avx512f_rndscale<mode>): Fix formatting.
1593 2013-10-15 Martin Jambor <mjambor@suse.cz>
1595 * ipa-utils.h (ipa_edge_within_scc): Declare.
1596 * ipa-cp.c (edge_within_scc): Moved...
1597 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1599 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1600 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1601 Sergey Lega <sergey.s.lega@intel.com>
1602 Anna Tikhonova <anna.tikhonova@intel.com>
1603 Ilya Tocar <ilya.tocar@intel.com>
1604 Andrey Turetskiy <andrey.turetskiy@intel.com>
1605 Ilya Verbin <ilya.verbin@intel.com>
1606 Kirill Yukhin <kirill.yukhin@intel.com>
1607 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1609 * config/i386/predicates.md (const_8_to_15_operand): New.
1610 (const_16_to_31_operand): Ditto.
1611 * config/i386/sse.md (V8FI): New.
1613 (reduc_splus_v8df): Ditto.
1614 (reduc_splus_v16sf): Ditto.
1615 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1616 (vec_extract_hi_<mode>): Ditto.
1617 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1618 (vec_set_lo_<mode>): Ditto.
1619 (vec_set_hi_<mode>): Ditto.
1620 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1621 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1622 (avx512f_pshufd_1): Ditto.
1623 (avx512f_broadcast<mode>): Ditto.
1624 (avx512f_broadcast<mode>): Ditto.
1625 (define_split): Split vec_extract_lo into move.
1626 (ssequartermode): Ditto.
1627 (ssedoublemode): Extened with wider modes.
1628 (vec_extract_lo_<mode>): Ditto.
1630 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1631 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1632 Sergey Lega <sergey.s.lega@intel.com>
1633 Anna Tikhonova <anna.tikhonova@intel.com>
1634 Ilya Tocar <ilya.tocar@intel.com>
1635 Andrey Turetskiy <andrey.turetskiy@intel.com>
1636 Ilya Verbin <ilya.verbin@intel.com>
1637 Kirill Yukhin <kirill.yukhin@intel.com>
1638 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1640 * config/i386/predicates.md (register_or_constm1_operand): New.
1641 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1642 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1643 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1644 UNSPEC_SCATTER_PREFETCH
1646 (avx512f_ucmp<mode>3): Ditto.
1647 (avx512f_vternlog<mode>): Ditto.
1648 (avx512f_align<mode>): Ditto.
1649 (<shift_insn><mode>3): Ditto.
1650 (avx512f_<rotate>v<mode>): Ditto.
1651 (avx512f_<rotate><mode>): Ditto.
1652 (avx512f_eq<mode>3): Ditto.
1653 (avx512f_eq<mode>3_1): Ditto.
1654 (avx512f_gt<mode>3): Ditto.
1655 (avx512f_testm<mode>3): Ditto.
1656 (avx512f_testnm<mode>3): Ditto.
1657 (avx512pf_gatherpf<mode>): Ditto.
1658 (*avx512pf_gatherpf<mode>_mask): Ditto.
1659 (*avx512pf_gatherpf<mode>): Ditto.
1660 (avx512pf_scatterpf<mode>): Ditto.
1661 (*avx512pf_scatterpf<mode>_mask): Ditto.
1662 (*avx512pf_scatterpf<mode>): Ditto.
1663 (avx512f_vec_dup_gpr<mode>): Ditto.
1664 (clz<mode>2): Ditto.
1665 (conflict<mode>): Ditto.
1666 (REDUC_SMINMAX_MODE): Extened with wider modes.
1667 (reduc_<code>_<mode>): Ditto.
1668 (vlshr<mode>3): Ditto.
1669 (vashl<mode>3): Ditto.
1671 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1672 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1673 Sergey Lega <sergey.s.lega@intel.com>
1674 Anna Tikhonova <anna.tikhonova@intel.com>
1675 Ilya Tocar <ilya.tocar@intel.com>
1676 Andrey Turetskiy <andrey.turetskiy@intel.com>
1677 Ilya Verbin <ilya.verbin@intel.com>
1678 Kirill Yukhin <kirill.yukhin@intel.com>
1679 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1681 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1684 (avx512fmaskmode): Ditto.
1685 (bcstscalarsuff): Ditto.
1686 (avx512f_blendm<mode>): Ditto.
1687 (cmp_imm_predicate): Ditto.
1688 (avx512f_cmp<mode>3): Ditto.
1689 (avx512f_vec_dup<mode>): Ditto.
1690 (avx512f_vec_dup_mem<mode>): Ditto.
1691 (avx512f_vpermi2var<mode>3): Ditto.
1692 (avx512f_vpermt2var<mode>3): Ditto.
1693 (vec_init<mode>): Ditto.
1694 (avx512f_gathersi<mode>): Ditto.
1695 (*avx512f_gathersi<mode>): Ditto.
1696 (*avx512f_gathersi<mode>_2): Ditto.
1697 (avx512f_gatherdi<mode>): Ditto.
1698 (*avx512f_gatherdi<mode>): Ditto.
1699 (*avx512f_gatherdi<mode>_2): Ditto.
1700 (avx512f_scattersi<mode>): Ditto.
1701 (*avx512f_scattersi<mode>): Ditto.
1702 (avx512f_scatterdi<mode>): Ditto.
1703 (*avx512f_scatterdi<mode>): Ditto.
1704 (sseintprefix): Extened with wider modes.
1705 (VEC_GATHER_IDXSI): Ditto.
1706 (VEC_GATHER_IDXDI): Ditto.
1707 (VEC_GATHER_SRCDI): Ditto.
1709 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1710 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1712 * config/arm/t-aprofile: New file.
1713 * config.gcc: Handle --with-multilib-list option.
1715 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1717 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1720 2013-10-15 Richard Biener <rguenther@suse.de>
1722 * tree-loop-distribution.c (build_empty_rdg): Inline into
1724 (rdg_flag_vertex): Inline into single user.
1725 (rdg_flag_vertex_and_dependent): Likewise.
1726 (build_rdg_partition_for_vertex): Remove processed bitmap.
1727 (rdg_build_partitions): Simplify.
1729 2013-10-15 Richard Biener <rguenther@suse.de>
1731 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1732 Restructure forwarding through conversions and copies to
1733 avoid performing copy-propagation the wrong way. Adjust
1734 recursion invocations.
1735 (forward_propagate_addr_expr): Add argument stating if we
1736 are recursing from a single-use.
1737 (ssa_forward_propagate_and_combine): Adjust.
1739 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1741 * dumpfile.h (gcc::dump_manager): New class, to hold state
1742 relating to dumpfile management.
1743 (get_dump_file_name): Remove in favor of method of dump_manager.
1744 (dump_initialized_p): Likewise.
1745 (dump_start): Likewise.
1746 (dump_finish): Likewise.
1747 (dump_switch_p): Likewise.
1748 (dump_register): Likewise.
1749 (get_dump_file_info): Likewise.
1750 * context.c (gcc::context::context): Construct the dump_manager
1752 * context.h (gcc::context::get_dumps): New.
1753 (gcc::context::m_dumps): New.
1754 * coverage.c (coverage_init): Port to dump_manager API.
1755 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1756 (extra_dump_files_in_use): Likewise.
1757 (extra_dump_files_alloced): Likewise.
1758 (gcc::dump_manager::dump_manager): New.
1759 (dump_register): Convert to...
1760 (gcc::dump_manager::dump_register): ...method, replacing
1761 function-static next_dump with m_next_dump field.
1762 (get_dump_file_info): Convert to...
1763 (gcc::dump_manager::get_dump_file_info): ...method.
1764 (get_dump_file_name): Convert to...
1765 (gcc::dump_manager::get_dump_file_name): ...method.
1766 (dump_start): Convert to...
1767 (gcc::dump_manager::dump_start): ...method.
1768 (dump_finish): Convert to...
1769 (gcc::dump_manager::dump_finish): ...method.
1770 (dump_begin): Replace body with...
1771 (gcc::dump_manager::dump_begin): ...new method.
1772 (dump_phase_enabled_p): Convert to...
1773 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1774 (dump_phase_enabled_p): Convert to...
1775 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1776 (dump_initialized_p): Convert to...
1777 (gcc::dump_manager::dump_initialized_p): ...method.
1778 (dump_flag_name): Replace body with...
1779 (gcc::dump_manager::dump_flag_name): ...new method.
1780 (dump_enable_all): Convert to...
1781 (gcc::dump_manager::dump_enable_all): ...new method.
1782 (opt_info_enable_passes): Convert to...
1783 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1784 (dump_switch_p_1): Convert to...
1785 (gcc::dump_manager::dump_switch_p_1): ...new method.
1786 (dump_switch_p): Convert to...
1787 (gcc::dump_manager::dump_switch_p): ...new method.
1788 (opt_info_switch_p): Port to dump_manager API.
1789 (enable_rtl_dump_file): Likewise.
1790 * opts-global.c (handle_common_deferred_options): Port to new
1792 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1793 (pass_manager::register_one_dump_file): Likewise.
1794 (pass_manager::register_pass): Likewise.
1795 (pass_init_dump_file): Likewise.
1796 (pass_fini_dump_file): Likewise.
1797 * statistics.c (statistics_early_init): Likewise.
1799 2013-10-14 Richard Biener <rguenther@suse.de>
1801 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1802 iterative_hash_canonical_type, gimple_canonical_type_hash,
1803 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1804 gimple_register_canonical_type, print_gimple_types_stats,
1805 free_gimple_type_tables): Move to lto/lto.c
1806 (gt-gimple.h): Do not include.
1807 * gimple.h (gimple_register_canonical_type,
1808 print_gimple_types_stats, free_gimple_type_tables): Remove.
1809 * Makefile.in (GTFILES): Remove gimple.c.
1811 2013-10-14 Travis Snoozy <quandary@remstate.com>
1814 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1815 scanning for msp430x targets.
1817 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1820 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1821 (registers_ok_for_ldd_peep): Move around.
1822 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1823 * config/sparc/sparc.md (widening peepholes): Use it.
1825 2013-10-14 Richard Biener <rguenther@suse.de>
1829 * gimple.c (iterative_hash_canonical_type): Make sure to
1830 record the hash into the correct hashtable slot.
1832 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1834 PR rtl-optimization/58662
1835 * combine.c (try_combine): Take into account death nodes on I2 when
1836 splitting a PARALLEL of two independent SETs. Fix dump message.
1838 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1841 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1842 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1843 * config/sh/t-sh (sh_treg_combine.o): New entry.
1844 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1845 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1846 (register_sh_passes): New function. Register sh_treg_combine pass.
1847 (sh_option_override): Invoke it.
1848 (sh_canonicalize_comparison): Handle op0_preserve_value.
1849 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1850 opportunities. Canonicalize branch condition.
1851 (nott): Allow only if pseudos can be created for non-SH2A.
1853 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1856 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1857 (ix86_expand_movmem): Replace copy_addr_to_reg with
1858 ix86_copy_addr_to_reg.
1859 (ix86_expand_setmem): Likewise.
1861 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1863 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1864 provided by ix86_fp_compare_mode instead of CCFPUmode.
1866 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1868 * config/aarch64/arm_neon.h
1869 (vtbx<1,3>_<psu>8): Fix register constriants.
1871 2013-10-11 Jeff Law <law@redhat.com>
1873 PR tree-optimization/58640
1874 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1875 threading paths that cross over two loop entry points.
1877 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1879 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1880 handle vector float as well.
1881 (*vsx_le_perm_load_v4si): Likewise.
1882 (*vsx_le_perm_store_v2di): Likewise.
1883 (*vsx_le_perm_store_v4si): Likewise.
1885 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1887 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1888 directly to circumvent subtract from splat{31} workaround.
1889 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1891 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1892 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1893 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1894 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1895 separate big and little endian logic.
1896 (*altivec_vperm_<mode>_internal): New define_insn.
1897 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1898 separate big and little endian logic.
1899 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1900 (vec_permv16qi): Add little endian logic.
1902 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1904 * doc/extend.texi (returns_nonnull): Remove arguments.
1906 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1907 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1908 Sergey Lega <sergey.s.lega@intel.com>
1909 Anna Tikhonova <anna.tikhonova@intel.com>
1910 Ilya Tocar <ilya.tocar@intel.com>
1911 Andrey Turetskiy <andrey.turetskiy@intel.com>
1912 Ilya Verbin <ilya.verbin@intel.com>
1913 Kirill Yukhin <kirill.yukhin@intel.com>
1914 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1916 * config/i386/sse.md (VI48F_256_512): New.
1917 (avx2_permvar<mode>): Change to ...
1918 (<avx2_avx512f>_permvar<mode>): This.
1920 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1921 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1922 Sergey Lega <sergey.s.lega@intel.com>
1923 Anna Tikhonova <anna.tikhonova@intel.com>
1924 Ilya Tocar <ilya.tocar@intel.com>
1925 Andrey Turetskiy <andrey.turetskiy@intel.com>
1926 Ilya Verbin <ilya.verbin@intel.com>
1927 Kirill Yukhin <kirill.yukhin@intel.com>
1928 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1930 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1931 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1932 * config/i386/sse.md (VI4_AVX): New.
1933 (sf2simodelower): Ditto.
1934 (sse2_cvtps2dq): Change to ...
1935 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1937 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1938 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1939 Sergey Lega <sergey.s.lega@intel.com>
1940 Anna Tikhonova <anna.tikhonova@intel.com>
1941 Ilya Tocar <ilya.tocar@intel.com>
1942 Andrey Turetskiy <andrey.turetskiy@intel.com>
1943 Ilya Verbin <ilya.verbin@intel.com>
1944 Kirill Yukhin <kirill.yukhin@intel.com>
1945 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1947 * config/i386/sse.md (V_512): New.
1949 (vcond<V_512:mode><VF_512:mode>): Ditto.
1950 (vcond<V_512:mode><VI_512:mode>): Ditto.
1951 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1953 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1954 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1955 Sergey Lega <sergey.s.lega@intel.com>
1956 Anna Tikhonova <anna.tikhonova@intel.com>
1957 Ilya Tocar <ilya.tocar@intel.com>
1958 Andrey Turetskiy <andrey.turetskiy@intel.com>
1959 Ilya Verbin <ilya.verbin@intel.com>
1960 Kirill Yukhin <kirill.yukhin@intel.com>
1961 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1963 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1964 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1966 (fma<mode>4): Removed condition.
1967 (fms<mode>4): Ditto.
1968 (fnma<mode>4): Ditto.
1969 (fnms<mode>4): Ditto.
1970 (fma4i_fmadd_<mode>): Ditto.
1971 (*fma_fmadd_<mode>): Ditto.
1972 (*fma_fmsub_<mode>): Ditto.
1973 (*fma_fnmadd_<mode>): Ditto.
1974 (*fma_fnmsub_<mode>): Ditto.
1975 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1976 (*fma_fmaddsub_<mode>): Ditto.
1977 (*fma_fmsubadd_<mode>): Ditto.
1978 (*fmai_fmadd_<mode>): Ditto.
1979 (*fmai_fmsub_<mode>): Ditto.
1980 (*fmai_fnmadd_<mode>): Ditto.
1981 (*fmai_fnmsub_<mode>): Ditto.
1983 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1984 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1985 Sergey Lega <sergey.s.lega@intel.com>
1986 Anna Tikhonova <anna.tikhonova@intel.com>
1987 Ilya Tocar <ilya.tocar@intel.com>
1988 Andrey Turetskiy <andrey.turetskiy@intel.com>
1989 Ilya Verbin <ilya.verbin@intel.com>
1990 Kirill Yukhin <kirill.yukhin@intel.com>
1991 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1993 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1994 (VI124_256): Changed to ...
1995 (VI124_256_48_512): This.
1996 (ssepackmode): Extended with wider modes.
1997 (<code><mode>3): Changed iterator.
1998 (*avx2_<code><mode>3): Ditto.
1999 (vec_pack_trunc_<mode>): Ditto.
2001 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2002 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2003 Sergey Lega <sergey.s.lega@intel.com>
2004 Anna Tikhonova <anna.tikhonova@intel.com>
2005 Ilya Tocar <ilya.tocar@intel.com>
2006 Andrey Turetskiy <andrey.turetskiy@intel.com>
2007 Ilya Verbin <ilya.verbin@intel.com>
2008 Kirill Yukhin <kirill.yukhin@intel.com>
2009 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2011 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2012 (VI8F_256_512): Ditto.
2013 (abs<mode>2): Changed iterator.
2014 (avx2_perm<mode>): Changed to ...
2015 (<avx2_avx512f>_perm<mode>): This.
2016 (avx2_perm<mode>_1): Changed to ...
2017 (<avx2_avx512f>_perm<mode>_1): This.
2019 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2020 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2021 Sergey Lega <sergey.s.lega@intel.com>
2022 Anna Tikhonova <anna.tikhonova@intel.com>
2023 Ilya Tocar <ilya.tocar@intel.com>
2024 Andrey Turetskiy <andrey.turetskiy@intel.com>
2025 Ilya Verbin <ilya.verbin@intel.com>
2026 Kirill Yukhin <kirill.yukhin@intel.com>
2027 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2029 * config/i386/sse.md (VI48_AVX512F): New.
2030 (VI48_AVX2): Changed to ...
2031 (VI48_AVX2_48_AVX512F): This.
2032 (avx2_ashrv<mode>): Changed to ...
2033 (<avx2_avx512f>_ashrv<mode>): This.
2034 (avx2_<shift_insn>v<mode>): Changed to ...
2035 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2037 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2038 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2039 Sergey Lega <sergey.s.lega@intel.com>
2040 Anna Tikhonova <anna.tikhonova@intel.com>
2041 Ilya Tocar <ilya.tocar@intel.com>
2042 Andrey Turetskiy <andrey.turetskiy@intel.com>
2043 Ilya Verbin <ilya.verbin@intel.com>
2044 Kirill Yukhin <kirill.yukhin@intel.com>
2045 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2047 * config/i386/sse.md (VI4_AVX512F): New.
2048 (VI8_AVX2_AVX512F): Ditto.
2049 (mul<mode>3): Extended with wider modes.
2050 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2051 (mul<mode>3): Ditto.
2052 (vec_widen_<s>mult_odd_<mode>): Ditto.
2054 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2055 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2056 Sergey Lega <sergey.s.lega@intel.com>
2057 Anna Tikhonova <anna.tikhonova@intel.com>
2058 Ilya Tocar <ilya.tocar@intel.com>
2059 Andrey Turetskiy <andrey.turetskiy@intel.com>
2060 Ilya Verbin <ilya.verbin@intel.com>
2061 Kirill Yukhin <kirill.yukhin@intel.com>
2062 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2064 * config/i386/sse.md (VI2_AVX512F): New.
2065 (VI124_AVX512F): Ditto.
2066 (sseunpackmode): Extended with wider modes.
2067 (sseunpackfltmode): Ditto.
2068 (vec_unpacks_float_hi_<mode>): Ditto.
2069 (vec_unpacks_float_lo_<mode>): Ditto.
2070 (vec_unpacku_float_hi_<mode>): Ditto.
2071 (vec_unpacku_float_lo_<mode>): Ditto.
2072 (vec_unpacks_lo_<mode>): Ditto.
2073 (vec_unpacks_hi_<mode>): Ditto.
2074 (vec_unpacku_lo_<mode>): Ditto.
2075 (vec_unpacku_hi_<mode>): Ditto.
2077 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2078 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2079 Sergey Lega <sergey.s.lega@intel.com>
2080 Anna Tikhonova <anna.tikhonova@intel.com>
2081 Ilya Tocar <ilya.tocar@intel.com>
2082 Andrey Turetskiy <andrey.turetskiy@intel.com>
2083 Ilya Verbin <ilya.verbin@intel.com>
2084 Kirill Yukhin <kirill.yukhin@intel.com>
2085 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2087 * config/i386/i386.md (multdiv): New.
2088 (multdiv_mnemonic): Ditto.
2089 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2090 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2091 (<sse>_vmdiv<mode>3): Removed.
2093 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2094 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2095 Sergey Lega <sergey.s.lega@intel.com>
2096 Anna Tikhonova <anna.tikhonova@intel.com>
2097 Ilya Tocar <ilya.tocar@intel.com>
2098 Andrey Turetskiy <andrey.turetskiy@intel.com>
2099 Ilya Verbin <ilya.verbin@intel.com>
2100 Kirill Yukhin <kirill.yukhin@intel.com>
2101 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2103 * config/i386/sse.md (V): Extended with wider modes.
2105 (ssehalfvecmode): Ditto.
2107 (ssepackfltmode): Ditto.
2108 (avx_vec_concat<mode>): Ditto.
2109 (V_256_512): New iterator.
2110 (VF2_512_256): Ditto.
2111 (si2dfmode): New attribute.
2112 (si2dfmodelower): Ditto.
2114 (concat_tg_mode): Ditto.
2115 (floatv4siv4df2): Changed to ...
2116 (float<si2dfmodelower><mode>2): This.
2117 (avx_cvtps2pd256): Changed to ...
2118 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2119 (vec_pack_trunc_v4df): Changed to ...
2120 (vec_pack_trunc_<mode>): This.
2121 (avx_vpermil<mode>): Changed to ...
2122 (<sse2_avx_avx512f>_vpermil<mode>): This.
2123 (<fixsuffix>fix_truncv8dfv8si2): New.
2124 (vec_pack_sfix_trunc_v8df): Ditto.
2125 (avx512f_rndscale<mode>): Ditto.
2126 (avx512f_roundpd512): Ditto.
2127 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2129 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2130 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2131 Sergey Lega <sergey.s.lega@intel.com>
2132 Anna Tikhonova <anna.tikhonova@intel.com>
2133 Ilya Tocar <ilya.tocar@intel.com>
2134 Andrey Turetskiy <andrey.turetskiy@intel.com>
2135 Ilya Verbin <ilya.verbin@intel.com>
2136 Kirill Yukhin <kirill.yukhin@intel.com>
2137 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2139 * config/i386/i386.md (any_fix): New iterator.
2140 (fixsuffix): New attribute.
2141 * config/i386/sse.md (VF1): Extened with wider modes.
2145 (sseintvecmodelower): Ditto.
2146 (ssescalarmode): Ditto.
2147 (ssescalarnum): Ditto.
2149 (ssexmmmode): Ditto.
2150 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2151 (<sse>_rcp<mode>2): Change iterator.
2152 (rsqrt<mode>2): Ditto.
2153 (<sse>_rsqrt<mode>2): Ditto.
2154 (avx2_vec_dup<mode>): Ditto.
2155 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2156 (round<mode>2_sfix): Ditto.
2157 (avx2_pbroadcast<mode>): Ditto.
2158 (*andnot<mode>3): Handle XI mode.
2159 (*<code><mode>3): Ditto.
2160 (AVXTOSSEMODE): Removed.
2161 (avx_vpermil<mode>): Changed to ...
2162 (<sse2_avx_avx512f>_vpermil<mode>): This.
2164 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2165 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2166 Sergey Lega <sergey.s.lega@intel.com>
2167 Anna Tikhonova <anna.tikhonova@intel.com>
2168 Ilya Tocar <ilya.tocar@intel.com>
2169 Andrey Turetskiy <andrey.turetskiy@intel.com>
2170 Ilya Verbin <ilya.verbin@intel.com>
2171 Kirill Yukhin <kirill.yukhin@intel.com>
2172 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2174 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2175 (<sse>_comi): Ditto.
2176 (<sse>_ucomi): Ditto.
2177 (sse_cvtss2siq_2): Ditto.
2178 (sse2_cvtsd2si): Ditto.
2179 (sse2_cvtsd2siq): Ditto.
2180 (sse2_cvttsd2si): Ditto.
2181 (sse2_cvttsd2siq): Ditto.
2182 (<shift_insn><mode>3): Ditto.
2183 (sse2_cvtsi2sdq): Update constraint and prefix.
2184 (sse_cvtsi2ss): Update prefix.
2185 (sse_cvtsi2ssq): Ditto.
2187 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2189 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2190 ignore internal calls.
2192 2013-10-11 Richard Biener <rguenther@suse.de>
2194 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2195 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2196 TREE_OVERFLOW is set.
2198 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2200 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2202 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2203 * gimple.h: Likewise.
2205 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2207 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2210 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2212 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2213 * configure: Regenerate.
2215 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2217 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2218 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2219 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2220 similarly pp_right_paren.
2221 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2222 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2223 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2224 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2226 (omp_declare_simd_clauses_equal,
2227 omp_remove_redundant_declare_simd_attrs): New functions.
2228 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2229 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2230 * tree.h (OMP_LOOP_CHECK): Define.
2231 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2232 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2233 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2234 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2235 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2236 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2237 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2238 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2239 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2240 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2241 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2242 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2243 (omp_remove_redundant_declare_simd_attrs): New prototype.
2244 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2245 GIMPLE_OMP_TEAMS): New codes.
2246 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2247 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2249 (target_nesting_level): New variable.
2250 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2251 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2252 for collapse > 1 static schedule unless ordered.
2253 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2254 (determine_parallel_type): Adjust get_ws_args_for caller.
2255 (install_var_field): Handle mask & 4 for double indirection.
2256 (scan_sharing_clauses): Ignore shared clause on teams construct.
2257 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2258 (create_omp_child_function): If inside target or declare target
2259 constructs, set "omp declare target" attribute on the child function.
2260 (find_combined_for): New function.
2261 (scan_omp_parallel): Handle combined loops.
2262 (scan_omp_target, scan_omp_teams): New functions.
2263 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2264 restrictions and set ctx->cancellable for cancellable constructs.
2265 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2266 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2267 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2268 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2269 (omp_clause_aligned_alignment): New function.
2270 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2271 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2272 on teams constructs. Handle user defined reductions and new
2274 (lower_reduction_clauses): Don't set placeholder to address of ref
2275 if it has already the right type.
2276 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2277 (expand_parallel_call): Use the new non-_start suffixed builtins,
2278 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2279 and GOMP_parallel_end after the call.
2280 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2281 (expand_omp_for_init_counts): Handle combined loops.
2282 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2284 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2285 end of cancellable loops.
2286 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2287 Likewise. Handle collapse > 1 loops.
2288 (expand_omp_simd): Handle combined loops.
2289 (expand_omp_for): Add inner_stmt argument, adjust callers of
2290 expand_omp_for* functions, use expand_omp_for_static*chunk even
2291 for collapse > 1 unless ordered.
2292 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2293 of cancellable sections.
2294 (expand_omp_single): Remove need_barrier variable, just rely on
2295 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2296 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2297 (expand_omp_atomic_load, expand_omp_atomic_store,
2298 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2299 (expand_omp_target): New function.
2300 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2301 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2302 (build_omp_regions_1): Immediately close region for
2303 GF_OMP_TARGET_KIND_UPDATE.
2304 (maybe_add_implicit_barrier_cancel): New function.
2305 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2307 (lower_omp_single): Likewise. Add clobber after the barrier.
2308 (lower_omp_taskgroup): New function.
2309 (lower_omp_for): Handle combined loops. Adjust
2310 lower_rec_input_clauses caller. Handle cancellation.
2311 (lower_depend_clauses): New function.
2312 (lower_omp_taskreg): Lower depend clauses. Adjust
2313 lower_rec_input_clauses caller. Add clobber after the call. Handle
2315 (lower_omp_target, lower_omp_teams): New functions.
2316 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2317 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2318 and GOMP_cancellation_point calls.
2319 (lower_omp): Fold stmts inside of target region.
2320 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2321 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2322 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2323 (BT_FN_VOID_OMPFN_PTR_UINT,
2324 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2325 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2326 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2327 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2328 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2329 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2330 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2331 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2332 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2333 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2334 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2335 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2336 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2337 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2338 BUILT_IN_GOMP_PARALLEL_END.
2339 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2340 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2341 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2342 GF_OMP_FOR_KIND_DISTRIBUTE.
2343 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2344 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2345 (dump_gimple_omp_return): Print lhs if it has any.
2346 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2347 gimple_omp_atomic_seq_cst_p.
2348 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2349 and GIMPLE_OMP_TEAMS.
2350 * langhooks.c (lhd_omp_mappable_type): New function.
2351 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2352 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2354 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2355 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2356 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2357 (struct gimplify_omp_ctx): Add combined_loop field.
2358 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2359 on stmts inside of target region.
2360 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2361 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2362 ORT_TARGET and ORT_TARGET_DATA.
2363 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2365 (omp_notice_threadprivate_variable): Complain about threadprivate
2366 variables in target region.
2367 (omp_notice_variable): Complain about vars with non-mappable type
2368 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2369 (omp_check_private): Ignore ORT_TARGET* regions.
2370 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2371 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2372 (find_combined_omp_for): New function.
2373 (gimplify_omp_for): Handle gimplification of combined loops.
2374 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2376 (gimplify_omp_target_update): New function.
2377 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2378 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2379 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2380 (gimplify_body): If fndecl has "omp declare target" attribute, add
2381 implicit ORT_TARGET context around it.
2382 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2383 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2384 * tree-nested.c (convert_nonlocal_reference_stmt,
2385 convert_local_reference_stmt, convert_gimple_call): Handle
2386 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2387 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2388 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2389 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2390 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2391 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2392 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2393 BUILT_IN_GOMP_LOOP_END_CANCEL,
2394 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2395 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2396 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2397 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2398 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2399 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2400 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2401 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2402 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2403 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2404 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2405 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2406 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2407 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2408 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2409 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2410 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2411 gimple_build_omp_teams): New functions.
2412 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2413 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2414 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2415 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2416 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2417 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2418 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2419 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2420 GF_OMP_ATOMIC_SEQ_CST): New.
2421 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2422 gimple_build_omp_teams): New prototypes.
2423 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2424 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2425 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2426 GIMPLE_OMP_SINGLE as end of range.
2427 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2428 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2429 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2430 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2431 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2432 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2433 gimple_omp_target_kind, gimple_omp_target_set_kind,
2434 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2435 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2436 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2437 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2438 gimple_omp_teams_set_clauses): New inlines.
2439 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2440 and GIMPLE_OMP_TASKGROUP.
2441 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2442 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2443 enum omp_clause_proc_bind_kind): New.
2444 (union omp_clause_subcode): Add depend_kind, map_kind and
2445 proc_bind_kind fields.
2446 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2447 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2448 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2449 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2450 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2452 2013-10-10 Teresa Johnson <tejohnson@google.com>
2454 * predict.c (tree_estimate_probability): Add new parameter
2455 for estimate_bb_frequencies.
2456 (estimate_bb_frequencies): Add new parameter to force estimation.
2457 (rebuild_frequencies): When max frequency in function is small,
2458 recompute counts from frequencies.
2459 * predict.h (estimate_bb_frequencies): New parameter.
2461 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2463 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2464 optimizations are disabled.
2466 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2468 * coverage.c (coverage_finish): Fix leak of da_file_name.
2470 2013-10-10 Jan Hubicka <jh@suse.cz>
2472 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2473 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2476 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2479 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2480 if any labels are in FALLTHRU_BB, use a special label emitted
2481 immediately after the asm goto insn rather than label_rtx
2483 (expand_asm_stmt): Adjust caller.
2484 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2485 edge if the last insn in predecessor is a jump with single successor,
2486 but it isn't simplejump_p.
2488 2013-10-10 Richard Biener <rguenther@suse.de>
2490 PR tree-optimization/58656
2491 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2493 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2495 * gimplify.c: Include expr.h and tm_p.h for targets with special
2496 va-arg padding requirements.
2498 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2500 * tree-flow.h: Move some prototypes to gimple.h.
2501 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2502 * gimple.h: Relocate some prototypes from tree-flow.h
2503 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2505 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2506 (build_va_arg_indirect_ref): Relocate and make static.
2507 (std_gimplify_va_arg_expr): Relocate here.
2508 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2509 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2511 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2513 * doc/md.texi: Document the mnemonic attribute.
2515 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2518 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2519 (set_attr_alternative x ...) when searching for user defined
2522 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2524 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2526 2013-10-09 Easwaran Raman <eraman@google.com>
2528 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2529 * cfgexpand.c (defer_stack_allocation): ...use here
2530 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2532 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2534 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2535 (operand_equal_for_value_replacement): New function, extracted from
2536 value_replacement and enhanced to catch more cases.
2537 (value_replacement): Use operand_equal_for_value_replacement.
2539 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2541 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2542 get_max_loop_iterations.
2544 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2546 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2547 Place comment above function.
2549 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2551 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2552 are not related to tree-cfg.c.
2553 * tree-ssa-address.h: New file. Relocate prototypes.
2554 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2555 (addr_for_mem_ref): New. Combine call to get_address_description and
2556 return addr_for_mem_ref.
2557 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2558 * tree-ssa-live.h: Adjust prototypes.
2559 * passes.c: Include tree-ssa-live.h.
2560 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2561 * graphite.c (graphite_transform_loops): Make static.
2562 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2563 make_pass_graphite, pass_data_graphite_transforms,
2564 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2565 * ipa-pure-const.c (warn_function_noreturn): Make static.
2566 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2567 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2568 Relocate from tree-cfg.c
2569 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2571 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2572 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2573 Move to ipa-pure-const.c.
2574 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2575 Relocate from tree-optimize.c.
2576 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2577 make_pass_fixup_cfg): Move to tree-cfg.c.
2578 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2579 Relocate some prototypes.
2580 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2581 * tree-dump.c (dump_function_to_file): Remove prototype.
2582 Add tree-flow.h to the include file.
2583 * tree-dump.h: Remove prototype.
2584 * tree-parloops.h: New File. Add prototypes.
2585 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2586 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2587 from tree-ssa-loop.c.
2588 * tree-predcom.c (run_tree_predictive_commoning,
2589 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2590 Relocate here from tree-ssa-loop.c.
2591 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2592 ssa_name_values.release ().
2593 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2594 (ssa_name_values): Relocate from tree-flow.h.
2595 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2596 * tree-ssa-loop.c (run_tree_predictive_commoning,
2597 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2598 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2599 make_pass_graphite, pass_data_graphite_transforms,
2600 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2601 tree_parallelize_loops, pass_data_parallelize_loops,
2602 make_pass_parallelize_loops): Move to other files.
2603 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2605 * tree.h: Remove prototypes from tree-address.c.
2607 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2609 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2610 (struct int_tree_map): Move to tree-hasher.h
2611 (SCALE, LABEL, PERCENT): Move to gimple.h
2612 * tree-flow-inline.h: Delete. Move functions to other files.
2613 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2614 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2615 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2616 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2617 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2618 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2619 * tree-hasher.h: Don't include tree-flow.h.
2620 (struct int_tree_map): Relocate from tree-flow.h.
2621 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2622 tree-flow-inline.h and make static.
2623 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2625 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2626 tree-flow-inline.h and make static.
2627 * tree.h (is_global_var, may_be_aliased): Relocate from
2629 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2630 * value-prof.c: No longer include tree-flow-inline.h.
2631 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2633 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2635 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2636 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2637 (enum move_pos): Move to tree-ssa-loop-im.h
2638 * cfgloop.h: Move some prototypes.
2639 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2640 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2641 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2642 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2643 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2644 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2645 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2646 (enum move_pos): Relocate here.
2647 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2649 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2650 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2651 make_pass_lim): Relocate here from tree-ssa-loop.c.
2652 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2654 (loop_edge_to_cancel, unloop_loops): Make static.
2655 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2656 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2657 (tree_complete_unroll, gate_tree_complete_unroll,
2658 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2659 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2660 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2661 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2662 (stmt_invariant_in_loop_p): Remove unused function.
2663 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2664 * tree-ssa-loop-manip.h: New file. Add prototypes.
2665 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2666 (gcov_type_to_double_int): Move to cfgloop.h.
2667 (double_int_cmp, bound_index,
2668 estimate_numbers_of_iterations_loop): Make static.
2669 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2670 (max_loop_iterations): Factor out get_max_loop_iterations.
2671 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2673 * tree-ssa-loop-niter.h: New file. Add prototypes.
2674 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2675 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2676 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2677 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2678 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2679 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2680 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2681 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2682 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2683 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2684 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2685 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2686 pass_data_complete_unroll, make_pass_complete_unroll,
2687 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2688 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2689 tree-ssa-loop-ivcanon.c.
2690 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2691 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2692 tree-ssa-loop-prefetch.c.
2693 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2695 (get_lsm_tmp_name): Relocate and add suffix parameter.
2696 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2697 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2698 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2699 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2700 (get_estimated_loop_iterations): Factor out accessor from
2701 estimated_loop_iterations in tree-ssa-loop-niter.c.
2702 (get_max_loop_iterations): Factor out accessor from
2703 _max_loop_iterations in tree-ssa-niter.c.
2704 * loop-unroll.c (decide_unroll_constant_iterations,
2705 decide_unroll_runtime_iterations, decide_peel_simple,
2706 decide_unroll_stupid): Use new get_* accessors.
2708 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2710 PR tree-optimization/20318
2711 * doc/extend.texi (returns_nonnull): New function attribute.
2712 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2714 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2715 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2717 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2720 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2721 false if both components are bitfields.
2723 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2725 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2726 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2727 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2728 * config/aarch64/aarch64.h
2729 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2730 * config/aarch64/aarch64-simd-builtins.def
2731 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2733 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2735 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2737 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2739 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2741 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2743 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2744 (vneg_f64): New intrinsic.
2745 (vneg_s8): Asm replaced with C.
2746 (vneg_s16): Likewise.
2747 (vneg_s32): Likewise.
2748 (vneg_s64): New intrinsic.
2749 (vnegq_f32): Asm replaced with C.
2750 (vnegq_f64): Likewise.
2751 (vnegq_s8): Likewise.
2752 (vnegq_s16): Likewise.
2753 (vnegq_s32): Likewise.
2754 (vnegq_s64): Likewise.
2756 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2758 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2760 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2762 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2763 packed stack special handling.
2764 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2765 back to fixed stack slots for FPRs saved due to stdarg.
2767 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2769 * config/s390/s390.c (s390_frame_info): Restructure function.
2771 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2773 * config/s390/s390.c (struct s390_frame_layout): New field
2775 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2776 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2777 regs_ever_clobbered to char*.
2778 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2779 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2780 (s390_register_info): Enable FPR save slots. Move/Copy some
2781 functionality into ...
2782 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2783 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2785 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2786 not imply a stack frame.
2787 (s390_init_frame_layout): Remove variable clobbered_regs.
2788 (s390_update_register_info): Remove function.
2789 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2790 cannot be used for register renaming.
2791 (s390_hard_regno_scratch_ok): New function.
2792 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2793 (s390_initial_elimination_offset): Change offset calculation of
2794 the return address pointer.
2795 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2796 (restore_gprs): Set frame related flag.
2797 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2798 (s390_emit_prologue): Call s390_register_info instead of
2799 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2800 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2801 (s390_optimize_prologue): Call s390_optimize_register_info.
2802 Try to remove also FPR slot save/restore INSNs. Remove frame
2803 related flags from restore INSNs.
2805 2013-10-08 DJ Delorie <dj@redhat.com>
2807 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2810 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2811 avoid conflict with the MI use of %c.
2812 * config/rl78/rl78-real.md: change %c to %C throughout.
2813 * config/rl78/rl78-virt.md: Likewise.
2815 2013-10-08 Jan Hubicka <jh@suse.cz>
2817 * config/i386/i386.c (ix86_option_override_internal): Switch
2818 to SSE math for -ffast-math when target ISA supports SSE2.
2820 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2822 * tree-flow.h: Remove some prototypes.
2823 * tree.h: Remove some protypes, add a couple.
2824 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2825 using_eh_for_cleanups_p): Add interface routines for front ends.
2826 * tree-eh.h: New file. Add protoptyes.
2827 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2828 (add_stmt_to_eh_lp_fn): Make static.
2829 (lower_try_finally): Use new using_eh_for_cleanups_p.
2830 * emit-rtl.c: Include tree-eh.h.
2831 * gimple.h: Include tree-eh.h.
2833 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2835 PR tree-optimization/58480
2836 * tree-vrp.c (infer_nonnull_range): New function.
2837 (infer_value_range): Call infer_nonnull_range.
2839 2013-10-08 Dehao Chen <dehao@google.com>
2841 PR tree-optimization/58619
2842 * tree-inline.c (copy_phis_for_bb): Combine location data
2845 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2848 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2849 RTX_FRAME_RELATED_P on INSN.
2851 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2853 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2854 (altivec_expand_vec_perm_const): Call it.
2856 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2858 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2859 sequences for LE VSX loads and stores at expand time.
2860 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2862 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2863 (rs6000_gen_le_vsx_permute): New.
2864 (rs6000_gen_le_vsx_load): New.
2865 (rs6000_gen_le_vsx_store): New.
2866 (rs6000_gen_le_vsx_move): New.
2867 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2868 (*vsx_le_perm_load_v4si): New.
2869 (*vsx_le_perm_load_v8hi): New.
2870 (*vsx_le_perm_load_v16qi): New.
2871 (*vsx_le_perm_store_v2di): New.
2872 (*vsx_le_perm_store_v4si): New.
2873 (*vsx_le_perm_store_v8hi): New.
2874 (*vsx_le_perm_store_v16qi): New.
2875 (*vsx_xxpermdi2_le_<mode>): New.
2876 (*vsx_xxpermdi4_le_<mode>): New.
2877 (*vsx_xxpermdi8_le_V8HI): New.
2878 (*vsx_xxpermdi16_le_V16QI): New.
2879 (*vsx_lxvd2x2_le_<mode>): New.
2880 (*vsx_lxvd2x4_le_<mode>): New.
2881 (*vsx_lxvd2x8_le_V8HI): New.
2882 (*vsx_lxvd2x16_le_V16QI): New.
2883 (*vsx_stxvd2x2_le_<mode>): New.
2884 (*vsx_stxvd2x4_le_<mode>): New.
2885 (*vsx_stxvd2x8_le_V8HI): New.
2886 (*vsx_stxvd2x16_le_V16QI): New.
2888 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2890 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2892 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2894 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2895 loop to work also for 31bit ABI.
2896 Save the stack pointer for frame_size > 0.
2898 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2900 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2901 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2902 constraint letters from expanders.
2903 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2904 retry count to general_operand.
2905 ("tabort"): Give operand 0 a mode.
2906 ("tabort_1"): Add mode and constraint letter for operand 0.
2907 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2909 2013-10-04 Jeff Law <law@redhat.com>
2911 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2913 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2915 (thread_across_edge): Here. Call it.
2917 2013-10-04 Cary Coutant <ccoutant@google.com>
2919 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2920 discarding a location list expression (or a piece of one).
2922 2013-10-03 Jan Hubicka <jh@suse.cz>
2924 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2925 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2926 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2928 2013-10-03 Jan Hubicka <jh@suse.cz>
2930 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2931 accumulate-outgoing-args when producing unwind info.
2933 2013-10-03 Wei Mi <wmi@google.com>
2935 * lra-constraints.c (insert_move_for_subreg): New function
2936 extracted from simplify_operand_subreg.
2937 (simplify_operand_subreg): Add reload for paradoxical subreg.
2939 2013-10-03 Rong Xu <xur@google.com>
2941 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2942 the candidate of builtin_expect such that we should fix the
2943 size/time estimation.
2944 (estimate_function_body_sizes): Do the acutally size/time fix-up
2947 2013-10-03 Rong Xu <xur@google.com>
2949 * predict.c (tree_predict_by_opcode): Get the probability
2950 for builtin_expect from param builtin_expect_probability.
2951 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2952 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2953 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2955 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2958 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2959 * calls.c (alloca_call_p): Use get_callee_fndecl.
2960 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2961 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2963 (vrp_visit_stmt): Remove duplicated code.
2965 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2967 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2968 ceildf2, btruncdf2, instead of vsx_* name.
2970 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2971 iterators to only do V2DF and V4SF here. Move the DF code to
2972 rs6000.md where it is combined with SF mode. Replace <VSv> with
2973 just 'v' since only vector operations are handled with these insns
2974 after moving the DF support to rs6000.md.
2975 (vsx_sub<mode>3): Likewise.
2976 (vsx_mul<mode>3): Likewise.
2977 (vsx_div<mode>3): Likewise.
2978 (vsx_fre<mode>2): Likewise.
2979 (vsx_neg<mode>2): Likewise.
2980 (vsx_abs<mode>2): Likewise.
2981 (vsx_nabs<mode>2): Likewise.
2982 (vsx_smax<mode>3): Likewise.
2983 (vsx_smin<mode>3): Likewise.
2984 (vsx_sqrt<mode>2): Likewise.
2985 (vsx_rsqrte<mode>2): Likewise.
2986 (vsx_fms<mode>4): Likewise.
2987 (vsx_nfma<mode>4): Likewise.
2988 (vsx_copysign<mode>3): Likewise.
2989 (vsx_btrunc<mode>2): Likewise.
2990 (vsx_floor<mode>2): Likewise.
2991 (vsx_ceil<mode>2): Likewise.
2992 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2993 (vsx_sminsf3): Likewise.
2994 (vsx_fmadf4): Likewise.
2995 (vsx_fmsdf4): Likewise.
2996 (vsx_nfmadf4): Likewise.
2997 (vsx_nfmsdf4): Likewise.
2998 (vsx_cmpdf_internal1): Likewise.
3000 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3001 simpler to select whether a target has SPE or traditional floating
3002 point support in iterators.
3003 (TARGET_DF_SPE): Likewise.
3004 (TARGET_SF_FPR): Likewise.
3005 (TARGET_DF_FPR): Likewise.
3006 (TARGET_SF_INSN): Macros to say whether floating point support
3007 exists for a given operation for expanders.
3008 (TARGET_DF_INSN): Likewise.
3010 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3011 combining of SF/DF mode operations, using both traditional and VSX
3019 (abs<mode>2): Combine SF/DF modes using traditional floating point
3020 instructions. Add support for using the upper DF registers with
3021 VSX support, and SF registers with power8-vector support. Update
3022 expanders for operations supported by both the SPE and traditional
3023 floating point units.
3024 (abs<mode>2_fpr): Likewise.
3025 (nabs<mode>2): Likewise.
3026 (nabs<mode>2_fpr): Likewise.
3027 (neg<mode>2): Likewise.
3028 (neg<mode>2_fpr): Likewise.
3029 (add<mode>3): Likewise.
3030 (add<mode>3_fpr): Likewise.
3031 (sub<mode>3): Likewise.
3032 (sub<mode>3_fpr): Likewise.
3033 (mul<mode>3): Likewise.
3034 (mul<mode>3_fpr): Likewise.
3035 (div<mode>3): Likewise.
3036 (div<mode>3_fpr): Likewise.
3037 (sqrt<mode>3): Likewise.
3038 (sqrt<mode>3_fpr): Likewise.
3039 (fre<Fs>): Likewise.
3040 (rsqrt<mode>2): Likewise.
3041 (cmp<mode>_fpr): Likewise.
3042 (smax<mode>3): Likewise.
3043 (smin<mode>3): Likewise.
3044 (smax<mode>3_vsx): Likewise.
3045 (smin<mode>3_vsx): Likewise.
3046 (negsf2): Delete SF operations that are merged with DF.
3053 (fmasf4_fpr): Likewise.
3054 (fmssf4_fpr): Likewise.
3055 (nfmasf4_fpr): Likewise.
3056 (nfmssf4_fpr): Likewise.
3057 (sqrtsf2): Likewise.
3058 (rsqrtsf_internal1): Likewise.
3059 (smaxsf3): Likewise.
3060 (sminsf3): Likewise.
3061 (cmpsf_internal1): Likewise.
3062 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3063 (negdf2): Delete DF operations that are merged with SF.
3065 (nabsdf2): Likewise.
3071 (rsqrtdf_internal1): Likewise.
3072 (fmadf4_fpr): Likewise.
3073 (fmsdf4_fpr): Likewise.
3074 (nfmadf4_fpr): Likewise.
3075 (nfmsdf4_fpr): Likewise.
3076 (sqrtdf2): Likewise.
3077 (smaxdf3): Likewise.
3078 (smindf3): Likewise.
3079 (cmpdf_internal1): Likewise.
3080 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3081 (btrunc<mode>2): Delete separate expander, and combine with the
3082 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3083 (btrunc<mode>2_fpr): Likewise.
3084 (ceil<mode>2): Likewise.
3085 (ceil<mode>2_fpr): Likewise.
3086 (floor<mode>2): Likewise.
3087 (floor<mode>2_fpr): Likewise.
3088 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3089 Add support for using the upper registers with VSX and
3090 power8-vector. Move insns to be closer to the define_expands. On
3091 VSX systems, prefer the traditional form of FMA over the VSX
3092 version, since the traditional form allows the target not to
3093 overlap with the inputs.
3094 (fms<mode>4_fpr): Likewise.
3095 (nfma<mode>4_fpr): Likewise.
3096 (nfms<mode>4_fpr): Likewise.
3098 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3099 Richard Earnshaw <richard.earnshaw@arm.com>
3101 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3102 (struct mult_cost_table): Likewise.
3103 (struct mem_cost_table): Likewise.
3104 (struct fp_cost_table): Likewise.
3105 (struct vector_cost_table): Likewise.
3106 (cpu_cost_table): Likewise.
3107 * config/arm/arm.opt (mold-rts-costs): New option.
3108 (mnew-generic-costs): Likewise.
3109 * config/arm/arm.c (generic_extra_costs): New table.
3110 (cortexa15_extra_costs): Likewise.
3111 (arm_slowmul_tune): Use NULL as new costs.
3112 (arm_fastmul_tune): Likewise.
3113 (arm_strongarm_tune): Likewise.
3114 (arm_xscale_tune): Likewise.
3115 (arm_9e_tune): Likewise.
3116 (arm_v6t2_tune): Likewise.
3117 (arm_cortex_a5_tune): Likewise.
3118 (arm_cortex_a9_tune): Likewise.
3119 (arm_v6m_tune): Likewise.
3120 (arm_fa726te_tune): Likewise.
3121 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3122 (arm_cortex_tune): Use generict_extra_costs.
3123 (shifter_op_p): New function.
3124 (arm_unspec_cost): Likewise.
3125 (LIBCALL_COST): Define.
3126 (arm_new_rtx_costs): New function.
3127 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3128 table is available. Use old costs otherwise unless mnew-generic-costs
3130 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3131 (cpu_cost_table): Declare.
3133 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3136 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3137 (*subs_mul_imm_<mode>)
3138 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3139 (*sub_<shift>_<mode>)
3140 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3141 Remove k constraint.
3143 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3145 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3147 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3149 2013-10-02 Teresa Johnson <tejohnson@google.com>
3151 * predict.c (probably_never_executed): New function.
3152 (probably_never_executed_bb_p): Invoke probably_never_executed.
3153 (probably_never_executed_edge_p): Ditto.
3154 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3155 Treat profile insanities conservatively.
3157 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3159 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3161 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3163 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3164 use it. Use smaller increase for scratch. Don't increase reject
3165 for early clobber scratch.
3166 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3167 setting eliminated regs except setting fp from hfp.
3168 (lra_eliminate): Check lra_insn_recog_data on NULL.
3170 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3173 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3174 setting -mvsx-timode by default until the underlying problem is fixed.
3175 (RS6000_CPU, power7 defaults): Likewise.
3177 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3179 * config/x-linux (host-linux.o): Remove header dependencies.
3180 Use $(COMPILE) and $(POSTCOMPILE).
3181 * config/t-linux-android (linux-android.o): Ditto.
3183 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3185 * Makefile.in (expmed.o-warn): Remove.
3187 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3189 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3190 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3192 2013-10-02 Teresa Johnson <tejohnson@google.com>
3194 * dojump.c (do_jump_1): Divide probability between
3195 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3197 2013-10-02 Tom Tromey <tromey@redhat.com>
3199 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3201 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3203 * tree-flow.h: Remove some prototypes.
3204 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3205 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3206 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3207 mark_virtual_phi_result_for_renaming): Relocate here.
3208 * tree-into-ssa.h: Add prototypes.
3209 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3210 single_pred_before_succ_order.
3211 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3212 (nonfreeing_call_p) Move to gimple.c.
3213 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3215 * basic-block.h (single_pred_before_succ_order): Add prototype.
3216 * gimple.c (nonfreeing_call_p): Relocate here.
3217 * gimple.h: Add prototype.
3218 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3219 * tree-ssa-dom.h: New file. Relocate prototypes here.
3220 * tree-ssa.h: Include tree-ssa-dom.h.
3222 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3224 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3225 Use $(COMPILE) and $(POSTCOMPILE).
3227 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3229 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3231 * tree-flow.h: Remove some prototypes.
3232 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3233 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3234 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3235 propagate_tree_value*): Move from here to...
3236 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3237 propagate_tree_value*): Relocate here.
3238 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3239 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3240 * gimple-fold.c: Remove gimple-fold.h from include list.
3241 * tree-vrp.c: Remove gimple-fold.h from include list.
3242 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3243 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3244 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3245 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3246 * sese.c: Add tree-ssa-propagate.h to include list.
3248 2013-10-02 Richard Biener <rguenther@suse.de>
3250 * tree-loop-distribution.c: Include tree-vectorizer.h for
3252 (enum partition_kind): Remove PKIND_REDUCTION.
3253 (struct partition_s): Remove has_writes member, add reduction_p member.
3254 (partition_alloc): Adjust.
3255 (partition_builtin_p): Likewise.
3256 (partition_has_writes): Remove.
3257 (partition_reduction_p): New function.
3258 (partition_merge_into): Likewise.
3259 (generate_code_for_partition): Commonize builtin partition
3261 (rdg_cannot_recompute_vertex_p): Remove.
3262 (already_processed_vertex_p): Likewise.
3263 (rdg_flag_vertex): Do not set has_writes.
3264 (classify_partition): Adjust.
3265 (rdg_build_partitions): Do not set has_writes, treat all
3266 partitions as useful.
3267 (distribute_loop): Record number of library calls generated. Adjust.
3268 (tree_loop_distribution): Report number of loops and library
3269 calls generated as opt-info.
3271 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3273 * tree-flow.h: Include new .h files. Move prototypes.
3274 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3275 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3276 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3277 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3278 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3279 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3280 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3281 * tree.h (get_ref_base_and_extent): Move prototype out.
3282 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3284 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3285 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3286 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3287 * tree-scalar-evolution.c: Include tree.h.
3288 * sese.c: Include tree.h.
3289 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3290 * dwarf2out.c: Include tree-dfa.h.
3291 * tree-chrec.c: Include tree.h.
3292 * tree-data-ref.c: Include tree.h.
3294 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3296 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3299 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3301 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3302 Use $(COMPILE) and $(POSTCOMPILE).
3303 (sol2-cxx.o): Likewise.
3304 (sol2-stubs.o): Likewise.
3306 * config/x-solaris (host-solaris.o): Likewise.
3308 * config/sparc/t-sparc (sparc.o): Remove.
3309 (sparc-c.o): Remove header dependencies.
3310 Use $(COMPILE) and $(POSTCOMPILE).
3311 * config/sparc/x-sparc: Likewise.
3313 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3315 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3316 * config/arc/arc700.md: Likewise.
3317 * config/arc/arc-modes.def: Likewise.
3318 * config/arc/arc-simd.h: Likewise.
3319 * config/arc/t-arc-uClibc: Likewise.
3320 * config/arc/t-arc-newlib: Likewise.
3322 2013-10-02 Renlin Li <renlin.li@arm.com>
3324 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3326 (aarch64_expand_epilogue): Likewise.
3328 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3329 Yufeng Zhang <yufeng.zhang@arm.com>
3331 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3333 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3334 'base_in' represent a conversion and legal_cast_p_1 holds; set
3335 'base_in' with the returned value from get_unwidened.
3337 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3339 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3340 plus_constant is not used.
3342 2013-10-01 Wei Mi <wmi@google.com>
3344 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3345 * config/i386/i386.md: Add define_peephole2 to
3346 break partial reg stall for cvtss2sd/cvtsd2ss.
3348 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3350 * config/arc/arc.c (pass_arc_ifcvt::clone):
3351 Update for ctxt_ -> m_ctxt change.
3353 2013-10-01 Jeff Law <law@redhat.com>
3355 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3356 outgoing_edge and intermediate_edge fields. Instead store the path.
3357 (redirection_data::hash): Hash on the last edge's destination index.
3358 (redirection_data::equal): Check the entire thread path.
3359 (lookup_redirectio_data): Corresponding changes.
3360 (create_edge_and_update_destination_phis): Likewise.
3361 (thread_single_edge): Likewise.
3363 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3364 Diego Novillo <dnovillo@google.com>
3366 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3367 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3368 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3369 versions of these patterns.
3371 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3372 (__builtin_arc_aligned): Likewise.
3374 * config/arc/arc.md: Expand adc_0 comment stating the intended
3375 purpose and why it isn't ready.
3376 Replace commented out call_value_via_label_mixed with a
3377 plain comment about bl_s.
3379 * config/arc/arc.c (stdio.h): Don't include directly.
3380 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3381 Remove if (1) condition.
3382 (arc_encode_section_info): Fix comment.
3384 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3386 * config/arc/arc.c (arc_conditional_register_usage):
3387 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3388 Also set reg_alloc_order for DMA config regs.
3390 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3391 Jeremy Bennett <jeremy.bennett@embecosm.com>
3393 * doc/install.texi (--with-cpu): Mention ARC.
3394 (arc-*-elf32): New paragraph.
3395 (arc-linux-uclibc): Likewise.
3396 * doc/md.texi (Machine Constraints): Add ARC part.
3397 * doc/invoke.texi: (menu): Add ARC Options.
3398 (Machine Dependent Options) <ARC Options>: Add synopsis.
3399 (node ARC Options): Add.
3400 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3401 (ARC Built-in Functions): New section defining
3402 generic ARC built-in functions.
3403 (ARC SIMD Built-in Functions): New section defining SIMD specific
3405 (Declaring Attributes of Functions): Extended
3406 description of short_call and long_call attributes for ARC and
3407 added index entries.
3409 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3410 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3411 Joern Rennecke <joern.rennecke@embecosm.com>
3412 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3413 Brendan Kehoe <brendan@zen.org>
3414 Michael Eager <eager@eagercon.com>
3415 Simon Cook <simon.cook@embecosm.com>
3416 Jeremy Bennett <jeremy.bennett@embecosm.com>
3418 * config/arc, common/config/arc: New directories.
3420 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3421 Brendan Kehoe <brendan@zen.org>
3422 Simon Cook <simon.cook@embecosm.com>
3424 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3426 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3428 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3429 * tree-ssa-coalesce.h: New. Move prototype to here.
3430 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3431 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3432 (gimple_can_coalesce_p): Move to...
3433 * gimple.c (gimple_can_coalesce_p): Here.
3435 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3437 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3438 (dump_decl_set): Move to gimple.c.
3439 * gimple.h: Don't include tree-ssa-operands.h.
3440 (dump_decl_set): Add prototype.
3441 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3442 Move to gimple-ssa.h.
3443 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3444 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3445 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3446 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3447 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3449 (dump_decl_set): Relocate here.
3450 * gimple-ssa.h: New file.
3451 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3452 Relocate from gimple.h.
3453 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3454 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3455 swap_tree_operands and remove non-ssa path.
3456 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3457 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3459 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3460 vect_is_simple_reduction_1): Use swap_ssa_operands.
3461 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3462 (enum need_phi_state): Move to tree-into-ssa.c.
3463 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3464 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3465 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3466 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3467 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3468 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3469 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3470 num_imm_uses): Move to ssa-iterators.h.
3471 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3472 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3474 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3475 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3476 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3477 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3478 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3479 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3480 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3481 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3482 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3483 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3484 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3485 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3486 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3487 (set_phi_nodes): Move to tree-phinodes.h.
3488 * tree-ssa-operands.h (enum ssa_op_iter_type,
3489 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3490 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3491 (dump_decl_set): Remove prototype.
3492 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3493 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3494 (set_phi_nodes): Relocate from tree-flow-inline.h.
3495 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3497 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3498 include list. Temporarily add gimple.h to include list.
3499 * ssa-iterators.h: New file.
3500 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3501 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3502 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3503 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3504 Relocate from tree-ssa-operands.h.
3505 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3506 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3507 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3508 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3509 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3510 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3511 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3512 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3513 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3514 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3515 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3516 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3517 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3518 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3519 Relocate from tree-flow-inline.h.
3520 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3522 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3525 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3526 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3527 Insert '\t' to output template.
3528 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3529 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3530 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3531 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3533 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3535 * doc/install.texi (Host/target specific installation notes for GCC):
3536 Put @anchor before @heading.
3537 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3538 Use @email for email addresses.
3540 2013-10-01 Jeff Law <law@redhat.com>
3542 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3543 a vec. Only delete the path if we create one without successfully
3544 registering a jump thread.
3545 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3547 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3548 (paths): New vector of jump threading paths.
3549 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3550 (THREAD_PATH): New accessor macro for the entire thread path.
3551 (lookup_redirection_data): Get intermediate and final outgoing edge
3552 from the thread path.
3553 (create_edge_and_update_destination_phis): Copy the threading path.
3554 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3555 jump threading path.
3556 (ssa_redirect_edges): Get edges and block types from the jump threading
3557 path. Free the path vector.
3558 (thread_block): Get edges from the jump threading path. Look at the
3559 entire path to see if we thread to a loop exit. If we cancel a jump
3560 thread request, then free the path vector.
3561 (thread_single_edge): Get edges and block types from the jump threading
3562 path. Free the path vector.
3563 (thread_through_loop_header): Get edges and block types from the jump
3564 threading path. Free the path vector.
3565 (mark_threaded_blocks): Iterate over the vector of paths and store
3566 the path on the appropriate edge. Get edges and block types from the
3567 jump threading path.
3568 (mark_threaded_blocks): Get edges and block types from the jump
3569 threading path. Free the path vector.
3570 (thread_through_all_blocks): Use the vector of paths rather than
3571 a vector of 3-edge sets.
3572 (register_jump_thread): Accept pointer to a path vector rather
3573 than the path vector itself. Store the path vector for later use.
3576 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3577 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3580 * config/s390/s390.c (s390_split_branches): Modify check for table
3582 (s390_chunkify_start): Rearrange table jump insn check in order to
3583 deal with compare and branch insns correctly.
3585 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3589 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3590 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3591 define_insn_and_split.
3592 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3593 (shiftsi3_compare): New pattern.
3595 * config/arm/unspecs.md (UNSPEC_RRX): New.
3597 2013-10-01 Alan Modra <amodra@gmail.com>
3599 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3600 casing inout operands.
3602 2013-10-01 Richard Biener <rguenther@suse.de>
3604 PR tree-optimization/58553
3605 * tree-loop-distribution.c (struct partition_s): Add niter member.
3606 (classify_partition): Populate niter member for the partition
3607 and properly identify whether the relevant store happens before
3608 or after the loop exit.
3609 (generate_memset_builtin): Use niter member from the partition.
3610 (generate_memcpy_builtin): Likewise.
3612 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3614 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3615 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3617 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3619 * basic-block.h (edge_list): Prefix member names with "m_".
3620 * context.h (context): Likewise.
3621 * domwalk.h (dom_walker): Likewise.
3622 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3623 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3624 * hash-table.h (hash_table): Likewise.
3625 * machmode.h (bit_field_mode_iterator): Likewise.
3626 * pass_manager.h (pass_list): Likewise.
3627 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3628 * tree-pass.h (pass_data): Likewise.
3629 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3630 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3631 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3632 * asan.c (pass_data_asan): Update accordingly.
3633 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3634 (control_dependences::control_dependences): Likewise.
3635 (control_dependences::~control_dependences): Likewise.
3636 (control_dependences::~control_dependences): Likewise.
3637 (control_dependences::get_edges_dependent_on): Likewise.
3638 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3639 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3640 * context.c (gcc::context::context): Likewise.
3641 * cprop.c (pass_rtl_cprop::clone): Likewise.
3642 * domwalk.c (dom_walker::walk): Likewise.
3643 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3644 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3645 * mode-switching.c (pass_mode_switching::clone): Likewise.
3646 * passes.c (opt_pass::opt_pass): Likewise.
3647 (pass_manager::pass_manager): Likewise.
3648 * predict.c (pass_strip_predict_hints::clone): Likewise.
3649 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3650 (pass_split_all_insns::clone): Likewise.
3651 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3653 (bit_field_mode_iterator::next_mode): Likewise.
3654 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3655 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3656 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3657 * tree-complex.c (pass_lower_complex::clone): Likewise.
3658 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3659 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3660 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3661 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3662 (pass_fold_builtins::clone): Likewise.
3663 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3664 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3665 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3666 (pass_cd_dce::clone): Likewise.
3667 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3668 (pass_phi_only_cprop::clone): Likewise.
3669 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3670 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3671 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3672 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3673 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3674 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3675 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3676 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3677 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3678 * tree-vrp.c (pass_vrp::clone): Likewise.
3679 * tsan.c (pass_tsan::clone): Likewise.
3681 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3684 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3685 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3688 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3689 optimization, punt if sign_bit_p looked through any zero extension.
3691 2013-09-30 Teresa Johnson <tejohnson@google.com>
3693 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3694 Update redirected out edge count in joiner case.
3695 (ssa_redirect_edges): Common the joiner and non-joiner cases
3696 so that joiner case gets profile updates.
3698 2013-09-30 Richard Biener <rguenther@suse.de>
3700 PR tree-optimization/58554
3701 * tree-loop-distribution.c (classify_partition): Require
3702 unconditionally executed stores for memcpy and memset recognition.
3703 (tree_loop_distribution): Calculate dominance info.
3705 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3707 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3708 (NO_PROFILE_COUNTERS): Likewise.
3709 (PROFILE_HOOK): Likewise.
3710 (FUNCTION_PROFILER): Likewise.
3711 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3713 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3715 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3716 calls and defines in TARGET_MACHO conditional.
3717 (load_macho_picbase_di): Likewise.
3718 (reload_macho_picbase): Likewise.
3719 (reload_macho_picbase_si): Likewise.
3720 (reload_macho_picbase_di): Likewise.
3721 (nonlocal_goto_receiver): Likewise.
3723 2013-09-30 Nick Clifton <nickc@redhat.com>
3725 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3726 that use the MSP430X ISA.
3727 (msp430_option_override): Scan -mmcu command line option for any
3728 MCU name that supports the MSP430X ISA.
3729 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3730 -mmcu options which enable the MSP430X ISA.
3732 2013-09-30 Richard Biener <rguenther@suse.de>
3735 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3736 before looking for setjmp-like calls.
3738 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3741 * config/darwin-protos.h (machopic_get_function_picbase): New.
3742 * config/darwin.c (machopic_get_function_picbase): New.
3743 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3744 label for a new func. (load_macho_picbase_di): Likewise.
3745 (reload_macho_picbase): New expand.
3746 (reload_macho_picbase_si): New insn.
3747 (reload_macho_picbase_di): New insn.
3748 (nonlocal_goto_receiver): New define and split.
3749 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3750 (unspecv enum): Add UNSPECV_NLGR.
3752 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3754 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3755 that altivec registers are correctly sized on Darwin.
3757 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3759 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3760 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3761 * config/x-darwin (host-darwin.o): Likewise.
3762 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3763 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3764 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3766 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3768 * doc/invoke.texi: Fix usage of @tie{} command.
3770 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3772 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3774 2013-09-28 Jan Hubicka <jh@suse.cz>
3776 * config/i386/x86-tune.def: Add documentation for each of the options;
3779 2013-09-28 Jan Hubicka <jh@suse.cz>
3781 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3783 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3784 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3785 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3786 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3787 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3789 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3791 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3792 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3793 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3794 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3795 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3796 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3797 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3798 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3799 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3800 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3801 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3802 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3803 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3804 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3805 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3806 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3807 gimple.h, godump.c, graphite-clast-to-gimple.c,
3808 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3809 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3810 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3811 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3812 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3813 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3814 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3815 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3816 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3817 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3818 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3819 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3820 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3821 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3822 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3823 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3824 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3825 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3826 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3827 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3828 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3829 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3830 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3831 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3832 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3833 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3834 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3835 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3836 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3837 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3838 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3839 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3840 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3841 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3842 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3843 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3844 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3845 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3846 whitespace before "(".
3848 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3850 * expr.h (extract_bit_field): Remove packedp parameter.
3851 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3852 from forward declaration.
3853 (store_split_bit_field): Remove packedp arg from calls to
3854 extract_fixed_bit_field.
3855 (extract_bit_field_1): Remove packedp parameter and packedp
3856 argument from recursive calls and calls to extract_fixed_bit_field.
3857 (extract_bit_field): Remove packedp parameter and corresponding
3858 arg to extract_bit_field_1.
3859 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3861 (extract_split_bit_field): Remove packedp arg from call to
3862 extract_fixed_bit_field.
3863 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3864 (copy_blkmode_from_reg): Likewise.
3865 (copy_blkmode_to_reg): Likewise.
3866 (read_complex_part): Likewise.
3867 (store_field): Likewise.
3868 (expand_expr_real_1): Likewise.
3869 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3870 to extract_bit_field.
3871 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3872 call to extract_bit_field.
3873 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3874 call to extract_bit_field.
3875 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3876 and special packedp behavior from -fstrict-volatile-bitfields
3879 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3881 * lra-eliminations.c (init_elim_table): Guard value_p.
3883 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3885 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3886 DFmode, DImode, and SFmode in the upper VSX registers based on the
3887 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3888 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3889 of the VSX register handling.
3891 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3893 (zero_extendsidi2_lfiwzx): Add support for loading into the
3894 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3895 only do VSX memory options on Altivec registers.
3896 (extendsidi2_lfiwax): Likewise.
3897 (extendsfdf2_fpr): Likewise.
3898 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3899 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3900 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3901 (movdi_internal64): Likewise.
3903 2013-09-27 Xinliang David Li <davidxl@google.com>
3905 * opts.c (finish_options): Adjust parameters
3906 according to vect cost model.
3907 (common_handle_option): Set dynamic vect cost
3909 targhooks.c (default_add_stmt_cost): Compute stmt cost
3911 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3912 Use helper function.
3913 * tree-vectorizer.h (unlimited_cost_model): New function.
3914 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3915 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3917 (vect_enhance_data_refs_alignment): Ditto.
3918 * flag-types.h: New enum.
3919 * common/config/i386/i386-common.c (ix86_option_init_struct):
3920 No need to initialize vect_cost_model flag.
3921 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3924 2013-09-27 Diego Novillo <dnovillo@google.com>
3926 * gimple.h (enum ssa_mode): Remove.
3928 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3930 * cfgloop.h (number_of_loops): Fix typo in check for null.
3932 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3935 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3936 are children of outermost saved_cfun's loop, and set it up to
3937 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3938 if loop != loop0 and SESE region contains bbs that belong to loop0.
3940 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3942 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3943 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3944 (set_address_segment, set_address_base, set_address_index)
3945 (set_address_disp): Accept the argument unconditionally.
3946 (baseness): Remove must_be_base_p and must_be_index_p checks.
3947 (decompose_normal_address): Classify as much as possible in the
3950 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3952 * cse.c (count_reg_usage): Handle INT_LIST.
3953 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3954 * reginfo.c (reg_scan_mark_refs): Likewise.
3955 * reload1.c (eliminate_regs_1): Likewise.
3957 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3960 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3963 2013-09-27 Richard Biener <rguenther@suse.de>
3965 PR tree-optimization/58459
3966 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3967 restriction not propagating into loops.
3969 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3971 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3972 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3973 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3975 2013-09-26 Richard Biener <rguenther@suse.de>
3977 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3979 2013-09-26 Richard Biener <rguenther@suse.de>
3981 * alias.h (component_uses_parent_alias_set): Rename to ...
3982 (component_uses_parent_alias_set_from): ... this.
3983 * alias.c (component_uses_parent_alias_set): Rename to ...
3984 (component_uses_parent_alias_set_from): ... this and return
3986 (reference_alias_ptr_type_1): Use the result from
3987 component_uses_parent_alias_set_from instead of stripping
3988 components one at a time.
3989 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3991 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3993 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3994 Move prototypes to...
3995 * tree-ssa-ter.h: New File. Move prototypes here.
3996 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3997 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3998 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3999 from is_replaceable_p.
4000 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4001 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4002 (process_replaceable): Use ter_is_replaceable_p.
4003 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4004 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4005 newly refactored ssa_is_replaceable_p.
4006 * cfgexpand.c: Include tree-outof-ssa.h.
4007 * ssaexpand.h: Delete.
4009 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4011 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4012 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4014 (create_gimple_tmp): Delete.
4015 (get_expr_type, build_assign, build_type_cast): Move to...
4016 * gimple-builder.c: New File.
4017 (get_expr_type): Relocate from gimple.c.
4018 (build_assign, build_type_cast): Change to only create ssanames.
4019 * gimple.h: Move prototypes to...
4020 * gimple-builder.h: New File. Here.
4021 * tree-ssa.h: And here.
4022 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4023 count_uses_and_derefs): Relocate from gimple.c.
4024 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4025 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4026 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4027 gimple_replace_ssa_lhs.
4028 * asan.c: Include gimple-builder.h.
4029 * Makefile.in: Add gimple-builder.o.
4031 2013-09-26 Richard Biener <rguenther@suse.de>
4033 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4034 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4035 (loe_visit_block): Use gcc_checking_assert.
4036 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4037 gimple_assign_ssa_name_copy_p.
4038 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4041 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4043 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4044 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4046 2013-09-26 Richard Biener <rguenther@suse.de>
4048 PR tree-optimization/58539
4049 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4050 the fact that debug statements are not taking part in loop-closed
4053 2013-09-26 Nick Clifton <nickc@redhat.com>
4055 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4056 time warning message.
4057 (msp430_print_operand_raw): Delete unused letter parameter.
4058 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4059 (msp430_print_operand_address): New function.
4060 (msp430_print_operand): Move address printing code from here to
4062 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4064 (zero_extendpsisi2): Likewise.
4065 (extendpsisi2): New pattern.
4066 (andneghi3): New pattern.
4068 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4070 * config/aarch64/aarch64.opt (mlra): New option.
4071 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4072 (TARGET_LRA_P): Define.
4074 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4076 * expr.c (expand_assignment): Remove obsolete comment.
4078 2013-09-25 Jeff Law <law@redhat.com>
4080 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4081 tree-ssa-threadupdate.h.
4082 (register_jump_thread): Similarly.
4083 * tree-ssa-threadupdate.h: New header file.
4084 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4085 * tree-vrp.c: Likewise.
4086 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4087 (thread_around_empty_blocks): Change type of path vector argument to
4088 an edge,type pair from just an edge. Initialize both elements when
4089 appending to a jump threading path. Tweak references to elements
4091 (thread_across_edge): Similarly. Release memory for the elements
4093 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4094 (dump_jump_thread_path): New function broken out from
4095 register_jump_thread.
4096 (register_jump_thread): Use dump_jump_thread_path. Change type of
4097 path vector entries. Search the path for NULL edges and dump
4098 the path if one is found. Tweak the conversion of path to 3-edge
4099 form to use the block copy type information embedded in the path.
4101 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4103 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4105 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4106 Vladimir Makarov <vmakarov@redhat.com>
4108 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4109 from the least significant bit.
4110 (strip_address_mutations): Add bitfield operations handling.
4111 (must_be_index_p): Add shifting and rotate operations handling.
4112 (set_address_base): Use must_be_base_p predicate.
4113 (set_address_index): Use must_be_index_p predicate.
4115 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4116 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4117 Sergey Lega <sergey.s.lega@intel.com>
4118 Anna Tikhonova <anna.tikhonova@intel.com>
4119 Ilya Tocar <ilya.tocar@intel.com>
4120 Andrey Turetskiy <andrey.turetskiy@intel.com>
4121 Ilya Verbin <ilya.verbin@intel.com>
4122 Kirill Yukhin <kirill.yukhin@intel.com>
4123 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4125 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4127 (ix86_expand_vector_move_misalign): Support new unaligned load and
4128 stores and use new names.
4129 (CODE_FOR_sse2_storedqu): Rename to ...
4130 (CODE_FOR_sse2_storedquv16qi): ... this.
4131 (CODE_FOR_sse2_loaddqu): Rename to ...
4132 (CODE_FOR_sse2_loaddquv16qi): ... this.
4133 (CODE_FOR_avx_loaddqu256): Rename to ...
4134 (CODE_FOR_avx_loaddquv32qi): ... this.
4135 (CODE_FOR_avx_storedqu256): Rename to ...
4136 (CODE_FOR_avx_storedquv32qi): ... this.
4137 * config/i386/i386.md (fpint_logic): New.
4138 * config/i386/sse.md (VMOVE): Extend for AVX512.
4142 (VI_UNALIGNED_LOADSTORE): Ditto.
4143 (sse2_avx_avx512f): Ditto.
4144 (sse2_avx2): Extend for AVX512.
4145 (sse4_1_avx2): Ditto.
4146 (avx2_avx512f): New.
4147 (sse): Extend for AVX512.
4150 (avxsizesuffix): Ditto.
4151 (sseintvecmode): Ditto.
4153 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4154 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4155 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4156 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4157 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4158 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4159 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4160 (STORENT_MODE): Extend for AVX512.
4161 (*absneg<mode>2): Replace constraint "x" with "v".
4162 (*mul<mode>3): Ditto.
4163 (*ieee_smin<mode>3): Ditto.
4164 (*ieee_smax<mode>3): Ditto.
4165 (avx_cmp<mode>3): Replace VF with VF_128_256.
4166 (*<sse>_maskcmp<mode>3_comm): Ditto.
4167 (<sse>_maskcmp<mode>3): Ditto.
4168 (<sse>_andnot<mode>3): Extend for AVX512.
4169 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4170 (<code><mode>3, fpint_logic): New.
4171 (*<code><mode>3): Extend for AVX512.
4172 (avx512flogicsuff): New.
4173 (avx512f_<logic><mode>): Ditto.
4174 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4176 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4177 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4178 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4179 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4180 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4181 (xop_vpermil2<mode>3): Ditto.
4182 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4183 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4184 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4185 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4187 2013-09-25 Tom Tromey <tromey@redhat.com>
4189 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4190 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4191 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4192 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4193 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4194 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4195 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4196 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4197 (GRAPHITE_HTAB_H): Remove.
4199 2013-09-25 Tom Tromey <tromey@redhat.com>
4201 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4203 2013-09-25 Tom Tromey <tromey@redhat.com>
4205 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4207 2013-09-25 Tom Tromey <tromey@redhat.com>
4209 * config/i386/t-i386 (i386.o): Remove.
4210 (i386-c.o): Use COMPILE and POSTCOMPILE.
4212 2013-09-25 Tom Tromey <tromey@redhat.com>
4214 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4216 2013-09-25 Tom Tromey <tromey@redhat.com>
4218 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4219 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4220 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4221 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4222 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4223 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4224 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4225 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4226 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4227 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4228 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4229 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4230 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4231 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4232 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4233 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4234 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4235 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4236 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4237 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4238 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4239 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4240 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4241 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4242 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4243 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4244 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4245 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4246 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4247 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4248 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4249 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4250 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4251 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4252 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4253 (graphite-clast-to-gimple.o, graphite-dependences.o)
4254 (graphite-interchange.o, graphite-poly.o)
4255 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4256 (graphite-optimize-isl.o, tree-vect-loop.o)
4257 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4258 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4259 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4260 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4261 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4262 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4263 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4264 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4265 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4266 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4267 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4268 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4269 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4270 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4271 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4272 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4273 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4274 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4275 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4276 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4277 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4278 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4279 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4280 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4281 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4282 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4283 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4284 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4285 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4286 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4287 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4288 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4289 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4290 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4291 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4292 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4293 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4294 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4295 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4296 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4297 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4298 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4299 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4300 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4301 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4302 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4303 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4304 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4305 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4306 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4307 (gcov-dump.o): Remove.
4308 (default-c.o): Use COMPILE and POSTCOMPILE.
4309 (CFLAGS-gcc.o): New variable.
4310 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4312 2013-09-25 Tom Tromey <tromey@redhat.com>
4314 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4315 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4316 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4317 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4318 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4319 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4320 (c-family/array-notation-common.o, c-family/stub-objc.o)
4321 (c-family/c-ubsan.o): Remove.
4323 2013-09-25 Tom Tromey <tromey@redhat.com>
4325 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4327 2013-09-25 Tom Tromey <tromey@redhat.com>
4329 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4330 to add -DENABLE_SHARED_LIBGCC.
4331 (gcc.o): Don't use subshell.
4333 2013-09-25 Tom Tromey <tromey@redhat.com>
4335 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4336 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4337 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4338 * configure, config.in: Rebuild.
4340 2013-09-25 Tom Tromey <tromey@redhat.com>
4342 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4343 (COMPILE, POSTCOMPILE): New variables.
4344 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4345 (DEPFILES): New variable.
4346 Include ".Po" files.
4347 * configure.ac: Add checks for dependency checking.
4348 * configure, aclocal.m4: Regenerate.
4350 2013-09-25 Tom Tromey <tromey@redhat.com>
4352 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4353 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4355 2013-09-25 Tom Tromey <tromey@redhat.com>
4357 * Makefile.in (generated_files): Add options.h,
4358 target-hooks-def.h, insn-opinit.h,
4359 common/common-target-hooks-def.h, pass-instances.def,
4360 c-family/c-target-hooks-def.h.
4362 2013-09-25 Jeff Law <law@redhat.com>
4364 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4365 than foo[foo.length () - 1] to access last member in a vec.
4366 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4368 2013-09-25 Richard Biener <rguenther@suse.de>
4371 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4373 2013-09-25 Jan Hubicka <jh@suse.cz>
4375 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4378 2013-09-25 Marek Polacek <polacek@redhat.com>
4381 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4382 when determining the type name.
4384 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4386 * config/sh/sh.md: Fix formatting.
4388 2013-09-24 Xinliang David Li <davidxl@google.com>
4390 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4391 max peel iterations parameter.
4392 * param.def: New parameter.
4393 * doc/invoke.texi: Document New parameter.
4395 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4397 * gimple-pretty-print.c: Various whitespace tweaks.
4398 * tree-core.h: Likewise.
4399 * tree-pretty-print.c: Likewise.
4400 * tree-ssa-alias.c: Likewise.
4401 * tree-ssa-copy.c: Likewise.
4402 * tree-ssanames.c: Likewise.
4403 * tree-ssanames.h: Likewise.
4404 * tree-vrp.c: Likewise.
4406 2013-09-24 Alan Modra <amodra@gmail.com>
4410 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4411 for output operands that disallow regs. Don't use EXPAND_WRITE on
4414 2013-09-24 Richard Biener <rguenther@suse.de>
4417 * tree.c (reference_alias_ptr_type): Move ...
4418 * alias.c (reference_alias_ptr_type): ... here and implement
4419 in terms of the new reference_alias_ptr_type_1.
4420 (ref_all_alias_ptr_type_p): New helper.
4421 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4422 use ref_all_alias_ptr_type_p.
4423 (get_deref_alias_set): Add flag_strict_aliasing check here.
4424 (reference_alias_ptr_type_1): New function, split out from ...
4425 (get_alias_set): ... here.
4426 (alias_ptr_types_compatible_p): New function.
4427 * alias.h (reference_alias_ptr_type): Declare.
4428 (alias_ptr_types_compatible_p): Likewise.
4429 * tree.h (reference_alias_ptr_type): Remove.
4430 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4431 to compare MEM_REF alias types.
4433 2013-09-24 Richard Biener <rguenther@suse.de>
4435 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4437 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4439 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4440 reload helper function arrays into a single array reg_addr.
4441 (reload_fpr_gpr): Likewise.
4442 (reload_gpr_vsx): Likewise.
4443 (reload_vsx_gpr): Likewise.
4444 (struct rs6000_reg_addr): Likewise.
4445 (reg_addr): Likewise.
4446 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4447 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4448 (rs6000_init_hard_regno_mode_ok): Likewise.
4449 (rs6000_secondary_reload_direct_move): Likewise.
4450 (rs6000_secondary_reload): Likewise.
4452 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4453 constraints: wu, ww, and wy. Repurpose wv constraint added during
4454 power8 changes. Put wg constraint in alphabetical order.
4456 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4457 for future work to add ISA 2.07 VSX single precision support.
4458 (-mvsx-scalar-double): Change default from -1 to 1, update
4459 documentation comment.
4460 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4461 (-mupper-regs-df): New debug switch to control whether DF values
4462 can go in the traditional Altivec registers.
4463 (-mupper-regs-sf): New debug switch to control whether SF values
4464 can go in the traditional Altivec registers.
4466 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4468 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4469 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4470 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4471 (rs6000_debug_legitimate_address_p): Print if we are running
4472 before, during, or after reload.
4473 (rs6000_secondary_reload): Add a comment.
4474 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4476 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4477 constraints. Update documentation string.
4478 (wd constraint): Likewise.
4479 (wf constraint): Likewise.
4480 (wg constraint): Likewise.
4481 (wn constraint): Likewise.
4482 (ws constraint): Likewise.
4483 (wt constraint): Likewise.
4484 (wx constraint): Likewise.
4485 (wz constraint): Likewise.
4486 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4488 (ww constraint): Likewise.
4489 (wy constraint): Likewise.
4490 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4491 in the previous submissions.
4492 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4494 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4496 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4497 (REG_BR_PROB): Say that the probability is stored in an int_list.
4498 * reg-notes.def: Update commentary to mention INT_LIST.
4499 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4500 (INT_LIST): New rtx.
4501 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4502 * rtlanal.c (int_reg_note_p): New function.
4503 (alloc_reg_note): Assert that the note does not have an int argument.
4504 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4505 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4506 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4507 rather than an INSN_LIST. Handle INT_LIST.
4508 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4509 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4510 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4511 Manipulate them as ints rather than rtxes.
4512 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4513 * regmove.c (copy_src_to_dest): Likewise.
4514 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4516 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4517 into the cases that need it.
4518 * config/arm/arm.c (arm_unwind_emit): Likewise.
4520 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4521 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4522 * loop-doloop.c (add_test, doloop_modify): Likewise.
4523 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4524 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4525 * predict.c (combine_predictions_for_insn): Likewise.
4526 * print-rtl.c (print_rtx): Handle INT_LIST.
4527 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4528 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4529 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4530 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4531 (ix86_split_fp_branch, predict_jump): Likewise.
4532 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4533 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4534 * config/spu/spu.c (ea_load_store_inline): Likewise.
4536 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4537 value of a REG_BR_PROB note.
4538 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4539 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4540 * emit-rtl.c (try_split): Likewise.
4541 * predict.c (br_prob_note_reliable_p): Likewise.
4542 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4543 * reorg.c (mostly_true_jump): Likewise.
4544 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4545 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4546 * config/i386/i386.c (ix86_print_operand): Likewise.
4547 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4548 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4549 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4550 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4551 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4552 * config/sparc/sparc.c (output_cbranch): Likewise.
4553 * config/spu/spu.c (get_branch_target): Likewise.
4554 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4555 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4557 2013-09-23 Jan Hubicka <jh@suse.cz>
4559 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4561 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4562 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4564 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4565 (gimple_fold_call): Dump inconsistent devirtualizations; add
4566 sanity check for type based devirtualizations.
4567 * ipa-prop.c: Include ipa-utils.h
4568 (ipa_intraprocedural_devirtualization): Add sanity check.
4569 (try_make_edge_direct_virtual_call): Likewise.
4571 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4573 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4574 assignment statements.
4576 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4578 * gimple-pretty-print.c (dump_ssaname_info): New function.
4579 (dump_gimple_phi): Call it.
4580 (pp_gimple_stmt_1): Likewise.
4581 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4582 (range_info_def): Declare.
4583 * tree-pretty-print.c (pp_double_int): New function.
4584 (dump_generic_node): Call it.
4585 * tree-pretty-print.h (pp_double_int): Declare.
4586 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4587 * tree-ssanames.h (range_info_def): New structure.
4588 (value_range_type): Move definition here.
4589 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4591 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4593 (set_range_info): New function.
4594 (get_range_info): Likewise.
4595 (duplicate_ssa_name_range_info): Likewise.
4596 (duplicate_ssa_name_fn): Check pointer type and call
4597 duplicate_ssa_name_range_info.
4598 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4599 * tree-vrp.c (value_range_type): Remove definition, now in
4601 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4602 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4603 (SSA_NAME_RANGE_INFO): New macro.
4605 2013-09-23 Richard Biener <rguenther@suse.de>
4607 PR tree-optimization/58464
4608 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4609 (phi_trans_add): Change to add conditionally on being not present.
4610 (phi_translate_1): Remove recursion detection here.
4611 (phi_translate): Pre-seed the cache with NULL to catch
4612 recursion here in a more generic way.
4613 (bitmap_find_leader): Adjust comment.
4614 (get_representative_for): Dump value-numbers.
4615 (create_expression_by_pieces): Likewise.
4616 (insert_into_preds_of_block): Likewise.
4618 2013-09-23 Christian Bruel <christian.bruel@st.com>
4621 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4622 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4624 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4627 2013-09-20 Renlin Li <renlin.li@arm.com>
4629 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4631 (aarch64_expand_epilogue): Likewise.
4632 (aarch64_legitimize_reload_address): Likewise.
4634 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4636 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4637 NULL_TREE before pushing them onto the vector. Likewise for labels.
4639 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4641 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4642 comparison operators when -fno-trapping-math is in effect.
4643 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4644 comparison operators in TFmode and assert that unsupported operators
4646 (ia64_print_operand): Likewise.
4648 2013-09-21 Jan Hubicka <jh@suse.cz>
4650 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4651 (sse_typeless_stores): Enable for core
4652 (sse_load0_by_pxor): Likewise.
4653 (four_jump_limit): Disable for core.
4654 (pad_returns): Likewise.
4655 (avoid_vector_decode): Likewise.
4656 (fuse_cmp_and_branch): Enable for cores.
4657 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4659 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4662 * config/pa/pa.c (pa_option_override): Disable auto increment and
4663 decrement instructions until reload is completed.
4665 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4666 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4667 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4669 2013-09-20 DJ Delorie <dj@redhat.com>
4670 Nick Clifton <nickc@redhat.com>
4672 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4673 (need_to_save): Save bank 0 on interrupts.
4674 (characterize_address): Strip far address wrappers.
4675 (rl78_as_legitimate_address): Likewise.
4676 (transcode_memory_rtx): Likewise.
4677 (rl78_peep_movhi_p): Disable this peephole after devirt.
4678 (rl78_propogate_register_origins): Forget all origins when a
4680 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4681 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4683 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4684 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4685 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4687 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4689 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4691 * config/pa/pa.md: In "scc" insn patterns, change output template to
4692 handle const0_rtx in reg_or_0_operand operands.
4694 2013-09-20 Martin Husemann <martin@NetBSD.org>
4697 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4698 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4700 2013-09-20 Richard Biener <rguenther@suse.de>
4703 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4704 remembering SSA name version and block index.
4705 (new_scev_info_str): Adjust.
4706 (hash_scev_info): Likewise. Also hash the block index.
4707 (eq_scev_info): Adjust.
4708 (find_var_scev_info): Likewise.
4709 (struct instantiate_cache_entry): Remove.
4710 (struct instantiate_cache_type): Use a htab to map name, block
4712 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4713 (get_instantiated_value_entry): Likewise.
4714 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4715 (instantiate_scev_name): Adjust.
4717 2013-09-20 Jeff Law <law@redhat.com>
4719 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4721 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4723 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4724 Call aarch64_simd_expand_args to update op[argc].
4726 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4728 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4731 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4733 * gengtype.c (file_rules): Added rule for *.cc files.
4734 (get_output_file_with_visibility): Give fatal message when no
4737 2013-09-20 Renlin Li <renlin.li@arm.com>
4739 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4740 (aarch64_expand_epilogue): Likewise.
4741 (aarch64_legitimize_reload_address): Likewise.
4743 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4746 * expr.c (expand_assignment): Remove misalignp code path.
4748 2013-09-20 Marek Polacek <polacek@redhat.com>
4751 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4752 TYPE_PRECISION. Add asserts.
4754 2013-09-20 Richard Biener <rguenther@suse.de>
4756 PR tree-optimization/58453
4757 * tree-loop-distribution.c (distribute_loop): Apply the cost
4758 model for -ftree-loop-distribute-patterns, too.
4760 2013-09-20 Richard Biener <rguenther@suse.de>
4763 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4764 make type comparison less strict.
4766 2013-09-20 Alan Modra <amodra@gmail.com>
4768 * configure: Regenerate.
4769 * aclocal.m4: Regenerate.
4771 2013-09-20 Marek Polacek <polacek@redhat.com>
4774 * doc/extend.texi: Document that attribute used is meant to be used
4775 on variables with static storage duration.
4777 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4779 PR tree-optimization/58472
4780 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4781 simd_lane_access set inv_p = false.
4782 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4783 the simduid magic VAR_DECL.
4785 2013-09-19 Jan Hubicka <jh@suse.cz>
4787 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4789 2013-09-17 Jeff Law <law@redhat.com>
4791 * tree-ssa-dom.c (record_temporary_equivalences): New function
4792 split out of dom_opt_dom_walker::after_dom_children.
4793 (dom_opt_dom_walker::thread_across_edge): Move common code
4794 in here from dom_opt_dom_walker::after_dom_children.
4795 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4797 2013-09-19 Jan Hubicka <jh@suse.cz>
4799 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4800 (TARGET_GENERIC): Use PROCESOR_GENERIC
4801 (enum processor_type): Unify generic32 and 64.
4802 * i386.md (cpu): Likewise.
4803 * x86-tune.def (use_leave): Enable for generic32.
4804 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4805 * athlon.md: Change generic64 to generic in all occurences.
4806 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4807 (ix86_target_macros_internal): Likewise.
4808 * driver-i386.c (host_detect_local_cpu): Likewise.
4809 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4811 (generic_memcpy, generic_memset, generic_cost): This one.
4812 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4813 (m_GENERIC32, m_GENERIC64): Remove.
4814 (m_GENERIC): Turn into one flag.
4815 (processor_target): Unify generic tunnings.
4816 (ix86_option_override_internal): Replace generic32/64 by generic.
4817 (ix86_issue_rate): Likewise.
4818 (ix86_adjust_cost): Likewise.
4820 2013-09-19 Jan Hubicka <jh@suse.cz>
4822 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4823 of speculative flag.
4825 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4827 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4828 GOMP_sections_start, even if !exit_reachable.
4830 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4832 * lra-constraints.c (need_for_all_save_p): Use macro
4833 HARD_REGNO_CALL_PART_CLOBBERED.
4834 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4835 set up pseudo conflict hard regs.
4837 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4840 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4843 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4845 PR rtl-optimization/58438
4846 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4847 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4848 from previous subpasses.
4850 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4852 * arm.c (arm_get_frame_offsets): Validate architecture supports
4853 LDRD/STRD before accepting the tuning preference.
4854 (arm_expand_prologue): Likewise.
4855 (arm_expand_epilogue): Likewise.
4857 2013-09-18 Richard Biener <rguenther@suse.de>
4859 PR tree-optimization/58417
4860 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4861 have chrecs with symbols defined in the loop as operands.
4862 (chrec_fold_multiply): Likewise.
4863 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4864 parameters before folding binary operations.
4865 (struct instantiate_cache_entry_hasher): Remove.
4866 (struct instantiate_cache_type): Use a pointer-map.
4867 (instantiate_cache_type::instantiate_cache_type): New function.
4868 (instantiate_cache_type::get): Likewise.
4869 (instantiate_cache_type::set): Likewise.
4870 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4871 (get_instantiated_value_entry): Likewise.
4872 (global_cache): New global.
4873 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4874 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4875 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4876 (instantiate_scev_name): Adjust.
4877 (instantiate_scev): Construct global instead of local cache.
4878 (resolve_mixers): Likewise.
4880 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4881 Paolo Carlini <paolo.carlini@oracle.com>
4884 * doc/implement-cxx.texi: Fix references to the C++ standards.
4886 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4888 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4889 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4890 simd_lane_access drs, update also DR_ALIGNED_TO.
4892 2013-09-18 Marek Polacek <polacek@redhat.com>
4895 * doc/extend.texi: Document no_sanitize_undefined attribute.
4896 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4897 no_sanitize_undefined attribute.
4899 2013-09-18 Nick Clifton <nickc@redhat.com>
4901 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4902 (ASM_DECLARE_FUNCTION_NAME): Define.
4904 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4906 * compare-elim.c (find_comparison_dom_walker): New class
4907 (find_comparisons_in_bb): Rename to
4908 find_comparison_dom_walker::before_dom_children
4909 (find_comparisons): Adjust
4910 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4912 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4913 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4914 (init_walk_dominator_tree): Remove declaration.
4915 (fini_walk_dominator_tree): Remove declaration.
4916 * fwprop.c (single_def_use_dom_walker): New class
4917 (single_def_use_enter_block): Convert to
4918 single_def_use_dom_walker::before_dom_children.
4919 (single_def_use_leave_block): Convert to
4920 single_def_use_dom_walker::after_dom_children.
4921 (build_single_def_use_links): Adjust.
4922 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4924 (find_candidates_in_block): Convert to
4925 find_candidates_dom_walker::before_dom_children.
4926 (execute_strength_reduction): Adjust.
4927 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4928 (sese_dom_walker): New class.
4929 (sese_dom_walker::sese_dom_walker): New constructor.
4930 (sese_dom_walker::~sese_dom_walker): New destructor.
4931 (build_sese_conditions_before): Convert to
4932 sese_dom_walker::before_dom_children.
4933 (build_sese_conditions_after): Convert to
4934 sese_dom_walker::after_dom_children.
4935 (build_poly_scop): Adjust
4936 * tree-into-ssa.c (rewrite_dom_walker): New class
4937 (rewrite_enter_block): Convert to
4938 rewrite_dom_walker::before_dom_children.
4939 (rewrite_leave_block): Convert to
4940 rewrite_dom_walker::after_dom_children.
4941 (rewrite_update_dom_walker): New class.
4942 (rewrite_update_enter_block): Convert to
4943 rewrite_update_dom_walker::before_dom_children.
4944 (rewrite_update_leave_block): Convert to
4945 rewrite_update_dom_walker::after_dom_children.
4946 (rewrite_blocks, rewrite_into_ssa): Adjust.
4947 (mark_def_dom_walker): New class.
4948 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4949 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4950 (mark_def_sites_blocks): Convert to
4951 mark_def_dom_walker::before_dom_children.
4952 (mark_def_site_blocks): Remove.
4953 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4954 (tree_ssa_dominator_optimize): Adjust.
4955 (dom_thread_across_edge): Convert to method
4956 dom_opt_dom_walker::thread_across_edge.
4957 (dom_opt_enter_block): Convert to member function
4958 dom_opt_dom_walker::before_dom_children.
4959 (dom_opt_leave_block): Convert to member function
4960 dom_opt_dom_walker::after_dom_children.
4961 * tree-ssa-dse.c (dse_dom_walker): New class.
4962 (dse_enter_block): Convert to member function
4963 dse_dom_walker::before_dom_children.
4964 (tree_ssa_dse): Adjust.
4965 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4966 (determine_invariantness_stmt): Convert to method
4967 invariantness_dom_walker::before_dom_children.
4968 (determine_invariantness): Remove
4969 (move_computations_dom_walker): New class.
4970 (move_computations_stmt): Convert to method
4971 move_computations_dom_walker::before_dom_children.
4972 (move_computations, tree_ssa_lim): Adjust.
4973 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4974 (nt_init_block): Make method
4975 notrappping_dom_walker::before_dom_children.
4976 (nt_fini_block): Make
4977 method nontrapping_dom_walker::after_dom_children.
4978 (get_non_trapping): Adjust.
4979 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4980 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4981 (eliminate_leave_block): Make method.
4982 eliminate_dom_walker::after_dom_children.
4984 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4985 (strlen_enter_block): Make method
4986 strlen_dom_walker::before_dom_children.
4987 (strlen_leave_block): Make
4988 method strlen_dom_walker::after_dom_children.
4989 (tree_ssa_strlen): Adjust.
4990 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4991 (tree_ssa_uncprop): Adjust.
4992 (uncprop_leave_block): Make method
4993 uncprop_dom_walker::after_dom_children.
4994 (uncprop_leave_block): Make method
4995 uncprop_dom_walker::before_dom_children.
4997 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4999 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5000 branch in same basic block. Check both src and dest of the move insn.
5002 2013-09-17 Nick Clifton <nickc@redhat.com>
5004 * config/rl78/rl78-real.md (bf): New pattern.
5006 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5007 (rl78_print_operand): Do not put a # before a %B.
5008 * config/rl78/rl78.opt: Tweak doc strings.
5010 2013-09-17 DJ Delorie <dj@redhat.com>
5012 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5013 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5014 (rl78_option_override): Likewise, if -mallregs.
5015 (is_virtual_register): Likewise.
5016 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5017 (REGNO_OK_FOR_BASE_P): Likewise.
5018 * config/rl78/rl78.opt (-mallregs): New.
5020 2013-09-17 Nick Clifton <nickc@redhat.com>
5022 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5023 For interrupt functions: save all call clobbered registers if the
5024 interrupt handler is not a leaf function.
5025 (rl78_expand_prologue): Always recompute the frame information.
5026 For interrupt functions: only select bank 0 if one of the bank 0
5027 registers is going to be psuhed.
5029 2013-09-17 DJ Delorie <dj@redhat.com>
5031 * config/rl78/constraints.md: For each W* constraint, rename to C*
5032 and create a W* constraint that checks for an optional ES: prefix
5034 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5035 (es_addr): New. Used to wrap far addresses.
5036 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5037 (rl78_es_base): New.
5038 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5039 wrapped far addresses.
5040 (rl78_print_operand_1): Unwrap far addresses before processing.
5041 (rl78_lo16): Wrap far addresses in unspecs.
5042 (rl78_es_addr): New.
5043 (rl78_es_base): New.
5044 (insn_ok_now): Check for not-yet-wrapped far addresses.
5045 (transcode_memory_rtx): Properly re-wrap far addresses.
5047 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5049 * config/sparc/t-rtems: Add leon3 multilibs.
5051 2013-09-17 Cong Hou <congh@google.com>
5053 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5054 when checking the dot production pattern. The type of rhs operand
5055 of multiply is now checked correctly.
5057 2013-09-17 Jeff Law <law@redhat.com>
5059 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5060 edge implied equivalences into successor phis.
5061 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5062 here from tree-ssa-threadedge.c.
5063 (mark_threaded_blocks): When threading through a joiner, if both
5064 successors of the joiner's clone reach the same block, verify the
5065 PHI arguments are equal. If not, cancel the jump threading request.
5066 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5067 tree-ssa-threadupdate.c
5068 (thread_across_edge): Don't check PHI argument equality when
5069 threading through joiner block here.
5071 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5073 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5074 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5075 (warn_uninit, warn_uninitialized_vars,
5076 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5077 Move to tree-ssa-uninit.c.
5078 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5079 (has_undefined_value_p): New. Pass dependant parts of
5080 ssa_undefined_value_p.
5081 (uninit_undefined_value_p): Use has_undefined_value_p.
5082 (warn_uninit, warn_uninitialized_vars,
5083 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5084 Move from tree-ssa.c.
5085 * tree-ssa.h: Adjust prototypes.
5087 2013-09-17 Jan Hubicka <jh@suse.cz>
5090 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5091 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5092 FUNCTION_NOT_OPTIMIZED.
5093 * ipa-inline-analysis.c (compute_inline_parameters): Function
5094 not optimized is not inlinable unless it is alwaysinline.
5095 (inline_analyze_function): Force calls in not optimized
5096 function not inlinable.
5098 2013-09-17 Jan Hubicka <jh@suse.cz>
5101 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5103 * ipa.c (function_and_variable_visibility): Likewise.
5104 * ipa-profile.c (ipa_profile): Likewise.
5106 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5109 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5110 attribute, and don't inline the resolver in this case.
5112 2013-09-17 Teresa Johnson <tejohnson@google.com>
5114 * coverage.c (get_coverage_counts): Add missing newline.
5116 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5118 PR tree-optimization/58088
5119 * fold-const.c (mask_with_trailing_zeros): New function.
5120 (fold_binary_loc): Make sure we don't recurse infinitely
5121 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5122 Use mask_with_trailing_zeros where appropriate.
5124 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5126 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5127 of 'prev' var to get better distance estimation.
5129 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5131 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5132 * tree-inline.c (is_parameter_of): New predicate.
5133 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5134 a parameter has been remapped.
5135 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5136 (optimize_inline_calls): Initialize transform_parameter.
5137 (copy_gimple_seq_and_replace_locals): Likewise.
5138 (tree_function_versioning): Likewise.
5139 (maybe_inline_call_in_expr): Likewise.
5141 2013-09-17 Nick Clifton <nickc@redhat.com>
5143 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5144 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5146 (is_attr_func): New function.
5147 (msp430_is_interrupt_func): New function.
5148 (is_naked_func): New function.
5149 (is_reentrant_func): New function.
5150 (is_critical_func): New function.
5151 (msp430_start_function): Add annotations for function attributes.
5152 (msp430_attr): New function.
5153 (msp430_attribute_table): New.
5154 (msp430_function_section): New function.
5155 (TARGET_ASM_FUNCTION_SECTION): Define.
5156 (msp430_builtin): New enum.
5157 (msp430_init_builtins): New function.
5158 (msp430_builtin_devl): New function.
5159 (msp430_expand_builtin): New function.
5160 (TARGET_INIT_BUILTINS): Define.
5161 (TARGET_EXPAND_BUILTINS): Define.
5162 (TARGET_BUILTIN_DECL): Define.
5163 (msp430_expand_prologue): Add support for naked, interrupt,
5164 critical and reentrant functions.
5165 (msp430_expand_epilogue): Likewise.
5166 (msp430_print_operand): Handle 'O' character.
5167 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5169 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5170 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5171 (pushm): Use a 'n' rather than an 'i' constraint.
5172 (msp_return): Add generation of the interrupt return instruction.
5173 (disable_interrupts): New pattern.
5174 (enable_interrupts): New pattern.
5175 (push_intr_state): New pattern.
5176 (pop_intr_state): New pattern.
5177 (bic_SR): New pattern.
5178 (bis_SR): New pattern.
5179 * doc/extend.texi: Document MSP430 function attributes and builtin
5182 2013-09-17 Richard Biener <rguenther@suse.de>
5184 PR tree-optimization/58432
5185 * tree-loop-distribution.c (tree_loop_distribution): Also
5186 scan PHIs for outside loop uses and seed a partition from them.
5188 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5190 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5191 (restructure_reference): Call backtrace_base_for_ref.
5193 2013-09-17 Alan Modra <amodra@gmail.com>
5196 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5199 2013-09-16 DJ Delorie <dj@redhat.com>
5201 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5202 vregs location for RL78/G10.
5203 (rl78_expand_prologue): Avoid SEL on G10.
5204 (rl78_expand_epilogue): Likewise.
5205 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5206 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5207 __RL78_G10__ when appropriate.
5208 (ASM_SPEC): Pass -mg10 along to the assembler.
5209 * config/rl78/rl78.md (sel_rb): Disable for G10.
5210 * config/rl78/rl78.opt: Add -mg10 option.
5211 * config/rl78/t-rl78: Add -mg10 multilib.
5213 2013-09-16 Xinliang David Li <davidxl@google.com>
5215 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5216 * omp-low.c (omp_max_vf): Ditto.
5217 (expand_omp_simd): Ditto.
5218 * tree-vectorizer.c (vectorize_loops): Ditto.
5219 (gate_vect_slp): Ditto.
5220 (gate_increase_alignment): Ditto.
5221 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5222 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5223 (gate_tree_vectorize): Name change.
5224 (tree_vectorize): Ditto.
5225 (pass_vectorize::gate): Call new function.
5226 (pass_vectorize::execute): Ditto.
5227 * opts.c: O3 default setting change.
5228 (finish_options): Check new flag.
5229 * doc/invoke.texi: Document new flags.
5230 * common.opt: New flags.
5232 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5234 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5235 (Condition Code): Don't reference it.
5236 * doc/tm.texi: Regenerate.
5238 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5241 * lra-constraints.c (undo_optional_reloads): Consider all optional
5242 reload even if it did not get a hard reg.
5244 2013-09-16 Teresa Johnson <tejohnson@google.com>
5246 * dumpfile.c (dump_loc): Remove newline emission.
5247 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5248 emission to dump_printf_loc calls where missing.
5249 (vect_mark_for_runtime_alias_test): Ditto.
5250 (vect_analyze_data_ref_dependence): Ditto.
5251 (vect_analyze_data_ref_dependences): Ditto.
5252 (vect_slp_analyze_data_ref_dependence): Ditto.
5253 (vect_slp_analyze_data_ref_dependences): Ditto.
5254 (vect_compute_data_ref_alignment): Ditto.
5255 (vect_update_misalignment_for_peel): Ditto.
5256 (vect_verify_datarefs_alignment): Ditto.
5257 (vector_alignment_reachable_p): Ditto.
5258 (vect_get_data_access_cost): Ditto.
5259 (vect_enhance_data_refs_alignment): Ditto.
5260 (vect_find_same_alignment_drs): Ditto.
5261 (vect_analyze_data_refs_alignment): Ditto.
5262 (vect_analyze_group_access): Ditto.
5263 (vect_analyze_data_ref_access): Ditto.
5264 (vect_analyze_data_ref_accesses): Ditto.
5265 (vect_prune_runtime_alias_test_list): Ditto.
5266 (vect_analyze_data_refs): Ditto.
5267 (vect_create_addr_base_for_vector_ref): Ditto.
5268 (vect_create_data_ref_ptr): Ditto.
5269 (vect_grouped_store_supported): Ditto.
5270 (vect_grouped_load_supported): Ditto.
5271 * value-prof.c (check_counter): Ditto.
5272 (check_ic_target): Ditto.
5273 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5274 (vect_recog_widen_mult_pattern): Ditto.
5275 (vect_recog_widen_sum_pattern): Ditto.
5276 (vect_recog_over_widening_pattern): Ditto.
5277 (vect_recog_widen_shift_pattern): Ditto.
5278 (vect_recog_rotate_pattern): Ditto.
5279 (vect_recog_vector_vector_shift_pattern): Ditto.
5280 (vect_recog_divmod_pattern): Ditto.
5281 (vect_recog_mixed_size_cond_pattern): Ditto.
5282 (vect_recog_bool_pattern): Ditto.
5283 (vect_pattern_recog_1): Ditto.
5284 (vect_pattern_recog): Ditto.
5285 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5286 (vect_is_simple_iv_evolution): Ditto.
5287 (vect_analyze_scalar_cycles_1): Ditto.
5288 (vect_get_loop_niters): Ditto.
5289 (vect_analyze_loop_1): Ditto.
5290 (vect_analyze_loop_form): Ditto.
5291 (vect_analyze_loop_operations): Ditto.
5292 (vect_analyze_loop_2): Ditto.
5293 (vect_analyze_loop): Ditto.
5294 (report_vect_op): Ditto.
5295 (vect_is_slp_reduction): Ditto.
5296 (vect_is_simple_reduction_1): Ditto.
5297 (vect_get_known_peeling_cost): Ditto.
5298 (vect_estimate_min_profitable_iters): Ditto.
5299 (vect_model_reduction_cost): Ditto.
5300 (vect_model_induction_cost): Ditto.
5301 (get_initial_def_for_induction): Ditto.
5302 (vect_create_epilog_for_reduction): Ditto.
5303 (vectorizable_reduction): Ditto.
5304 (vectorizable_induction): Ditto.
5305 (vectorizable_live_operation): Ditto.
5306 (vect_loop_kill_debug_uses): Ditto.
5307 (vect_transform_loop): Ditto.
5308 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5309 (vect_stmt_relevant_p): Ditto.
5310 (process_use): Ditto.
5311 (vect_mark_stmts_to_be_vectorized): Ditto.
5312 (vect_model_simple_cost): Ditto.
5313 (vect_model_promotion_demotion_cost): Ditto.
5314 (vect_model_store_cost): Ditto.
5315 (vect_get_store_cost): Ditto.
5316 (vect_model_load_cost): Ditto.
5317 (vect_get_load_cost): Ditto.
5318 (vect_init_vector_1): Ditto.
5319 (vect_get_vec_def_for_operand): Ditto.
5320 (vect_finish_stmt_generation): Ditto.
5321 (vectorizable_call): Ditto.
5322 (vectorizable_conversion): Ditto.
5323 (vectorizable_assignment): Ditto.
5324 (vectorizable_shift): Ditto.
5325 (vectorizable_operation): Ditto.
5326 (vectorizable_store): Ditto.
5327 (vectorizable_load): Ditto.
5328 (vectorizable_condition): Ditto.
5329 (vect_analyze_stmt): Ditto.
5330 (vect_transform_stmt): Ditto.
5331 (vect_is_simple_use): Ditto.
5332 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5333 (vect_can_advance_ivs_p): Ditto.
5334 (vect_update_ivs_after_vectorizer): Ditto.
5335 (vect_do_peeling_for_loop_bound): Ditto.
5336 (vect_gen_niters_for_prolog_loop): Ditto.
5337 (vect_update_inits_of_drs): Ditto.
5338 (vect_create_cond_for_alias_checks): Ditto.
5339 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5340 (vect_build_slp_tree_1): Ditto.
5341 (vect_supported_load_permutation_p): Ditto.
5342 (vect_analyze_slp_instance): Ditto.
5343 (vect_analyze_slp): Ditto.
5344 (vect_make_slp_decision): Ditto.
5345 (vect_detect_hybrid_slp): Ditto.
5346 (vect_bb_vectorization_profitable_p): Ditto.
5347 (vect_slp_analyze_bb_1): Ditto.
5348 (vect_update_slp_costs_according_to_vf): Ditto.
5349 (vect_get_mask_element): Ditto.
5350 (vect_transform_slp_perm_load): Ditto.
5351 (vect_schedule_slp_instance): Ditto.
5352 (vect_schedule_slp): Ditto.
5353 (vect_slp_transform_bb): Ditto.
5354 * profile.c (read_profile_edge_counts): Ditto.
5355 (compute_branch_probabilities): Ditto.
5356 * coverage.c (get_coverage_counts): Ditto.
5358 2013-09-16 Diego Novillo <dnovillo@google.com>
5360 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5362 2013-09-16 Jan Hubicka <jh@suse.cz>
5364 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5365 abstract functions; for static functions check the presence of body.
5367 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5369 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5370 * config/aarch64/aarch64-simd.md
5371 (aarch64_mla_elt<mode>): New.
5372 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5373 (aarch64_mls_elt<mode>): Likewise.
5374 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5375 (aarch64_fma4_elt<mode>): Likewise.
5376 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5377 (aarch64_fma4_elt_to_128v2df): Likewise.
5378 (aarch64_fma4_elt_to_64df): Likewise.
5379 (fnma<mode>4): Likewise.
5380 (aarch64_fnma4_elt<mode>): Likewise.
5381 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5382 (aarch64_fnma4_elt_to_128v2df): Likewise.
5383 (aarch64_fnma4_elt_to_64df): Likewise.
5384 * config/aarch64/iterators.md (VDQSF): New.
5385 * config/aarch64/arm_neon.h
5386 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5387 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5389 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5391 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5392 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5393 (aarch64_mul3_elt_to_128df): Likewise.
5394 (aarch64_mul3_elt_to_64v2df): Likewise.
5395 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5397 (VMUL_CHANGE_NLANES) Likewise.
5400 * config/aarch64/arm_neon.h
5401 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5403 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5405 * config/aarch64/arm_neon.h
5406 (vcvtx_high_f32_f64): Fix parameters.
5408 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5409 Uros Bizjak <ubizjak@gmail.com>
5411 * config/alpha.c: Include tree-ssanames.h.
5413 2013-09-16 Richard Biener <rguenther@suse.de>
5415 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5416 (dot_rdg_1): Handle control_dd.
5417 (create_edge_for_control_dependence): New function.
5418 (create_rdg_edges): Add control dependences if asked for.
5419 (build_rdg): Likewise.
5420 (generate_loops_for_partition): If there are not necessary
5421 control stmts remove all their dependencies.
5422 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5423 (distribute_loop): Pass on control dependences.
5424 (tree_loop_distribution): Compute control dependences and remove
5425 restriction on number of loop nodes.
5427 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5429 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5432 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5434 * cse.c (try_const_anchors): Punt on CC modes.
5436 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5438 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5440 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5443 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5446 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5449 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5451 (darwin_objc1_section): Likewise.
5452 (darwin_file_end): Emit Image Info section when required.
5454 2013-09-14 Jan Hubicka <jh@suse.cz>
5456 * tree-into-ssa.c (gate_into_ssa): New.
5457 (pass_data_build_ssa): Use it.
5458 * cgraph.h (expand_thunk): Update prototype.
5459 * cgraphunit.c (analyze_function): Expand thunks early.
5460 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5461 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5462 set lowered flag; do not add new function.
5463 (assemble_thunks_and_aliases): Update.
5464 * tree-ssa.c (gate_init_datastructures): New gate.
5465 (pass_data_init_datastructures): Use it.
5467 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5470 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5471 xmm register set described in the psABI.
5473 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5475 * dwarf2out.c (should_emit_struct_debug): Add check
5476 for type_decl variable is not NULL.
5478 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5480 * config.gcc: Use new winnt-c.c target hooks
5481 * config/t-winnt: New file
5482 * config/winnt-c.c: New file
5483 * doc/tm.texi.in: Document new hook
5484 * doc/tm.texi: Regenerated
5486 2013-09-13 Jan Hubicka <jh@suse.cz>
5489 * ipa-inline.c (check_callers): New function.
5490 (check_caller_edge): Remove.
5491 (want_inline_function_to_all_callers_p): Also permit alises that are
5493 (inline_to_all_callers): Terminate the walk when devirtualization
5494 introduce new calls.
5496 2013-09-13 Jan Hubicka <jh@suse.cz>
5498 * ipa-inline-analysis.c (struct growth_data): Add node.
5499 (do_estimate_growth_1): Fix detection of recursion.
5501 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5503 PR tree-optimization/58392
5504 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5505 to avoid shadowing of outer loop variable. If
5506 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5507 replace_by_duplicate_decl simduid of loops that have it set and
5508 set dest_cfun->has_simduid_loops and/or
5509 dest_cfun->has_force_vect_loops.
5510 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5511 instead of maybe_lookup_decl.
5512 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5513 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5514 call. Copy over force_vect and copy and remap simduid. Set
5515 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5516 (copy_cfg_body): Remove blocks_to_copy argument. Use
5517 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5518 caller. Don't set cfun->has_simduid_loops and/or
5519 cfun->has_force_vect_loops here.
5520 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5522 (expand_call_inline, tree_function_versioning): Adjust copy_body
5525 2013-09-13 Martin Jambor <mjambor@suse.cz>
5528 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5529 the assert if the edge was a speculative one.
5531 2013-09-13 Richard Biener <rguenther@suse.de>
5533 * tree-data-ref.h (known_dependences_p): Move here ...
5534 * tree-loop-distribution.c (known_dependences_p): ... from here.
5535 (dump_rdg_component, debug_rdg_component): Remove.
5537 (generate_loops_for_partition): Use gimple_uid instead of
5538 relying on matching stmt visit order.
5539 (rdg_build_partitions): Take starting stmt vector.
5540 (ldist_gen): Merge into ...
5541 (distribute_loop): ... this function. Do not compute starting
5543 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5545 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5547 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5548 Set type attribute correctly. Set predicable_short_it attribute.
5549 (cmpsi_shiftsi): Remove %? from output template.
5551 2013-09-13 Richard Biener <rguenther@suse.de>
5553 * tree-loop-distribution.c (struct rdg_component,
5554 rdg_defs_used_in_other_loops_p, free_rdg_components,
5555 rdg_build_components): Remove.
5556 (stmts_from_loop): Do not record virtual PHIs.
5557 (generate_loops_for_partition): Skip virtual PHIs.
5558 (build_rdg_partition_for_component): Rename to ...
5559 (build_rdg_partition_for_vertex): ... this and adjust.
5560 (rdg_build_partitions): Take a vector of starting vertices
5561 instead of components. Remove unnecessary leftover handling.
5562 (ldist_gen): Do not build components or record other stores.
5563 (distribute_loop): Do not distribute loops containing stmts
5566 2013-09-13 Christian Bruel <christian.bruel@st.com>
5569 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5571 2013-09-13 Kai Tietz <ktietz@redhat.com>
5573 * config.gcc: Separate cases for mingw and cygwin targets,
5574 and add 64-bit cygwin target case.
5576 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5577 dll-export inline-functions.
5578 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5581 2013-09-13 Jeff Law <law@redhat.com>
5585 2013-09-06 Jeff Law <law@redhat.com>
5587 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5588 edge implied equivalences into successor phis.
5590 2013-09-12 DJ Delorie <dj@redhat.com>
5592 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5595 2013-09-12 Brooks Moses <bmoses@google.com>
5598 * Makefile.in: Do not install driver binaries in $(target)/bin.
5600 2013-09-12 DJ Delorie <dj@redhat.com>
5602 * config/rl78/rl78.opt (mrelax): New.
5603 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5604 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5606 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5608 (rl78_expand_epilogue): Likewise.
5610 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5613 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5614 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5615 pointer elimination with using remove_reg_equal_offset_note.
5617 2013-09-12 DJ Delorie <dj@redhat.com>
5619 * config/msp430/: New port.
5620 * config.gcc (msp430): Added.
5621 * doc/invoke.texi: Document MSP430 options.
5622 * doc/install.texi: Document msp430-elf
5623 * doc/md.texi: Document msp430-elf
5624 * doc/contrib.texi: Document msp430-elf
5626 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5629 2013-09-12 Martin Jambor <mjambor@suse.cz>
5632 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5633 reference descriptor is NULL.
5634 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5637 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5639 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5640 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5641 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5642 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5643 make_temp_ssa_name): move to tree-ssanames.h
5644 * tree-ssa-alias.h: Move prototype.
5645 * tree-ssa.h: Include tree-ssanames.h.
5646 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5647 * tree-ssanames.h: New. Move items from tree-flow*.h
5648 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5650 2013-09-12 Richard Biener <rguenther@suse.de>
5652 PR tree-optimization/58404
5653 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5654 propagate non-invariant addresses into dereferences wrapped
5655 in component references.
5657 2013-09-12 Richard Biener <rguenther@suse.de>
5659 PR tree-optimization/58402
5660 * passes.def: Move pass_late_warn_uninitialized later.
5662 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5664 * tree-ssa.h: New. Move content from tree-flow.h and
5666 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5667 Move prototypes belonging to tree-ssa.c.
5668 * tree-flow-inline.h (redirect_edge_var_map_def,
5669 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5671 * gimple.h: Adjust prototypes.
5672 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5674 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5675 * tree.h: Move prototype to tree-ssa.h.
5676 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5677 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5678 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5679 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5680 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5681 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5682 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5683 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5684 graphite-blocking.c, graphite-clast-to-gimple.c,
5685 graphite-dependences.c, graphite-interchange.c,
5686 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5687 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5688 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5689 ipa-reference.c, ipa-split.c, ipa-utils.c,
5690 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5691 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5692 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5693 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5694 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5695 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5696 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5697 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5698 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5699 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5700 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5701 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5702 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5703 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5704 value-prof.c, var-tracking.c,
5705 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5707 2013-09-12 Richard Biener <rguenther@suse.de>
5709 PR tree-optimization/58396
5710 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5711 (build_rdg): Take a loop-nest parameter, fix memleaks.
5712 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5714 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5716 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5719 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5721 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5722 The FMA instruction names should have a 'v' prefix.
5724 2013-09-12 Richard Biener <rguenther@suse.de>
5726 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5727 (dot_rdg): Use popen instead of system in optional code.
5728 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5729 (already_processed_vertex_p): Adjust.
5730 (has_anti_or_output_dependence, predecessor_has_mem_write,
5731 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5732 rdg_flag_uses): Remove.
5733 (rdg_flag_vertex): Simplify.
5734 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5736 (build_rdg_partition_for_component): Process the first vertex
5737 of a component only.
5738 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5740 2013-09-12 Alan Modra <amodra@gmail.com>
5742 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5744 2013-09-11 DJ Delorie <dj@redhat.com>
5745 Nick Clifton <nickc@redhat.com>
5747 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5748 (rl78_stack_based_mem): New.
5749 * config/rl78/constraints.md (Iv08): New.
5758 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5760 (movsi): Change from expand to insn-and-split.
5761 (ashrsi3): Clobber AX.
5765 * config/rl78/rl78.md (CC_REG): Fix.
5766 (addsi3): Allow memory and immediate operands.
5767 (addsi3_internal): Split into...
5768 (addsi3_internal_virt): ...new, and ...
5769 (addsi3_internal_real): ...new.
5771 (subsi3_internal_virt): New.
5772 (subsi3_internal_real): New.
5773 (mulsi3): Add memory operand.
5774 (mulsi3_rl78): Likewise.
5775 (mulsi3_g13): Likewise.
5776 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5777 (cbranchqi4_real): Add more constraint options.
5778 (cbranchhi4_real): Expand pattern.
5779 (cbranchhi4_real_signed): New.
5780 (cbranchhi4_real_inverted): New.
5781 (cbranchsi4_real_lt): New.
5782 (cbranchsi4_real_ge): New.
5783 (cbranchsi4_real_signed): New.
5784 (cbranchsi4_real): New.
5786 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5788 (lshrsi3_virt): Likewise.
5789 (ashlsi3_virt): Likewise.
5790 (cbranchqi4_virt_signed): New.
5791 (cbranchhi4_virt_signed): New.
5792 (cbranchsi4_virt): New.
5793 * config/rl78/rl78.c: Whitespace fixes throughout.
5794 (move_elim_pass): New.
5795 (pass_data_rl78_move_elim): New.
5796 (pass_rl78_move_elim): New.
5797 (make_pass_rl78_move_elim): New.
5798 (rl78_devirt_info): Run devirt earlier.
5799 (rl78_move_elim_info): New.
5800 (rl78_asm_file_start): Register it.
5801 (rl78_split_movsi): New.
5802 (rl78_as_legitimate_address): Allow virtual base registers when
5804 (rl78_addr_space_convert): Remove spurious debug stuff.
5805 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5806 (rl78_print_operand): More cases for not printing '#'.
5807 (rl78_expand_compare): Remove most of the logic.
5808 (content_memory): New.
5809 (clear_content_memory): New.
5810 (get_content_index): New.
5811 (get_content_name): New.
5812 (display_content_memory): New.
5813 (update_content): New.
5814 (record_content): New.
5815 (already_contains): New.
5816 (insn_ok_now): Re-recog insns with virtual registers.
5817 (add_postponed_content_update): New.
5818 (process_postponed_content_update): New.
5819 (gen_and_emit_move): New.
5820 (transcode_memory_rtx): Record new location content.
5821 Use gen_and_emit_move.
5822 (force_into_acc): New.
5823 (move_to_acc): Use gen_and_emit_move.
5824 (move_from_acc): Likewise.
5825 (move_acc_to_reg): Likewise.
5826 (move_to_x): Likewise.
5827 (move_to_hl): Likewise.
5828 (move_to_de): Likewise.
5829 (rl78_alloc_physical_registers_op1): Record location content.
5830 (has_constraint): New.
5831 (rl78_alloc_physical_registers_op2): Record location content.
5833 (rl78_alloc_physical_registers_ro1): Likewise.
5834 (rl78_alloc_physical_registers_cmp): Likewise.
5835 (rl78_alloc_physical_registers_umul): Likewise.
5836 (rl78_alloc_address_registers_macax): New.
5837 (rl78_alloc_physical_registers): Initialize and set location
5838 content memory as needed.
5839 (rl78_reorg): Make sure split2 is called.
5840 (rl78_rtx_costs): New.
5842 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5844 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5845 for (not (neg ...)) and (neg (not ...)) cases.
5847 2013-09-11 Richard Biener <rguenther@suse.de>
5850 * passes.def: Split critical edges before late uninit warning passes.
5851 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5853 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5855 PR tree-optimization/58385
5856 * fold-const.c (build_range_check): If both low and high are NULL,
5857 use omit_one_operand_loc to preserve exp side-effects.
5859 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5861 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5863 2013-09-11 Richard Biener <rguenther@suse.de>
5865 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5866 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5867 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5868 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5869 stmts_from_loop, known_dependences_p, build_empty_rdg,
5870 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5871 * tree-loop-distribution.c: ... here.
5872 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5873 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5874 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5875 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5876 * tree-loop-distribution.c: ... here.
5877 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5878 (struct partition_s): Add loops member.
5879 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5880 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5881 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5883 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5884 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5885 Sergey Lega <sergey.s.lega@intel.com>
5886 Anna Tikhonova <anna.tikhonova@intel.com>
5887 Ilya Tocar <ilya.tocar@intel.com>
5888 Andrey Turetskiy <andrey.turetskiy@intel.com>
5889 Ilya Verbin <ilya.verbin@intel.com>
5890 Kirill Yukhin <kirill.yukhin@intel.com>
5891 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5893 * config/i386/constraints.md (k): New.
5895 * config/i386/i386.c (const regclass_map): Add new mask registers.
5896 (dbx_register_map): Ditto.
5897 (dbx64_register_map): Ditto.
5898 (svr4_dbx_register_map): Ditto.
5899 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5901 (ix86_preferred_reload_class): Disable constants for mask registers.
5902 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5903 (ix86_hard_regno_mode_ok): Support new mask registers.
5904 (x86_order_regs_for_local_alloc): Ditto.
5905 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5906 (FIXED_REGISTERS): Add new mask registers.
5907 (CALL_USED_REGISTERS): Ditto.
5908 (REG_ALLOC_ORDER): Ditto.
5909 (VALID_MASK_REG_MODE): New.
5910 (FIRST_MASK_REG): Ditto.
5911 (LAST_MASK_REG): Ditto.
5912 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5913 (MAYBE_MASK_CLASS_P): New.
5914 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5915 (REG_CLASS_CONTENTS): Ditto.
5916 (MASK_REGNO_P): New.
5917 (ANY_MASK_REG_P): Ditto.
5918 (HI_REGISTER_NAMES): Add new mask registers.
5919 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5920 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5922 (attribute "type"): Add mskmov, msklog.
5923 (attribute "length_immediate"): Support them.
5924 (attribute "memory"): Ditto.
5925 (attribute "prefix_0f"): Ditto.
5926 (*movhi_internal): Support new mask registers.
5927 (*movqi_internal): Ditto.
5928 (define_split): Split out clobber pattern is a logic
5929 insn on mask registers.
5930 (*k<logic><mode>): New.
5931 (*andhi_1): Extend to support mask regs.
5932 (*andqi_1): Extend to support mask regs.
5934 (define_split): Split and-not to and and not if operands
5936 (*<code><mode>_1): Separate HI mode to new pattern...
5937 (*<code>hi_1): This.
5938 (*<code>qi_1): Extend to support mask regs.
5940 (kortestzhi): Ditto.
5941 (kortestchi): Ditto.
5943 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5944 (*one_cmplhi2_1): ...Here, now with mask registers support.
5945 (*one_cmplqi2_1): Support new mask registers.
5946 (HI/QImode arithmetics splitter): Don't split if mask registers
5948 (HI/QImode not splitter): Ditto.
5949 * config/i386/predicated.md (mask_reg_operand): New.
5950 (general_reg_operand): Ditto.
5952 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5954 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5955 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5957 2013-09-10 Jeff Law <law@redhat.com>
5959 PR tree-optimization/58380
5960 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5961 of threading through a buried loop header.
5963 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5964 return value for single successor case.
5966 2013-09-10 Jan Hubicka <jh@suse.cz>
5968 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5970 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5973 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5974 support conditional execution.
5975 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5977 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5979 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5981 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5982 Check destination too to check move insn.
5983 (undo_optional_reloads): Add check that the original peudo did not
5984 changed its allocation and the optional reload was inherited on last
5985 inheritance pass. Break loop after deciding to keep optional reload.
5986 (lra_undo_inheritance): Add check that inherited pseudo still in
5989 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5991 * config/aarch64/aarch64.md (generic_sched): New.
5992 * config/aarch64/aarch64-generic.md (load): Make conditional
5993 on generic_sched attribute.
5994 (nonload): Likewise.
5996 2013-09-10 Jan Hubicka <jh@suse.cz>
5998 * lto-cgraph.c: Include ipa-utils.h.
5999 (compute_ltrans_boundary): Also add possible targets into the boundary.
6001 2013-09-10 Jan Hubicka <jh@suse.cz>
6003 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6004 VAR_DECL of vtable rather than full expression.
6006 2013-09-10 Jan Hubicka <jh@suse.cz>
6007 Paolo Carlini <paolo.carlini@oracle.com>
6009 * cgraphunit.c (analyze_functions): Save input_location, set it
6010 to UNKNOWN_LOCATION and restore it at the end.
6012 2013-09-10 Martin Jambor <mjambor@suse.cz>
6014 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6015 represented by a thunk.
6017 2013-09-10 Jeff Law <law@redhat.com>
6019 PR tree-optimization/58343
6020 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6021 jump threading requests through loop headers buried in the middle
6022 of a jump threading path.
6024 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6025 in return value/type.
6027 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6029 PR rtl-optimization/58365
6030 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6031 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6034 2013-09-10 Richard Biener <rguenther@suse.de>
6036 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6037 * tree-data-ref.c (create_rdg_vertices): Collect all data
6038 references, signal failure to the caller, use data-ref API.
6039 (build_rdg): Compute data references only once. Maintain lifetime
6040 of data references and data dependences from within RDG.
6041 (free_rdg): Free dependence relations.
6042 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6043 inventing extra dependences.
6044 (distribute_loop): Update for RDG API changes.
6046 2013-09-10 Kai Tietz <ktietz@redhat.com>
6048 * doc/invoke.texi (fms-extensions): Document changed
6049 behavior for ms-abi targets.
6050 * config/i386/i386.c (ix86_option_override_internal):
6051 Set default value of option -fms-extension for ms-abi targets.
6053 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6055 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6057 2013-09-10 Alan Modra <amodra@gmail.com>
6060 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6062 2013-09-10 Alan Modra <amodra@gmail.com>
6064 * config/rs6000/predicates.md (add_cint_operand): New.
6065 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6066 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6067 using add_cint_operand.
6068 (largetoc_high_plus_aix): Likewise.
6070 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6072 PR tree-optimization/58364
6073 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6074 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6075 the current range can't be an unconditional true or false.
6077 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6079 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6081 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6083 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6085 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6088 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6090 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6092 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6093 NO_REGS for immediate that can't be moved directly into FP_REGS.
6095 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6097 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6098 comparison with negated operand.
6099 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6102 2013-09-09 Richard Biener <rguenther@suse.de>
6105 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6106 of a subloop record all its block as affecting loop-closed SSA form.
6108 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6110 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6111 of an rtx/bitpos pair.
6112 (store_fixed_bit_field): Update accordingly.
6114 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6116 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6118 * builtins.c (expand_errno_check): Likewise.
6119 * dwarf2cfi.c (init_return_column_size): Likewise.
6120 * except.c (sjlj_mark_call_sites): Likewise.
6121 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6122 * lra-constraints.c (emit_inc): Likewise.
6123 * ree.c (combine_set_extension): Likewise.
6124 * regmove.c (fixup_match_2): Likewise.
6125 * reload1.c (inc_for_reload): Likewise.
6127 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6129 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6130 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6131 Use gen_int_mode with the mode of the associated simplify_* call.
6132 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6133 * expmed.c (expand_shift_1): Likewise.
6134 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6135 * loop-iv.c (iv_number_of_iterations): Likewise.
6136 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6137 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6139 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6141 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6142 of the associated expand_* call.
6143 (asan_emit_stack_protection): Likewise.
6144 * builtins.c (round_trampoline_addr): Likewise.
6145 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6146 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6147 (emit_store_flag): Likewise.
6148 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6149 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6151 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6152 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6153 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6154 * modulo-sched.c (generate_prolog_epilog): Likewise.
6155 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6156 (expand_ctz, expand_ffs, expand_unop): Likewise.
6158 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6160 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6161 of the associated gen_rtx_* call.
6162 * caller-save.c (init_caller_save): Likewise.
6163 * combine.c (find_split_point, make_extraction): Likewise.
6164 (make_compound_operation): Likewise.
6165 * dwarf2out.c (mem_loc_descriptor): Likewise.
6166 * explow.c (plus_constant, probe_stack_range): Likewise.
6167 * expmed.c (expand_mult_const): Likewise.
6168 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6169 * reload1.c (init_reload): Likewise.
6170 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6171 * var-tracking.c (adjust_mems): Likewise.
6172 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6173 rather than gen_rtx_fmt_ee.
6175 2013-09-09 Jan Hubicka <jh@suse.cz>
6178 * value-prof.c (gimple_ic): Copy also abnormal edges.
6180 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6182 * asan.c (asan_shadow_cst): Use gen_int_mode.
6184 2013-09-08 Jan Hubicka <jh@suse.cz>
6186 * ipa-profile.c: Add toplevel comment.
6187 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6188 (contains_hot_call_p): New function.
6189 (ipa_propagate_frequency): Set frequencies based on counts when
6191 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6192 profile; do not tamper with profile after inlining if it is read.
6194 2013-09-08 Jan Hubicka <jh@suse.cz>
6196 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6199 2013-09-08 Jan Hubicka <jh@suse.cz>
6201 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6204 2013-09-08 Jeff Law <law@redhat.com>
6207 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6210 2013-09-08 Andi Kleen <ak@linux.intel.com>
6212 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6214 2013-09-08 Jan Hubicka <jh@suse.cz>
6216 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6218 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6220 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6221 for non-debug insns.
6222 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6223 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6224 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6227 2013-09-08 Jan Hubicka <jh@suse.cz>
6229 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6230 targets and devirtualize to BUILT_IN_UNREACHABLE.
6231 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6232 * ipa.c (walk_polymorphic_call_targets): New function.
6233 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6234 functions; use the new timevar.
6235 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6236 was removed from the program.
6237 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6238 not consider it in the walk when its vtable is dead.
6239 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6241 (devirt_variable_node_removal_hook): New function.
6242 (possible_polymorphic_call_targets): Also register
6243 devirt_variable_node_removal_hook.
6244 (ipa_devirt): Do not do non-speculative devirtualization.
6245 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6247 2013-09-08 Jan Hubicka <jh@suse.cz>
6249 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6250 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6251 varpool_remove_variable_insertion_hook): Declare.
6252 * varpool.c (varpool_node_hook_list): New structure.
6253 (first_varpool_node_removal_hook,
6254 first_varpool_variable_insertion_hook): New variables.
6255 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6256 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6257 varpool_remove_variable_insertion_hook,
6258 varpool_call_variable_insertion_hooks): New functions.
6259 (varpool_remove_node): Use it.
6261 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6264 * diagnostic.c (diagnostic_build_prefix): When s.file is
6265 "<built-in>" don't output line and column numbers.
6267 2013-09-06 Jan Hubicka <jh@suse.cz>
6269 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6270 * lto-streamer-out.c: Stream thunks, too.
6271 * lto-streamer-in.c (input_function): Pop cfun here
6272 (lto_read_body): Instead of here.
6274 2013-09-06 Caroline Tice <cmtice@google.com>
6276 * doc/install.texi: Add documentation for the --enable-vtable-verify
6277 and the --disable-libvtv configure options.
6279 2013-09-06 Jeff Law <law@redhat.com>
6281 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6282 edge implied equivalences into successor phis.
6284 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6286 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6288 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6290 * resource.c (mark_target_live_regs): Compute resources taking
6291 into account if a call is predicated or not.
6293 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6295 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6296 the compiler in the function names.
6298 2013-09-06 Jan Hubicka <jh@suse.cz>
6301 * ipa-inline.c (has_caller_p): New function.
6302 (want_inline_function_to_all_callers_p): Use it.
6303 (sum_callers, inline_to_all_callers): Break out from ...
6304 (ipa_inline): ... here.
6306 2013-09-06 Jan Hubicka <jh@suse.cz>
6308 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6309 only when AVX is enabled.
6311 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6313 * config/aarch64/aarch64.md
6314 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6315 is fpsimd_<load/store>2.
6316 (load_pair<mode>): Likewise.
6317 (store_pair<mode>): Likewise.
6319 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6321 * config/arm/types.md (type): Add "mrs" type.
6322 * config/aarch64/aarch64.md
6323 (aarch64_load_tp_hard): Make type "mrs".
6325 (load_tp_hard): Make type "mrs".
6326 * config/arm/cortex-a15.md: Update with new attributes.
6327 * config/arm/cortex-a5.md: Update with new attributes.
6328 * config/arm/cortex-a53.md: Update with new attributes.
6329 * config/arm/cortex-a7.md: Update with new attributes.
6330 * config/arm/cortex-a8.md: Update with new attributes.
6331 * config/arm/cortex-a9.md: Update with new attributes.
6332 * config/arm/cortex-m4.md: Update with new attributes.
6333 * config/arm/cortex-r4.md: Update with new attributes.
6334 * config/arm/fa526.md: Update with new attributes.
6335 * config/arm/fa606te.md: Update with new attributes.
6336 * config/arm/fa626te.md: Update with new attributes.
6337 * config/arm/fa726te.md: Update with new attributes.
6339 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6341 * config/aarch64/aarch64.md
6342 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6343 (*movtf_aarch64): Likewise.
6345 (thumb1_movdi_insn): Use "multiple" for type where more than one
6346 instruction is used for a move.
6347 (*arm32_movhf): Likewise.
6348 (*thumb_movdf_insn): Likewise.
6350 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6352 * config/arm/types.md (type): Rename fcpys to fmov.
6354 (*arm_movsi_vfp): Rename type fcpys as fmov.
6355 (*thumb2_movsi_vfp): Likewise
6356 (*movhf_vfp_neon): Likewise
6357 (*movhf_vfp): Likewise
6358 (*movsf_vfp): Likewise
6359 (*thumb2_movsf_vfp): Likewise
6360 (*movsfcc_vfp): Likewise
6361 (*thumb2_movsfcc_vfp): Likewise
6362 * config/aarch64/aarch64-simd.md
6363 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6364 * config/aarch64/aarch64.md
6365 (*movsi_aarch64): Replace type mov_reg with fmovs.
6366 (*movdi_aarch64): Likewise
6367 (*movsf_aarch64): Likewise
6368 (*movdf_aarch64): Likewise
6370 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6371 * config/arm/iwmmxt.md
6372 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6373 * config/arm/arm1020e.md: Update with new attributes.
6374 * config/arm/cortex-a15-neon.md: Update with new attributes.
6375 * config/arm/cortex-a5.md: Update with new attributes.
6376 * config/arm/cortex-a53.md: Update with new attributes.
6377 * config/arm/cortex-a7.md: Update with new attributes.
6378 * config/arm/cortex-a8-neon.md: Update with new attributes.
6379 * config/arm/cortex-a9.md: Update with new attributes.
6380 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6381 * config/arm/cortex-r4f.md: Update with new attributes.
6382 * config/arm/marvell-pj4.md: Update with new attributes.
6383 * config/arm/vfp11.md: Update with new attributes.
6385 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6387 * config/aarch64/aarch64.md
6388 (*madd<mode>): Fix type attribute.
6389 (*maddsi_uxtw): Likewise.
6390 (*msub<mode>): Likewise.
6391 (*msubsi_uxtw): Likewise.
6392 (<su_optab>maddsidi4): Likewise.
6393 (<su_optab>msubsidi4): Likewise.
6395 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6397 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6398 * config/arm/arm.md (core_cycles): Remove fdiv.
6399 * config/arm/vfp.md:
6400 (*sqrtsf2_vfp): Update for attribute changes.
6401 (*sqrtdf2_vfp): Likewise.
6402 * config/aarch64/aarch64.md:
6403 (sqrt<mode>2): Update for attribute changes.
6404 * config/arm/arm1020e.md: Update with new attributes.
6405 * config/arm/cortex-a15-neon.md: Update with new attributes.
6406 * config/arm/cortex-a5.md: Update with new attributes.
6407 * config/arm/cortex-a53.md: Update with new attributes.
6408 * config/arm/cortex-a7.md: Update with new attributes.
6409 * config/arm/cortex-a8-neon.md: Update with new attributes.
6410 * config/arm/cortex-a9.md: Update with new attributes.
6411 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6412 * config/arm/cortex-r4f.md: Update with new attributes.
6413 * config/arm/marvell-pj4.md: Update with new attributes.
6414 * config/arm/vfp11.md: Update with new attributes.
6416 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6418 * config/arm/types.md
6419 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6420 * config/aarch64/aarch64.md
6421 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6423 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6424 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6425 (float<GPI:mode><GPF:mode>2): Likewise.
6427 (*truncsisf2_vfp): Update with new attributes.
6428 (*truncsidf2_vfp): Likewise.
6429 (fixuns_truncsfsi2): Likewise.
6430 (fixuns_truncdfsi2): Likewise.
6431 (*floatsisf2_vfp): Likewise.
6432 (*floatsidf2_vfp): Likewise.
6433 (floatunssisf2): Likewise.
6434 (floatunssidf2): Likewise.
6435 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6436 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6437 * config/arm/arm1020e.md: Update with new attributes.
6438 * config/arm/cortex-a15-neon.md: Update with new attributes.
6439 * config/arm/cortex-a5.md: Update with new attributes.
6440 * config/arm/cortex-a53.md: Update with new attributes.
6441 * config/arm/cortex-a7.md: Update with new attributes.
6442 * config/arm/cortex-a8-neon.md: Update with new attributes.
6443 * config/arm/cortex-a9.md: Update with new attributes.
6444 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6445 * config/arm/cortex-r4f.md: Update with new attributes.
6446 * config/arm/marvell-pj4.md: Update with new attributes.
6447 * config/arm/vfp11.md: Update with new attributes.
6449 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6451 * config/aarch64/arm_neon.h
6452 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6453 (vqtbx<1,2,3,4><q>_s8): Likewise.
6455 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6457 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6458 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6459 (add<mode>3): Add type attribute.
6460 (add<mode>3): Likewise.
6461 (usadd<mode>3): Likewise.
6462 (ssadd<mode>3): Likewise.
6463 (sub<mode>3): Likewise.
6464 (sub<mode>3): Likewise.
6465 (ussub<mode>3): Likewise.
6466 (sssub<mode>3): Likewise.
6467 (ssmulsa3): Likewise.
6468 (usmulusa3): Likewise.
6469 (arm_usatsihi): Likewise.
6471 (*movdi_vfp): Add types for all instructions.
6472 (*movdi_vfp_cortexa8): Likewise.
6473 (*movhf_vfp_neon): Likewise.
6474 (*movhf_vfp): Likewise.
6475 (*movdf_vfp): Likewise.
6476 (*thumb2_movdf_vfp): Likewise.
6477 (*thumb2_movdfcc_vfp): Likewise.
6478 * config/arm/arm.md: Add type attribute to all insn patterns.
6479 (*thumb1_adddi3): Add type attribute.
6480 (*arm_adddi3): Likewise.
6481 (*adddi_sesidi_di): Likewise.
6482 (*adddi_zesidi_di): Likewise.
6483 (*thumb1_addsi3): Likewise.
6484 (addsi3_compare0): Likewise.
6485 (*addsi3_compare0_scratch): Likewise.
6486 (*compare_negsi_si): Likewise.
6487 (cmpsi2_addneg): Likewise.
6488 (*addsi3_carryin_<optab>): Likewise.
6489 (*addsi3_carryin_alt2_<optab>): Likewise.
6490 (*addsi3_carryin_clobercc_<optab>): Likewise.
6491 (*subsi3_carryin): Likewise.
6492 (*subsi3_carryin_const): Likewise.
6493 (*subsi3_carryin_compare): Likewise.
6494 (*subsi3_carryin_compare_const): Likewise.
6495 (*arm_subdi3): Likewise.
6496 (*thumb_subdi3): Likewise.
6497 (*subdi_di_zesidi): Likewise.
6498 (*subdi_di_sesidi): Likewise.
6499 (*subdi_zesidi_di): Likewise.
6500 (*subdi_sesidi_di): Likewise.
6501 (*subdi_zesidi_ze): Likewise.
6502 (thumb1_subsi3_insn): Likewise.
6503 (*arm_subsi3_insn): Likewise.
6504 (*anddi3_insn): Likewise.
6505 (*anddi_zesidi_di): Likewise.
6506 (*anddi_sesdi_di): Likewise.
6507 (*ne_zeroextracts): Likewise.
6508 (*ne_zeroextracts): Likewise.
6509 (*ite_ne_zeroextr): Likewise.
6510 (*ite_ne_zeroextr): Likewise.
6511 (*anddi_notdi_di): Likewise.
6512 (*anddi_notzesidi): Likewise.
6513 (*anddi_notsesidi): Likewise.
6514 (andsi_notsi_si): Likewise.
6515 (thumb1_bicsi3): Likewise.
6516 (*iordi3_insn): Likewise.
6517 (*iordi_zesidi_di): Likewise.
6518 (*iordi_sesidi_di): Likewise.
6519 (*thumb1_iorsi3_insn): Likewise.
6520 (*xordi3_insn): Likewise.
6521 (*xordi_zesidi_di): Likewise.
6522 (*xordi_sesidi_di): Likewise.
6523 (*arm_xorsi3): Likewise.
6524 (*andsi_iorsi3_no): Likewise.
6525 (*smax_0): Likewise.
6526 (*smax_m1): Likewise.
6527 (*arm_smax_insn): Likewise.
6528 (*smin_0): Likewise.
6529 (*arm_smin_insn): Likewise.
6530 (*arm_umaxsi3): Likewise.
6531 (*arm_uminsi3): Likewise.
6532 (*minmax_arithsi): Likewise.
6533 (*minmax_arithsi_): Likewise.
6534 (*satsi_<SAT:code>): Likewise.
6535 (arm_ashldi3_1bit): Likewise.
6536 (arm_ashrdi3_1bit): Likewise.
6537 (arm_lshrdi3_1bit): Likewise.
6538 (*arm_negdi2): Likewise.
6539 (*thumb1_negdi2): Likewise.
6540 (*arm_negsi2): Likewise.
6541 (*thumb1_negsi2): Likewise.
6542 (*negdi_extendsid): Likewise.
6543 (*negdi_zero_extend): Likewise.
6544 (*arm_abssi2): Likewise.
6545 (*thumb1_abssi2): Likewise.
6546 (*arm_neg_abssi2): Likewise.
6547 (*thumb1_neg_abss): Likewise.
6548 (one_cmpldi2): Likewise.
6549 (extend<mode>di2): Likewise.
6550 (*compareqi_eq0): Likewise.
6551 (*arm_extendhisi2addsi): Likewise.
6552 (*arm_movdi): Likewise.
6553 (*thumb1_movdi_insn): Likewise.
6554 (*arm_movt): Likewise.
6555 (*thumb1_movsi_insn): Likewise.
6556 (pic_add_dot_plus_four): Likewise.
6557 (pic_add_dot_plus_eight): Likewise.
6558 (tls_load_dot_plus_eight): Likewise.
6559 (*thumb1_movhi_insn): Likewise.
6560 (*thumb1_movsf_insn): Likewise.
6561 (*movdf_soft_insn): Likewise.
6562 (*thumb_movdf_insn): Likewise.
6563 (cbranchsi4_insn): Likewise.
6564 (cbranchsi4_scratch): Likewise.
6565 (*negated_cbranchsi4): Likewise.
6566 (*tbit_cbranch): Likewise.
6567 (*tlobits_cbranch): Likewise.
6568 (*tstsi3_cbranch): Likewise.
6569 (*cbranchne_decr1): Likewise.
6570 (*addsi3_cbranch): Likewise.
6571 (*addsi3_cbranch_scratch): Likewise.
6572 (*arm_cmpdi_insn): Likewise.
6573 (*arm_cmpdi_unsig): Likewise.
6574 (*arm_cmpdi_zero): Likewise.
6575 (*thumb_cmpdi_zero): Likewise.
6576 (*deleted_compare): Likewise.
6577 (*mov_scc): Likewise.
6578 (*mov_negscc): Likewise.
6579 (*mov_notscc): Likewise.
6580 (*cstoresi_eq0_thumb1_insn): Likewise.
6581 (cstoresi_nltu_thumb1): Likewise.
6582 (cstoresi_ltu_thu): Likewise.
6583 (thumb1_addsi3_addgeu): Likewise.
6584 (*arm_jump): Likewise.
6585 (*thumb_jump): Likewise.
6586 (*check_arch2): Likewise.
6587 (arm_casesi_internal): Likewise.
6588 (thumb1_casesi_dispatch): Likewise.
6589 (*arm_indirect_jump): Likewise.
6590 (*thumb1_indirect_jump): Likewise.
6592 (*and_scc): Likewise.
6593 (*ior_scc): Likewise.
6594 (*compare_scc): Likewise.
6595 (*cond_move): Likewise.
6596 (*cond_arith): Likewise.
6597 (*cond_sub): Likewise.
6598 (*cmp_ite0): Likewise.
6599 (*cmp_ite1): Likewise.
6600 (*cmp_and): Likewise.
6601 (*cmp_ior): Likewise.
6602 (*ior_scc_scc): Likewise.
6603 (*ior_scc_scc_cmp): Likewise.
6604 (*and_scc_scc): Likewise.
6605 (*and_scc_scc_cmp): Likewise.
6606 (*and_scc_scc_nod): Likewise.
6607 (*negscc): Likewise.
6608 (movcond_addsi): Likewise.
6609 (movcond): Likewise.
6610 (*ifcompare_plus_move): Likewise.
6611 (*if_plus_move): Likewise.
6612 (*ifcompare_move_plus): Likewise.
6613 (*if_move_plus): Likewise.
6614 (*ifcompare_arith_arith): Likewise.
6615 (*if_arith_arith): Likewise.
6616 (*ifcompare_arith_move): Likewise.
6617 (*if_arith_move): Likewise.
6618 (*ifcompare_move_arith): Likewise.
6619 (*if_move_arith): Likewise.
6620 (*ifcompare_move_not): Likewise.
6621 (*if_move_not): Likewise.
6622 (*ifcompare_not_move): Likewise.
6623 (*if_not_move): Likewise.
6624 (*ifcompare_shift_move): Likewise.
6625 (*if_shift_move): Likewise.
6626 (*ifcompare_move_shift): Likewise.
6627 (*if_move_shift): Likewise.
6628 (*ifcompare_shift_shift): Likewise.
6629 (*ifcompare_not_arith): Likewise.
6630 (*ifcompare_arith_not): Likewise.
6631 (*if_arith_not): Likewise.
6632 (*ifcompare_neg_move): Likewise.
6633 (*if_neg_move): Likewise.
6634 (*ifcompare_move_neg): Likewise.
6635 (*if_move_neg): Likewise.
6636 (prologue_thumb1_interwork): Likewise.
6637 (*cond_move_not): Likewise.
6638 (*sign_extract_onebit): Likewise.
6639 (*not_signextract_onebit): Likewise.
6640 (stack_tie): Likewise.
6641 (align_4): Likewise.
6642 (align_8): Likewise.
6643 (consttable_end): Likewise.
6644 (consttable_1): Likewise.
6645 (consttable_2): Likewise.
6646 (consttable_4): Likewise.
6647 (consttable_8): Likewise.
6648 (consttable_16): Likewise.
6649 (*thumb1_tablejump): Likewise.
6650 (prefetch): Likewise.
6651 (force_register_use): Likewise.
6652 (thumb_eh_return): Likewise.
6653 (load_tp_hard): Likewise.
6654 (load_tp_soft): Likewise.
6655 (tlscall): Likewise.
6656 (*arm_movtas_ze): Likewise.
6657 (*arm_rev): Likewise.
6658 (*arm_revsh): Likewise.
6659 (*arm_rev16): Likewise.
6660 * config/arm/thumb2.md
6661 (*thumb2_smaxsi3): Likewise.
6662 (*thumb2_sminsi3): Likewise.
6663 (*thumb32_umaxsi3): Likewise.
6664 (*thumb2_uminsi3): Likewise.
6665 (*thumb2_negdi2): Likewise.
6666 (*thumb2_abssi2): Likewise.
6667 (*thumb2_neg_abss): Likewise.
6668 (*thumb2_movsi_insn): Likewise.
6669 (tls_load_dot_plus_four): Likewise.
6670 (*thumb2_movhi_insn): Likewise.
6671 (*thumb2_mov_scc): Likewise.
6672 (*thumb2_mov_negs): Likewise.
6673 (*thumb2_mov_negs): Likewise.
6674 (*thumb2_mov_nots): Likewise.
6675 (*thumb2_mov_nots): Likewise.
6676 (*thumb2_movsicc_): Likewise.
6677 (*thumb2_movsfcc_soft_insn): Likewise.
6678 (*thumb2_indirect_jump): Likewise.
6679 (*thumb2_and_scc): Likewise.
6680 (*thumb2_ior_scc): Likewise.
6681 (*thumb2_ior_scc_strict_it): Likewise.
6682 (*thumb2_cond_move): Likewise.
6683 (*thumb2_cond_arith): Likewise.
6684 (*thumb2_cond_ari): Likewise.
6685 (*thumb2_cond_sub): Likewise.
6686 (*thumb2_negscc): Likewise.
6687 (*thumb2_movcond): Likewise.
6688 (thumb2_casesi_internal): Likewise.
6689 (thumb2_casesi_internal_pic): Likewise.
6690 (*thumb2_alusi3_short): Likewise.
6691 (*thumb2_mov<mode>_shortim): Likewise.
6692 (*thumb2_addsi_short): Likewise.
6693 (*thumb2_subsi_short): Likewise.
6694 (thumb2_addsi3_compare0): Likewise.
6695 (*thumb2_cbz): Likewise.
6696 (*thumb2_cbnz): Likewise.
6697 (*thumb2_one_cmplsi2_short): Likewise.
6698 (*thumb2_negsi2_short): Likewise.
6699 (*orsi_notsi_si): Likewise.
6700 * config/arm/arm1020e.md: Update with new attributes.
6701 * config/arm/arm1026ejs.md: Update with new attributes.
6702 * config/arm/arm1136jfs.md: Update with new attributes.
6703 * config/arm/arm926ejs.md: Update with new attributes.
6704 * config/arm/cortex-a15.md: Update with new attributes.
6705 * config/arm/cortex-a5.md: Update with new attributes.
6706 * config/arm/cortex-a53.md: Update with new attributes.
6707 * config/arm/cortex-a7.md: Update with new attributes.
6708 * config/arm/cortex-a8.md: Update with new attributes.
6709 * config/arm/cortex-a9.md: Update with new attributes.
6710 * config/arm/cortex-m4.md: Update with new attributes.
6711 * config/arm/cortex-r4.md: Update with new attributes.
6712 * config/arm/fa526.md: Update with new attributes.
6713 * config/arm/fa606te.md: Update with new attributes.
6714 * config/arm/fa626te.md: Update with new attributes.
6715 * config/arm/fa726te.md: Update with new attributes.
6717 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6719 * config/aarch64/aarch64-simd.md
6720 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6721 <vwx> iterator to ensure correct register choice.
6722 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6723 (aarch64_sqdmull_n<mode>): Likewise.
6724 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6725 * config/aarch64/arm_neon.h
6726 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6727 (vml<as><q>_n_<su>16): Likewise.
6728 (vml<as>l_high_lane<q>_<su>16): Likewise.
6729 (vml<as>l_high_n_<su>16): Likewise.
6730 (vml<as>l_lane<q>_<su>16): Likewise.
6731 (vml<as>l_n_<su>16): Likewise.
6732 (vmul<q>_lane<q>_<su>16): Likewise.
6733 (vmul<q>_n_<su>16): Likewise.
6734 (vmull_lane<q>_<su>16): Likewise.
6735 (vmull_n_<su>16): Likewise.
6736 (vmull_high_lane<q>_<su>16): Likewise.
6737 (vmull_high_n_<su>16): Likewise.
6738 (vqrdmulh<q>_n_s16): Likewise.
6740 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6742 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6743 have the correct lane parameter.
6745 2013-09-06 Richard Biener <rguenther@suse.de>
6747 * cfganal.c (control_dependences::~control_dependences):
6748 Properly free all of the vector.
6750 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6753 * config/i386/i386.c (ix86_conditional_register_usage):
6754 Proper initialize extended SSE registers.
6756 2013-09-06 Jan Hubicka <jh@suse.cz>
6758 PR tree-optimization/58311
6759 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6761 2013-09-06 Jan Hubicka <jh@suse.cz>
6763 * Makefile.in (tree-sra.o): Update dependencies.
6764 * tree-sra.c: Include ipa-utils.h
6765 (scan_function): Use recursive_call_p.
6766 (has_caller_p): New function.
6767 (cgraph_for_node_and_aliases): Count also callers of aliases.
6769 2013-09-06 Jan Hubicka <jh@suse.cz>
6772 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6773 * tree-tailcall.c: Include ipa-utils.h.
6774 (find_tail_calls): Use it.
6775 * ipa-pure-const.c (check_call): Likewise.
6776 * ipa-utils.c (recursive_call_p): New function.
6777 * ipa-utils.h (recursive_call_p): Dclare.
6778 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6779 (symtab_semantically_equivalent_p): New function.
6780 * Makefile.in (tree-tailcall.o): Update dependencies.
6782 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6784 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6787 2013-09-06 Richard Biener <rguenther@suse.de>
6789 * lto-streamer.h (lto_global_var_decls): Remove.
6790 * Makefile.in (OBJS): Remove lto-symtab.o.
6791 (lto-symtab.o): Remove.
6792 (GTFILES): Remove lto-symtab.c
6793 * lto-symtab.c: Move to lto/
6795 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6797 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6798 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6799 (UNSPEC_FPINT_RINT): New constant definitions.
6800 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6801 definition with 2 attributes.
6802 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6803 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6806 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6808 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6809 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6810 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6811 attribute to "z196_alone".
6812 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6815 2013-09-06 Richard Biener <rguenther@suse.de>
6817 * basic-block.h (class control_dependences): New.
6818 * tree-ssa-dce.c (control_dependence_map): Remove.
6820 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6821 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6822 find_pdom, find_control_dependence, find_all_control_dependences):
6824 (mark_control_dependent_edges_necessary,
6825 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6826 tree_dce_done, perform_tree_ssa_dce): Adjust.
6827 * cfganal.c (set_control_dependence_map_bit,
6828 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6829 find_all_control_dependences): Move from tree-ssa-dce.c and
6830 implement as methods of control_dependences class.
6831 (control_dependences::control_dependences): New.
6832 (control_dependences::~control_dependences): Likewise.
6833 (control_dependences::get_edges_dependent_on): Likewise.
6834 (control_dependences::get_edge): Likewise.
6836 2013-09-04 Jan Hubicka <jh@suse.cz>
6838 * tree.c (types_same_for_odr): Drop overactive check.
6839 * ipa-devirt.c (hash_type_name): Likewise.
6841 2013-09-04 Jan Hubicka <jh@suse.cz>
6843 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6844 (analyze_functions): ... here.
6846 2013-09-04 Jan Hubicka <jh@suse.cz>
6849 * cgraphunit.c (analyze_functions): Clear AUX fields
6850 after processing; initialize assembler name has.
6852 2013-09-05 Jeff Law <law@redhat.com>
6854 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6855 from thread_around_empty_block. Record threading path into PATH.
6856 Recurse if threading through the initial block is successful.
6857 (thread_across_edge): Corresponding changes to slightly simplify.
6859 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6861 * config/aarch64/aarch64.md
6862 (type): Remove frecpe, frecps, frecpx.
6863 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6864 fix to be a TARGET_SIMD instruction.
6865 (aarch64_frecps): Remove.
6866 * config/aarch64/aarch64-simd.md
6867 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6868 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6870 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6871 Sofiane Naci <sofiane.naci@arm.com>
6873 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6874 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6875 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6876 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6877 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6878 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6879 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6880 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6881 into "clz, "rbit". Rename "shift" to "shift_imm".
6882 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6883 changes. Update for attribute changes all occurrences of arlo_* and
6885 * config/arm/arm-fixed.md: Update for attribute changes
6886 all occurrences of arlo_* types.
6887 * config/arm/thumb2.md: Update for attribute changes all occurrences
6889 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6890 (cortexa7_older_only): Likewise.
6891 (cortexa7_younger): Likewise.
6892 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6893 (1020alu_shift_op): Likewise.
6894 (1020alu_shift_reg_op): Likewise.
6895 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6896 (alu_shift_op): Likewise.
6897 (alu_shift_reg_op): Likewise.
6898 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6899 (11_alu_shift_op): Likewise.
6900 (11_alu_shift_reg_op): Likewise.
6901 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6902 (9_alu_shift_reg_op): Likewise.
6903 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6905 (cortex_a15_alu_shift): Likewise.
6906 (cortex_a15_alu_shift_reg): Likewise.
6907 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6909 (cortex_a5_alu_shift): Likewise.
6910 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6912 (cortex_a53_alu_shift): Likewise.
6913 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6915 (cortex_a7_alu_reg): Likewise.
6916 (cortex_a7_alu_shift): Likewise.
6917 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6919 (cortex_a8_alu_shift): Likewise.
6920 (cortex_a8_alu_shift_reg): Likewise.
6921 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6922 (cortex_a9_dp_shift): Likewise.
6923 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6925 * config/arm/cortex-r4.md
6926 (cortex_r4_alu): Update for attribute changes.
6927 (cortex_r4_mov): Likewise.
6928 (cortex_r4_alu_shift_reg): Likewise.
6929 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6930 (526_alu_shift_op): Likewise.
6931 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6932 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6933 (626te_alu_shift_op): Likewise.
6934 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6935 (726te_alu_shift_op): Likewise.
6936 (726te_alu_shift_reg_op): Likewise.
6937 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6938 (mp626_alu_shift_op): Likewise.
6939 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6940 (pj4_alu_conds): Likewise.
6941 (pj4_shift): Likewise.
6942 (pj4_shift_conds): Likewise.
6943 (pj4_alu_shift): Likewise.
6944 (pj4_alu_shift_conds): Likewise.
6945 * config/aarch64/aarch64.md: Update for attribute change
6946 all occurrences of arlo_* and shift* types.
6948 2013-09-05 Mike Stump <mikestump@comcast.net>
6950 * tree.h: Move documentation for tree_function_decl to tree-core.h
6951 with the declaration.
6953 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6956 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6957 looking for widest mode.
6959 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6961 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6963 2013-09-05 Richard Biener <rguenther@suse.de>
6965 PR tree-optimization/58137
6966 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6967 Do not create vectors of pointers.
6968 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6969 types for the components of the vector initializer.
6970 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6971 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6973 2013-09-05 Martin Jambor <mjambor@suse.cz>
6975 * ipa-prop.c (remove_described_reference): Accept missing references,
6976 return false if that hppens, otherwise return true.
6977 (cgraph_node_for_jfunc): New function.
6978 (try_decrement_rdesc_refcount): Likewise.
6979 (try_make_edge_direct_simple_call): Use them.
6980 (ipa_edge_removal_hook): Remove references from rdescs.
6981 (ipa_edge_duplication_hook): Clone rdescs and their references
6982 when the new edge has the same caller as the old one.
6983 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6984 reference before removing any edges.
6986 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6988 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6990 * thumb2.md (thumb2_storewb_parisi): New pattern.
6992 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6994 * config/aarch64/aarch64-option-extensions.def: Add
6995 AARCH64_OPT_EXTENSION of 'crc'.
6996 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6997 (AARCH64_ISA_CRC): Ditto.
6998 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6999 description of the CRC extension.
7001 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7003 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7004 * config/rs6000/linux.h: Ditto.
7005 * alpha/linux.h: Ditto.
7006 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7007 no_c99_libc_has_function.
7008 * config/c6x/uclinux-elf.h: Ditto.
7009 * config/lm32/uclinux-elf.h: Ditto.
7010 * config/m68k/uclinux.h: Ditto.
7011 * config/moxie/uclinux.h: Ditto.
7012 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7013 (crisv32-*-linux*, cris-*-linux*): Ditto.
7014 * config/bfin/bfin.c: Include "tm_p.h".
7016 2013-09-05 Richard Biener <rguenther@suse.de>
7018 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7019 check for a definition without a basic-block.
7021 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7022 Sofiane Naci <sofiane.naci@arm.com>
7024 * config/aarch64/aarch64.md
7025 (*movti_aarch64): Rename r_2_f and f_2_r.
7026 (*movsf_aarch64): Likewise.
7027 (*movdf_aarch64): Likewise.
7028 (*movtf_aarch64): Likewise.
7029 (aarch64_movdi_<mode>low): Likewise.
7030 (aarch64_movdi_<mode>high): Likewise.
7031 (aarch64_mov<mode>high_di): Likewise.
7032 (aarch64_mov<mode>low_di): Likewise.
7033 (aarch64_movtilow_tilow): Likewise.
7034 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7035 values to config/arm/types.md
7036 (attribute "conds"): Update for attribute change.
7037 (anddi3_insn): Likewise.
7038 (iordi3_insn): Likewise.
7039 (xordi3_insn): Likewise.
7040 (one_cmpldi2): Likewise.
7041 * config/arm/types.md (type): Add Neon types.
7042 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7043 use "type" attribute.
7044 (movmisalign<mode>_neon_store): Likewise.
7045 (movmisalign<mode>_neon_load): Likewise.
7046 (vec_set<mode>_internal): Likewise.
7047 (vec_setv2di_internal): Likewise.
7048 (vec_extract<mode>): Likewise.
7049 (vec_extractv2di): Likewise.
7050 (add<mode>3_neon): Likewise.
7051 (adddi3_neon): Likewise.
7052 (sub<mode>3_neon): Likewise.
7053 (subdi3_neon): Likewise.
7054 (mul<mode>3_neon): Likewise.
7055 (mul<mode>3add<mode>_neon): Likewise.
7056 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7057 (fma<VCVTF:mode>4)): Likewise.
7058 (fma<VCVTF:mode>4_intrinsic): Likewise.
7059 (fmsub<VCVTF:mode>4)): Likewise.
7060 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7061 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7062 (ior<mode>3): Likewise.
7063 (and<mode>3): Likewise.
7064 (anddi3_neon): Likewise.
7065 (orn<mode>3_neon): Likewise.
7066 (orndi3_neon): Likewise.
7067 (bic<mode>3_neon): Likewise.
7068 (bicdi3_neon): Likewise.
7069 (xor<mode>3): Likewise.
7070 (one_cmpl<mode>2): Likewise.
7071 (abs<mode>2): Likewise.
7072 (neg<mode>2): Likewise.
7073 (umin<mode>3_neon): Likewise.
7074 (umax<mode>3_neon): Likewise.
7075 (smin<mode>3_neon): Likewise.
7076 (smax<mode>3_neon): Likewise.
7077 (vashl<mode>3): Likewise.
7078 (vashr<mode>3_imm): Likewise.
7079 (vlshr<mode>3_imm): Likewise.
7080 (ashl<mode>3_signed): Likewise.
7081 (ashl<mode>3_unsigned): Likewise.
7082 (neon_load_count): Likewise.
7083 (ashldi3_neon_noclobber): Likewise.
7084 (signed_shift_di3_neon): Likewise.
7085 (unsigned_shift_di3_neon): Likewise.
7086 (ashrdi3_neon_imm_noclobber): Likewise.
7087 (lshrdi3_neon_imm_noclobber): Likewise.
7088 (widen_ssum<mode>3): Likewise.
7089 (widen_usum<mode>3): Likewise.
7090 (quad_halves_<code>v4si): Likewise.
7091 (quad_halves_<code>v4sf): Likewise.
7092 (quad_halves_<code>v8hi): Likewise.
7093 (quad_halves_<code>v16qi): Likewise.
7094 (reduc_splus_v2di): Likewise.
7095 (neon_vpadd_internal<mode>): Likewise.
7096 (neon_vpsmin<mode>): Likewise.
7097 (neon_vpsmax<mode>): Likewise.
7098 (neon_vpumin<mode>): Likewise.
7099 (neon_vpumax<mode>): Likewise.
7100 (ss_add<mode>_neon): Likewise.
7101 (us_add<mode>_neon): Likewise.
7102 (ss_sub<mode>_neon): Likewise.
7103 (us_sub<mode>_neon): Likewise.
7104 (neon_vadd<mode>_unspec): Likewise.
7105 (neon_vaddl<mode>): Likewise.
7106 (neon_vaddw<mode>): Likewise.
7107 (neon_vhadd<mode>): Likewise.
7108 (neon_vqadd<mode>): Likewise.
7109 (neon_vaddhn<mode>): Likewise.
7110 (neon_vmul<mode>): Likewise.
7111 (neon_vmla<mode>): Likewise.
7112 (neon_vmlal<mode>): Likewise.
7113 (neon_vmls<mode>): Likewise.
7114 (neon_vmlsl<mode>): Likewise.
7115 (neon_vqdmulh<mode>): Likewise.
7116 (neon_vqdmlal<mode>): Likewise.
7117 (neon_vqdmlsl<mode>): Likewise.
7118 (neon_vmull<mode>): Likewise.
7119 (neon_vqdmull<mode>): Likewise.
7120 (neon_vsub<mode>_unspec): Likewise.
7121 (neon_vsubl<mode>): Likewise.
7122 (neon_vsubw<mode>): Likewise.
7123 (neon_vqsub<mode>): Likewise.
7124 (neon_vhsub<mode>): Likewise.
7125 (neon_vsubhn<mode>): Likewise.
7126 (neon_vceq<mode>): Likewise.
7127 (neon_vcge<mode>): Likewise.
7128 (neon_vcgeu<mode>): Likewise.
7129 (neon_vcgt<mode>): Likewise.
7130 (neon_vcgtu<mode>): Likewise.
7131 (neon_vcle<mode>): Likewise.
7132 (neon_vclt<mode>): Likewise.
7133 (neon_vcage<mode>): Likewise.
7134 (neon_vcagt<mode>): Likewise.
7135 (neon_vtst<mode>): Likewise.
7136 (neon_vabd<mode>): Likewise.
7137 (neon_vabdl<mode>): Likewise.
7138 (neon_vaba<mode>): Likewise.
7139 (neon_vabal<mode>): Likewise.
7140 (neon_vmax<mode>): Likewise.
7141 (neon_vmin<mode>): Likewise.
7142 (neon_vpaddl<mode>): Likewise.
7143 (neon_vpadal<mode>): Likewise.
7144 (neon_vpmax<mode>): Likewise.
7145 (neon_vpmin<mode>): Likewise.
7146 (neon_vrecps<mode>): Likewise.
7147 (neon_vrsqrts<mode>): Likewise.
7148 (neon_vqabs<mode>): Likewise.
7149 (neon_vqneg<mode>): Likewise.
7150 (neon_vcls<mode>): Likewise.
7151 (clz<mode>2): Likewise.
7152 (popcount<mode>2): Likewise.
7153 (neon_vrecpe): Likewise.
7154 (neon_vrsqrte): Likewise.
7155 (neon_vget_lane<mode>_sext_internal): Likewise.
7156 (neon_vget_lane<mode>_zext_internal): Likewise.
7157 (neon_vdup_n<mode>): Likewise.
7158 (neon_vdup_nv2di): Likewise.
7159 (neon_vdpu_lane<mode>_internal): Likewise.
7160 (neon_vswp<mode>): Likewise.
7161 (float<mode><V_cvtto>2): Likewise.
7162 (floatuns<mode><V_cvtto>2): Likewise.
7163 (fix_trunc<mode><V_cvtto>)2): Likewise
7164 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7165 (neon_vcvt<mode>): Likewise.
7166 (neon_vcvtv4sfv4hf): Likewise.
7167 (neon_vcvtv4hfv4sf): Likewise.
7168 (neon_vcvt_n<mode>): Likewise.
7169 (neon_vmovn<mode>): Likewise.
7170 (neon_vqmovn<mode>): Likewise.
7171 (neon_vqmovun<mode>): Likewise.
7172 (neon_vmovl<mode>): Likewise.
7173 (neon_vmul_lane<mode>): Likewise.
7174 (neon_vmull_lane<mode>): Likewise.
7175 (neon_vqdmull_lane<mode>): Likewise.
7176 (neon_vqdmulh_lane<mode>): Likewise.
7177 (neon_vmla_lane<mode>): Likewise.
7178 (neon_vmlal_lane<mode>): Likewise.
7179 (neon_vqdmlal_lane<mode>): Likewise.
7180 (neon_vmls_lane<mode>): Likewise.
7181 (neon_vmlsl_lane<mode>): Likewise.
7182 (neon_vqdmlsl_lane<mode>): Likewise.
7183 (neon_vext<mode>): Likewise.
7184 (neon_vrev64<mode>): Likewise.
7185 (neon_vrev32<mode>): Likewise.
7186 (neon_vrev16<mode>): Likewise.
7187 (neon_vbsl<mode>_internal): Likewise.
7188 (neon_vshl<mode>): Likewise.
7189 (neon_vqshl<mode>): Likewise.
7190 (neon_vshr_n<mode>): Likewise.
7191 (neon_vshrn_n<mode>): Likewise.
7192 (neon_vqshrn_n<mode>): Likewise.
7193 (neon_vqshrun_n<mode>): Likewise.
7194 (neon_vshl_n<mode>): Likewise.
7195 (neon_vqshl_n<mode>): Likewise.
7196 (neon_vqshlu_n<mode>): Likewise.
7197 (neon_vshll_n<mode>): Likewise.
7198 (neon_vsra_n<mode>): Likewise.
7199 (neon_vsri_n<mode>): Likewise.
7200 (neon_vsli_n<mode>): Likewise.
7201 (neon_vtbl1v8qi): Likewise.
7202 (neon_vtbl2v8qi): Likewise.
7203 (neon_vtbl3v8qi): Likewise.
7204 (neon_vtbl4v8qi): Likewise.
7205 (neon_vtbx1v8qi): Likewise.
7206 (neon_vtbx2v8qi): Likewise.
7207 (neon_vtbx3v8qi): Likewise.
7208 (neon_vtbx4v8qi): Likewise.
7209 (neon_vtrn<mode>_internal): Likewise.
7210 (neon_vzip<mode>_internal): Likewise.
7211 (neon_vuzp<mode>_internal): Likewise.
7212 (neon_vld1<mode>): Likewise.
7213 (neon_vld1_lane<mode>): Likewise.
7214 (neon_vld1_dup<mode>): Likewise.
7215 (neon_vld1_dupv2di): Likewise.
7216 (neon_vst1<mode>): Likewise.
7217 (neon_vst1_lane<mode>): Likewise.
7218 (neon_vld2<mode>): Likewise.
7219 (neon_vld2_lane<mode>): Likewise.
7220 (neon_vld2_dup<mode>): Likewise.
7221 (neon_vst2<mode>): Likewise.
7222 (neon_vst2_lane<mode>): Likewise.
7223 (neon_vld3<mode>): Likewise.
7224 (neon_vld3qa<mode>): Likewise.
7225 (neon_vld3qb<mode>): Likewise.
7226 (neon_vld3_lane<mode>): Likewise.
7227 (neon_vld3_dup<mode>): Likewise.
7228 (neon_vst3<mode>): Likewise.
7229 (neon_vst3qa<mode>): Likewise.
7230 (neon_vst3qb<mode>): Likewise.
7231 (neon_vst3_lane<mode>): Likewise.
7232 (neon_vld4<mode>): Likewise.
7233 (neon_vld4qa<mode>): Likewise.
7234 (neon_vld4qb<mode>): Likewise.
7235 (neon_vld4_lane<mode>): Likewise.
7236 (neon_vld4_dup<mode>): Likewise.
7237 (neon_vst4<mode>): Likewise.
7238 (neon_vst4qa<mode>): Likewise.
7239 (neon_vst4qb<mode>): Likewise.
7240 (neon_vst4_lane<mode>): Likewise.
7241 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7242 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7243 (neon_vec_<US>mult_lo_<mode>): Likewise.
7244 (neon_vec_<US>mult_hi_<mode>): Likewise.
7245 (neon_vec_<US>shiftl_<mode>): Likewise.
7246 (neon_unpack<US>_<mode>): Likewise.
7247 (neon_vec_<US>mult_<mode>): Likewise.
7248 (vec_pack_trunc_<mode>): Likewise.
7249 (neon_vec_pack_trunk_<mode>): Likewise.
7250 (neon_vabd<mode>_2): Likewise.
7251 (neon_vabd<mode>_3): Likewise.
7252 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7253 (thumb2_movsi_vfp): Likewise.
7254 (movdi_vfp): Likewise.
7255 (movdi_vfp_cortexa8): Likewise.
7256 (movhf_vfp_neon): Likewise.
7257 (movhf_vfp): Likewiwse.
7258 (movsf_vfp): Likewiwse.
7259 (thumb2_movsf_vfp): Likewiwse.
7260 (movdf_vfp): Likewise.
7261 (thumb2_movdf_vfp): Likewise.
7262 (movsfcc_vfp): Likewise.
7263 (thumb2_movsfcc_vfp): Likewise.
7264 (movdfcc_vfp): Likewise.
7265 (thumb2_movdfcc_vfp): Likewise.
7266 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7267 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7268 (v10_v2c): Likewise.
7269 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7271 (cortex_a15_neon_int_2): Likewise.
7272 (cortex_a15_neon_int_3): Likewise.
7273 (cortex_a15_neon_int_4): Likewise.
7274 (cortex_a15_neon_int_5): Likewise.
7275 (cortex_a15_neon_vqneg_vqabs): Likewise.
7276 (cortex_a15_neon_vmov): Likewise.
7277 (cortex_a15_neon_vaba): Likewise.
7278 (cortex_a15_neon_vaba_qqq): Likewise.
7279 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7280 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7281 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7282 scalar_64_32_long_scalar): Likewise.
7283 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7284 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7285 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7286 lotype_qdd_64_32_long): Likewise.
7287 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7288 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7289 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7290 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7291 (cortex_a15_neon_shift_1): Likewise.
7292 (cortex_a15_neon_shift_2): Likewise.
7293 (cortex_a15_neon_shift_3): Likewise.
7294 (cortex_a15_neon_vshl_ddd): Likewise.
7295 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7296 (cortex_a15_neon_vsra_vrsra): Likewise.
7297 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7298 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7299 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7300 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7301 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7302 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7303 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7304 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7305 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7306 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7307 (cortex_a15_neon_bp_simple): Likewise.
7308 (cortex_a15_neon_bp_2cycle): Likewise.
7309 (cortex_a15_neon_bp_3cycle): Likewise.
7310 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7311 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7312 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7313 (cortex_a15_neon_vld2_4_regs): Likewise.
7314 (cortex_a15_neon_vld3_vld4): Likewise.
7315 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7316 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7317 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7318 (cortex_a15_neon_vst3_vst4): Likewise.
7319 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7320 (cortex_a15_neon_vld3_vld4_lane" 10
7321 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7322 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7323 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7324 (cortex_a15_neon_ldm_2): Likewise.0
7325 (cortex_a15_neon_stm_2): Likewise.
7326 (cortex_a15_neon_mcr): Likewise.
7327 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7328 (cortex_a15_neon_mrc): Likewise.
7329 (cortex_a15_neon_mrrc): Likewise.
7330 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7332 (cortex_a15_alu_shift): Likewise.
7333 (cortex_a15_alu_shift_reg): Likewise.
7334 (cortex_a15_mult32): Likewise.
7335 (cortex_a15_mult64): Likewise.
7336 (cortex_a15_block): Likewise.
7337 (cortex_a15_branch): Likewise.
7338 (cortex_a15_load1): Likewise.
7339 (cortex_a15_load3): Likewise.
7340 (cortex_a15_store1): Likewise.
7341 (cortex_a15_store3): Likewise.
7342 (cortex_a15_call): Likewise.
7343 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7344 (cortex_a5_f2r): Likewise.
7345 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7347 (cortex_a53_f2r): Likewise.
7348 * config/arm/cortex-a7.md
7349 (cortex_a7_branch): Update for attribute change.
7350 (cortex_a7_call): Likewise.
7351 (cortex_a7_alu_imm): Likewise.
7352 (cortex_a7_alu_reg): Likewise.
7353 (cortex_a7_alu_shift): Likewise.
7354 (cortex_a7_mul): Likewise.
7355 (cortex_a7_load1): Likewise.
7356 (cortex_a7_store1): Likewise.
7357 (cortex_a7_load2): Likewise.
7358 (cortex_a7_store2): Likewise.
7359 (cortex_a7_load3): Likewise.
7360 (cortex_a7_store3): Likewise.
7361 (cortex_a7_load4): Likewise.
7362 (cortex_a7_store4): Likewise.
7363 (cortex_a7_fpalu): Likewise.
7364 (cortex_a7_fconst): Likewise.
7365 (cortex_a7_fpmuls): Likewise.
7366 (cortex_a7_neon_mul): Likewise.
7367 (cortex_a7_fpmacs): Likewise.
7368 (cortex_a7_neon_mla: Likewise.
7369 (cortex_a7_fpmuld: Likewise.
7370 (cortex_a7_fpmacd: Likewise.
7371 (cortex_a7_fpfmad: Likewise.
7372 (cortex_a7_fdivs: Likewise.
7373 (cortex_a7_fdivd: Likewise.
7374 (cortex_a7_r2f: Likewise.
7375 (cortex_a7_f2r: Likewise.
7376 (cortex_a7_f_flags: Likewise.
7377 (cortex_a7_f_loads: Likewise.
7378 (cortex_a7_f_loadd: Likewise.
7379 (cortex_a7_f_stores: Likewise.
7380 (cortex_a7_f_stored: Likewise.
7381 (cortex_a7_neon): Likewise.
7382 * config/arm/cortex-a8-neon.md
7383 (cortex_a8_neon_mrc): Update for attribute change.
7384 (cortex_a8_neon_mrrc): Likewise.
7385 (cortex_a8_neon_int_1): Likewise.
7386 (cortex_a8_neon_int_2): Likewise.
7387 (cortex_a8_neon_int_3): Likewise.
7388 (cortex_a8_neon_int_4): Likewise.
7389 (cortex_a8_neon_int_5): Likewise.
7390 (cortex_a8_neon_vqneg_vqabs): Likewise.
7391 (cortex_a8_neon_vmov): Likewise.
7392 (cortex_a8_neon_vaba): Likewise.
7393 (cortex_a8_neon_vaba_qqq): Likewise.
7394 (cortex_a8_neon_vsma): Likewise.
7395 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7396 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7397 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7398 long_scalar): Likewise.
7399 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7400 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7401 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7402 long_scalar_qdd_64_32_long): Likewise.
7403 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7404 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7405 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7406 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7407 (cortex_a8_neon_shift_1): Likewise.
7408 (cortex_a8_neon_shift_2): Likewise.
7409 (cortex_a8_neon_shift_3): Likewise.
7410 (cortex_a8_neon_vshl_ddd): Likewise.
7411 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7412 (cortex_a8_neon_vsra_vrsra): Likewise.
7413 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7414 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7415 (cortex_a8_neon_fp_vsum): Likewise.
7416 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7417 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7418 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7419 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7420 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7421 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7422 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7423 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7424 (cortex_a8_neon_bp_simple): Likewise.
7425 (cortex_a8_neon_bp_2cycle): Likewise.
7426 (cortex_a8_neon_bp_3cycle): Likewise.
7427 (cortex_a8_neon_ldr): Likewise.
7428 (cortex_a8_neon_str): Likewise.
7429 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7430 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7431 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7432 (cortex_a8_neon_vld2_4_regs): Likewise.
7433 (cortex_a8_neon_vld3_vld4): Likewise.
7434 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7435 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7436 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7437 (cortex_a8_neon_vst3_vst4): Likewise.
7438 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7439 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7440 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7441 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7442 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7443 (cortex_a8_neon_mcr): Likewise.
7444 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7445 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7446 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7448 (ca9_neon_mrrc): Likewise.
7449 (cortex_a9_neon_int_1): Likewise.
7450 (cortex_a9_neon_int_2): Likewise.
7451 (cortex_a9_neon_int_3): Likewise.
7452 (cortex_a9_neon_int_4): Likewise.
7453 (cortex_a9_neon_int_5): Likewise.
7454 (cortex_a9_neon_vqneg_vqabs): Likewise.
7455 (cortex_a9_neon_vmov): Likewise.
7456 (cortex_a9_neon_vaba): Likewise.
7457 (cortex_a9_neon_vaba_qqq): Likewise.
7458 (cortex_a9_neon_vsma): Likewise.
7459 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7460 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7461 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7462 long_scalar): Likewise.
7463 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7464 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7465 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7466 long_scalar_qdd_64_32_long): Likewise.
7467 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7468 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7469 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7470 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7471 (cortex_a9_neon_shift_1): Likewise.
7472 (cortex_a9_neon_shift_2): Likewise.
7473 (cortex_a9_neon_shift_3): Likewise.
7474 (cortex_a9_neon_vshl_ddd): Likewise.
7475 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7476 (cortex_a9_neon_vsra_vrsra): Likewise.
7477 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7478 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7479 (cortex_a9_neon_fp_vsum): Likewise.
7480 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7481 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7482 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7483 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7484 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7485 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7486 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7487 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7488 (cortex_a9_neon_bp_simple): Likewise.
7489 (cortex_a9_neon_bp_2cycle): Likewise.
7490 (cortex_a9_neon_bp_3cycle): Likewise.
7491 (cortex_a9_neon_ldr): Likewise.
7492 (cortex_a9_neon_str): Likewise.
7493 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7494 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7495 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7496 (cortex_a9_neon_vld2_4_regs): Likewise.
7497 (cortex_a9_neon_vld3_vld4): Likewise.
7498 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7499 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7500 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7501 (cortex_a9_neon_vst3_vst4): Likewise.
7502 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7503 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7504 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7505 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7506 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7507 (cortex_a9_neon_mcr): Likewise.
7508 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7509 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7510 (cortex_a9_fps): Likewise.
7511 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7513 (cortex_m4_fmuls): Likewise.
7514 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7516 (cortex_r4_mrc): Likewise.
7517 * config/arm/iterators.md: Update comment referring to neon_type.
7518 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7519 (iwmmxt_movsi_insn): Likewise.
7520 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7522 (pj4_core_to_vfp): Likewise.
7523 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7525 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7526 (vfp_fstore): Likewise.
7527 * doc/md.texi: Change references to neon_type to refer to type.
7529 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7531 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7533 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7535 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7538 2013-09-04 Easwaran Raman <eraman@google.com>
7541 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7542 (build_and_add_sum): Use it.
7543 (appears_later_in_bb): Simplify code.
7545 2013-09-04 Teresa Johnson <tejohnson@google.com>
7547 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7549 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7551 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7553 2013-09-04 Jan Hubicka <jh@suse.cz>
7555 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7556 * ipa-devirt.c: Include diganostic.h
7557 (odr_type_d): Add types and types_set.
7558 (hash_type_name): Work for types with vtables during LTO.
7559 (odr_hasher::remove): Fix comment; destroy types_set.
7560 (add_type_duplicate): New function,
7561 (get_odr_type): Use it.
7562 (dump_type_inheritance_graph): Dump type duplicates.
7563 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7565 * tree.c (types_same_for_odr): Give exact answers on types with
7568 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7570 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7571 explaining their differences.
7573 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7575 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7577 2013-09-03 Jeff Law <law@redhat.com>
7579 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7580 when not threading through a joiner block. Pass joiner/no joiner
7581 state to register_jump_thread.
7582 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7583 state from argument rather than implying on path length.
7584 Dump the entire jump thread path into debugging dump.
7585 * tree-flow.h (register_jump_thread): Update prototype.
7587 2013-08-29 Xinliang David Li <davidxl@google.com>
7589 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7590 Remove a trivial gcc_assert.
7592 2013-08-29 Xinliang David Li <davidxl@google.com>
7594 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7595 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7596 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7597 Delay base decl alignment adjustment.
7598 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7599 * tree-vectorizer.h: New data structure.
7600 (set_dr_misalignment): New function.
7601 (dr_misalignment): Ditto.
7602 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7603 (vectorizable_load): Ditto.
7604 (ensure_base_align): New function.
7605 (vectorize_loops): Add dbg_cnt support.
7606 (execute_vect_slp): Ditto.
7607 * dbgcnt.def: New debug counter.
7608 * Makefile: New dependency.
7610 2013-09-03 Meador Inge <meadori@codesourcery.com>
7614 2013-08-30 Meador Inge <meadori@codesourcery.com>
7616 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7618 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7620 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7621 function descriptor.
7623 2013-09-03 Richard Biener <rguenther@suse.de>
7625 * tree-affine.c (add_elt_to_tree): Fix association issue,
7626 avoid useless converts and make sure to always return a
7627 properly typed result.
7629 2013-09-03 Richard Biener <rguenther@suse.de>
7632 * fold-const.c (negate_expr_p): Fix division case.
7633 (negate_expr): Likewise.
7635 2013-09-03 Richard Biener <rguenther@suse.de>
7638 * tree-streamer-out.c: Include tm.h.
7639 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7641 2013-09-03 Jan Hubicka <jh@suse.cz>
7643 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7645 2013-09-03 Alan Modra <amodra@gmail.com>
7647 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7648 multilibs to big-endian target and vice versa.
7649 * config/rs6000/t-linux64: Use := assignment on all vars.
7650 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7651 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7652 * config/rs6000/t-linux64le: New file.
7653 * config/rs6000/t-linux64bele: New file.
7654 * config/rs6000/t-linux64lebe: New file.
7656 2013-09-02 Jan Hubicka <jh@suse.cz>
7658 * ipa-inline-transform.c (inline_transform): Do not
7659 optimize_inline_calls when not optimizing.
7661 2013-09-02 Jan Hubicka <jh@suse.cz>
7663 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7664 duplicated nodes for assembler names.
7665 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7666 hard registers from assembler name hash.
7668 2013-09-02 Jan Hubicka <jh@suse.cz>
7670 * ipa-split.c (execute_split_functions): Split externally visible
7671 functions called once.
7673 2013-09-02 Martin Jambor <mjambor@suse.cz>
7676 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7677 linked list. When finding the correct duplicate, also consider also
7678 the caller in additon to its inlined_to node.
7680 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7682 * config/aarch64/aarch64-simd-builtins.def
7683 (dup_lane_scalar): Remove.
7684 * config/aarch64/aarch64-simd.md
7685 (aarch64_simd_dup): Add 'w->w' alternative.
7686 (aarch64_dup_lane<mode>): Allow for VALL.
7687 (aarch64_dup_lane_scalar<mode>): Remove.
7688 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7689 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7690 (aarch64_get_lane_unsigned<mode>): Likewise.
7691 (aarch64_get_lane<mode>): Likewise.
7692 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7693 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7694 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7695 (VCON): Change container of V2SF.
7696 (vswap_width_name): Likewise.
7697 * config/aarch64/arm_neon.h
7698 (__aarch64_vdup_lane_any): New.
7699 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7700 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7701 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7703 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7706 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7707 the source or the destination is a single hard register.
7709 2013-09-02 Richard Biener <rguenther@suse.de>
7712 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7715 2013-09-02 Richard Biener <rguenther@suse.de>
7717 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7718 arithmetic to sizetype.
7720 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7722 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7723 Find auto-increment use both before and after candidate.
7725 2013-09-02 Marek Polacek <polacek@redhat.com>
7727 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7729 2013-09-01 Jan Hubicka <jh@suse.cz>
7731 * Makefile.in: Add ipa-profile.o
7732 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7733 * cgraph.c (struct cgraph_propagate_frequency_data,
7734 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7735 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7736 * cgraph.h (cgraph_propagate_frequency): Remove.
7737 * ipa-inline-analysis.c: Include ipa-utils.h;
7738 drop duplicated cfgloop.h.
7739 (inline_update_callee_summaries): Update.
7740 * ipa-profile.c: New file.
7741 * ipa-utils.h (ipa_propagate_frequency): Declare.
7742 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7743 data-streamer.h, value-prof.h.
7744 (symtab_remove_unreachable_nodes): Update profile.
7745 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7746 account_time_size, cmp_counts, dump_histogram,
7747 ipa_profile_generate_summary, ipa_profile_write_summary,
7748 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7749 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7750 Move to ipa-profile.c.
7752 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7754 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7756 2013-09-01 Jan Hubicka <jh@suse.cz>
7758 * common.opt (fdevirtualize-speculatively): New function.
7759 * invoke.texi (fdevirtualize-speculatively): Document.
7760 * ipa-devirt.c: Include ipa-inline.h
7761 (likely_target_p): New function.
7762 (ipa_devirt): New function.
7763 (gate_ipa_devirt): New function.
7764 (pass_data_ipa_devirt): New static var.
7765 (pass_ipa_devirt): Likewise.
7766 (make_pass_ipa_devirt): New function.
7767 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7768 (common_handle_option): Disable devirtualization when
7769 value range profiling is available.
7770 * passes.def (pass_ipa_devirt): Add.
7771 * timever.def (TV_IPA_DEVIRT): New timevar.
7772 * tree-pass.h (make_pass_ipa_devirt):
7774 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7776 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7777 include sanitize(undefined).
7779 2013-08-31 Diego Novillo <dnovillo@google.com>
7781 * Makefile.in (TREE_CORE_H): Define.
7783 (GTFILES): Add tree-core.h.
7784 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7786 * tree-core.h: New file.
7787 Move all data structures, enum, typedefs, global
7788 declarations and constants from ...
7791 2013-08-31 Jan Hubicka <jh@suse.cz>
7793 * bulitins.c (expand_builtin): Do not early exit for gcov
7794 instrumented functions.
7796 2013-08-31 Marek Polacek <polacek@redhat.com>
7798 * ubsan.c: Include tm_p.h.
7800 2013-08-31 Jan Hubicka <jh@suse.cz>
7802 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7805 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7806 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7807 * tree-cfg.c (verify_gimple_label): ... here.
7808 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7809 (ipa_merge_profiles): New function.
7810 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7811 (lto_input_function_body): Likewise.
7812 * ipa-utils.h (ipa_merge_profiles): Declare.
7813 * lto-streamer.h (lto_input_function_body): Update prototype.
7814 (emit_label_in_global_context_p): Remove.
7815 * lto-symtab.c: Include ipa-utils.h
7816 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7818 2013-08-31 Jan Hubicka <jh@suse.cz>
7820 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7822 2013-08-31 Jan Hubicka <jh@suse.cz>
7824 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7826 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7828 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7829 "cmp" RTX before signed_comparison_operator check to account
7832 2013-08-30 Jan Hubicka <jh@suse.cz>
7834 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7835 (detect_type_change_1): Rename to ...
7836 (detect_type_change): ... this one; early return on non-polymorphic
7838 (detect_type_change_ssa): Add comp_type parameter; update
7839 use of detect_type_change.
7840 (compute_complex_assign_jump_func): Add param_type parameter;
7841 update use of detect_type_change_ssa.
7842 (compute_complex_ancestor_jump_func): Likewise.
7843 (ipa_get_callee_param_type): New function.
7844 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7845 update calls to the jump function computation functions.
7847 2013-08-30 Teresa Johnson <tejohnson@google.com>
7848 Steven Bosscher <steven@gcc.gnu.org>
7850 * cfgrtl.c (fixup_new_cold_bb): New routine.
7851 (commit_edge_insertions): Invoke fixup_partitions.
7852 (find_partition_fixes): New routine.
7853 (fixup_partitions): Ditto.
7854 (verify_hot_cold_block_grouping): Update comments.
7855 (rtl_verify_edges): Invoke find_partition_fixes.
7856 (rtl_verify_bb_pointers): Update comments.
7857 (rtl_verify_bb_layout): Ditto.
7858 * basic-block.h (probably_never_executed_edge_p): Declare.
7859 (fixup_partitions): Ditto.
7860 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7861 * bb-reorder.c (sanitize_hot_paths): New function.
7862 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7864 * predict.c (probably_never_executed_edge_p): New routine.
7865 * cfg.c (check_bb_profile): Add partition insanity warnings.
7867 2013-08-30 Meador Inge <meadori@codesourcery.com>
7869 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7871 2013-08-30 Marek Polacek <polacek@redhat.com>
7873 * Makefile.in (ubsan.o): Add.
7874 (c-family/c-ubsan.o): Add.
7875 (builtins.o): Add ubsan.h dependency.
7876 * ubsan.h: New file.
7877 * ubsan.c: New file.
7878 * common.opt: Add -fsanitize=undefined option.
7879 (flag_sanitize): Add variable.
7880 (fsanitize=): Add option. Add Driver.
7881 (fsanitize=thread): Remove option.
7882 (fsanitize=address): Likewise.
7883 (static-libubsan): New option.
7884 * doc/invoke.texi: Document the new flag and -static-libubsan.
7885 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7886 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7887 * builtin-attrs.def (ATTR_COLD): Define.
7888 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7889 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7890 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7891 * flag-types.h (sanitize_code): New enum.
7892 * opts.c (common_handle_option): Parse command line arguments
7893 of -fsanitize=. Add -fsanitize=unreachable option.
7894 * varasm.c (get_variable_section): Adjust.
7895 (assemble_noswitch_variable): Likewise.
7896 (assemble_variable): Likewise.
7897 (output_constant_def_contents): Likewise.
7898 (categorize_decl_for_section): Likewise.
7899 (place_block_symbol): Likewise.
7900 (output_object_block): Likewise.
7901 * builtins.def: Likewise.
7902 * toplev.c (compile_file): Likewise.
7903 (process_options): Likewise.
7904 * cppbuiltin.c: Likewise.
7905 * tsan.c (tsan_pass): Likewise.
7906 (tsan_gate): Likewise.
7907 (tsan_gate_O0): Likewise.
7908 * cfgexpand.c (partition_stack_vars): Likewise.
7909 (expand_stack_vars): Likewise.
7910 (defer_stack_allocation): Likewise.
7911 (expand_used_vars): Likewise.
7912 * cfgcleanup.c (old_insns_match_p): Likewise.
7913 * asan.c (asan_finish_file): Likewise.
7914 (asan_instrument): Likewise.
7915 (gate_asan): Likewise.
7916 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7917 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7918 (asan_global_struct): Use pointer_sized_int_node instead
7919 calling build_nonstandard_integer_type.
7920 (initialize_sanitizer_builtins): Likewise.
7921 (asan_finish_file): Likewise.
7922 * gcc.c: Document %{%:function(args):X}.
7923 (static_spec_functions): Add sanitize.
7924 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7925 store funcval != NULL there.
7926 (do_spec_1): Adjust handle_spec_function caller.
7927 (handle_braces): Allow %:function(args) as condition.
7928 (sanitize_spec_function): New function.
7929 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7930 (LIBUBSAN_SPEC): Likewise.
7931 (LIBUBSAN_EARLY_SPEC): Likewise.
7932 (SANITIZER_SPEC): Handle libubsan.
7933 (SANITIZER_EARLY_SPEC): Likewise.
7934 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7935 instead of fsanitize=address.
7936 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7937 instead of fsanitize=address*.
7938 * builtins.c: Include ubsan.h.
7939 (fold_builtin_0): Instrument __builtin_unreachable.
7940 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7941 instead of flag_asan.
7942 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7943 (pointer_sized_int_node): Define.
7944 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7946 2013-08-30 Mike Stump <mikestump@comcast.net>
7948 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7951 2013-08-29 Jan Hubicka <jh@suse.cz>
7953 * cgraph.c (cgraph_function_body_availability): Handle weakref
7955 * passes.def: Remove pass_fixup_cfg.
7956 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7957 track when we need to remove functions.
7958 (gate_ipa_inline): Execute inlining always; add comment why.
7959 (pass_data_ipa_inline): Remove TODO_remove_functions.
7960 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7961 do not produce summaries.
7962 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7963 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7964 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7965 constant pool and vtable.
7967 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7969 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7970 New arm_neon.h's internal macros to specify 64-bit constants.
7971 Avoid using stdint.h's macros.
7973 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7975 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7976 before applying REG_P.
7978 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7980 PR tree-optimization/58277
7981 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7982 after seeing too many stmts with vdef in between dombb and current
7983 bb, invalidate everything.
7985 2013-08-30 Richard Biener <rguenther@suse.de>
7987 * fold-const.c (fold_single_bit_test): Fix overflow test.
7989 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7991 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7992 and which can live in a register, always retrieve the value on entry.
7993 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7994 passed by invisible reference specially.
7995 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7996 (vt_add_function_parameter): Correctly deal with a parameter passed by
7997 invisible reference.
7999 2013-08-30 Jan Hubicka <jh@suse.cz>
8001 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8003 2013-08-30 Richard Biener <rguenther@suse.de>
8005 PR tree-optimization/58228
8006 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8007 allow invariant loads in nested loop vectorization.
8009 2013-08-30 Richard Biener <rguenther@suse.de>
8011 PR tree-optimization/58223
8012 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8013 (has_anti_or_output_dependence): ... this and adjust to also
8014 look for output dependences.
8015 (mark_nodes_having_upstream_mem_writes): Adjust.
8016 (rdg_flag_uses): Likewise.
8018 2013-08-30 Richard Biener <rguenther@suse.de>
8020 PR tree-optimization/58010
8021 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8022 assert that we have a loop-closed PHI.
8024 2013-08-29 Jan Hubicka <jh@suse.cz>
8026 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8027 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8028 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8029 (lto_free_function_in_decl_state_for_node): New function.
8031 2013-08-29 Xinliang David Li <davidxl@google.com>
8033 * loop-unroll.c (report_unroll_peel): Minor message change.
8034 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8035 Emit alignment peeling message with default -fopt-info.
8036 (vect_loop_versioning): Emit loop version info message.
8037 * tree-vectorizer.c (vectorize_loops): Minor message change.
8038 (execute_vect_slp): Ditto.
8040 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8042 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8043 of the clone from the DECL_NAME of the original function.
8045 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8047 * passes.c (register_pass): Add overload.
8048 * tree-pass.h (register_pass): Forward declare it. Add comment.
8050 2013-08-29 Jan Hubicka <jh@suse.cz>
8052 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8053 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8054 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8055 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8056 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8057 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8058 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8059 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8061 2013-08-29 Teresa Johnson <tejohnson@google.com>
8063 * dumpfile.c (dump_loc): Output column number.
8064 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8065 * doc/invoke.texi: Document optall -fopt-info flag.
8066 * profile.c (read_profile_edge_counts): Use new dump framework.
8067 (compute_branch_probabilities): Ditto.
8068 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8069 when pass not in any opt group.
8070 * pass_manager.h (pass_manager::get_pass_profile): New method.
8071 * value-prof.c (check_counter): Use new dump framework.
8072 (check_ic_target): Ditto.
8073 * coverage.c (get_coverage_counts): Ditto.
8074 (coverage_init): Setup new dump framework.
8076 2013-08-29 Richard Biener <rguenther@suse.de>
8078 PR tree-optimization/58246
8079 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8080 handle the dominance check inside a basic-block.
8082 2013-08-29 Richard Biener <rguenther@suse.de>
8085 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8086 of default defs that appear in abnormal PHI nodes.
8088 2013-08-29 Richard Biener <rguenther@suse.de>
8090 PR tree-optimization/57685
8091 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8092 single-use operands to avoid exponential complexity.
8094 2013-08-28 Dehao Chen <dehao@google.com>
8096 * ipa-inline.c (edge_badness): Fix integer underflow.
8098 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8100 * gtm-builtins.def (_ITM_free): Declare leaf.
8102 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8105 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8106 (*tls_local_dynamic_base_64_largepic): Likewise.
8107 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8108 Remove predicate from call operand.
8109 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8110 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8112 2013-08-28 Jeff Law <law@redhat.com>
8114 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8115 checks for the number of predecessors and successors allowed.
8116 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8117 which require copying a joiner block if there is a request which
8118 is a subpath that requires no joiner block copying.
8120 2013-08-28 Jan Hubicka <jh@suse.cz>
8122 * lto-streamer-out.c (DFS_write_tree_body): Drop
8123 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8124 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8125 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8126 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8127 Do not read DECL_ERROR_ISSUED.
8128 (unpack_ts_decl_with_vis_value_fields): Do not read
8130 (lto_input_ts_binfo_tree_pointers): Do not read
8131 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8132 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8133 write DECL_ERROR_ISSUED..
8134 (pack_ts_decl_with_vis_value_fields): Do not write
8136 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8137 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8138 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8139 * tree.h (tree_decl_common): Update comment.
8140 (DECL_ERROR_ISSUED): Remove.
8142 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8145 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8147 2013-08-28 Jan Hubicka <jh@suse.cz>
8149 * builtins.def (free): Declare leaf.
8151 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8154 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8155 * configure: Regenerate.
8157 2013-08-27 Martin Jambor <mjambor@suse.cz>
8159 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8160 (ipa_ancestor_jf_data): Likewise.
8161 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8162 (ipa_get_jf_pass_through_type_preserved): New function.
8163 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8164 (ipa_get_jf_ancestor_type_preserved): New function.
8165 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8166 (ipa_get_jf_ancestor_result): Likewise.
8167 (propagate_vals_accross_pass_through): Use
8168 ipa_get_jf_pass_through_result to do all the value mappings.
8169 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8170 type_preserved flag.
8171 (ipa_set_jf_cst_copy): New function.
8172 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8173 (ipa_set_jf_arith_pass_through): Likewise.
8174 (ipa_set_ancestor_jf): Likewise.
8175 (compute_complex_assign_jump_func): Set type_preserved instead of
8177 (ipa_compute_jump_functions_for_edge): Likewise.
8178 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8179 (update_jump_functions_after_inlining): Update type_preserved.
8180 Explicitely create jump functions when combining one with pass_through.
8181 (ipa_write_jump_function): Stream the type_preserved flags.
8182 (ipa_read_jump_function): Likewise.
8184 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8185 Aldy Hernandez <aldyh@redhat.com>
8187 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8188 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8189 * function.h (struct function): Add has_force_vect_loops and
8191 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8192 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8194 * gimple.def: Update CLAUSES comments.
8195 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8196 (gimple_build_omp_for): Add argument to prototype.
8197 (gimple_omp_for_kind): New.
8198 (gimple_omp_for_set_kind): New.
8199 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8200 GOVD_DATA_SHARE_CLASS.
8201 (enum omp_region_type): Add ORT_SIMD.
8202 (gimple_add_tmp_var): Handle ORT_SIMD.
8203 (gimplify_var_or_parm_decl): Same.
8204 (is_gimple_stmt): Same.
8205 (omp_firstprivatize_variable): Same.
8206 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8207 (omp_notice_variable): Handle ORT_SIMD.
8208 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8209 (omp_check_private): Handle ORT_SIMD.
8210 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8212 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8213 Handle OMP_CLAUSE_LASTPRIVATE.
8214 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8216 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8217 (gimplify_expr): Handle OMP_SIMD.
8218 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8219 (expand_GOMP_SIMD_VF): New.
8220 (expand_GOMP_SIMD_LAST_LANE): New.
8221 * internal-fn.def (GOMP_SIMD_LANE): New.
8222 (GOMP_SIMD_VF): New.
8223 (GOMP_SIMD_LAST_LANE): New.
8224 * omp-low.c: Include target.h.
8225 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8227 (check_omp_nesting_restrictions): Same.
8229 (lower_rec_simd_input_clauses): New.
8230 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8232 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8233 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8234 (expand_omp_build_assign): New.
8235 (expand_omp_for_init_counts): New.
8236 (expand_omp_for_init_vars): New.
8237 (extract_omp_for_update_vars): New.
8238 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8239 and rewrite accordingly.
8240 (expand_omp_simd): New.
8241 (expand_omp_for): Use expand_omp_simd.
8242 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8243 (lower_omp_for): Do not lower the body.
8244 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8246 * tree-flow.h (find_omp_clause): Remove prototype.
8247 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8248 forcing vectorization of the loop, or if flag_tree_vectorize.
8249 (gate_tree_if_conversion): Similarly.
8250 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8251 gimple_build_omp_for.
8252 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8253 * tree-parloops (create_parallel_loop): Pass kind argument to
8254 gimple_build_omp_for.
8255 * tree-pretty-print.c (dump_omp_clause): Add cases for
8256 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8257 OMP_CLAUSE__SIMDUID_.
8258 (dump_generic_node): Handle OMP_SIMD.
8259 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8260 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8261 unroll OMP_SIMD loops here.
8262 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8263 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8265 (vect_analyze_data_refs): Handle simd loops.
8266 * tree-vect-loop.c (vectorizable_live_operation): Handle
8268 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8269 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8270 (vectorizable_load): Same.
8271 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8272 (struct simduid_to_vf): New.
8273 (simduid_to_vf::hash): New.
8274 (simduid_to-vf::equal): New.
8275 (struct simd_array_to_simduid): New.
8276 (simd_array_to_simduid::hash): New.
8277 (simd_array_to_simduid::equal): New.
8278 (adjust_simduid_builtins): New.
8279 (struct note_simd_array_uses_struct): New.
8280 (note_simd_array_uses_cb): New.
8281 (note_simd_array_uses): New.
8282 (vectorize_loops): Handle simd hints and adjust simd builtins
8284 * tree-vectorizer.h (struct _stmt_vec_info): Add
8285 simd_lane_access_p field.
8286 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8287 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8288 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8289 (omp_clause_code_name): Same.
8290 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8291 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8292 * tree.def (OMP_SIMD): New entry.
8293 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8294 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8295 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8296 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8297 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8298 (OMP_CLAUSE_LINEAR_STEP): New.
8299 (OMP_CLAUSE_SAFELEN_EXPR): New.
8300 (OMP_CLAUSE__SIMDUID__DECL): New.
8301 (find_omp_clause): New prototype.
8303 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8305 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8306 Haswell processor detection.
8308 2013-08-27 Christian Widmer <shadow@umbrox.de>
8311 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8312 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8313 AVX2 capable processors.
8315 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8317 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8318 of vget_low_* with implementations in terms of other intrinsics.
8320 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8323 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8324 values to -1, 0 and 1.
8326 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8328 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8329 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8330 (<optab><mode>3_insn): Remove.
8331 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8332 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8333 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8334 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8335 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8336 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8337 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8338 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8339 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8340 (ror<mode>3_insn): Likewise.
8341 * config/aarch64/predicates.md (aarch64_simd_register): New.
8343 2013-08-27 Richard Biener <rguenther@suse.de>
8345 PR tree-optimization/57521
8346 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8347 one edge is non-critical.
8348 (find_phi_replacement_condition): Make sure to use a non-critical
8349 edge. Cleanup and remove old bug workarounds.
8350 (bb_postdominates_preds): Remove.
8351 (if_convertible_loop_p_1): Do not compute post-dominators.
8352 (combine_blocks): Do not free post-dominators.
8353 (main_tree_if_conversion): Likewise.
8354 (pass_data_if_conversion): Add TODO_verify_ssa.
8356 2013-08-27 DJ Delorie <dj@redhat.com>
8358 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8360 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8362 * function.c (assign_parm_find_data_types): Set passed_mode and
8363 nominal_mode to the TYPE_MODE of nominal_type for the built
8364 pointer type in case of the struct-pass-by-reference.
8366 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8368 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8369 (UINT16_TYPE): Change default to "unsigned int".
8371 * config/avr/avr.opt (mfract-convert-truncate): New option.
8372 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8373 is set, round negative fractional integers according to n1169
8374 when converting to integer types.
8376 2013-08-26 Jan Hubicka <jh@suse.cz>
8378 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8379 methods can not be called indirectly when their address is not taken.
8381 2013-08-26 Jan Hubicka <jh@suse.cz>
8383 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8386 2013-08-26 Jan Hubicka <jh@suse.cz>
8388 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8391 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8393 * reload.c (find_valid_class): Allow classes that do not include
8394 FIRST_PSEUDO_REGISTER - 1.
8396 2013-08-26 Jan Hubicka <jh@suse.cz>
8398 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8399 fix edge count/frequency when speculation failed; fix type check
8400 for the direct call.
8402 2013-08-26 Jan Hubicka <jh@suse.cz>
8404 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8406 2013-08-26 Jan Hubicka <jh@suse.cz>
8408 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8409 to be changed by edge redirection.
8411 2013-08-26 Jan Hubicka <jh@suse.cz>
8413 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8414 formating; add sanity check.
8415 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8416 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8417 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8418 (resolve_noninline_speculation): Update callee keys, too.
8420 2013-08-26 Jan Hubicka <jh@suse.cz>
8422 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8423 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8425 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8427 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8430 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8432 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8433 (x86_64_elf_select_section): Ditto.
8434 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8436 (ix86_pass_by_reference): Ditto.
8437 (output_set_got): Ditto.
8438 (ix86_unary_operator_ok): Ditto.
8439 (ix86_expand_builtin): Ditto.
8441 2013-08-23 Jan Hubicka <jh@suse.cz>
8443 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8445 2013-08-23 Jan Hubicka <jh@suse.cz>
8447 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8448 (tree_decl_with_vis): Add FINAL field.
8450 2013-08-23 Jeff Law <law@redhat.com>
8452 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8453 the debugging dump when the expression is fully redundant.
8455 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8457 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8458 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8459 * pretty-print.c (pp_formatted_text_data): Likewise.
8460 (pp_write_text_to_stream): Likewise.
8461 (pp_write_text_as_dot_label_to_stream): Likewise.
8462 (pp_append_r): Likewise.
8463 (pp_format): Likewise.
8464 (pp_flush): Likewise.
8465 (pp_clear_output_area): Likewise.
8466 (pp_append_text): Likewise.
8467 (pp_formatted_text): Likewise.
8468 (pp_remaining_character_count_for_line): Likewise.
8469 (pp_newline): Likewise.
8470 (pp_character): Likewise.
8471 (output_buffer::~output_buffer): Define.
8472 (pretty_printer::~pretty_printer): Destruct output buffer.
8473 * pretty-print.h (output_buffer::~output_buffer): Declare.
8474 (pretty_printer::~pretty_printer): Declare virtual.
8476 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8479 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8480 HOST_WIDE_INT_M1U): New macros.
8481 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8482 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8483 unsigned -1 for lshift.
8484 * cse.c (cse_insn): Likewise.
8485 * double-int.c (rshift_double, lshift_double): Likewise.
8486 * builtins.c (fold_builtin_bitop): Likewise.
8487 * combine.c (force_to_mode): Likewise.
8488 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8489 * simplify-rtx.c (simplify_const_unary_operation,
8490 simplify_const_binary_operation): Likewise.
8491 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8492 check_va_list_escapes): Likewise.
8493 * rtlanal.c (nonzero_bits1): Likewise.
8494 * expmed.c (expand_smod_pow2): Likewise.
8495 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8497 2013-08-23 Jan Hubicka <jh@suse.cz>
8499 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8500 as having address taken.
8502 2013-08-23 Jan Hubicka <jh@suse.cz>
8504 * ipa-utils.h (method_class_type): Declare.
8505 * ipa-devirt.c (method_class_type): Export.
8507 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8508 do not walk base classes of anonymous types.
8510 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8512 PR rtl-optimization/58220
8514 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8515 handle SEQUENCE insns properly.
8517 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8519 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8521 (pp_newline_and_indent): Likewise.
8522 (pp_separate_with): Likewise.
8523 * pretty-print.c (pp_newline_and_flush): Define.
8524 (pp_newline_and_indent): Likewise.
8525 (pp_separate_with): Likewise.
8527 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8530 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8531 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8533 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8535 * config/i386/predicates.md (ext_sse_reg_operand): New.
8536 * config/i386/i386.md (*movti_internal): Use
8537 predicate to determine if EVEX is needed.
8538 (*movsi_internal): Ditto.
8539 (*movdf_internal): Ditto.
8540 (*movsf_internal): Ditto.
8541 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8543 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8545 PR tree-optimization/58209
8546 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8547 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8548 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8549 emit POINTER_PLUS_EXPR.
8550 (create_tailcall_accumulator): For pointer result type accumulate in
8553 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8555 * configure.ac: Add backslashes missing from the last change.
8556 * configure: Regenerate.
8558 2013-08-22 Jan Hubicka <jh@susue.cz>
8560 * ipa.c (function_and_variable_visibility): First remember function
8561 was global and then make it local.
8563 2013-08-22 Julian Brown <julian@codesourcery.com>
8565 * configure.ac: Add aarch64 to list of arches which use "nop" in
8567 * configure: Regenerate.
8569 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8571 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8572 gnu_libc_has_function.
8573 * config/s390/tpf.h: Likewise.
8575 2013-08-22 Jan Hubicka <jh@susue.cz>
8577 * timevar.c (validate_phases): Add cast.
8579 2013-08-22 Jan Hubicka <jh@susue.cz>
8581 * timevar.c (validate_phases): Use size_t for memory.
8582 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8584 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8586 * pretty-print.h (output_buffer::output_buffer): Declare.
8587 (pretty_printer::pretty_printer): Likewise.
8588 (pp_construct): Remove.
8589 * pretty-print.c (output_buffer::output_buffer): Define.
8590 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8591 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8592 (print_gimple_expr): Likewise.
8593 (print_gimple_seq): Likewise.
8594 (gimple_dump_bb): Likewise.
8595 * sched-vis.c (dump_value_slim): Likewise.
8596 (dump_insn_slim): Likewise.
8597 (dump_rtl_slim): Likewise.
8598 (str_pattern_slim): Likewise.
8599 * tree-mudflap.c (mf_varname_tree): Likewise.
8600 * graph.c (print_graph_cfg): Likewise.
8601 (start_graph_dump): Likewise.
8602 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8604 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8605 pretty printer initialization.
8606 * coretypes.h (diagnostic_context): Remove superflous type alias
8608 (pretty_printer): Likewise. Declare directly as a class.
8609 (pretty_print_info): Remove declaration as class.
8610 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8611 and pp_clear_output_area.
8612 (asan_add_global): Likewise.
8614 2013-08-22 Jan Hubicka <jh@suse.cz>
8616 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8617 * ipa-utils.h (update_type_inheritance_graph): Declare.
8618 (possible_polymorphic_call_target_p): Declare.
8619 (possible_polymorphic_call_target_p): New.
8620 * ipa-devirt.c: Update toplevel comments.
8621 (cached_polymorphic_call_targets): Move up.
8622 (odr_type_d): Move ID down.
8623 (polymorphic_type_binfo_p): Update comment.
8624 (odr_hasher::remove): Likewise;
8625 (get_odr_type): Set anonymous_namespace.
8626 (dump_odr_type): Dump it.
8627 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8628 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8629 (record_binfo): Add comment.
8630 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8632 (devirt_node_removal_hook): Do not iCE when cache is freed.
8633 (possible_polymorphic_call_target_p): New predicate.
8634 (update_type_inheritance_graph): New function.
8636 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8637 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8638 Sergey Lega <sergey.s.lega@intel.com>
8639 Anna Tikhonova <anna.tikhonova@intel.com>
8640 Ilya Tocar <ilya.tocar@intel.com>
8641 Andrey Turetskiy <andrey.turetskiy@intel.com>
8642 Ilya Verbin <ilya.verbin@intel.com>
8643 Kirill Yukhin <kirill.yukhin@intel.com>
8644 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8646 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8647 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8648 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8649 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8650 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8651 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8652 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8653 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8654 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8655 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8656 OPT_mavx512pf, OPT_mavx512er cases.
8657 * config/i386/constraints.md (v): New constraint.
8658 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8659 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8660 (bit_AVX512CD): New.
8661 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8662 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8663 * config/i386/i386-c.c (ix86_target_macros_internal):
8664 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8666 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8667 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8668 * config/i386/i386.c (regclass_map, dbx_register_map)
8669 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8670 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8671 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8673 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8674 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8675 -mavx512pf options. Fix formatting.
8676 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8677 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8678 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8679 -mavx512cd, -mavx512pf options.
8680 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8681 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8682 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8684 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8685 (ix86_set_reg_reg_cost): Ditto.
8686 (x86_order_regs_for_local_alloc): Ditto.
8687 (MAX_VECT_LEN): Extend to 64-byte.
8688 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8689 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8690 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8691 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8692 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8693 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8694 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8695 (SSE_REG_MODE_P): Support new modes.
8696 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8697 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8698 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8699 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8700 (REG_CLASS_CONTENTS): Add new registers.
8701 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8702 (EXT_REX_SSE_REGNO_P): New.
8703 (HI_REGISTER_NAMES): Add new registers.
8704 * config/i386/i386.md: Define constants for new registers.
8705 (mode): Add new 512-bit modes.
8706 (prefix): Support evex prefix.
8707 (isa): Support avx512f, noavx512f, fma_avx512f.
8708 (ssemodesuffix): Add new 512-bit modes.
8710 (*movxi_internal_avx512f): Ditto.
8711 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8713 (*movsi_internal): Likewise.
8714 (*movdf_internal): Likewise.
8715 (*movsf_internal): Likewise.
8716 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8717 (<code><mode>3): Likewise.
8718 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8720 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8721 with the new constraint "v".
8722 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8724 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8725 with the new constraint "v".
8726 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8727 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8728 (*<plusminus_insn><mode>3): Likewise.
8729 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8730 (*mul<mode>3): Likewise.
8731 (<sse>_vmmul<mode>3): Likewise.
8732 (<sse>_div<mode>3): Likewise.
8733 (<sse>_vmdiv<mode>3): Likewise.
8734 (<sse>_sqrt<mode>2): Likewise.
8735 (<sse>_vmsqrt<mode>2): Likewise.
8736 (*<code><mode>3_finite): Likewise.
8737 (*<code><mode>3) <smaxmin>: Likewise.
8738 (<sse>_vm<code><mode>3): Likewise.
8739 (*<code><mode>3) <any_logic>: Likewise.
8740 (*fma_fmadd_<mode>): Likewise.
8741 (*fma_fmsub_<mode>): Likewise.
8742 (*fma_fnmadd_<mode>): Likewise.
8743 (*fma_fnmsub_<mode>): Likewise.
8744 (*fma_fmaddsub_<mode>): Likewise.
8745 (*fma_fmsubadd_<mode>): Likewise.
8746 (*fmai_fmadd_<mode>): Likewise.
8747 (*fmai_fmsub_<mode>): Likewise.
8748 (*fmai_fnmadd_<mode>): Likewise.
8749 (*fmai_fnmsub_<mode>): Likewise.
8750 (sse_cvtsi2ss): Likewise.
8751 (sse_cvtsi2ssq): Likewise.
8752 (sse_cvtss2si): Likewise.
8753 (sse_cvtss2si_2): Likewise.
8754 (sse_cvtss2siq): Likewise.
8755 (sse_cvtss2siq_2): Likewise.
8756 (sse_cvttss2si): Likewise.
8757 (sse_cvtss2siq_2): Likewise.
8758 (float<sseintvecmodelower><mode>2): Likewise.
8759 (sse2_cvtsd2si_2): Likewise.
8760 (sse2_cvtsd2siq_2): Likewise.
8761 (*<plusminus_insn><mode>3): Likewise.
8762 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8763 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8764 (ashr<mode>3): Likewise.
8765 (<shift_insn><mode>3): Likewise.
8766 (avx2_<code><mode>3): Likewise.
8767 (*avx2_<code><mode>3): Likewise.
8768 (*andnot<mode>3): Likewise.
8769 (*<code><mode>3) <any_logic>: Likewise.
8770 (abs<mode>2): Likewise.
8771 (avx2_permvar<mode>): Likewise.
8772 (avx2_perm<mode>_1): Likewise.
8773 (*avx_vpermilp<mode>): Likewise.
8774 (avx_vpermilvar<mode>3): Likewise.
8775 (avx2_ashrv<mode>): Likewise.
8776 (avx2_<shift_insn>v<mode>): Likewise.
8777 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8779 * doc/rtl.texi: Document XImode.
8781 2013-08-21 Jeff Law <law@redhat.com>
8783 * tree-flow.h (register_jump_thread): Pass vector of edges
8784 instead of each important edge.
8785 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8786 thread path into a vector and pass that to register_jump_thread.
8787 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8788 passed in edge vector to the current 3-edge form.
8791 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8793 * dce.c (fini_dce): Call df_analyze again just in case
8794 delete_unmarked_insns removed anything.
8796 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8798 * reload.h (struct reg_equivs): Rename to ..
8799 (struct reg_equivs_s): .. this.
8801 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8803 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8805 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8807 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8809 2013-08-21 Jeff Law <law@redhat.com>
8811 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8812 simplify assignments too. If the RHS collapses to a singleton
8813 range, then return the value for the range.
8815 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8817 * config/i386/sse.md (V16): Rename to...
8819 (mov<mode>): Update iterator name.
8820 (*mov<mode>_internal): Ditto.
8821 (push<mode>1): Ditto.
8822 (movmisalign<mode>): Ditto.
8824 2013-08-20 Jan Hubicka <jh@suse.cz>
8827 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8828 entry for direct edges.
8829 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8831 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8833 Revert my last two changes, r201865 and r201864:
8836 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8838 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8839 instances can own GC refs.
8841 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8842 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8843 (gcc::context::gt_pch_nx): Likewise.
8844 (gcc::context::gt_pch_nx): Likewise.
8845 * ggc.h (gt_ggc_mx <T>): New.
8846 (gt_pch_nx_with_op <T>): New.
8847 (gt_pch_nx <T>): New.
8848 * passes.c (opt_pass::gt_ggc_mx): New.
8849 (opt_pass::gt_pch_nx): New.
8850 (opt_pass::gt_pch_nx_with_op): New.
8851 (pass_manager::gt_ggc_mx): New.
8852 (pass_manager::gt_pch_nx): New.
8853 (pass_manager::gt_pch_nx_with_op): New.
8854 (pass_manager::operator new): Use
8855 ggc_internal_cleared_alloc_stat rather than xcalloc.
8856 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8857 (pass_manager::gt_ggc_mx): New.
8858 (pass_manager::gt_pch_nx): New.
8859 (pass_manager::gt_pch_nx_with_op): New.
8860 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8861 (opt_pass::operator new): New.
8862 (opt_pass::gt_ggc_mx): New.
8863 (opt_pass::gt_pch_nx): New.
8864 (opt_pass::gt_pch_nx_with_op): New.
8867 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8869 * Makefile.in (GTFILES): Add context.h.
8870 * context.c (gcc::context::operator new): New.
8871 (gcc::context::gt_ggc_mx): New.
8872 (gcc::context::gt_pch_nx): New.
8873 (gcc::context::gt_pch_nx): New.
8874 * context.h (gcc::context): Add GTY((user)) marking.
8875 (gcc::context::operator new): New.
8876 (gcc::context::gt_ggc_mx): New.
8877 (gcc::context::gt_pch_nx): New.
8878 (gcc::context::gt_pch_nx): New.
8879 (g): Add GTY marking.
8880 (gt_ggc_mx (gcc::context *)): New.
8881 (gt_pch_nx (gcc::context *)): New.
8882 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8883 void *cookie)): New.
8884 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8886 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8888 * dce.c (fini_dce): Call df_analyze again just in case
8889 delete_unmarked_insns removed anything.
8891 2013-08-20 Teresa Johnson <tejohnson@google.com>
8893 PR rtl-optimizations/57451
8894 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8895 from crossing split section boundaries.
8897 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8899 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8900 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8901 with MULTLIB_DEFAULTS.
8903 2013-08-20 Nick Clifton <nickc@redhat.com>
8905 * target.def (narrow_volatile_bitfield): Note that the default
8906 value is false, not !TARGET_STRICT_ALIGN.
8907 * doc/tm.texi: Regenerate.
8909 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8911 Fix LIB_SPEC for systems without libpthread.
8913 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8914 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8916 * config/i386/linux-common.h: Likewise.
8917 * config/mips/linux-common.h: Likewise.
8919 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8921 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8924 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8926 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8927 instances can own GC refs.
8929 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8930 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8931 (gcc::context::gt_pch_nx): Likewise.
8932 (gcc::context::gt_pch_nx): Likewise.
8933 * ggc.h (gt_ggc_mx <T>): New.
8934 (gt_pch_nx_with_op <T>): New.
8935 (gt_pch_nx <T>): New.
8936 * passes.c (opt_pass::gt_ggc_mx): New.
8937 (opt_pass::gt_pch_nx): New.
8938 (opt_pass::gt_pch_nx_with_op): New.
8939 (pass_manager::gt_ggc_mx): New.
8940 (pass_manager::gt_pch_nx): New.
8941 (pass_manager::gt_pch_nx_with_op): New.
8942 (pass_manager::operator new): Use
8943 ggc_internal_cleared_alloc_stat rather than xcalloc.
8944 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8945 (pass_manager::gt_ggc_mx): New.
8946 (pass_manager::gt_pch_nx): New.
8947 (pass_manager::gt_pch_nx_with_op): New.
8948 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8949 (opt_pass::operator new): New.
8950 (opt_pass::gt_ggc_mx): New.
8951 (opt_pass::gt_pch_nx): New.
8952 (opt_pass::gt_pch_nx_with_op): New.
8954 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8956 * Makefile.in (GTFILES): Add context.h.
8957 * context.c (gcc::context::operator new): New.
8958 (gcc::context::gt_ggc_mx): New.
8959 (gcc::context::gt_pch_nx): New.
8960 (gcc::context::gt_pch_nx): New.
8961 * context.h (gcc::context): Add GTY((user)) marking.
8962 (gcc::context::operator new): New.
8963 (gcc::context::gt_ggc_mx): New.
8964 (gcc::context::gt_pch_nx): New.
8965 (gcc::context::gt_pch_nx): New.
8966 (g): Add GTY marking.
8967 (gt_ggc_mx (gcc::context *)): New.
8968 (gt_pch_nx (gcc::context *)): New.
8969 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8970 void *cookie)): New.
8971 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8973 2013-08-20 Alan Modra <amodra@gmail.com>
8976 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8977 (rs6000_emit_epilogue): Likewise.
8979 2013-08-19 Dehao Chen <dehao@google.com>
8981 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8983 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8984 Jakub Jelinek <jakub@redhat.com>
8986 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8987 (BUILT_IN_FABSD64): Likewise.
8988 (BUILT_IN_FABSD128): Likewise.
8989 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8990 (fold_builtin_1): Likewise.
8991 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8992 destination and source operands.
8993 (*abstd2_fpr): Likewise.
8994 (*nabstd2_fpr): Likewise.
8996 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8998 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9001 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9003 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9006 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9008 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9009 * builtins.c (default_libc_has_function): New.
9010 (gnu_libc_has_function): Ditto.
9011 (no_c99_libc_has_function): Ditto.
9012 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9013 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9014 (fold_builtin_sincos): Likewise.
9015 (fold_builtin_cexp): Likewise.
9016 * builtins.def (DEF_C94_BUILTIN): Likewise.
9017 (DEF_C99_BUILTIN): Likewise.
9018 (DEF_C99_C90RES_BUILTIN): Likewise.
9019 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9020 definitions to using this define.
9021 * config/darwin-protos.h (darwin_libc_has_function): New.
9022 * config/darwin.c (darwin_libc_has_function): Ditto.
9023 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9024 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9025 * config/darwin.h: Ditto.
9026 * config/elfos.h: Ditto.
9027 * config/freebsd.h: Ditto.
9028 * config/i386/cygming.h: Ditto.
9029 * config/i386/djgpp.h: Ditto.
9030 * config/i386/i386-interix.h: Ditto.
9031 * config/microblaze/microblaze.h: Ditto.
9032 * config/mmix/mmix.h: Ditto.
9033 * config/gnu-user.h: Ditto.
9034 * config/ia64/hpux.h: Ditto.
9035 * config/pa/pa-hpux.h: Ditto.
9036 * config/pdp11/pdp11.h: Ditto.
9037 * config/picochip/picochip.h: Ditto.
9038 * config/linux.h: Ditto.
9039 * config/netbsd.h: Ditto.
9040 * config/openbsd.h: Ditto.
9041 * config/rs6000/aix43.h: Ditto.
9042 * config/rs6000/aix51.h: Ditto.
9043 * config/rs6000/aix52.h: Ditto.
9044 * config/rs6000/aix53.h: Ditto.
9045 * config/rs6000/aix61.h: Ditto.
9046 * config/rs6000/darwin.h: Ditto.
9047 * config/rs6000/linux.h: Ditto.
9048 * config/rs6000/linux64.h: Ditto.
9049 * config/s390/tpf.h: Ditto.
9050 * config/sol2-10.h: Ditto.
9051 * config/sol2.h: Ditto.
9052 * config/vms/vms.h: Ditto.
9053 * config/vxworks.h: Ditto.
9054 * config/linux-android.c (linux_android_libc_has_function):
9055 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9056 * config/linux-protos.h (linux_android_libc_has_function):
9058 * config/i386/i386.c (ix86_libc_has_function): New.
9059 * config/i386/i386-protos.h
9060 (ix86_libc_has_function): New declaration.
9061 * config/i386/i386.md
9062 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9063 ("isinf<mode>2): Likewise.
9064 * convert.c (convert_to_integer): Using new target hook
9065 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9066 TARGET_C99_FUNCTIONS.
9067 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9068 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9069 * coretypes.h (function_class): New enum for different
9070 classes of functions.
9071 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9072 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9073 (TARGET_HAS_SINCOS): Likewise.
9074 (TARGET_LIBC_HAS_FUNCTION): New.
9075 * doc/tm.texi: Regenerated.
9076 * targhooks.h (default_libc_has_function): New declaration.
9077 (no_c99_libc_has_function): Ditto.
9078 (gnu_libc_has_function): Ditto.
9079 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9080 and TARGET_HAS_SINCOS.
9082 2013-08-18 Jan Hubicka <jh@suse.cz>
9084 * Makeifle-in (ipa-devirt.o): New.
9085 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9086 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9087 (analyze_functions): Look into possible targets of polymorphic call.
9088 * dumpfile.c (dump_files): Add type-inheritance dump.
9089 * dumpfile.h (TDI_inheritance): New.
9090 * ipa-devirt.c: New file.
9091 * ipa-utils.h (odr_type_d): Forward declare.
9092 (odr_type): New type.
9093 (build_type_inheritance_graph): Declare.
9094 (possible_polymorphic_call_targets): Declare and introduce inline
9095 variant when only edge is pased.
9096 (dump_possible_polymorphic_call_targets): Likewise.
9097 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9098 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9099 (types_same_for_odr): ... here.
9100 * tree.h (type_in_anonymous_namespace_p): Declare.
9102 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9104 PR tree-optimization/58006
9105 * tree-parloops.c (take_address_of): Don't ICE if get_name
9107 (eliminate_local_variables_stmt): Remove clobber stmts.
9109 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9111 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9112 error message is issued for an alias to undefined symbol.
9114 2013-08-18 Jan Hubicka <jh@suse.cz>
9116 * cgraph.c (cgraph_create_indirect_edge): Discover
9117 polymorphic calls and record basic info into indirect_info.
9118 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9119 devirtualization, ignore objc function calls.
9120 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9121 call with no parm index info.
9122 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9123 * tree.c (virtual_method_call_p): New function.
9124 * tree.h (virtual_method_call_p): Declare.
9126 2013-08-16 Jan Hubicka <jh@suse.cz>
9129 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9131 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9133 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9136 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9138 * gengtype.c (type_for_name): Add special-case support for
9139 locating types within the "gcc::" namespace.
9140 (open_base_files): Emit a "using namespace gcc" directive.
9142 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9145 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9146 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9148 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9149 array instead of each individual operand as a separate argument.
9150 (emit_fusion_gpr_load): Likewise.
9151 (expand_fusion_gpr_load): Add new function declaration.
9153 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9154 signature to have the operands passed as an array, instead of as
9155 separate arguments. Allow ZERO_EXTEND to be in the memory
9156 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9157 depend on the register live/dead flags when peepholes are run.
9158 (expand_fusion_gpr_load): New function to be called from the
9159 peephole2 pass, to change the register that addis sets to be the
9161 (emit_fusion_gpr_load): Change the calling signature to have the
9162 operands passed as an array, instead of as separate arguments.
9163 Allow ZERO_EXTEND to be in the memory address, and also
9164 SIGN_EXTEND if -mpower8-fusion-sign.
9166 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9168 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9169 adjust the register addis loads up in the peephole2 pass. Do not
9170 depend on the register live/dead state when the peephole pass is done.
9172 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9174 * gengtype.c (create_user_defined_type): Ensure that the kind
9175 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9176 declaration is seen before the GTY((user)) marking.
9178 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9181 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9183 2013-08-16 Jan Hubicka <jh@suse.cz>
9185 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9186 arugment expected_type.
9187 (gimple_fold_call): Use it.
9188 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9189 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9190 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9191 (try_make_edge_direct_virtual_call): Likewise.
9192 * tree.c (obj_type_ref_class): New.
9193 * tree.h (obj_type_ref_class): Use it.
9195 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9197 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9198 (rtl_slim_pp): Likewise.
9199 (init_rtl_slim_pretty_print): Likewise.
9200 (dump_value_slim): Don't call it. Use local pretty printer.
9201 (dump_insn_slim): Likewise.
9202 (dump_rtl_slim): Likewise.
9203 (str_pattern_slim): Likewise.
9204 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9207 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9209 PR tree-optimization/58164
9210 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9211 walk gimple_goto_dest of GIMPLE_GOTO.
9213 PR tree-optimization/58165
9214 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9215 bi_call must be the last stmt in a bb, don't split_block, instead
9216 use fallthru edge from it and give up if there is none.
9217 Release conds vector when returning early.
9219 2013-08-14 Xinliang David Li <davidxl@google.com>
9221 * config/i386/i386.c (ix86_option_override_internal):
9222 Remove unused variable and field.
9224 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9227 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9228 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9229 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9230 and Linux, correct BLKmode alignment when 128-bit alignment is
9231 required and compatibility flag is not set.
9232 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9233 for zero-size arguments when compatibility flag is not set.
9235 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9237 PR tree-optimization/58145
9238 * tree-sra.c (build_ref_for_offset): If prev_base has
9239 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9241 2013-08-14 Xinliang David Li <davidxl@google.com>
9243 * config/i386/i386.c (ix86_option_override_internal):
9244 Fix uninitialized variable error.
9246 2013-08-14 Xinliang David Li <davidxl@google.com>
9248 * config/i386/i386.opt: Define two new options.
9249 * config/i386/x86-tune.def: Add arch selector field in macros.
9250 * config/i386/i386.h: Adjust macro definition.
9251 * config/i386/i386.c (ix86_option_override_internal):
9253 (parse_mtune_ctrl_str): New function.
9254 (set_ix86_tune_features): New function.
9255 (ix86_function_specific_restore): Call the new helper function.
9257 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9259 PR rtl-optimization/57662
9260 * sel-sched.c (code_motion_process_successors): When the current insn
9261 is removed after the recursive traversal, break from the loop.
9262 Add comments and debug printouts.
9264 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9265 Alexandre Oliva <aoliva@redhat.com>
9268 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9269 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9270 there also UNSPEC_PLTOFF.
9272 2013-08-14 Marek Polacek <polacek@redhat.com>
9274 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9275 NUM_CONDITIONS bit positions.
9277 2013-08-13 Cary Coutant <ccoutant@google.com>
9279 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9280 (attr_checksum): Hash vector contents instead of pointer.
9281 (attr_checksum_ordered): Likewise.
9283 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9285 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9286 when Pmode != word_mode. Add length_address attribute.
9287 (sse3_monitor_<mode>): Merge from sse3_monitor and
9288 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9289 Pmode != word_mode. Update insn length attribute.
9290 * config/i386/i386.c (ix86_option_override_internal): Update
9291 ix86_gen_monitor selection for merged sse3_monitor insn.
9293 2013-08-13 Julian Brown <julian@codesourcery.com>
9295 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9296 perform invalid legitimization on greater-than-word-size modes for
9299 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9301 * ira.c (setup_class_translate_array): Use aclass instead of cl
9302 for classes not fully covered by allocno classes.
9304 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9306 PR tree-optimization/57661
9307 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9308 * tree-inline.c (tree_function_versioning): Initialize it.
9309 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9310 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9311 that is not being copied.
9314 * asan.c (instrument_strlen_call): Fix typo in comment.
9315 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9317 2013-08-13 Steve Ellcey <sellcey@mips.com>
9319 * config/mips/mips.md (prefetch): Use lw instead of ld on
9320 loongson in 32bit mode.
9322 2013-08-13 Nick Clifton <nickc@redhat.com>
9324 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9326 2013-08-13 Jan Hubicka <jh@suse.cz>
9328 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9329 introduced edge; fix typo in sanity check.
9330 (cgraph_resolve_speculation): Export; improve diagnostic.
9331 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9332 speculation at type mismatch.
9333 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9334 (cgraph_resolve_speculation): Declare.
9335 (symtab_can_be_discarded): New function.
9336 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9337 * ipa-inline-transform.c (speculation_removed): New global var.
9338 (clone_inlined_nodes): See if speculation can be removed.
9339 (inline_call): If speculations was removed, we growths may not match.
9340 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9341 (speculation_useful_p): New function.
9342 (resolve_noninline_speculation): New function.
9343 (inline_small_functions): Resolve useless speculations.
9344 * ipa-inline.h (speculation_useful_p): Declare
9345 * ipa.c (can_replace_by_local_alias): Simplify.
9346 (ipa_profile): Produce speculative calls in non-lto, too;
9347 add simple cost model; produce local aliases.
9349 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9351 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9354 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9356 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9358 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9360 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9361 iterator instead of X87MODEF.
9363 2013-08-12 Perez Read <netfirewall@gmail.com>
9366 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9367 operand 0 for intel asm alternative.
9368 (*movabs<mode>_2): Ditto for operand 1.
9370 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9372 * config/aarch64/arm_none.h
9373 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9375 2013-08-12 Nick Clifton <nickc@redhat.com>
9377 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9379 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9381 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9382 expand for QI/HImode operand to produce more effictive code for
9383 unsigned char(short) --> float(double) conversion.
9385 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9387 * doc/invoke.texi: Mention that -ftls-model does not force the final
9390 2013-08-12 Marek Polacek <polacek@redhat.com>
9391 Marc Glisse <marc.glisse@inria.fr>
9393 PR tree-optimization/57980
9394 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9395 when creating -1 constant.
9397 2013-08-10 Jan Hubicka <jh@suse.cz>
9399 Workaround binutils PR14342.
9400 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9401 (gimple_init_edge_profiler): Likewise.
9402 (gimple_gen_ic_func_profiler): Likewise.
9404 2013-08-09 Jan Hubicka <jh@suse.cz>
9406 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9408 2013-08-09 Xinliang David Li <davidxl@google.com>
9410 * config/i386/stringop.def: New file.
9411 * config/i386/stringop.opt: New file.
9412 * config/i386/i386-opts.h: Include stringopt.def.
9413 * config/i386/i386.opt: Include stringopt.opt.
9414 * config/i386/i386.c (ix86_option_override_internal):
9415 Override default size based stringop inline strategies with options.
9416 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9419 2013-08-09 Jan Hubicka <jh@suse.cz>
9421 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9423 2013-08-09 Jan Hubicka <jh@suse.cz>
9425 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9427 (dump_cgraph_node): Dump profile-id.
9428 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9429 and common_target_probability.
9430 * lto-cgraph.c (lto_output_edge): Stream common targets.
9431 (lto_output_node): Stream profile ids.
9432 (input_node): Stream profile ids.
9433 (input_edge): Stream common targets.
9434 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9435 * ipa.c: Include value-prof.h
9436 (ipa_profile_generate_summary): Turn indirect call statement histograms
9437 into common targets.
9438 (ipa_profile): Turn common targets into speculative edges.
9440 2013-08-09 Jan Hubicka <jh@suse.cz>
9442 * cgraph.h (cgraph_node): Add profile_id.
9443 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9444 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9445 (del_node_map): Update.
9446 (find_func_by_funcdef_no): Replace by ...
9447 (find_func_by_profile_id): ... this one.
9448 (gimple_ic_transform): Do not remove useful histograms when
9449 speculation is not done; dump info when indirect call removal
9451 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9452 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9453 (__gcov_indirect_call_profiler_v2): .. this one.
9454 * profile.h (init_node_map): Update.
9455 * coverage.c (coverage_compute_profile_id): New function.
9456 * coverage.h (coverage_compute_profile_id): Declare.
9457 * tree-profile.c (init_ic_make_global_vars): Make
9458 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9459 (gimple_init_edge_profiler): Update prototype of
9460 __gcov_indirect_call_profiler.
9461 (gimple_gen_ic_func_profiler): Simplify.
9462 (tree_profiling): Use init_node_map
9464 2013-08-09 Jan Hubicka <jh@suse.cz>
9466 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9467 non-speculative refs.
9468 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9469 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9470 (cgraph_set_call_stmt): Likewise.
9471 (cgraph_create_edge_1): Fix release checking compilatoin;
9473 (cgraph_free_edge): Free indirect info.
9474 (cgraph_turn_edge_to_speculative): New function.
9475 (cgraph_speculative_call_info): New function.
9476 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9477 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9478 (dump_cgraph_node): Dump speculation.
9479 (verify_edge_count_and_frequency): Accept speculative edges.
9480 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9481 (verify_cgraph_node): Handle speculation.
9482 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9483 (cgraph_set_call_stmt): Update prototype.
9484 (cgraph_make_edge_direct): Update prototype.
9485 (cgraph_speculative_call_info): Declare.
9486 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9487 to change; update call of ipa_find_references.
9488 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9489 lto_stmt_uid and speculative flags.
9490 (ipa_dump_references): Dump speculation.
9491 (ipa_clone_references): Clone speculative flag.
9492 (ipa_clone_referring): Likewise.
9493 (ipa_clone_ref): New function.
9494 (ipa_find_reference): Look into lto_stmt_uids
9495 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9496 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9497 (ipa_find_reference): Update declaration.
9498 (ipa_clone_ref): Declare.
9499 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9500 stream speculative flag.
9501 (lto_output_ref): Stream statements uids and speculation.
9502 (input_ref): Likewise.
9503 (input_edge): Stream speuclation.
9504 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9505 (cgraph_set_call_stmt_including_clones): Handle speculation.
9506 * ipa-inline.c (heap_edge_removal_hook): New function.
9507 (inline_small_functions): Register it.
9508 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9509 also initialize refs.
9510 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9512 (try_make_edge_direct_simple_call): Likewise.
9513 (try_make_edge_direct_simple_call): Likewise.
9514 (update_indirect_edges_after_inlining): Likewise.
9515 (remove_described_reference): Look proper lto_stmt_uid.
9516 (propagate_controlled_uses): Likewise.
9517 (propagate_controlled_uses): Liekwise.
9518 * tree-inline.c (copy_bb): Copy speculative edges.
9519 (redirect_all_calls): New function.
9520 (copy_cfg_body): Do redirection after loop info is updated.
9521 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9523 2013-08-09 Jan Hubicka <jh@suse.cz>
9525 * lto-streamer-out.c (output_function): Renumber PHIs.
9526 * lto-streamer-in.c (input_function): Likewise.
9528 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9530 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9531 (get_lane_unsigned): Likewise.
9532 (dup_lane_scalar): Likewise.
9533 (get_lane): enable for VALL.
9534 * config/aarch64/aarch64-simd.md
9535 (aarch64_dup_lane_scalar<mode>): Remove.
9536 (aarch64_get_lane_signed<mode>): Likewise.
9537 (aarch64_get_lane_unsigned<mode>): Likewise.
9538 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9539 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9540 (aarch64_get_lane<mode>): Enable for all vector modes.
9541 (aarch64_get_lanedi): Remove misleading constraints.
9542 * config/aarch64/arm_neon.h
9543 (__aarch64_vget_lane_any): Define.
9544 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9545 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9546 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9547 * config/aarch64/iterators.md (VDQQH): New.
9551 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9553 * configure.ac: Add GAS check for LEON instructions on SPARC.
9554 * configure: Regenerate.
9555 * config.in: Likewise.
9556 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9558 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9559 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9561 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9562 (AS_LEON_FLAG): New macro.
9563 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9564 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9565 Deal with LEON and LEON3 for the memory model.
9566 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9567 (atomic_compare_and_swap<mode>_1): Likewise.
9568 (*atomic_compare_and_swap<mode>_1): Likewise.
9570 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9572 * config/arm/neon.md (vcond): Fix floating-point vector
9573 comparisons against 0.
9575 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9577 * lra-constraints.c (emit_spill_move): Remove assert.
9578 (process_alt_operands): Add more debugging
9579 output. Increase reject for spilling into memory. Decrease
9580 reject for reloading scratch.
9581 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9583 2013-08-08 Steve Ellcey <sellcey@mips.com>
9585 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9586 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9587 micromips incompatible. Add nan2008.
9588 (MULTILIB_DIRNAMES): Add nan2008.
9589 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9590 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9591 and micromips incompatible. Add nan2008.
9592 (MULTILIB_DIRNAMES): Add nan2008.
9593 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9595 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9597 PR rtl-optimization/58079
9598 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9599 simplify_comparison has widened a comparison with an integer.
9601 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9603 * config/arm/neon.md (movmisalign<mode>): Disable when we
9604 don't allow unaligned accesses.
9605 (*movmisalign<mode>_neon_store): Likewise.
9606 (*movmisalign<mode>_neon_load): Likewise.
9607 (*movmisalign<mode>_neon_store): Likewise.
9608 (*movmisalign<mode>_neon_load): Likewise.
9610 2013-08-08 Jan Hubicka <jh@suse.cz>
9612 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9613 (make_pass_rebuild_cgraph_edges): Also clear references.
9614 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9615 * ipa-inline-transform.c (inline_transform): Remove all references
9617 * cgraphunit.c (expand_function): Remove all references after
9619 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9620 (ipa_find_reference): Rewrite to iterator.
9621 (remove_stmt_references): Likewise.
9622 (ipa_clear_stmts_in_references): New function.
9623 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9624 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9626 * ipa-split.c (split_function): Remove references in split function.
9628 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9631 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9632 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9634 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9637 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9638 suggested mode for the assignment isn't compatible with the
9641 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9644 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9646 2013-08-07 Xinliang David Li <davidxl@google.com>
9648 * config/i386/i386.opt: New option -mtune-ctrl=.
9649 * config/i386/x86-tune.def: New file.
9650 * config/i386/i386.h: include x86-tune.def.
9651 * config/i386/i386.c (ix86_option_override_internal):
9652 Parsing -mtune-ctrl= option and set tune features.
9654 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9657 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9658 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9661 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9663 * rtl.h (update_alignments): Declare.
9664 * final.c (grow_label_align): New function extracted from...
9665 (shorten_branches): ...here. Call it.
9666 (update_alignments): New function.
9667 * reorg.c (sibling_labels): New variable.
9668 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9669 the new label along with it onto the sibling_labels vector.
9670 (fill_simple_delay_slots): Adjust call to get_label_before.
9671 (fill_slots_from_thread): Likewise.
9672 (relax_delay_slots): Likewise.
9673 (make_return_insns): Likewise.
9674 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9676 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9678 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9679 document its semantics.
9680 (diagnostic_report_diagnostic): Adjust accordingly.
9682 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9684 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9685 (sparc_option_override): ...and port to new C++ pass API.
9686 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9688 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9690 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9692 2013-08-06 Caroline Tice <cmtice@google.com>
9694 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9695 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9696 * tree-pass.h: Add pass_vtable_verify.
9697 * varasm.c (assemble_variable): Add code to properly set the comdat
9698 section and name for the .vtable_map_vars section.
9699 (assemble_vtyv_preinit_initializer): New function.
9700 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9702 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9703 * vtable-verify.c: New file.
9704 * vtable-verify.h: New file.
9705 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9706 initialiation levels.
9707 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9708 * passes.def: Insert pass_vtable_verify.
9709 * aclocal.m4: Reorder includes.
9710 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9711 -fvtv-counts options.
9712 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9713 as appropriate, if -fvtable-verify=... is used.
9714 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9715 -fvtable-verify=... is used.
9716 * Makefile.in (OBJS): Add vtable-verify.o to list.
9717 (vtable-verify.o): Add new build rule.
9718 (GTFILES): Add vtable-verify.c to list.
9719 * common.opt (fvtable-verify=): New flag.
9720 (vtv_priority): Values for fvtable-verify= flag.
9721 (fvtv-counts): New flag.
9722 (fvtv-debug): New flag.
9723 * tree.h (save_vtable_map_decl): New extern function decl.
9725 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9727 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9728 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9729 (pass_data_rl78_devirt): ...new pass_data instance and...
9730 (make_pass_rl78_devirt): ...new function.
9731 (rl78_asm_file_start): Port pass registration to new C++ API.
9733 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9735 * coretypes.h (rtl_opt_pass): Add.
9736 (gcc::context): Add.
9737 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9738 (epiphany_init): Port to new C++ pass API.
9739 (epiphany_optimize_mode_switching): Likewise.
9740 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9741 (pass_manager::get_pass_mode_switching): New.
9742 (pass_manager::get_pass_peephole2): New.
9743 * mode-switching.c (pass_mode_switching): Add clone method.
9744 * recog.c (pass_peephole2): Add clone method.
9745 (pass_split_all_insns): Add clone method.
9747 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9749 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9750 (mips_option_override): ...here, porting to new C++ API for passes.
9752 2013-08-06 Jan Hubicka <jh@suse.cz>
9754 * cgraph.c (cgraph_get_body): New function based on lto.c
9756 * cgraph.h (cgraph_get_body): Declare.
9757 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9759 * cgraphunit.c (expand_function): Get body prior expanding.
9760 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9761 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9763 * passes.c (do_per_function_toporder): Get body.
9764 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9765 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9767 2013-08-06 Martin Jambor <mjambor@suse.cz>
9770 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9771 is not re-finalized. Rename second parameter to no_collect.
9773 2013-08-06 Martin Jambor <mjambor@suse.cz>
9776 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9777 MEM_REF has proper alignment information.
9779 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9782 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9784 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9785 rtx (*) (rtx, ...) with insn_gen_fn.
9786 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9787 insn_gen_fn::stored_funcptr. Add initializer braces.
9789 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9791 Rewrite how instances of passes are cloned to remove assumptions
9792 about their sizes (thus allowing pass subclasses to have
9793 additional data fields, albeit non-GC-managed ones at this point).
9795 * passes.c (make_pass_instance): Now that passes have clone
9796 methods, rewrite this function to eliminate XNEW and memcpy
9797 calls that used hardcoded sizes. Since this function no longer
9798 creates pass instances, rename it to...
9799 (add_pass_instance): ...this. Document the old way that passes were
9800 numbered and flagged, and rework this function to continue using it.
9801 (next_pass_1): Add an initial_pass argument for use by
9803 (position_pass): When adding multiple instances of a pass, use
9804 the pass's clone method, rather than relying on the XNEW/memcpy
9805 within the former make_pass_instance (now add_pass_instance).
9806 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9807 the initial instance of the current pass within the pass manager.
9809 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9811 This is the automated part of the conversion of passes from C
9812 structs to C++ classes.
9814 Patch autogenerated by refactor_passes.py from
9815 https://github.com/davidmalcolm/gcc-refactoring-scripts
9816 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9818 * asan.c (pass_asan): Convert from a global struct to a subclass of
9819 gimple_opt_pass along with...
9820 (pass_data_asan): ...new pass_data instance and...
9821 (make_pass_asan): ...new function.
9822 (pass_asan_O0): Convert from a global struct to a subclass of
9823 gimple_opt_pass along with...
9824 (pass_data_asan_O0): ...new pass_data instance and...
9825 (make_pass_asan_O0): ...new function.
9826 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9827 subclass of rtl_opt_pass along with...
9828 (pass_data_inc_dec): ...new pass_data instance and...
9829 (make_pass_inc_dec): ...new function.
9830 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9831 a subclass of rtl_opt_pass along with...
9832 (pass_data_reorder_blocks): ...new pass_data instance and...
9833 (make_pass_reorder_blocks): ...new function.
9834 (pass_duplicate_computed_gotos): Convert from a global struct to a
9835 subclass of rtl_opt_pass along with...
9836 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9837 (make_pass_duplicate_computed_gotos): ...new function.
9838 (pass_partition_blocks): Convert from a global struct to a subclass of
9839 rtl_opt_pass along with...
9840 (pass_data_partition_blocks): ...new pass_data instance and...
9841 (make_pass_partition_blocks): ...new function.
9842 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9843 struct to a subclass of rtl_opt_pass along with...
9844 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9846 (make_pass_branch_target_load_optimize1): ...new function.
9847 (pass_branch_target_load_optimize2): Convert from a global struct to a
9848 subclass of rtl_opt_pass along with...
9849 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9851 (make_pass_branch_target_load_optimize2): ...new function.
9852 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9853 of rtl_opt_pass along with...
9854 (pass_data_jump): ...new pass_data instance and...
9855 (make_pass_jump): ...new function.
9856 (pass_jump2): Convert from a global struct to a subclass of
9857 rtl_opt_pass along with...
9858 (pass_data_jump2): ...new pass_data instance and...
9859 (make_pass_jump2): ...new function.
9860 * cfgexpand.c (pass_expand): Convert from a global struct to a
9861 subclass of rtl_opt_pass along with...
9862 (pass_data_expand): ...new pass_data instance and...
9863 (make_pass_expand): ...new function.
9864 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9865 of rtl_opt_pass along with...
9866 (pass_data_free_cfg): ...new pass_data instance and...
9867 (make_pass_free_cfg): ...new function.
9868 (pass_into_cfg_layout_mode): Convert from a global struct to a
9869 subclass of rtl_opt_pass along with...
9870 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9871 (make_pass_into_cfg_layout_mode): ...new function.
9872 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9873 subclass of rtl_opt_pass along with...
9874 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9875 (make_pass_outof_cfg_layout_mode): ...new function.
9876 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9877 struct to a subclass of gimple_opt_pass along with...
9878 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9879 (make_pass_build_cgraph_edges): ...new function.
9880 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9881 subclass of gimple_opt_pass along with...
9882 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9883 (make_pass_rebuild_cgraph_edges): ...new function.
9884 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9885 subclass of gimple_opt_pass along with...
9886 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9888 (make_pass_remove_cgraph_callee_edges): ...new function.
9889 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9890 struct to a subclass of rtl_opt_pass along with...
9891 (pass_data_stack_adjustments): ...new pass_data instance and...
9892 (make_pass_stack_adjustments): ...new function.
9893 * combine.c (pass_combine): Convert from a global struct to a subclass
9894 of rtl_opt_pass along with...
9895 (pass_data_combine): ...new pass_data instance and...
9896 (make_pass_combine): ...new function.
9897 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9898 global struct to a subclass of rtl_opt_pass along with...
9899 (pass_data_compare_elim_after_reload): ...new pass_data instance
9901 (make_pass_compare_elim_after_reload): ...new function.
9902 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9903 of rtl_opt_pass along with...
9904 (pass_data_rtl_cprop): ...new pass_data instance and...
9905 (make_pass_rtl_cprop): ...new function.
9906 * cse.c (pass_cse): Convert from a global struct to a subclass of
9907 rtl_opt_pass along with...
9908 (pass_data_cse): ...new pass_data instance and...
9909 (make_pass_cse): ...new function.
9910 (pass_cse2): Convert from a global struct to a subclass of
9911 rtl_opt_pass along with...
9912 (pass_data_cse2): ...new pass_data instance and...
9913 (make_pass_cse2): ...new function.
9914 (pass_cse_after_global_opts): Convert from a global struct to a
9915 subclass of rtl_opt_pass along with...
9916 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9917 (make_pass_cse_after_global_opts): ...new function.
9918 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9919 of rtl_opt_pass along with...
9920 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9921 (make_pass_ud_rtl_dce): ...new function.
9922 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9923 rtl_opt_pass along with...
9924 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9925 (make_pass_fast_rtl_dce): ...new function.
9926 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9927 a subclass of rtl_opt_pass along with...
9928 (pass_data_df_initialize_opt): ...new pass_data instance and...
9929 (make_pass_df_initialize_opt): ...new function.
9930 (pass_df_initialize_no_opt): Convert from a global struct to a
9931 subclass of rtl_opt_pass along with...
9932 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9933 (make_pass_df_initialize_no_opt): ...new function.
9934 (pass_df_finish): Convert from a global struct to a subclass of
9935 rtl_opt_pass along with...
9936 (pass_data_df_finish): ...new pass_data instance and...
9937 (make_pass_df_finish): ...new function.
9938 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9939 rtl_opt_pass along with...
9940 (pass_data_rtl_dse1): ...new pass_data instance and...
9941 (make_pass_rtl_dse1): ...new function.
9942 (pass_rtl_dse2): Convert from a global struct to a subclass of
9943 rtl_opt_pass along with...
9944 (pass_data_rtl_dse2): ...new pass_data instance and...
9945 (make_pass_rtl_dse2): ...new function.
9946 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9947 subclass of rtl_opt_pass along with...
9948 (pass_data_dwarf2_frame): ...new pass_data instance and...
9949 (make_pass_dwarf2_frame): ...new function.
9950 * except.c (pass_set_nothrow_function_flags): Convert from a global
9951 struct to a subclass of rtl_opt_pass along with...
9952 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9954 (make_pass_set_nothrow_function_flags): ...new function.
9955 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9956 subclass of rtl_opt_pass along with...
9957 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9959 (make_pass_convert_to_eh_region_ranges): ...new function.
9960 * final.c (pass_compute_alignments): Convert from a global struct to a
9961 subclass of rtl_opt_pass along with...
9962 (pass_data_compute_alignments): ...new pass_data instance and...
9963 (make_pass_compute_alignments): ...new function.
9964 (pass_final): Convert from a global struct to a subclass of
9965 rtl_opt_pass along with...
9966 (pass_data_final): ...new pass_data instance and...
9967 (make_pass_final): ...new function.
9968 (pass_shorten_branches): Convert from a global struct to a subclass of
9969 rtl_opt_pass along with...
9970 (pass_data_shorten_branches): ...new pass_data instance and...
9971 (make_pass_shorten_branches): ...new function.
9972 (pass_clean_state): Convert from a global struct to a subclass of
9973 rtl_opt_pass along with...
9974 (pass_data_clean_state): ...new pass_data instance and...
9975 (make_pass_clean_state): ...new function.
9976 * function.c (pass_instantiate_virtual_regs): Convert from a global
9977 struct to a subclass of rtl_opt_pass along with...
9978 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9979 (make_pass_instantiate_virtual_regs): ...new function.
9980 (pass_leaf_regs): Convert from a global struct to a subclass of
9981 rtl_opt_pass along with...
9982 (pass_data_leaf_regs): ...new pass_data instance and...
9983 (make_pass_leaf_regs): ...new function.
9984 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9985 subclass of rtl_opt_pass along with...
9986 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9988 (make_pass_thread_prologue_and_epilogue): ...new function.
9989 (pass_match_asm_constraints): Convert from a global struct to a
9990 subclass of rtl_opt_pass along with...
9991 (pass_data_match_asm_constraints): ...new pass_data instance and...
9992 (make_pass_match_asm_constraints): ...new function.
9993 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9994 subclass of rtl_opt_pass along with...
9995 (pass_data_rtl_fwprop): ...new pass_data instance and...
9996 (make_pass_rtl_fwprop): ...new function.
9997 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9998 rtl_opt_pass along with...
9999 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10000 (make_pass_rtl_fwprop_addr): ...new function.
10001 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10002 rtl_opt_pass along with...
10003 (pass_data_rtl_pre): ...new pass_data instance and...
10004 (make_pass_rtl_pre): ...new function.
10005 (pass_rtl_hoist): Convert from a global struct to a subclass of
10006 rtl_opt_pass along with...
10007 (pass_data_rtl_hoist): ...new pass_data instance and...
10008 (make_pass_rtl_hoist): ...new function.
10009 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10010 subclass of gimple_opt_pass along with...
10011 (pass_data_lower_cf): ...new pass_data instance and...
10012 (make_pass_lower_cf): ...new function.
10013 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10014 from a global struct to a subclass of gimple_opt_pass along with...
10015 (pass_data_strength_reduction): ...new pass_data instance and...
10016 (make_pass_strength_reduction): ...new function.
10017 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10018 of rtl_opt_pass along with...
10019 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10020 (make_pass_rtl_ifcvt): ...new function.
10021 (pass_if_after_combine): Convert from a global struct to a subclass of
10022 rtl_opt_pass along with...
10023 (pass_data_if_after_combine): ...new pass_data instance and...
10024 (make_pass_if_after_combine): ...new function.
10025 (pass_if_after_reload): Convert from a global struct to a subclass of
10026 rtl_opt_pass along with...
10027 (pass_data_if_after_reload): ...new pass_data instance and...
10028 (make_pass_if_after_reload): ...new function.
10029 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10030 a subclass of rtl_opt_pass along with...
10031 (pass_data_initialize_regs): ...new pass_data instance and...
10032 (make_pass_initialize_regs): ...new function.
10033 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10034 of ipa_opt_pass_d along with...
10035 (pass_data_ipa_cp): ...new pass_data instance and...
10036 (make_pass_ipa_cp): ...new function.
10037 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10038 global struct to a subclass of gimple_opt_pass along with...
10039 (pass_data_inline_parameters): ...new pass_data instance and...
10040 (make_pass_inline_parameters): ...new function.
10041 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10042 subclass of gimple_opt_pass along with...
10043 (pass_data_early_inline): ...new pass_data instance and...
10044 (make_pass_early_inline): ...new function.
10045 (pass_ipa_inline): Convert from a global struct to a subclass of
10046 ipa_opt_pass_d along with...
10047 (pass_data_ipa_inline): ...new pass_data instance and...
10048 (make_pass_ipa_inline): ...new function.
10049 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10050 struct to a subclass of gimple_opt_pass along with...
10051 (pass_data_local_pure_const): ...new pass_data instance and...
10052 (make_pass_local_pure_const): ...new function.
10053 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10054 ipa_opt_pass_d along with...
10055 (pass_data_ipa_pure_const): ...new pass_data instance and...
10056 (make_pass_ipa_pure_const): ...new function.
10057 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10058 to a subclass of ipa_opt_pass_d along with...
10059 (pass_data_ipa_reference): ...new pass_data instance and...
10060 (make_pass_ipa_reference): ...new function.
10061 * ipa-split.c (pass_split_functions): Convert from a global struct to
10062 a subclass of gimple_opt_pass along with...
10063 (pass_data_split_functions): ...new pass_data instance and...
10064 (make_pass_split_functions): ...new function.
10065 (pass_feedback_split_functions): Convert from a global struct to a
10066 subclass of gimple_opt_pass along with...
10067 (pass_data_feedback_split_functions): ...new pass_data instance and...
10068 (make_pass_feedback_split_functions): ...new function.
10069 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10070 global struct to a subclass of simple_ipa_opt_pass along with...
10071 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10073 (make_pass_ipa_function_and_variable_visibility): ...new function.
10074 (pass_ipa_free_inline_summary): Convert from a global struct to a
10075 subclass of simple_ipa_opt_pass along with...
10076 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10077 (make_pass_ipa_free_inline_summary): ...new function.
10078 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10079 subclass of ipa_opt_pass_d along with...
10080 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10082 (make_pass_ipa_whole_program_visibility): ...new function.
10083 (pass_ipa_profile): Convert from a global struct to a subclass of
10084 ipa_opt_pass_d along with...
10085 (pass_data_ipa_profile): ...new pass_data instance and...
10086 (make_pass_ipa_profile): ...new function.
10087 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10088 ipa_opt_pass_d along with...
10089 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10090 (make_pass_ipa_cdtor_merge): ...new function.
10091 * ira.c (pass_ira): Convert from a global struct to a subclass of
10092 rtl_opt_pass along with...
10093 (pass_data_ira): ...new pass_data instance and...
10094 (make_pass_ira): ...new function.
10095 (pass_reload): Convert from a global struct to a subclass of
10096 rtl_opt_pass along with...
10097 (pass_data_reload): ...new pass_data instance and...
10098 (make_pass_reload): ...new function.
10099 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10100 subclass of rtl_opt_pass along with...
10101 (pass_data_cleanup_barriers): ...new pass_data instance and...
10102 (make_pass_cleanup_barriers): ...new function.
10103 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10104 of rtl_opt_pass along with...
10105 (pass_data_loop2): ...new pass_data instance and...
10106 (make_pass_loop2): ...new function.
10107 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10108 rtl_opt_pass along with...
10109 (pass_data_rtl_loop_init): ...new pass_data instance and...
10110 (make_pass_rtl_loop_init): ...new function.
10111 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10112 rtl_opt_pass along with...
10113 (pass_data_rtl_loop_done): ...new pass_data instance and...
10114 (make_pass_rtl_loop_done): ...new function.
10115 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10116 subclass of rtl_opt_pass along with...
10117 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10118 (make_pass_rtl_move_loop_invariants): ...new function.
10119 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10120 rtl_opt_pass along with...
10121 (pass_data_rtl_unswitch): ...new pass_data instance and...
10122 (make_pass_rtl_unswitch): ...new function.
10123 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10124 subclass of rtl_opt_pass along with...
10125 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10127 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10128 (pass_rtl_doloop): Convert from a global struct to a subclass of
10129 rtl_opt_pass along with...
10130 (pass_data_rtl_doloop): ...new pass_data instance and...
10131 (make_pass_rtl_doloop): ...new function.
10132 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10133 a subclass of rtl_opt_pass along with...
10134 (pass_data_lower_subreg): ...new pass_data instance and...
10135 (make_pass_lower_subreg): ...new function.
10136 (pass_lower_subreg2): Convert from a global struct to a subclass of
10137 rtl_opt_pass along with...
10138 (pass_data_lower_subreg2): ...new pass_data instance and...
10139 (make_pass_lower_subreg2): ...new function.
10140 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10141 struct to a subclass of ipa_opt_pass_d along with...
10142 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10143 (make_pass_ipa_lto_gimple_out): ...new function.
10144 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10145 of ipa_opt_pass_d along with...
10146 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10147 (make_pass_ipa_lto_finish_out): ...new function.
10148 * mode-switching.c (pass_mode_switching): Convert from a global struct
10149 to a subclass of rtl_opt_pass along with...
10150 (pass_data_mode_switching): ...new pass_data instance and...
10151 (make_pass_mode_switching): ...new function.
10152 * modulo-sched.c (pass_sms): Convert from a global struct to a
10153 subclass of rtl_opt_pass along with...
10154 (pass_data_sms): ...new pass_data instance and...
10155 (make_pass_sms): ...new function.
10156 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10157 subclass of gimple_opt_pass along with...
10158 (pass_data_expand_omp): ...new pass_data instance and...
10159 (make_pass_expand_omp): ...new function.
10160 (pass_lower_omp): Convert from a global struct to a subclass of
10161 gimple_opt_pass along with...
10162 (pass_data_lower_omp): ...new pass_data instance and...
10163 (make_pass_lower_omp): ...new function.
10164 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10165 of gimple_opt_pass along with...
10166 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10167 (make_pass_diagnose_omp_blocks): ...new function.
10168 * passes.c (pass_early_local_passes): Convert from a global struct to
10169 a subclass of simple_ipa_opt_pass along with...
10170 (pass_data_early_local_passes): ...new pass_data instance and...
10171 (make_pass_early_local_passes): ...new function.
10172 (pass_all_early_optimizations): Convert from a global struct to a
10173 subclass of gimple_opt_pass along with...
10174 (pass_data_all_early_optimizations): ...new pass_data instance and...
10175 (make_pass_all_early_optimizations): ...new function.
10176 (pass_all_optimizations): Convert from a global struct to a subclass
10177 of gimple_opt_pass along with...
10178 (pass_data_all_optimizations): ...new pass_data instance and...
10179 (make_pass_all_optimizations): ...new function.
10180 (pass_all_optimizations_g): Convert from a global struct to a subclass
10181 of gimple_opt_pass along with...
10182 (pass_data_all_optimizations_g): ...new pass_data instance and...
10183 (make_pass_all_optimizations_g): ...new function.
10184 (pass_rest_of_compilation): Convert from a global struct to a subclass
10185 of rtl_opt_pass along with...
10186 (pass_data_rest_of_compilation): ...new pass_data instance and...
10187 (make_pass_rest_of_compilation): ...new function.
10188 (pass_postreload): Convert from a global struct to a subclass of
10189 rtl_opt_pass along with...
10190 (pass_data_postreload): ...new pass_data instance and...
10191 (make_pass_postreload): ...new function.
10192 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10193 subclass of rtl_opt_pass along with...
10194 (pass_data_gcse2): ...new pass_data instance and...
10195 (make_pass_gcse2): ...new function.
10196 * postreload.c (pass_postreload_cse): Convert from a global struct to
10197 a subclass of rtl_opt_pass along with...
10198 (pass_data_postreload_cse): ...new pass_data instance and...
10199 (make_pass_postreload_cse): ...new function.
10200 * predict.c (pass_profile): Convert from a global struct to a subclass
10201 of gimple_opt_pass along with...
10202 (pass_data_profile): ...new pass_data instance and...
10203 (make_pass_profile): ...new function.
10204 (pass_strip_predict_hints): Convert from a global struct to a subclass
10205 of gimple_opt_pass along with...
10206 (pass_data_strip_predict_hints): ...new pass_data instance and...
10207 (make_pass_strip_predict_hints): ...new function.
10208 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10209 of rtl_opt_pass along with...
10210 (pass_data_peephole2): ...new pass_data instance and...
10211 (make_pass_peephole2): ...new function.
10212 (pass_split_all_insns): Convert from a global struct to a subclass of
10213 rtl_opt_pass along with...
10214 (pass_data_split_all_insns): ...new pass_data instance and...
10215 (make_pass_split_all_insns): ...new function.
10216 (pass_split_after_reload): Convert from a global struct to a subclass
10217 of rtl_opt_pass along with...
10218 (pass_data_split_after_reload): ...new pass_data instance and...
10219 (make_pass_split_after_reload): ...new function.
10220 (pass_split_before_regstack): Convert from a global struct to a
10221 subclass of rtl_opt_pass along with...
10222 (pass_data_split_before_regstack): ...new pass_data instance and...
10223 (make_pass_split_before_regstack): ...new function.
10224 (pass_split_before_sched2): Convert from a global struct to a subclass
10225 of rtl_opt_pass along with...
10226 (pass_data_split_before_sched2): ...new pass_data instance and...
10227 (make_pass_split_before_sched2): ...new function.
10228 (pass_split_for_shorten_branches): Convert from a global struct to a
10229 subclass of rtl_opt_pass along with...
10230 (pass_data_split_for_shorten_branches): ...new pass_data instance
10232 (make_pass_split_for_shorten_branches): ...new function.
10233 * ree.c (pass_ree): Convert from a global struct to a subclass of
10234 rtl_opt_pass along with...
10235 (pass_data_ree): ...new pass_data instance and...
10236 (make_pass_ree): ...new function.
10237 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10238 subclass of rtl_opt_pass along with...
10239 (pass_data_stack_regs): ...new pass_data instance and...
10240 (make_pass_stack_regs): ...new function.
10241 (pass_stack_regs_run): Convert from a global struct to a subclass of
10242 rtl_opt_pass along with...
10243 (pass_data_stack_regs_run): ...new pass_data instance and...
10244 (make_pass_stack_regs_run): ...new function.
10245 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10246 subclass of rtl_opt_pass along with...
10247 (pass_data_cprop_hardreg): ...new pass_data instance and...
10248 (make_pass_cprop_hardreg): ...new function.
10249 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10250 subclass of rtl_opt_pass along with...
10251 (pass_data_reginfo_init): ...new pass_data instance and...
10252 (make_pass_reginfo_init): ...new function.
10253 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10254 of rtl_opt_pass along with...
10255 (pass_data_regmove): ...new pass_data instance and...
10256 (make_pass_regmove): ...new function.
10257 * regrename.c (pass_regrename): Convert from a global struct to a
10258 subclass of rtl_opt_pass along with...
10259 (pass_data_regrename): ...new pass_data instance and...
10260 (make_pass_regrename): ...new function.
10261 * reorg.c (pass_delay_slots): Convert from a global struct to a
10262 subclass of rtl_opt_pass along with...
10263 (pass_data_delay_slots): ...new pass_data instance and...
10264 (make_pass_delay_slots): ...new function.
10265 (pass_machine_reorg): Convert from a global struct to a subclass of
10266 rtl_opt_pass along with...
10267 (pass_data_machine_reorg): ...new pass_data instance and...
10268 (make_pass_machine_reorg): ...new function.
10269 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10270 of rtl_opt_pass along with...
10271 (pass_data_sched): ...new pass_data instance and...
10272 (make_pass_sched): ...new function.
10273 (pass_sched2): Convert from a global struct to a subclass of
10274 rtl_opt_pass along with...
10275 (pass_data_sched2): ...new pass_data instance and...
10276 (make_pass_sched2): ...new function.
10277 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10278 to a subclass of rtl_opt_pass along with...
10279 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10280 (make_pass_stack_ptr_mod): ...new function.
10281 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10282 to a subclass of rtl_opt_pass along with...
10283 (pass_data_rtl_store_motion): ...new pass_data instance and...
10284 (make_pass_rtl_store_motion): ...new function.
10285 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10286 of gimple_opt_pass along with...
10287 (pass_data_tracer): ...new pass_data instance and...
10288 (make_pass_tracer): ...new function.
10289 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10290 to a subclass of gimple_opt_pass along with...
10291 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10292 (make_pass_diagnose_tm_blocks): ...new function.
10293 (pass_lower_tm): Convert from a global struct to a subclass of
10294 gimple_opt_pass along with...
10295 (pass_data_lower_tm): ...new pass_data instance and...
10296 (make_pass_lower_tm): ...new function.
10297 (pass_tm_init): Convert from a global struct to a subclass of
10298 gimple_opt_pass along with...
10299 (pass_data_tm_init): ...new pass_data instance and...
10300 (make_pass_tm_init): ...new function.
10301 (pass_tm_mark): Convert from a global struct to a subclass of
10302 gimple_opt_pass along with...
10303 (pass_data_tm_mark): ...new pass_data instance and...
10304 (make_pass_tm_mark): ...new function.
10305 (pass_tm_edges): Convert from a global struct to a subclass of
10306 gimple_opt_pass along with...
10307 (pass_data_tm_edges): ...new pass_data instance and...
10308 (make_pass_tm_edges): ...new function.
10309 (pass_tm_memopt): Convert from a global struct to a subclass of
10310 gimple_opt_pass along with...
10311 (pass_data_tm_memopt): ...new pass_data instance and...
10312 (make_pass_tm_memopt): ...new function.
10313 (pass_ipa_tm): Convert from a global struct to a subclass of
10314 simple_ipa_opt_pass along with...
10315 (pass_data_ipa_tm): ...new pass_data instance and...
10316 (make_pass_ipa_tm): ...new function.
10317 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10318 subclass of gimple_opt_pass along with...
10319 (pass_data_call_cdce): ...new pass_data instance and...
10320 (make_pass_call_cdce): ...new function.
10321 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10322 subclass of gimple_opt_pass along with...
10323 (pass_data_build_cfg): ...new pass_data instance and...
10324 (make_pass_build_cfg): ...new function.
10325 (pass_split_crit_edges): Convert from a global struct to a subclass of
10326 gimple_opt_pass along with...
10327 (pass_data_split_crit_edges): ...new pass_data instance and...
10328 (make_pass_split_crit_edges): ...new function.
10329 (pass_warn_function_return): Convert from a global struct to a
10330 subclass of gimple_opt_pass along with...
10331 (pass_data_warn_function_return): ...new pass_data instance and...
10332 (make_pass_warn_function_return): ...new function.
10333 (pass_warn_function_noreturn): Convert from a global struct to a
10334 subclass of gimple_opt_pass along with...
10335 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10336 (make_pass_warn_function_noreturn): ...new function.
10337 (pass_warn_unused_result): Convert from a global struct to a subclass
10338 of gimple_opt_pass along with...
10339 (pass_data_warn_unused_result): ...new pass_data instance and...
10340 (make_pass_warn_unused_result): ...new function.
10341 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10342 a subclass of gimple_opt_pass along with...
10343 (pass_data_merge_phi): ...new pass_data instance and...
10344 (make_pass_merge_phi): ...new function.
10345 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10346 a subclass of gimple_opt_pass along with...
10347 (pass_data_lower_complex): ...new pass_data instance and...
10348 (make_pass_lower_complex): ...new function.
10349 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10350 gimple_opt_pass along with...
10351 (pass_data_lower_complex_O0): ...new pass_data instance and...
10352 (make_pass_lower_complex_O0): ...new function.
10353 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10354 subclass of gimple_opt_pass along with...
10355 (pass_data_lower_eh): ...new pass_data instance and...
10356 (make_pass_lower_eh): ...new function.
10357 (pass_refactor_eh): Convert from a global struct to a subclass of
10358 gimple_opt_pass along with...
10359 (pass_data_refactor_eh): ...new pass_data instance and...
10360 (make_pass_refactor_eh): ...new function.
10361 (pass_lower_resx): Convert from a global struct to a subclass of
10362 gimple_opt_pass along with...
10363 (pass_data_lower_resx): ...new pass_data instance and...
10364 (make_pass_lower_resx): ...new function.
10365 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10366 of gimple_opt_pass along with...
10367 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10368 (make_pass_lower_eh_dispatch): ...new function.
10369 (pass_cleanup_eh): Convert from a global struct to a subclass of
10370 gimple_opt_pass along with...
10371 (pass_data_cleanup_eh): ...new pass_data instance and...
10372 (make_pass_cleanup_eh): ...new function.
10373 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10374 to a subclass of simple_ipa_opt_pass along with...
10375 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10376 (make_pass_ipa_lower_emutls): ...new function.
10377 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10378 a subclass of gimple_opt_pass along with...
10379 (pass_data_if_conversion): ...new pass_data instance and...
10380 (make_pass_if_conversion): ...new function.
10381 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10382 subclass of gimple_opt_pass along with...
10383 (pass_data_build_ssa): ...new pass_data instance and...
10384 (make_pass_build_ssa): ...new function.
10385 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10386 global struct to a subclass of gimple_opt_pass along with...
10387 (pass_data_loop_distribution): ...new pass_data instance and...
10388 (make_pass_loop_distribution): ...new function.
10389 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10390 subclass of gimple_opt_pass along with...
10391 (pass_data_mudflap_1): ...new pass_data instance and...
10392 (make_pass_mudflap_1): ...new function.
10393 (pass_mudflap_2): Convert from a global struct to a subclass of
10394 gimple_opt_pass along with...
10395 (pass_data_mudflap_2): ...new pass_data instance and...
10396 (make_pass_mudflap_2): ...new function.
10397 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10398 subclass of gimple_opt_pass along with...
10399 (pass_data_mudflap_1): ...new pass_data instance and...
10400 (make_pass_mudflap_1): ...new function.
10401 (pass_mudflap_2): Convert from a global struct to a subclass of
10402 gimple_opt_pass along with...
10403 (pass_data_mudflap_2): ...new pass_data instance and...
10404 (make_pass_mudflap_2): ...new function.
10405 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10406 gimple_opt_pass along with...
10407 (pass_data_nrv): ...new pass_data instance and...
10408 (make_pass_nrv): ...new function.
10409 (pass_return_slot): Convert from a global struct to a subclass of
10410 gimple_opt_pass along with...
10411 (pass_data_return_slot): ...new pass_data instance and...
10412 (make_pass_return_slot): ...new function.
10413 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10414 to a subclass of gimple_opt_pass along with...
10415 (pass_data_object_sizes): ...new pass_data instance and...
10416 (make_pass_object_sizes): ...new function.
10417 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10418 global struct to a subclass of gimple_opt_pass along with...
10419 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10421 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10422 (pass_fixup_cfg): Convert from a global struct to a subclass of
10423 gimple_opt_pass along with...
10424 (pass_data_fixup_cfg): ...new pass_data instance and...
10425 (make_pass_fixup_cfg): ...new function.
10426 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10427 (make_pass_mudflap_1): ...new function.
10428 (pass_mudflap_2): Replace declaration with that of...
10429 (make_pass_mudflap_2): ...new function.
10430 (pass_asan): Replace declaration with that of...
10431 (make_pass_asan): ...new function.
10432 (pass_asan_O0): Replace declaration with that of...
10433 (make_pass_asan_O0): ...new function.
10434 (pass_tsan): Replace declaration with that of...
10435 (make_pass_tsan): ...new function.
10436 (pass_tsan_O0): Replace declaration with that of...
10437 (make_pass_tsan_O0): ...new function.
10438 (pass_lower_cf): Replace declaration with that of...
10439 (make_pass_lower_cf): ...new function.
10440 (pass_refactor_eh): Replace declaration with that of...
10441 (make_pass_refactor_eh): ...new function.
10442 (pass_lower_eh): Replace declaration with that of...
10443 (make_pass_lower_eh): ...new function.
10444 (pass_lower_eh_dispatch): Replace declaration with that of...
10445 (make_pass_lower_eh_dispatch): ...new function.
10446 (pass_lower_resx): Replace declaration with that of...
10447 (make_pass_lower_resx): ...new function.
10448 (pass_build_cfg): Replace declaration with that of...
10449 (make_pass_build_cfg): ...new function.
10450 (pass_early_tree_profile): Replace declaration with that of...
10451 (make_pass_early_tree_profile): ...new function.
10452 (pass_cleanup_eh): Replace declaration with that of...
10453 (make_pass_cleanup_eh): ...new function.
10454 (pass_sra): Replace declaration with that of...
10455 (make_pass_sra): ...new function.
10456 (pass_sra_early): Replace declaration with that of...
10457 (make_pass_sra_early): ...new function.
10458 (pass_early_ipa_sra): Replace declaration with that of...
10459 (make_pass_early_ipa_sra): ...new function.
10460 (pass_tail_recursion): Replace declaration with that of...
10461 (make_pass_tail_recursion): ...new function.
10462 (pass_tail_calls): Replace declaration with that of...
10463 (make_pass_tail_calls): ...new function.
10464 (pass_tree_loop): Replace declaration with that of...
10465 (make_pass_tree_loop): ...new function.
10466 (pass_tree_loop_init): Replace declaration with that of...
10467 (make_pass_tree_loop_init): ...new function.
10468 (pass_lim): Replace declaration with that of...
10469 (make_pass_lim): ...new function.
10470 (pass_tree_unswitch): Replace declaration with that of...
10471 (make_pass_tree_unswitch): ...new function.
10472 (pass_predcom): Replace declaration with that of...
10473 (make_pass_predcom): ...new function.
10474 (pass_iv_canon): Replace declaration with that of...
10475 (make_pass_iv_canon): ...new function.
10476 (pass_scev_cprop): Replace declaration with that of...
10477 (make_pass_scev_cprop): ...new function.
10478 (pass_empty_loop): Replace declaration with that of...
10479 (make_pass_empty_loop): ...new function.
10480 (pass_record_bounds): Replace declaration with that of...
10481 (make_pass_record_bounds): ...new function.
10482 (pass_graphite): Replace declaration with that of...
10483 (make_pass_graphite): ...new function.
10484 (pass_graphite_transforms): Replace declaration with that of...
10485 (make_pass_graphite_transforms): ...new function.
10486 (pass_if_conversion): Replace declaration with that of...
10487 (make_pass_if_conversion): ...new function.
10488 (pass_loop_distribution): Replace declaration with that of...
10489 (make_pass_loop_distribution): ...new function.
10490 (pass_vectorize): Replace declaration with that of...
10491 (make_pass_vectorize): ...new function.
10492 (pass_slp_vectorize): Replace declaration with that of...
10493 (make_pass_slp_vectorize): ...new function.
10494 (pass_complete_unroll): Replace declaration with that of...
10495 (make_pass_complete_unroll): ...new function.
10496 (pass_complete_unrolli): Replace declaration with that of...
10497 (make_pass_complete_unrolli): ...new function.
10498 (pass_parallelize_loops): Replace declaration with that of...
10499 (make_pass_parallelize_loops): ...new function.
10500 (pass_loop_prefetch): Replace declaration with that of...
10501 (make_pass_loop_prefetch): ...new function.
10502 (pass_iv_optimize): Replace declaration with that of...
10503 (make_pass_iv_optimize): ...new function.
10504 (pass_tree_loop_done): Replace declaration with that of...
10505 (make_pass_tree_loop_done): ...new function.
10506 (pass_ch): Replace declaration with that of...
10507 (make_pass_ch): ...new function.
10508 (pass_ccp): Replace declaration with that of...
10509 (make_pass_ccp): ...new function.
10510 (pass_phi_only_cprop): Replace declaration with that of...
10511 (make_pass_phi_only_cprop): ...new function.
10512 (pass_build_ssa): Replace declaration with that of...
10513 (make_pass_build_ssa): ...new function.
10514 (pass_build_alias): Replace declaration with that of...
10515 (make_pass_build_alias): ...new function.
10516 (pass_build_ealias): Replace declaration with that of...
10517 (make_pass_build_ealias): ...new function.
10518 (pass_dominator): Replace declaration with that of...
10519 (make_pass_dominator): ...new function.
10520 (pass_dce): Replace declaration with that of...
10521 (make_pass_dce): ...new function.
10522 (pass_dce_loop): Replace declaration with that of...
10523 (make_pass_dce_loop): ...new function.
10524 (pass_cd_dce): Replace declaration with that of...
10525 (make_pass_cd_dce): ...new function.
10526 (pass_call_cdce): Replace declaration with that of...
10527 (make_pass_call_cdce): ...new function.
10528 (pass_merge_phi): Replace declaration with that of...
10529 (make_pass_merge_phi): ...new function.
10530 (pass_split_crit_edges): Replace declaration with that of...
10531 (make_pass_split_crit_edges): ...new function.
10532 (pass_pre): Replace declaration with that of...
10533 (make_pass_pre): ...new function.
10534 (pass_profile): Replace declaration with that of...
10535 (make_pass_profile): ...new function.
10536 (pass_strip_predict_hints): Replace declaration with that of...
10537 (make_pass_strip_predict_hints): ...new function.
10538 (pass_lower_complex_O0): Replace declaration with that of...
10539 (make_pass_lower_complex_O0): ...new function.
10540 (pass_lower_complex): Replace declaration with that of...
10541 (make_pass_lower_complex): ...new function.
10542 (pass_lower_vector): Replace declaration with that of...
10543 (make_pass_lower_vector): ...new function.
10544 (pass_lower_vector_ssa): Replace declaration with that of...
10545 (make_pass_lower_vector_ssa): ...new function.
10546 (pass_lower_omp): Replace declaration with that of...
10547 (make_pass_lower_omp): ...new function.
10548 (pass_diagnose_omp_blocks): Replace declaration with that of...
10549 (make_pass_diagnose_omp_blocks): ...new function.
10550 (pass_expand_omp): Replace declaration with that of...
10551 (make_pass_expand_omp): ...new function.
10552 (pass_expand_omp_ssa): Replace declaration with that of...
10553 (make_pass_expand_omp_ssa): ...new function.
10554 (pass_object_sizes): Replace declaration with that of...
10555 (make_pass_object_sizes): ...new function.
10556 (pass_strlen): Replace declaration with that of...
10557 (make_pass_strlen): ...new function.
10558 (pass_fold_builtins): Replace declaration with that of...
10559 (make_pass_fold_builtins): ...new function.
10560 (pass_stdarg): Replace declaration with that of...
10561 (make_pass_stdarg): ...new function.
10562 (pass_early_warn_uninitialized): Replace declaration with that of...
10563 (make_pass_early_warn_uninitialized): ...new function.
10564 (pass_late_warn_uninitialized): Replace declaration with that of...
10565 (make_pass_late_warn_uninitialized): ...new function.
10566 (pass_cse_reciprocals): Replace declaration with that of...
10567 (make_pass_cse_reciprocals): ...new function.
10568 (pass_cse_sincos): Replace declaration with that of...
10569 (make_pass_cse_sincos): ...new function.
10570 (pass_optimize_bswap): Replace declaration with that of...
10571 (make_pass_optimize_bswap): ...new function.
10572 (pass_optimize_widening_mul): Replace declaration with that of...
10573 (make_pass_optimize_widening_mul): ...new function.
10574 (pass_warn_function_return): Replace declaration with that of...
10575 (make_pass_warn_function_return): ...new function.
10576 (pass_warn_function_noreturn): Replace declaration with that of...
10577 (make_pass_warn_function_noreturn): ...new function.
10578 (pass_cselim): Replace declaration with that of...
10579 (make_pass_cselim): ...new function.
10580 (pass_phiopt): Replace declaration with that of...
10581 (make_pass_phiopt): ...new function.
10582 (pass_forwprop): Replace declaration with that of...
10583 (make_pass_forwprop): ...new function.
10584 (pass_phiprop): Replace declaration with that of...
10585 (make_pass_phiprop): ...new function.
10586 (pass_tree_ifcombine): Replace declaration with that of...
10587 (make_pass_tree_ifcombine): ...new function.
10588 (pass_dse): Replace declaration with that of...
10589 (make_pass_dse): ...new function.
10590 (pass_nrv): Replace declaration with that of...
10591 (make_pass_nrv): ...new function.
10592 (pass_rename_ssa_copies): Replace declaration with that of...
10593 (make_pass_rename_ssa_copies): ...new function.
10594 (pass_sink_code): Replace declaration with that of...
10595 (make_pass_sink_code): ...new function.
10596 (pass_fre): Replace declaration with that of...
10597 (make_pass_fre): ...new function.
10598 (pass_check_data_deps): Replace declaration with that of...
10599 (make_pass_check_data_deps): ...new function.
10600 (pass_copy_prop): Replace declaration with that of...
10601 (make_pass_copy_prop): ...new function.
10602 (pass_vrp): Replace declaration with that of...
10603 (make_pass_vrp): ...new function.
10604 (pass_uncprop): Replace declaration with that of...
10605 (make_pass_uncprop): ...new function.
10606 (pass_return_slot): Replace declaration with that of...
10607 (make_pass_return_slot): ...new function.
10608 (pass_reassoc): Replace declaration with that of...
10609 (make_pass_reassoc): ...new function.
10610 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10611 (make_pass_rebuild_cgraph_edges): ...new function.
10612 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10613 (make_pass_remove_cgraph_callee_edges): ...new function.
10614 (pass_build_cgraph_edges): Replace declaration with that of...
10615 (make_pass_build_cgraph_edges): ...new function.
10616 (pass_local_pure_const): Replace declaration with that of...
10617 (make_pass_local_pure_const): ...new function.
10618 (pass_tracer): Replace declaration with that of...
10619 (make_pass_tracer): ...new function.
10620 (pass_warn_unused_result): Replace declaration with that of...
10621 (make_pass_warn_unused_result): ...new function.
10622 (pass_diagnose_tm_blocks): Replace declaration with that of...
10623 (make_pass_diagnose_tm_blocks): ...new function.
10624 (pass_lower_tm): Replace declaration with that of...
10625 (make_pass_lower_tm): ...new function.
10626 (pass_tm_init): Replace declaration with that of...
10627 (make_pass_tm_init): ...new function.
10628 (pass_tm_mark): Replace declaration with that of...
10629 (make_pass_tm_mark): ...new function.
10630 (pass_tm_memopt): Replace declaration with that of...
10631 (make_pass_tm_memopt): ...new function.
10632 (pass_tm_edges): Replace declaration with that of...
10633 (make_pass_tm_edges): ...new function.
10634 (pass_split_functions): Replace declaration with that of...
10635 (make_pass_split_functions): ...new function.
10636 (pass_feedback_split_functions): Replace declaration with that of...
10637 (make_pass_feedback_split_functions): ...new function.
10638 (pass_strength_reduction): Replace declaration with that of...
10639 (make_pass_strength_reduction): ...new function.
10640 (pass_ipa_lower_emutls): Replace declaration with that of...
10641 (make_pass_ipa_lower_emutls): ...new function.
10642 (pass_ipa_function_and_variable_visibility): Replace declaration with
10644 (make_pass_ipa_function_and_variable_visibility): ...new function.
10645 (pass_ipa_tree_profile): Replace declaration with that of...
10646 (make_pass_ipa_tree_profile): ...new function.
10647 (pass_early_local_passes): Replace declaration with that of...
10648 (make_pass_early_local_passes): ...new function.
10649 (pass_ipa_whole_program_visibility): Replace declaration with that
10651 (make_pass_ipa_whole_program_visibility): ...new function.
10652 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10653 (make_pass_ipa_lto_gimple_out): ...new function.
10654 (pass_ipa_increase_alignment): Replace declaration with that of...
10655 (make_pass_ipa_increase_alignment): ...new function.
10656 (pass_ipa_inline): Replace declaration with that of...
10657 (make_pass_ipa_inline): ...new function.
10658 (pass_ipa_free_lang_data): Replace declaration with that of...
10659 (make_pass_ipa_free_lang_data): ...new function.
10660 (pass_ipa_free_inline_summary): Replace declaration with that of...
10661 (make_pass_ipa_free_inline_summary): ...new function.
10662 (pass_ipa_cp): Replace declaration with that of...
10663 (make_pass_ipa_cp): ...new function.
10664 (pass_ipa_reference): Replace declaration with that of...
10665 (make_pass_ipa_reference): ...new function.
10666 (pass_ipa_pure_const): Replace declaration with that of...
10667 (make_pass_ipa_pure_const): ...new function.
10668 (pass_ipa_pta): Replace declaration with that of...
10669 (make_pass_ipa_pta): ...new function.
10670 (pass_ipa_lto_finish_out): Replace declaration with that of...
10671 (make_pass_ipa_lto_finish_out): ...new function.
10672 (pass_ipa_tm): Replace declaration with that of...
10673 (make_pass_ipa_tm): ...new function.
10674 (pass_ipa_profile): Replace declaration with that of...
10675 (make_pass_ipa_profile): ...new function.
10676 (pass_ipa_cdtor_merge): Replace declaration with that of...
10677 (make_pass_ipa_cdtor_merge): ...new function.
10678 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10680 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10681 (pass_init_datastructures): Replace declaration with that of...
10682 (make_pass_init_datastructures): ...new function.
10683 (pass_fixup_cfg): Replace declaration with that of...
10684 (make_pass_fixup_cfg): ...new function.
10685 (pass_expand): Replace declaration with that of...
10686 (make_pass_expand): ...new function.
10687 (pass_instantiate_virtual_regs): Replace declaration with that of...
10688 (make_pass_instantiate_virtual_regs): ...new function.
10689 (pass_rtl_fwprop): Replace declaration with that of...
10690 (make_pass_rtl_fwprop): ...new function.
10691 (pass_rtl_fwprop_addr): Replace declaration with that of...
10692 (make_pass_rtl_fwprop_addr): ...new function.
10693 (pass_jump): Replace declaration with that of...
10694 (make_pass_jump): ...new function.
10695 (pass_jump2): Replace declaration with that of...
10696 (make_pass_jump2): ...new function.
10697 (pass_lower_subreg): Replace declaration with that of...
10698 (make_pass_lower_subreg): ...new function.
10699 (pass_cse): Replace declaration with that of...
10700 (make_pass_cse): ...new function.
10701 (pass_fast_rtl_dce): Replace declaration with that of...
10702 (make_pass_fast_rtl_dce): ...new function.
10703 (pass_ud_rtl_dce): Replace declaration with that of...
10704 (make_pass_ud_rtl_dce): ...new function.
10705 (pass_rtl_dce): Replace declaration with that of...
10706 (make_pass_rtl_dce): ...new function.
10707 (pass_rtl_dse1): Replace declaration with that of...
10708 (make_pass_rtl_dse1): ...new function.
10709 (pass_rtl_dse2): Replace declaration with that of...
10710 (make_pass_rtl_dse2): ...new function.
10711 (pass_rtl_dse3): Replace declaration with that of...
10712 (make_pass_rtl_dse3): ...new function.
10713 (pass_rtl_cprop): Replace declaration with that of...
10714 (make_pass_rtl_cprop): ...new function.
10715 (pass_rtl_pre): Replace declaration with that of...
10716 (make_pass_rtl_pre): ...new function.
10717 (pass_rtl_hoist): Replace declaration with that of...
10718 (make_pass_rtl_hoist): ...new function.
10719 (pass_rtl_store_motion): Replace declaration with that of...
10720 (make_pass_rtl_store_motion): ...new function.
10721 (pass_cse_after_global_opts): Replace declaration with that of...
10722 (make_pass_cse_after_global_opts): ...new function.
10723 (pass_rtl_ifcvt): Replace declaration with that of...
10724 (make_pass_rtl_ifcvt): ...new function.
10725 (pass_into_cfg_layout_mode): Replace declaration with that of...
10726 (make_pass_into_cfg_layout_mode): ...new function.
10727 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10728 (make_pass_outof_cfg_layout_mode): ...new function.
10729 (pass_loop2): Replace declaration with that of...
10730 (make_pass_loop2): ...new function.
10731 (pass_rtl_loop_init): Replace declaration with that of...
10732 (make_pass_rtl_loop_init): ...new function.
10733 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10734 (make_pass_rtl_move_loop_invariants): ...new function.
10735 (pass_rtl_unswitch): Replace declaration with that of...
10736 (make_pass_rtl_unswitch): ...new function.
10737 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10738 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10739 (pass_rtl_doloop): Replace declaration with that of...
10740 (make_pass_rtl_doloop): ...new function.
10741 (pass_rtl_loop_done): Replace declaration with that of...
10742 (make_pass_rtl_loop_done): ...new function.
10743 (pass_web): Replace declaration with that of...
10744 (make_pass_web): ...new function.
10745 (pass_cse2): Replace declaration with that of...
10746 (make_pass_cse2): ...new function.
10747 (pass_df_initialize_opt): Replace declaration with that of...
10748 (make_pass_df_initialize_opt): ...new function.
10749 (pass_df_initialize_no_opt): Replace declaration with that of...
10750 (make_pass_df_initialize_no_opt): ...new function.
10751 (pass_reginfo_init): Replace declaration with that of...
10752 (make_pass_reginfo_init): ...new function.
10753 (pass_inc_dec): Replace declaration with that of...
10754 (make_pass_inc_dec): ...new function.
10755 (pass_stack_ptr_mod): Replace declaration with that of...
10756 (make_pass_stack_ptr_mod): ...new function.
10757 (pass_initialize_regs): Replace declaration with that of...
10758 (make_pass_initialize_regs): ...new function.
10759 (pass_combine): Replace declaration with that of...
10760 (make_pass_combine): ...new function.
10761 (pass_if_after_combine): Replace declaration with that of...
10762 (make_pass_if_after_combine): ...new function.
10763 (pass_ree): Replace declaration with that of...
10764 (make_pass_ree): ...new function.
10765 (pass_partition_blocks): Replace declaration with that of...
10766 (make_pass_partition_blocks): ...new function.
10767 (pass_match_asm_constraints): Replace declaration with that of...
10768 (make_pass_match_asm_constraints): ...new function.
10769 (pass_regmove): Replace declaration with that of...
10770 (make_pass_regmove): ...new function.
10771 (pass_split_all_insns): Replace declaration with that of...
10772 (make_pass_split_all_insns): ...new function.
10773 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10774 (make_pass_fast_rtl_byte_dce): ...new function.
10775 (pass_lower_subreg2): Replace declaration with that of...
10776 (make_pass_lower_subreg2): ...new function.
10777 (pass_mode_switching): Replace declaration with that of...
10778 (make_pass_mode_switching): ...new function.
10779 (pass_sms): Replace declaration with that of...
10780 (make_pass_sms): ...new function.
10781 (pass_sched): Replace declaration with that of...
10782 (make_pass_sched): ...new function.
10783 (pass_ira): Replace declaration with that of...
10784 (make_pass_ira): ...new function.
10785 (pass_reload): Replace declaration with that of...
10786 (make_pass_reload): ...new function.
10787 (pass_clean_state): Replace declaration with that of...
10788 (make_pass_clean_state): ...new function.
10789 (pass_branch_prob): Replace declaration with that of...
10790 (make_pass_branch_prob): ...new function.
10791 (pass_value_profile_transformations): Replace declaration with that
10793 (make_pass_value_profile_transformations): ...new function.
10794 (pass_postreload_cse): Replace declaration with that of...
10795 (make_pass_postreload_cse): ...new function.
10796 (pass_gcse2): Replace declaration with that of...
10797 (make_pass_gcse2): ...new function.
10798 (pass_split_after_reload): Replace declaration with that of...
10799 (make_pass_split_after_reload): ...new function.
10800 (pass_branch_target_load_optimize1): Replace declaration with that
10802 (make_pass_branch_target_load_optimize1): ...new function.
10803 (pass_thread_prologue_and_epilogue): Replace declaration with that
10805 (make_pass_thread_prologue_and_epilogue): ...new function.
10806 (pass_stack_adjustments): Replace declaration with that of...
10807 (make_pass_stack_adjustments): ...new function.
10808 (pass_peephole2): Replace declaration with that of...
10809 (make_pass_peephole2): ...new function.
10810 (pass_if_after_reload): Replace declaration with that of...
10811 (make_pass_if_after_reload): ...new function.
10812 (pass_regrename): Replace declaration with that of...
10813 (make_pass_regrename): ...new function.
10814 (pass_cprop_hardreg): Replace declaration with that of...
10815 (make_pass_cprop_hardreg): ...new function.
10816 (pass_reorder_blocks): Replace declaration with that of...
10817 (make_pass_reorder_blocks): ...new function.
10818 (pass_branch_target_load_optimize2): Replace declaration with that
10820 (make_pass_branch_target_load_optimize2): ...new function.
10821 (pass_leaf_regs): Replace declaration with that of...
10822 (make_pass_leaf_regs): ...new function.
10823 (pass_split_before_sched2): Replace declaration with that of...
10824 (make_pass_split_before_sched2): ...new function.
10825 (pass_compare_elim_after_reload): Replace declaration with that of...
10826 (make_pass_compare_elim_after_reload): ...new function.
10827 (pass_sched2): Replace declaration with that of...
10828 (make_pass_sched2): ...new function.
10829 (pass_stack_regs): Replace declaration with that of...
10830 (make_pass_stack_regs): ...new function.
10831 (pass_stack_regs_run): Replace declaration with that of...
10832 (make_pass_stack_regs_run): ...new function.
10833 (pass_df_finish): Replace declaration with that of...
10834 (make_pass_df_finish): ...new function.
10835 (pass_compute_alignments): Replace declaration with that of...
10836 (make_pass_compute_alignments): ...new function.
10837 (pass_duplicate_computed_gotos): Replace declaration with that of...
10838 (make_pass_duplicate_computed_gotos): ...new function.
10839 (pass_variable_tracking): Replace declaration with that of...
10840 (make_pass_variable_tracking): ...new function.
10841 (pass_free_cfg): Replace declaration with that of...
10842 (make_pass_free_cfg): ...new function.
10843 (pass_machine_reorg): Replace declaration with that of...
10844 (make_pass_machine_reorg): ...new function.
10845 (pass_cleanup_barriers): Replace declaration with that of...
10846 (make_pass_cleanup_barriers): ...new function.
10847 (pass_delay_slots): Replace declaration with that of...
10848 (make_pass_delay_slots): ...new function.
10849 (pass_split_for_shorten_branches): Replace declaration with that of...
10850 (make_pass_split_for_shorten_branches): ...new function.
10851 (pass_split_before_regstack): Replace declaration with that of...
10852 (make_pass_split_before_regstack): ...new function.
10853 (pass_convert_to_eh_region_ranges): Replace declaration with that
10855 (make_pass_convert_to_eh_region_ranges): ...new function.
10856 (pass_shorten_branches): Replace declaration with that of...
10857 (make_pass_shorten_branches): ...new function.
10858 (pass_set_nothrow_function_flags): Replace declaration with that of...
10859 (make_pass_set_nothrow_function_flags): ...new function.
10860 (pass_dwarf2_frame): Replace declaration with that of...
10861 (make_pass_dwarf2_frame): ...new function.
10862 (pass_final): Replace declaration with that of...
10863 (make_pass_final): ...new function.
10864 (pass_rtl_seqabstr): Replace declaration with that of...
10865 (make_pass_rtl_seqabstr): ...new function.
10866 (pass_release_ssa_names): Replace declaration with that of...
10867 (make_pass_release_ssa_names): ...new function.
10868 (pass_early_inline): Replace declaration with that of...
10869 (make_pass_early_inline): ...new function.
10870 (pass_inline_parameters): Replace declaration with that of...
10871 (make_pass_inline_parameters): ...new function.
10872 (pass_update_address_taken): Replace declaration with that of...
10873 (make_pass_update_address_taken): ...new function.
10874 (pass_convert_switch): Replace declaration with that of...
10875 (make_pass_convert_switch): ...new function.
10876 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10877 to a subclass of simple_ipa_opt_pass along with...
10878 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10879 (make_pass_ipa_tree_profile): ...new function.
10880 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10881 subclass of gimple_opt_pass along with...
10882 (pass_data_sra_early): ...new pass_data instance and...
10883 (make_pass_sra_early): ...new function.
10884 (pass_sra): Convert from a global struct to a subclass of
10885 gimple_opt_pass along with...
10886 (pass_data_sra): ...new pass_data instance and...
10887 (make_pass_sra): ...new function.
10888 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10889 gimple_opt_pass along with...
10890 (pass_data_early_ipa_sra): ...new pass_data instance and...
10891 (make_pass_early_ipa_sra): ...new function.
10892 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10893 subclass of gimple_opt_pass along with...
10894 (pass_data_ccp): ...new pass_data instance and...
10895 (make_pass_ccp): ...new function.
10896 (pass_fold_builtins): Convert from a global struct to a subclass of
10897 gimple_opt_pass along with...
10898 (pass_data_fold_builtins): ...new pass_data instance and...
10899 (make_pass_fold_builtins): ...new function.
10900 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10901 subclass of gimple_opt_pass along with...
10902 (pass_data_copy_prop): ...new pass_data instance and...
10903 (make_pass_copy_prop): ...new function.
10904 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10905 global struct to a subclass of gimple_opt_pass along with...
10906 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10907 (make_pass_rename_ssa_copies): ...new function.
10908 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10909 subclass of gimple_opt_pass along with...
10910 (pass_data_dce): ...new pass_data instance and...
10911 (make_pass_dce): ...new function.
10912 (pass_dce_loop): Convert from a global struct to a subclass of
10913 gimple_opt_pass along with...
10914 (pass_data_dce_loop): ...new pass_data instance and...
10915 (make_pass_dce_loop): ...new function.
10916 (pass_cd_dce): Convert from a global struct to a subclass of
10917 gimple_opt_pass along with...
10918 (pass_data_cd_dce): ...new pass_data instance and...
10919 (make_pass_cd_dce): ...new function.
10920 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10921 subclass of gimple_opt_pass along with...
10922 (pass_data_dominator): ...new pass_data instance and...
10923 (make_pass_dominator): ...new function.
10924 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10925 gimple_opt_pass along with...
10926 (pass_data_phi_only_cprop): ...new pass_data instance and...
10927 (make_pass_phi_only_cprop): ...new function.
10928 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10929 subclass of gimple_opt_pass along with...
10930 (pass_data_dse): ...new pass_data instance and...
10931 (make_pass_dse): ...new function.
10932 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10933 a subclass of gimple_opt_pass along with...
10934 (pass_data_forwprop): ...new pass_data instance and...
10935 (make_pass_forwprop): ...new function.
10936 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10937 struct to a subclass of gimple_opt_pass along with...
10938 (pass_data_tree_ifcombine): ...new pass_data instance and...
10939 (make_pass_tree_ifcombine): ...new function.
10940 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10941 subclass of gimple_opt_pass along with...
10942 (pass_data_ch): ...new pass_data instance and...
10943 (make_pass_ch): ...new function.
10944 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10945 subclass of gimple_opt_pass along with...
10946 (pass_data_tree_loop): ...new pass_data instance and...
10947 (make_pass_tree_loop): ...new function.
10948 (pass_tree_loop_init): Convert from a global struct to a subclass of
10949 gimple_opt_pass along with...
10950 (pass_data_tree_loop_init): ...new pass_data instance and...
10951 (make_pass_tree_loop_init): ...new function.
10952 (pass_lim): Convert from a global struct to a subclass of
10953 gimple_opt_pass along with...
10954 (pass_data_lim): ...new pass_data instance and...
10955 (make_pass_lim): ...new function.
10956 (pass_tree_unswitch): Convert from a global struct to a subclass of
10957 gimple_opt_pass along with...
10958 (pass_data_tree_unswitch): ...new pass_data instance and...
10959 (make_pass_tree_unswitch): ...new function.
10960 (pass_predcom): Convert from a global struct to a subclass of
10961 gimple_opt_pass along with...
10962 (pass_data_predcom): ...new pass_data instance and...
10963 (make_pass_predcom): ...new function.
10964 (pass_vectorize): Convert from a global struct to a subclass of
10965 gimple_opt_pass along with...
10966 (pass_data_vectorize): ...new pass_data instance and...
10967 (make_pass_vectorize): ...new function.
10968 (pass_graphite): Convert from a global struct to a subclass of
10969 gimple_opt_pass along with...
10970 (pass_data_graphite): ...new pass_data instance and...
10971 (make_pass_graphite): ...new function.
10972 (pass_graphite_transforms): Convert from a global struct to a subclass
10973 of gimple_opt_pass along with...
10974 (pass_data_graphite_transforms): ...new pass_data instance and...
10975 (make_pass_graphite_transforms): ...new function.
10976 (pass_check_data_deps): Convert from a global struct to a subclass of
10977 gimple_opt_pass along with...
10978 (pass_data_check_data_deps): ...new pass_data instance and...
10979 (make_pass_check_data_deps): ...new function.
10980 (pass_iv_canon): Convert from a global struct to a subclass of
10981 gimple_opt_pass along with...
10982 (pass_data_iv_canon): ...new pass_data instance and...
10983 (make_pass_iv_canon): ...new function.
10984 (pass_scev_cprop): Convert from a global struct to a subclass of
10985 gimple_opt_pass along with...
10986 (pass_data_scev_cprop): ...new pass_data instance and...
10987 (make_pass_scev_cprop): ...new function.
10988 (pass_record_bounds): Convert from a global struct to a subclass of
10989 gimple_opt_pass along with...
10990 (pass_data_record_bounds): ...new pass_data instance and...
10991 (make_pass_record_bounds): ...new function.
10992 (pass_complete_unroll): Convert from a global struct to a subclass of
10993 gimple_opt_pass along with...
10994 (pass_data_complete_unroll): ...new pass_data instance and...
10995 (make_pass_complete_unroll): ...new function.
10996 (pass_complete_unrolli): Convert from a global struct to a subclass of
10997 gimple_opt_pass along with...
10998 (pass_data_complete_unrolli): ...new pass_data instance and...
10999 (make_pass_complete_unrolli): ...new function.
11000 (pass_parallelize_loops): Convert from a global struct to a subclass
11001 of gimple_opt_pass along with...
11002 (pass_data_parallelize_loops): ...new pass_data instance and...
11003 (make_pass_parallelize_loops): ...new function.
11004 (pass_loop_prefetch): Convert from a global struct to a subclass of
11005 gimple_opt_pass along with...
11006 (pass_data_loop_prefetch): ...new pass_data instance and...
11007 (make_pass_loop_prefetch): ...new function.
11008 (pass_iv_optimize): Convert from a global struct to a subclass of
11009 gimple_opt_pass along with...
11010 (pass_data_iv_optimize): ...new pass_data instance and...
11011 (make_pass_iv_optimize): ...new function.
11012 (pass_tree_loop_done): Convert from a global struct to a subclass of
11013 gimple_opt_pass along with...
11014 (pass_data_tree_loop_done): ...new pass_data instance and...
11015 (make_pass_tree_loop_done): ...new function.
11016 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11017 struct to a subclass of gimple_opt_pass along with...
11018 (pass_data_cse_reciprocals): ...new pass_data instance and...
11019 (make_pass_cse_reciprocals): ...new function.
11020 (pass_cse_sincos): Convert from a global struct to a subclass of
11021 gimple_opt_pass along with...
11022 (pass_data_cse_sincos): ...new pass_data instance and...
11023 (make_pass_cse_sincos): ...new function.
11024 (pass_optimize_bswap): Convert from a global struct to a subclass of
11025 gimple_opt_pass along with...
11026 (pass_data_optimize_bswap): ...new pass_data instance and...
11027 (make_pass_optimize_bswap): ...new function.
11028 (pass_optimize_widening_mul): Convert from a global struct to a
11029 subclass of gimple_opt_pass along with...
11030 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11031 (make_pass_optimize_widening_mul): ...new function.
11032 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11033 subclass of gimple_opt_pass along with...
11034 (pass_data_phiopt): ...new pass_data instance and...
11035 (make_pass_phiopt): ...new function.
11036 (pass_cselim): Convert from a global struct to a subclass of
11037 gimple_opt_pass along with...
11038 (pass_data_cselim): ...new pass_data instance and...
11039 (make_pass_cselim): ...new function.
11040 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11041 subclass of gimple_opt_pass along with...
11042 (pass_data_phiprop): ...new pass_data instance and...
11043 (make_pass_phiprop): ...new function.
11044 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11045 subclass of gimple_opt_pass along with...
11046 (pass_data_pre): ...new pass_data instance and...
11047 (make_pass_pre): ...new function.
11048 (pass_fre): Convert from a global struct to a subclass of
11049 gimple_opt_pass along with...
11050 (pass_data_fre): ...new pass_data instance and...
11051 (make_pass_fre): ...new function.
11052 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11053 subclass of gimple_opt_pass along with...
11054 (pass_data_reassoc): ...new pass_data instance and...
11055 (make_pass_reassoc): ...new function.
11056 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11057 subclass of gimple_opt_pass along with...
11058 (pass_data_sink_code): ...new pass_data instance and...
11059 (make_pass_sink_code): ...new function.
11060 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11061 subclass of gimple_opt_pass along with...
11062 (pass_data_strlen): ...new pass_data instance and...
11063 (make_pass_strlen): ...new function.
11064 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11065 struct to a subclass of gimple_opt_pass along with...
11066 (pass_data_build_alias): ...new pass_data instance and...
11067 (make_pass_build_alias): ...new function.
11068 (pass_build_ealias): Convert from a global struct to a subclass of
11069 gimple_opt_pass along with...
11070 (pass_data_build_ealias): ...new pass_data instance and...
11071 (make_pass_build_ealias): ...new function.
11072 (pass_ipa_pta): Convert from a global struct to a subclass of
11073 simple_ipa_opt_pass along with...
11074 (pass_data_ipa_pta): ...new pass_data instance and...
11075 (make_pass_ipa_pta): ...new function.
11076 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11077 subclass of gimple_opt_pass along with...
11078 (pass_data_uncprop): ...new pass_data instance and...
11079 (make_pass_uncprop): ...new function.
11080 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11081 global struct to a subclass of gimple_opt_pass along with...
11082 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11083 (make_pass_late_warn_uninitialized): ...new function.
11084 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11085 to a subclass of gimple_opt_pass along with...
11086 (pass_data_init_datastructures): ...new pass_data instance and...
11087 (make_pass_init_datastructures): ...new function.
11088 (pass_early_warn_uninitialized): Convert from a global struct to a
11089 subclass of gimple_opt_pass along with...
11090 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11091 (make_pass_early_warn_uninitialized): ...new function.
11092 (pass_update_address_taken): Convert from a global struct to a
11093 subclass of gimple_opt_pass along with...
11094 (pass_data_update_address_taken): ...new pass_data instance and...
11095 (make_pass_update_address_taken): ...new function.
11096 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11097 struct to a subclass of gimple_opt_pass along with...
11098 (pass_data_release_ssa_names): ...new pass_data instance and...
11099 (make_pass_release_ssa_names): ...new function.
11100 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11101 subclass of gimple_opt_pass along with...
11102 (pass_data_stdarg): ...new pass_data instance and...
11103 (make_pass_stdarg): ...new function.
11104 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11105 global struct to a subclass of gimple_opt_pass along with...
11106 (pass_data_convert_switch): ...new pass_data instance and...
11107 (make_pass_convert_switch): ...new function.
11108 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11109 to a subclass of gimple_opt_pass along with...
11110 (pass_data_tail_recursion): ...new pass_data instance and...
11111 (make_pass_tail_recursion): ...new function.
11112 (pass_tail_calls): Convert from a global struct to a subclass of
11113 gimple_opt_pass along with...
11114 (pass_data_tail_calls): ...new pass_data instance and...
11115 (make_pass_tail_calls): ...new function.
11116 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11117 struct to a subclass of gimple_opt_pass along with...
11118 (pass_data_lower_vector): ...new pass_data instance and...
11119 (make_pass_lower_vector): ...new function.
11120 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11121 gimple_opt_pass along with...
11122 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11123 (make_pass_lower_vector_ssa): ...new function.
11124 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11125 to a subclass of gimple_opt_pass along with...
11126 (pass_data_slp_vectorize): ...new pass_data instance and...
11127 (make_pass_slp_vectorize): ...new function.
11128 (pass_ipa_increase_alignment): Convert from a global struct to a
11129 subclass of simple_ipa_opt_pass along with...
11130 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11131 (make_pass_ipa_increase_alignment): ...new function.
11132 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11133 gimple_opt_pass along with...
11134 (pass_data_vrp): ...new pass_data instance and...
11135 (make_pass_vrp): ...new function.
11136 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11137 subclass of simple_ipa_opt_pass along with...
11138 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11139 (make_pass_ipa_free_lang_data): ...new function.
11140 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11141 gimple_opt_pass along with...
11142 (pass_data_tsan): ...new pass_data instance and...
11143 (make_pass_tsan): ...new function.
11144 (pass_tsan_O0): Convert from a global struct to a subclass of
11145 gimple_opt_pass along with...
11146 (pass_data_tsan_O0): ...new pass_data instance and...
11147 (make_pass_tsan_O0): ...new function.
11148 * var-tracking.c (pass_variable_tracking): Convert from a global
11149 struct to a subclass of rtl_opt_pass along with...
11150 (pass_data_variable_tracking): ...new pass_data instance and...
11151 (make_pass_variable_tracking): ...new function.
11152 * web.c (pass_web): Convert from a global struct to a subclass of
11153 rtl_opt_pass along with...
11154 (pass_data_web): ...new pass_data instance and...
11155 (make_pass_web): ...new function.
11156 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11157 declaration with that of...
11158 (make_pass_mode_switch_use): ...new function.
11159 (pass_resolve_sw_modes): Replace declaration with that of...
11160 (make_pass_resolve_sw_modes): ...new function.
11161 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11162 from a global struct to a subclass of rtl_opt_pass along with...
11163 (pass_data_mode_switch_use): ...new pass_data instance and...
11164 (make_pass_mode_switch_use): ...new function.
11165 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11166 from a global struct to a subclass of rtl_opt_pass along with...
11167 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11168 (make_pass_resolve_sw_modes): ...new function.
11169 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11170 struct to a subclass of rtl_opt_pass along with...
11171 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11172 (make_pass_insert_vzeroupper): ...new function.
11173 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11174 global struct to a subclass of rtl_opt_pass along with...
11175 (pass_data_work_around_errata): ...new pass_data instance and...
11176 (make_pass_work_around_errata): ...new function.
11177 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11178 struct to a subclass of rtl_opt_pass along with...
11179 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11180 (make_pass_mips_machine_reorg2): ...new function.
11182 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11184 * passes.c (pass_manager::operator new): New.
11186 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11188 Handwritten part of conversion of passes to C++ classes.
11190 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11191 (toplev.o): Add dep on PASS_MANAGER_H.
11192 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11193 of early local pases to reflect this moving from a global to a
11194 member of gcc::pass_manager.
11195 (cgraph_add_new_function): Likewise.
11196 * lto-cgraph.c (lto_output_node): Update for conversion of
11197 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11198 * passes.c (opt_pass::clone): New.
11199 (opt_pass::gate): New.
11200 (opt_pass::execute): New.
11201 (opt_pass::opt_pass): New.
11202 (pass_manager::execute_early_local_passes): New.
11203 (pass_manager::execute_pass_mode_switching): new.
11204 (finish_optimization_passes): Convert to...
11205 (pass_manager::finish_optimization_passes): ...this.
11206 (finish_optimization_passes): Update for conversion of passes to
11208 (register_dump_files_1): Use has_gate since we cannot portably
11209 check a vtable entry against NULL.
11210 (dump_one_pass): Likewise.
11211 (ipa_write_summaries_2): Likewise.
11212 (ipa_write_optimization_summaries_1): Likewise.
11213 (ipa_read_summaries_1): Likewise.
11214 (ipa_read_optimization_summaries_1): Likewise.
11215 (execute_ipa_stmt_fixups): Likewise.
11216 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11217 pass-creation functions rather than wiring up globals, and
11218 storing the results in fields of pass_manager generated using
11219 pass-instances.def.
11220 (pass_manager::dump_profile_report): Update for conversion of
11221 passes to C++ classes.
11222 (pass_manager::execute_ipa_summary_passes): Likewise.
11223 (execute_one_ipa_transform_pass): Likewise.
11224 (execute_one_pass): Use has_gate and has_execute since we cannot
11225 portably check a vtable entry against NULL.
11226 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11227 (pass_manager): Use pass-instances.def to add fields for the
11228 various pass instances.
11229 * toplev.c (finalize): Update for move of
11230 finish_optimization_passes to a method of gcc::pass_manager.
11231 * toplev.h (finish_optimization_passes): Move to method of class
11233 * tree-pass.h (struct pass_data): New.
11234 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11235 (opt_pass::gate): Convert to virtual function.
11236 (opt_pass::~opt_pass): New.
11237 (opt_pass::clone): New.
11238 (opt_pass::execute): Convert to virtual function.
11239 (opt_pass::opt_pass): New.
11240 (opt_pass::ctxt_): new.
11241 (gimple_opt_pass): Convert to subclass of opt_pass.
11242 (gimple_opt_pass::gimple_opt_pass): New.
11243 (rtl_opt_pass): Convert to subclass of opt_pass.
11244 (rtl_opt_pass::rtl_opt_pass): New.
11245 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11246 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11247 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11248 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11249 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11250 invocation of pass_mode_switching to reflect this moving from a
11251 global to a member of gcc::pass_manager.
11252 (ix86_option_override): Rework how pass_insert_vzeroupper is
11253 added to the pass_manager to reflect autogenerated changes.
11254 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11256 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11258 PR rtl-optimization/57708
11259 * recog.c (peep2_find_free_register): Validate all regs in a
11262 2013-08-05 Jan Hubicka <jh@suse.cz>
11265 * cgraph.c (verify_cgraph_node): Accept local flags from other
11267 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11268 (function_and_variable_visibility): Likewise.
11269 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11271 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11273 * graph.c (init_graph_slim_pretty_print): Remove.
11274 (print_graph_cfg): Do not call it. Use local pretty printer.
11275 (start_graph_dump): Likewise.
11277 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11279 * gimple-pretty-print.c (buffer): Remove.
11280 (initialized): Likewise.
11281 (maybe_init_pretty_print): Likewise.
11282 (print_gimple_stmt): Do not call it. Use non-static local
11283 pretty_printer variable.
11284 (print_gimple_expr): Likewise.
11285 (print_gimple_seq): Likewise.
11286 (gimple_dump_bb): Likewise.
11288 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11290 * asan.c (asan_pp): Remove.
11291 (asan_pp_initialized): Likewise.
11292 (asan_pp_initialize): Likewise.
11293 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11294 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11295 (asan_add_global): Likewise.
11297 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11299 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11300 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11301 * pretty-print.c (pp_formatted_text_data): Likewise.
11302 (pp_indent): Rename from pp_base_indent.
11303 (pp_format): Rename from pp_base_format.
11304 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11305 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11306 (pp_flush): Rename from pp_base_flush.
11307 (pp_set_line_maximum_length): Rename from
11308 pp_base_set_line_maximum_length.
11309 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11310 (pp_set_prefix): Rename from pp_base_set_prefix.
11311 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11312 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11313 (pp_append_text): Rename from pp_base_append_text.
11314 (pp_formatted_text): Rename from pp_base_formatted_text.
11315 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11316 (pp_remaining_character_count_for_line): Rename from
11317 pp_base_remaining_character_count_for_line.
11318 (pp_newline): Rename from pp_base_newline.
11319 (pp_character): Rename from pp_base_character.
11320 (pp_string): Rename from pp_base_string.
11321 (pp_maybe_space): Rename from pp_base_maybe_space.
11322 * asan.c (asan_pp_string): Adjust.
11323 (asan_emit_stack_protection): Likewise.
11324 (asan_add_global): Likewise.
11325 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11326 * tree-mudflap.c (mf_varname_tree): Likewise.
11327 * tree-pretty-print.c (pp_tree_identifier): Rename from
11328 pp_base_tree_identifier.
11329 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11330 Declare as function.
11332 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11334 * pretty-print.h (pp_bar_bar): New.
11335 (pp_ampersand_ampersand): Likewise.
11336 (pp_less_equal): Likewise.
11337 (pp_greater_equal): Likewise.
11338 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11339 printer functions instead of pp_string or operators and punctuators.
11340 (dump_gimple_call): Likewise.
11341 (dump_gimple_omp_for): Likewise.
11342 (dump_gimple_transaction): Likewise.
11343 (dump_gimple_phi): Likewise.
11344 (pp_gimple_stmt_1): Likewise.
11345 * sched-vis.c (print_insn): Likewise.
11346 * tree-mudflap.c (mf_varname_tree): Likewise.
11347 * tree-pretty-print.c (dump_block_node): Likewise.
11348 (dump_generic_node): Likewise.
11350 2013-08-02 Jan Hubicka <jh@suse.cz>
11352 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11354 * lto-streamer-out.c (tree_is_indexable): Results decls and
11355 parm decls are not indexable.
11356 (DFS_write_tree_body): Do not follow args and results.
11357 (hash_tree): Likewise.
11358 (output_functions): Rearrange so struct function is needed
11359 only when real body is output; be able to also ouptut abstract
11360 functions; output DECL_ARGUMENTS and DECL_RESULT.
11361 (lto_output): When not in WPA, ale store abstract functions.
11362 (write_symbol): Do not care about RESULT_DECL.
11363 (output_symbol_p): Handle correctly sbtract decls.
11364 * lto-streamer-in.c (input_function): Rearrange so struct
11365 function can be NULL at entry; allow streaming of
11366 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11367 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11368 sanity check during LTO.
11369 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11370 RESULT_DECl and DECL_ARGUMENTS.
11371 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11374 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11376 * pretty-print.h (pp_underscore): New.
11378 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11379 printer functions instead of pp_character.
11380 (dump_binary_rhs): Likewise.
11381 (dump_ternary_rhs): Likewise.
11382 (dump_gimple_call_args): Likewise.
11383 (pp_points_to_solution): Likewise.
11384 (dump_gimple_call): Likewise.
11385 (dump_gimple_switch): Likewise.
11386 (dump_gimple_cond): Likewise.
11387 (dump_gimple_bind): Likewise.
11388 (dump_gimple_try): Likewise.
11389 (dump_gimple_omp_for): Likewise.
11390 (dump_gimple_omp_continue): Likewise.
11391 (dump_gimple_omp_single): Likewise.
11392 (dump_gimple_omp_sections): Likewise.
11393 (dump_gimple_omp_block): Likewise.
11394 (dump_gimple_omp_critical): Likewise.
11395 (dump_gimple_transaction): Likewise.
11396 (dump_gimple_asm): Likewise.
11397 (dump_gimple_phi): Likewise.
11398 (dump_gimple_omp_parallel): Likewise.
11399 (dump_gimple_omp_task): Likewise.
11400 (dump_gimple_omp_atomic_load): Likewise.
11401 (dump_gimple_omp_atomic_store): Likewise.
11402 (dump_gimple_mem_ops): Likewise.
11403 (pp_gimple_stmt_1): Likewise.
11404 (pp_cfg_jump): Likewise.
11405 (dump_implicit_edges): Likewise.
11406 (gimple_dump_bb_for_graph): Likewise.
11407 * graph.c (draw_cfg_node): Likewise.
11408 * langhooks.c (lhd_print_error_function): Likewise.
11409 * sched-vis.c (print_exp): Likewise.
11410 (print_value): Likewise.
11411 (print_pattern): Likewise.
11412 (print_insn): Likewise.
11413 (rtl_dump_bb_for_graph): Likewise.
11414 * tree-pretty-print.c (dump_function_declaration): Likewise.
11415 (dump_array_domain): Likewise.
11416 (dump_omp_clause): Likewise.
11417 (dump_location): Likewise.
11418 (dump_generic_node): Likewise.
11419 (print_struct_decl): Likewise.
11420 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11422 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11424 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11425 candidate table when replacing a candidate statement.
11426 (replace_rhs_if_not_dup): Likewise.
11427 (replace_one_candidate): Likewise.
11429 2013-08-02 Jan Hubicka <jh@suse.cz>
11430 Martin Liska <marxin.liska@gmail.com>
11432 * cgraphunit.c (add_new_function): Fix logic when adding from
11434 (assemble_thunk): Rename to ...
11435 (expand_thunk); .. this one; export; get it working with
11436 general functions; make produced gimple valid.
11437 * cgraph.h (expand_thunk): Declare.
11439 2013-08-02 Jan Hubicka <jh@suse.cz>
11441 * ipa-cp.c (gather_context_independent_values): Use
11442 ipa_get_param_move_cost.
11443 (get_replacement_map): Remove PARAM; move parameter folding
11445 (create_specialized_node): Update.
11446 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11447 assert that we have gimple body; update move_cost.
11448 (count_formal_params): Assert that we have gimple body.
11449 (ipa_dump_param): New function.
11450 (ipa_alloc_node_params): Break out from ...
11451 (ipa_initialize_node_params): ... here.
11452 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11453 (ipa_write_node_info): Stream move costs.
11454 (ipa_read_node_info): Read move costs.
11455 (ipa_update_after_lto_read): Do not recompute node params.
11456 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11457 (ipa_get_param): Check we are not in WPA.
11458 (ipa_get_param_move_cost): New.
11459 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11460 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11461 parm numbers to be present.
11463 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11465 PR rtl-optimization/58048
11466 * lra-constraints.c (process_alt_operands): Don't check asm
11467 operand on register.
11469 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11471 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11472 the implied StoreLoad barrier for atomic operations if before.
11474 2013-08-02 Jan Hubicka <jh@suse.cz>
11475 Martin Liska <marxin.liska@gmail.com>
11477 * cgraph.c (cgraph_function_body_availability): Do not check
11479 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11480 symtab_node_availability): Declare.
11481 * ipa.c (can_replace_by_local_alias): New.
11482 (function_and_variable_visibility): Use it.
11483 * symtab.c (symtab_for_node_and_aliases,
11484 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11486 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11488 PR rtl-optimization/57963
11489 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11490 (lra_constraints): Use them.
11492 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11494 * config/arm/types.md (define_attr "type"): Add "load_acq"
11496 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11498 (cortex_a53_store1): Likewise.
11500 2013-08-01 Jan Hubicka <jh@suse.cz>
11502 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11503 partitions are not needed.
11505 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11507 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11508 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11509 MAYBE_NON_Q_CLASS_P where appropriate.
11511 2013-08-01 Jan Hubicka <jh@suse.cz>
11513 * cgraph.h (release_function_body): Declare.
11514 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11515 of unused delcarations.
11517 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11519 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11520 RTL form when subtracting a constant.
11522 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11524 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11525 Generate canonical plus rtx with negated immediate instead of minus
11527 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11529 2013-08-01 Jan Hubicka <jh@suse.cz>
11531 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11532 (cgraph_release_function_body): Likewise.
11533 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11534 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11535 to used_as_abstract_origin.
11536 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11537 symbols used as abstract origins.
11538 * cgraphunit.c (analyze_functions): Update.
11539 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11540 used_as_abstract_origin.
11541 * tree-inline.c (tree_function_versioning): Update
11542 used_as_abstract_origin; be ready for DECL_RESULT and
11543 DECL_ARGUMENTS to be NULL.
11545 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11546 for abstract functions.
11547 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11550 2013-08-01 Jan Hubicka <jh@suse.cz>
11552 * profile.c (compute_value_histograms): Fix thinko.
11554 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11556 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11557 aarch-common-protos.h to extra_headers.
11558 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11559 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11560 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11562 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11564 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11565 Include "../arm/types.md". Define "type" attribute for all patterns.
11566 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11569 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11571 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11572 to support power8 load fusion.
11573 (fusion_gpr_mem_load): Likewise.
11575 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11577 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11578 declarations for power8 load fusion.
11579 (emit_fusion_gpr_load): Likewise.
11581 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11582 tuning for power8, turn on fusion mode by default. Turn on sign
11583 extending fusion mode if normal fusion mode is on, and we are at
11585 (fusion_gpr_load_p): New function, return true if we can fuse an
11586 addis instruction with a dependent load to a GPR.
11587 (emit_fusion_gpr_load): Emit the instructions for power8 load
11590 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11591 (VSX load fusion peepholes): New peepholes to fuse together an
11592 addi instruction with a VSX load instruction.
11594 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11595 peepholes to fuse an addis instruction with a load to a GPR base
11596 register. If we are supporting sign extending fusions, convert
11597 sign extending loads to zero extending loads and add an explicit
11600 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11602 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11603 aarch-common-protos.h to extra_headers.
11604 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11605 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11606 (arm_early_store_addr_dep): Likewise.
11607 (arm_no_early_alu_shift_dep): Likewise.
11608 (arm_no_early_alu_shift_value_dep): Likewise.
11609 (arm_no_early_mul_dep): Likewise.
11610 (arm_no_early_store_addr_dep): Likewise.
11611 (arm_mac_accumulator_is_mul_result): Likewise.
11612 (arm_mac_accumulator_is_result): Likewise.
11613 * config/arm/aarch-common.c: ... here. New file.
11614 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11616 (arm_early_store_addr_dep): Likewise.
11617 (arm_no_early_alu_shift_dep): Likewise.
11618 (arm_no_early_alu_shift_value_dep): Likewise.
11619 (arm_no_early_mul_dep): Likewise.
11620 (arm_no_early_store_addr_dep): Likewise.
11621 (arm_mac_accumulator_is_mul_result): Likewise.
11622 (arm_mac_accumulator_is_result): Likewise.
11623 * config/arm/aarch-common-protos.h: ... here. New file.
11624 * config/arm/t-arm (aarch-common.o): Define.
11626 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11628 * config/arm/arm.md: Include new file "types.md".
11629 (define_attr "type"): Move from here to ...
11630 (define_attr "mul32"): Likewise.
11631 (define_attr "mul64"): Likewise.
11632 * config/arm/types.md: ... here. New file.
11634 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11636 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11637 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11639 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11641 * gen-pass-instances.awk: Fix offset of substr().
11643 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11645 * Makefile.in (pass-instances.def): New.
11646 (passes.o): Replace dependency on passes.def with one on
11649 * gen-pass-instances.awk: New.
11651 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11652 rather than passes.def, updating local definition of NEXT_PASS
11653 macro to add an extra NUM parameter (currently unused).
11655 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11657 * Makefile.in (PASS_MANAGER_H): New.
11658 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11659 (passes.o): Likewise.
11660 (statistics.o): Likewise.
11661 (cgraphunit.o): Likewise.
11662 (context.o): Depend on PASS_MANAGER_H.
11664 * pass_manager.h: New.
11666 * cgraphunit.c (cgraph_add_new_function): Update for moves
11667 of globals to fields of pass_manager.
11668 (analyze_function): Likewise.
11669 (expand_function): Likewise.
11670 (ipa_passes): Likewise.
11671 (compile): Likewise.
11673 * context.c (context::context): New.
11674 * context.h (context::context): New.
11675 (context::get_passes): New.
11676 (context::passes_): New.
11678 * lto-cgraph.c (input_node): Update for moves of globals to
11679 fields of pass_manager.
11681 * passes.c (all_passes): Remove, in favor of a field of the
11682 same name within the new class pass_manager.
11683 (all_small_ipa_passes): Likewise.
11684 (all_lowering_passes): Likewise.
11685 (all_regular_ipa_passes): Likewise.
11686 (all_late_ipa_passes): Likewise.
11687 (all_lto_gen_passes): Likewise.
11688 (passes_by_id): Likewise.
11689 (passes_by_id_size): Likewise.
11690 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11691 the new class pass_manager.
11692 (set_pass_for_id): Convert to...
11693 (pass_manager::set_pass_for_id): ...method.
11694 (get_pass_for_id): Convert to...
11695 (pass_manager::get_pass_for_id): ...method.
11696 (register_one_dump_file): Move body of implementation into...
11697 (pass_manager::register_one_dump_file): ...here.
11698 (register_dump_files_1): Convert to...
11699 (pass_manager::register_dump_files_1): ...method.
11700 (register_dump_files): Convert to...
11701 (pass_manager::register_dump_files): ...method.
11702 (create_pass_tab): Update for moves of globals to fields of
11704 (dump_passes): Move body of implementation into...
11705 (pass_manager::dump_passes): ...here.
11706 (register_pass): Move body of implementation into...
11707 (pass_manager::register_pass): ...here.
11708 (init_optimization_passes): Convert into...
11709 (pass_manager::pass_manager): ...constructor for new
11710 pass_manager class, and initialize the pass_lists array.
11711 (check_profile_consistency): Update for moves of globals to
11712 fields of pass_manager.
11713 (dump_profile_report): Move body of implementation into...
11714 (pass_manager::dump_profile_report): ...here.
11715 (ipa_write_summaries_1): Update for moves of pass lists from
11716 being globals to fields of pass_manager.
11717 (ipa_write_optimization_summaries): Likewise.
11718 (ipa_read_summaries): Likewise.
11719 (ipa_read_optimization_summaries): Likewise.
11720 (execute_all_ipa_stmt_fixups): Likewise.
11722 * statistics.c (statistics_fini): Update for moves of globals to
11723 fields of pass_manager.
11725 * toplev.c (general_init): Replace call to
11726 init_optimization_passes with construction of the pass_manager
11729 * tree-pass.h (all_passes): Remove, in favor of a field of the
11730 same name within the new class pass_manager.
11731 (all_small_ipa_passes): Likewise.
11732 (all_lowering_passes): Likewise.
11733 (all_regular_ipa_passes): Likewise.
11734 (all_lto_gen_passes): Likewise.
11735 (all_late_ipa_passes): Likewise.
11736 (passes_by_id): Likewise.
11737 (passes_by_id_size): Likewise.
11738 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11739 the new class pass_manager.
11740 (get_pass_for_id): Remove.
11742 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11744 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11747 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11749 * arm.md (mulhi3): New expand pattern.
11751 2013-07-30 Jan Hubicka <jh@suse.cz>
11752 Martin Liska <marxin.liska@gmail.com>
11754 * profile.c (compute_value_histograms): Do not ICE when
11755 there is mismatch only on some counters.
11757 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11759 PR rtl-optimization/57637
11760 * function.c (move_insn_for_shrink_wrap): Also check the
11761 GEN set of the LIVE problem for the liveness analysis
11762 if it exists, otherwise give up.
11764 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11766 PR tree-optimization/57993
11767 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11768 replaced statement in the candidate table.
11769 (phi_add_costs): Return infinite cost when the hidden basis does
11770 not dominate all phis on which the candidate is dependent.
11771 (replace_one_candidate): Record replaced statement in the
11774 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11776 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11777 (ashlv2si3): New expander.
11778 (*ashlv2si3_i): New define_insn_and_split.
11779 * predicates.md (float_operation): Allow patterns with three
11780 basic sub-patterns.
11782 PR rtl-optimization/58021
11783 * mode-switching.c (create_pre_exit): Always split off preceding
11784 insns if we are not at the basic block head.
11786 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11788 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11789 (UCLIBC_DYNAMIC_LINKER): New macro.
11790 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11792 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11793 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11795 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11796 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11797 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11798 for SF and DF modes. Use ieee_quad_format for TF mode.
11799 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11800 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11801 (mips_option_override): Handle `-mnan=legacy'.
11802 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11803 `-mabs=2008' and `-mnan=2008'.
11804 (OPTION_DEFAULT_SPECS): Add "nan" default.
11805 (ASM_SPEC): Handle `-mnan='.
11806 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11807 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11808 comment accordingly.
11809 (neg<mode>2): Likewise.
11810 * config/mips/mips.opt (mabs, mnan): New options.
11811 * doc/install.texi (Configuration): Document `--with-nan=' option.
11812 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11814 (MIPS Options): Document them.
11815 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11816 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11817 * configure: Regenerate.
11818 * config.in: Regenerate.
11820 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11822 * config/i386/i386.md (float post-reload splitters): Do not check
11823 for subregs of SSE registers.
11825 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11826 H.J. Lu <hongjiu.lu@intel.com>
11830 * config/i386/i386.md (post-reload splitter
11831 to avoid partial SSE reg dependency stalls): New pattern.
11833 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11835 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11837 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11839 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11840 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11842 2013-07-26 Cary Coutant <ccoutant@google.com>
11844 * dwarf2out.c (die_checksum_ordered): Don't include template
11845 instantiations in signature.
11846 (is_template_parameter): New function.
11847 (is_template_instantiation): New function.
11848 (generate_skeleton_bottom_up): Don't include template instantiations
11850 (generate_skeleton): Likewise.
11851 (break_out_comdat_types): Move recursive call to break out nested
11853 (prune_unused_types_mark_generic_parms_dies): Call
11854 is_template_parameter.
11856 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11858 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11859 uses vector registers.
11860 * config/aarch64/iterators.md: Add attributes rtn and vas.
11862 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11863 Richard Earnshaw <richard.earnshaw@arm.com>
11865 * combine.c (simplify_comparison): Re-canonicalize operands
11867 * config/arm/arm.md (movcond_addsi): New splitter.
11869 2013-07-25 Sterling Augustine <saugustine@google.com>
11871 * dwarf2out.c (size_of_pubnames): Move code to...
11872 (include_pubname_in_output): ...here. New.
11873 (want_pubnames): Rearrange.
11874 (output_pubnames): Call include_pubname_in_output. Move assertion.
11876 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11878 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11880 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11883 * doc/extend.texi: Remove obsolete builtins. Fix
11884 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11886 2013-07-25 Jan Hubicka <jh@suse.cz>
11888 * cgraph.c (release_function_body): Break out from ...
11889 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11890 and DECL_ARGUMENTS.
11891 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11892 old_tree in the map.
11893 (create_specialized_node): Update.
11894 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11896 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11897 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11898 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11899 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11902 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11904 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11905 addsi3_carryin_alt2_<optab>): Correct output template.
11907 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11909 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11910 Adjust for arm_restrict_it.
11911 Remove trailing whitespace.
11913 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11915 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11916 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11918 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11920 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11922 PR rtl-optimization/57960
11923 * lra-constraints.c (process_alt_operands): Use the right mode
11924 when checking strict_low.
11926 2013-07-25 Jan Hubicka <jh@suse.cz>
11928 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11929 * cgraph.c (cgraph_remove_node): Do not release function body
11930 when in cgraph streaming.
11931 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11932 in other partitions are not considered reachable; fix handling of
11935 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11937 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11939 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11944 * config/arm/arm.md ("*sibcall_insn): Replace use of
11945 Ss with US. Adjust output for v5 and v4t.
11946 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11948 * config/arm/constraints.md ("Ss"): Rename to US.
11950 2013-07-25 Terry Guo <terry.guo@arm.com>
11952 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11953 shift_add/shift_sub0/shift_sub1 RTXs.
11955 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11956 Anton Blanchard <anton@au1.ibm.com>
11958 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11959 (altivec_vpks<VI_char>ss): Likewise.
11960 (altivec_vpks<VI_char>us): Likewise.
11961 (altivec_vpku<VI_char>us): Likewise.
11962 (altivec_vpku<VI_char>um): Likewise.
11964 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11966 Introduce context class.
11968 * Makefile.in (CONTEXT_H): New.
11969 (OBJS): Add context.o.
11970 (toplev.o): Add CONTEXT_H to dependencies.
11973 * toplev.c (general_init): Create the singleton gcc::context instance.
11979 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11981 PR rtl-optimization/57968
11982 * mode-switching.c (create_pre_exit): Allow instructions that
11983 don't set a return register to need a non-exit mode.
11985 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11986 Anton Blanchard <anton@au1.ibm.com>
11988 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11989 operands to vperm for little endian.
11990 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11991 of lvsl to create the control mask for a vperm for little endian.
11993 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11994 Anton Blanchard <anton@au1.ibm.com>
11996 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11997 two operands for little-endian.
11999 2013-07-23 Steve Ellcey <sellcey@mips.com>
12001 * config/mips/mips.c (mips_case_values_threshold): New.
12002 (TARGET_CASE_VALUES_THRESHOLD): Define.
12004 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12005 Anton Blanchard <anton@au1.ibm.com>
12007 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12008 selection of field for vector splat in little endian mode.
12010 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12012 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12013 expanders to rs6000.md.
12014 (ior<mode>3): Likewise.
12015 (and<mode>3): Likewise.
12016 (one_cmpl<mode>2): Likewise.
12017 (nor<mode>3): Likewise.
12018 (andc<mode>3): Likewise.
12019 (eqv<mode>3): Likewise.
12020 (nand<mode>3): Likewise.
12021 (orc<mode>3): Likewise.
12023 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12026 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12027 to split multi-word logical operations.
12028 (rs6000_split_logical_di): Likewise.
12029 (rs6000_split_logical): Likewise.
12031 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12032 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12033 and allow TImode operations in 32-bit.
12034 (vsx_and<mode>3_64bit): Likewise.
12035 (vsx_ior<mode>3_32bit): Likewise.
12036 (vsx_ior<mode>3_64bit): Likewise.
12037 (vsx_xor<mode>3_32bit): Likewise.
12038 (vsx_xor<mode>3_64bit): Likewise.
12039 (vsx_one_cmpl<mode>2_32bit): Likewise.
12040 (vsx_one_cmpl<mode>2_64bit): Likewise.
12041 (vsx_nor<mode>3_32bit): Likewise.
12042 (vsx_nor<mode>3_64bit): Likewise.
12043 (vsx_andc<mode>3_32bit): Likewise.
12044 (vsx_andc<mode>3_64bit): Likewise.
12045 (vsx_eqv<mode>3_32bit): Likewise.
12046 (vsx_eqv<mode>3_64bit): Likewise.
12047 (vsx_nand<mode>3_32bit): Likewise.
12048 (vsx_nand<mode>3_64bit): Likewise.
12049 (vsx_orc<mode>3_32bit): Likewise.
12050 (vsx_orc<mode>3_64bit): Likewise.
12052 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12053 logical types in GPRs.
12055 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12056 logical insns to rs6000.md, and allow TImode operations in
12058 (altivec_ior<mode>3): Likewise.
12059 (altivec_xor<mode>3): Likewise.
12060 (altivec_one_cmpl<mode>2): Likewise.
12061 (altivec_nor<mode>3): Likewise.
12062 (altivec_andc<mode>3): Likewise.
12064 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12065 attributes for moving the 128-bit logical operations into
12067 (BOOL_REGS_OUTPUT): Likewise.
12068 (BOOL_REGS_OP1): Likewise.
12069 (BOOL_REGS_OP2): Likewise.
12070 (BOOL_REGS_UNARY): Likewise.
12071 (BOOL_REGS_AND_CR0): Likewise.
12072 (one_cmpl<mode>2): Add support for DI logical operations on
12073 32-bit, splitting the operations to 32-bit.
12074 (anddi3): Likewise.
12075 (iordi3): Likewise.
12076 (xordi3): Likewise.
12077 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12078 changes to combine the 32/64-bit code, allow logical operations on
12079 TI mode in 32-bit, and to use similar match_operator patterns like
12080 scalar mode uses. Combine the Altivec and VSX code for logical
12081 operations, and move it here.
12082 (ior<mode>3, 128-bit types): Likewise.
12083 (xor<mode>3, 128-bit types): Likewise.
12084 (one_cmpl<mode>3, 128-bit types): Likewise.
12085 (nor<mode>3, 128-bit types): Likewise.
12086 (andc<mode>3, 128-bit types): Likewise.
12087 (eqv<mode>3, 128-bit types): Likewise.
12088 (nand<mode>3, 128-bit types): Likewise.
12089 (orc<mode>3, 128-bit types): Likewise.
12090 (and<mode>3_internal): Likewise.
12091 (bool<mode>3_internal): Likewise.
12092 (boolc<mode>3_internal1): Likewise.
12093 (boolc<mode>3_internal2): Likewise.
12094 (boolcc<mode>3_internal1): Likewise.
12095 (boolcc<mode>3_internal2): Likewise.
12096 (eqv<mode>3_internal1): Likewise.
12097 (eqv<mode>3_internal2): Likewise.
12098 (one_cmpl1<mode>3_internal): Likewise.
12100 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12102 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12103 Rename flag_stack_usage to flag_stack_usage_info.
12105 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12107 * config/microblaze/sync.md: New file.
12108 * config/microblaze/microblaze.md: Include sync.md
12109 * config/microblaze/microblaze.c: Add print_operand 'y'.
12110 * config/microblaze/constraints.md: Add memory_contraint
12111 'Q' which is a single register.
12113 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12115 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12117 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12119 * reload.c (find_reloads): Exit loop once we find this operand
12120 cannot be reloaded somehow for this alternative.
12122 * reload.c (find_reloads): Exit loop once we find a hard register.
12124 * rtlanal.c (computed_jump_p): Exit loop once we find label
12127 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12129 * cfgloopmanip.c (remove_path): Exit loop after setting
12132 * gensupport.c (subst_dup): Avoid loop if code is not
12133 MATCH_DUP nor MATCH_OP_DUP.
12135 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12137 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12139 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12141 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12142 true for SP_REGNUM if mode == ptr_mode.
12143 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12144 with value R0_REGNUM + 31.
12146 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12148 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12149 pad pointer-typed argument downward.
12151 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12153 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12154 and __ILP32__ when the ILP32 model is in use.
12156 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12158 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12159 (aarch64_load_symref_appropriately): In the case of
12160 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12161 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12162 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12163 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12164 if 'mode' doesn't equal to 'ptr_mode'.
12165 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12166 'vcall_offset'; change to call aarch64_emit_move differently depending
12167 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12168 to calculate the upper bound of 'vcall_offset'.
12169 (aarch64_cannot_force_const_mem): Change to also return true if
12171 (aarch64_legitimize_reload_address): In the case of large
12172 displacements, add new local variable 'xmode' and an assertion
12173 based on it; change to use 'xmode' to generate the new rtx and
12175 (aarch64_asm_trampoline_template): Change to generate the template
12176 differently depending on TARGET_ILP32 or not; change to use
12177 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12178 (aarch64_trampoline_size): Removed.
12179 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12180 and replace immediate literals with it. Change to use 'ptr_mode'
12181 instead of 'DImode' and call convert_memory_address if the mode
12182 of 'fnaddr' doesn't equal to 'ptr_mode'.
12183 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12185 (aarch64_elf_asm_destructor): Likewise.
12186 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12187 on TARGET_ILP32 instead of aarch64_trampoline_size.
12188 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12189 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12190 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12191 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12192 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12193 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12194 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12195 depending on the value of 'mode'.
12196 (add_losym_<mode>): New.
12197 (ldr_got_small_<mode>): New, based on ldr_got_small.
12198 (ldr_got_small): Remove.
12199 (ldr_got_small_sidi): New.
12200 * config/aarch64/iterators.md (P): New.
12201 (PTR): Change to 'ptr_mode' in the condition.
12203 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12205 * config.gcc (aarch64*-*-*): Support --with-abi.
12206 (aarch64*-*-elf): Support --with-multilib-list.
12207 (aarch64*-*-linux*): Likewise.
12208 (supported_defaults): Add abi to aarch64*-*-*.
12209 * configure.ac: Mention AArch64 for --with-multilib-list.
12210 * configure: Re-generated.
12211 * config/aarch64/biarchilp32.h: New file.
12212 * config/aarch64/biarchlp64.h: New file.
12213 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12215 (MULTILIB_DEFAULTS): Ditto.
12216 (DRIVER_SELF_SPECS): Ditto.
12217 (ASM_SPEC): Update to also substitute -mabi.
12218 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12219 file whose name depends on -mabi= and -mbig-endian.
12220 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12222 (POINTER_SIZE): New define.
12223 (POINTERS_EXTEND_UNSIGNED): Ditto.
12224 (enum aarch64_abi_type): New enumeration tag.
12225 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12226 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12227 (TARGET_ILP32): New define.
12228 * config/aarch64/aarch64.opt (mabi): New.
12229 (aarch64_abi): New.
12230 (ilp32, lp64): New values for -mabi.
12231 * config/aarch64/t-aarch64 (comma): New define.
12232 (MULTILIB_OPTIONS): Ditto.
12233 (MULTILIB_DIRNAMES): Ditto.
12234 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12235 * doc/invoke.texi: Document -mabi for AArch64.
12237 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12239 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12241 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12242 Anton Blanchard <anton@au1.ibm.com>
12244 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12245 endianness when selecting field to splat.
12247 2013-07-22 Eric Christopher <echristo@gmail.com>
12249 * dwarf2out.c (die_odr_checksum): New function to use
12250 CHECKSUM_ macros and ULEB128 for DIE tag.
12251 (generate_type_signature): Use.
12253 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12255 * config.gcc (sparc*-*-*): Accept leon3 processor.
12256 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12257 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12258 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12259 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12260 (mfix-ut699): Adjust comment.
12261 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12262 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12263 (CPP_CPU_SPEC): Likewise.
12264 (ASM_CPU_SPEC): Likewise.
12265 * config/sparc/sparc.c (leon3_cost): New constant.
12266 (sparc_option_override): Add leon3 support.
12267 (mem_ref): New function.
12268 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12269 (sparc_do_work_around_errata): Look into the instruction in the delay
12270 slot and adjust accordingly. Add fix for the data cache nullify issues
12271 of the UT699. Change insertion position for the NOP.
12272 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12273 (leon3_load): New reservation.
12274 (leon_store): Bump latency to 2.
12275 (grfpu): New automaton.
12276 (grfpu_alu): New unit.
12277 (grfpu_ds): Likewise.
12278 (leon_fp_alu): Adjust.
12279 (leon_fp_mult): Delete.
12280 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12281 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12282 * config/sparc/sparc.md (cpu): Add leon3.
12283 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12284 (swapsi): Likewise.
12285 (atomic_test_and_set): Likewise.
12286 (ldstub): Likewise.
12288 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12290 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12291 default for R5900 targets.
12292 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12293 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12294 * config/mips/mips.c (mips_option_override): Report an error for
12295 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12296 for -march=r5900 -mhard-float.
12298 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12300 * df-problems.c (can_move_insns_across): Exit loop once we
12301 find a non-fixed, non-global register.
12303 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12306 * omega.c (omega_eliminate_red): Break after setting red_found.
12307 (omega_problem_has_red_equations): Similarly after setting found.
12308 (omega_query_variable): Similarly after setting coupled.
12310 2013-07-22 Marek Polacek <polacek@redhat.com>
12312 * gimplify.c: Don't include gimple.h twice.
12314 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12316 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12317 instead of TARGET_THUMB1.
12318 (Pz): New constraint.
12319 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12321 (compare_negsi_si): Likewise.
12322 (compare_addsi2_op0): Likewise.
12323 (compare_addsi2_op1): Likewise.
12324 (addsi3_carryin_<optab>): Likewise.
12325 (addsi3_carryin_alt2_<optab>): Likewise.
12326 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12327 for arm_restrict_it.
12328 (subsi3_carryin): Likewise.
12329 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12330 (minmax_arithsi): Disable for arm_restrict_it.
12331 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12332 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12333 (satsi_<SAT:code>_shift): Likewise.
12334 (arm_shiftsi3): Add alternative for 16-bit encoding.
12335 (arm32_movhf): Disable for arm_restrict_it.
12336 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12337 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12339 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12341 * config/arm/arm.md (attribute "insn"): Delete.
12342 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12343 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12344 (not_shiftsi): Update for attribute change.
12345 (not_shiftsi_compare0): Likewise.
12346 (not_shiftsi_compare0_scratch): Likewise.
12347 (arm_one_cmplsi2): Likewise.
12348 (thumb1_one_cmplsi2): Likewise.
12349 (notsi_compare0): Likewise.
12350 (notsi_compare0_scratch): Likewise.
12351 (thumb1_movdi_insn): Likewise.
12352 (arm_movsi_insn): Likewise.
12353 (movhi_insn_arch4): Likewise.
12354 (movhi_bytes): Likewise.
12355 (arm_movqi_insn): Likewise.
12356 (thumb1_movqi_insn): Likewise.
12357 (arm32_movhf): Likewise.
12358 (thumb1_movhf): Likewise.
12359 (arm_movsf_soft_insn): Likewise.
12360 (thumb1_movsf_insn): Likewise.
12361 (thumb_movdf_insn): Likewise.
12362 (movsicc_insn): Likewise.
12363 (movsfcc_soft_insn): Likewise.
12364 (and_scc): Likewise.
12365 (cond_move): Likewise.
12366 (if_move_not): Likewise.
12367 (if_not_move): Likewise.
12368 (if_shift_move): Likewise.
12369 (if_move_shift): Likewise.
12370 (if_shift_shift): Likewise.
12371 (if_not_arith): Likewise.
12372 (if_arith_not): Likewise.
12373 (cond_move_not): Likewise.
12374 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12375 (neon_mov<mode>): Likewise.
12376 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12377 (thumb2_movsi_vfp): Likewise.
12378 (movsf_vfp): Likewise.
12379 (thumb2_movsf_vfp): Likewise.
12380 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12382 (cortexa7_older_only): Likewise.
12383 (cortexa7_younger): Likewise.
12384 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12385 (1020alu_shift_op): Likewise.
12386 (1020alu_shift_reg_op): Likewise.
12387 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12388 (alu_shift_op): Likewise.
12389 (alu_shift_reg_op): Likewise.
12390 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12391 (11_alu_shift_op): Likewise.
12392 (11_alu_shift_reg_op): Likewise.
12393 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12394 (9_alu_shift_reg_op): Likewise.
12395 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12397 (cortex_a15_alu_shift): Likewise.
12398 (cortex_a15_alu_shift_reg): Likewise.
12399 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12400 (cortex_a5_alu_shift): Likewise.
12401 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12403 (cortex_a53_alu_shift): Likewise.
12404 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12406 (cortex_a7_alu_reg): Likewise.
12407 (cortex_a7_alu_shift): Likewise.
12408 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12409 (cortex_a8_alu_shift): Likewise.
12410 (cortex_a8_alu_shift_reg): Likewise.
12411 (cortex_a8_mov): Likewise.
12412 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12413 (cortex_a9_dp_shift): Likewise.
12414 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12415 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12416 (cortex_r4_mov): Likewise.
12417 (cortex_r4_alu_shift): Likewise.
12418 (cortex_r4_alu_shift_reg): Likewise.
12419 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12420 (526_alu_shift_op): Likewise.
12421 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12422 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12423 (626te_alu_shift_op): Likewise.
12424 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12425 (726te_alu_op): Likewise.
12426 (726te_alu_shift_op): Likewise.
12427 (726te_alu_shift_reg_op): Likewise.
12428 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12429 (mp626_alu_shift_op): Likewise.
12430 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12431 (pj4_alu_e1_conds): Likewise.
12432 (pj4_alu): Likewise.
12433 (pj4_alu_conds): Likewise.
12434 (pj4_shift): Likewise.
12435 (pj4_shift_conds): Likewise.
12436 (pj4_alu_shift): Likewise.
12437 (pj4_alu_shift_conds): Likewise.
12439 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12441 * config/arm/predicates.md (shiftable_operator_strict_it):
12443 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12444 Disable cond_exec version for arm_restrict_it.
12445 (thumb2_smaxsi3): Convert to generate cond_exec.
12446 (thumb2_sminsi3): Likewise.
12447 (thumb32_umaxsi3): Likewise.
12448 (thumb2_uminsi3): Likewise.
12449 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12450 (thumb2_neg_abssi2): Likewise.
12451 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12452 (thumb2_movsicc_insn): Adjust alternatives.
12453 (thumb2_mov_negscc): Disable for arm_restrict_it.
12454 (thumb2_mov_negscc_strict_it): New pattern.
12455 (thumb2_mov_notscc_strict_it): New pattern.
12456 (thumb2_mov_notscc): Disable for arm_restrict_it.
12457 (thumb2_ior_scc): Likewise.
12458 (thumb2_ior_scc_strict_it): New pattern.
12459 (thumb2_cond_move): Adjust for arm_restrict_it.
12460 (thumb2_cond_arith): Disable for arm_restrict_it.
12461 (thumb2_cond_arith_strict_it): New pattern.
12462 (thumb2_cond_sub): Adjust for arm_restrict_it.
12463 (thumb2_movcond): Likewise.
12464 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12465 (thumb2_zero_extendhisi2_v6): Likewise.
12466 (thumb2_zero_extendqisi2_v6): Likewise.
12467 (orsi_notsi_si): Likewise.
12468 (orsi_not_shiftsi_si): Likewise.
12470 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12472 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12473 instruction sequence is 1 byte shorter.
12475 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12477 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12478 it is not needed after split.
12480 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12483 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12484 second label for nonlocal goto receivers. Don't output pic base labels
12485 unless we're producing PIC; mark that action unreachable().
12486 (ix86_save_reg): If the function contains a nonlocal label, save the
12488 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12489 * config/darwin.c (emitted_pic_label_num): New GTY.
12490 (update_pic_label_number_if_needed): New.
12491 (machopic_output_function_base_name): Adjust for nonlocal receiver
12493 (machopic_should_output_picbase_label): New.
12494 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12495 (nonlocal_goto_receiver): New insn and split.
12497 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12499 * config/aarch64/aarch64-builtins.c
12500 (aarch64_fold_builtin): Fold abs in all modes.
12501 * config/aarch64/aarch64-simd-builtins.def
12502 (abs): Enable for all modes.
12503 * config/aarch64/arm_neon.h
12504 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12505 (vabs_f64): Add missing intrinsic.
12507 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12509 * config/aarch64/arm_neon.h (vabs_s64): New function
12511 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12514 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12515 * config/avr/avr.md (adjust_len): Add `round'.
12516 * config/avr/avr-protos.h (avr_out_round): New prototype.
12517 (avr_out_plus): Add `out_label' argument.
12518 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12519 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12520 Handle the case where `insn' is just a pattern.
12521 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12522 (avr_out_round): New function.
12523 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12525 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12527 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12528 Add check for flag_stack_usage to handle -fstack-usage support
12530 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12532 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12533 interaction for new Power8 flags and VSX.
12535 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12537 PR middle-end/57698
12538 * tree-inline.c (expand_call_inline): Emit errors during
12539 early_inlining only if optimization is not turned on.
12541 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12545 * passes.c (init_optimization_passes): Move the construction of
12546 the pass hierarchy into a new passes.def file.
12548 * Makefile.in (passes.o): Add dependency on passes.def.
12550 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12552 * passes.c (init_optimization_passes): Introduce macros for
12553 constructing the tree of passes (INSERT_PASSES_AFTER,
12554 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12555 TERMINATE_PASS_LIST).
12557 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12558 Wei Mi <wmi@google.com>
12560 PR rtl-optimization/57878
12561 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12563 (reload_pseudo_compare_func): Check nregs first for reload
12566 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12568 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12570 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12572 * read-rtl.c (validate_const_int): Once an invalid character is
12573 seen, quit the loop.
12575 * gengtype.c (write_roots): Similarly once we find the "deletable"
12576 or "if_marked" option.
12578 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12580 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12581 "xtab" and "sat". Move value "clz" from here to ...
12582 (attriubte "type"): ... here.
12583 (satsi_<SAT:code>): Delete "insn" attribute.
12584 (satsi_<SAT:code>_shift): Likewise.
12585 (arm_zero_extendqisi2addsi): Likewise.
12586 (arm_extendqisi2addsi): Likewise.
12587 (clzsi2): Update for attribute changes.
12588 (rbitsi2): Likewise.
12589 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12591 (arm_usatsihi): Likewise.
12592 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12594 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12596 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12597 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12598 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12599 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12600 in alphabetical order.
12601 (attribute "core_cycles"): Update for attribute changes.
12602 (arm_addsi3): Likewise.
12603 (addsi3_compare0): Likewise.
12604 (addsi3_compare0_scratch): Likewise.
12605 (addsi3_compare_op1): Likewise.
12606 (addsi3_compare_op2): Likewise.
12607 (compare_addsi2_op0): Likewise.
12608 (compare_addsi2_op1): Likewise.
12609 (addsi3_carryin_shift_<optab>): Likewise.
12610 (subsi3_carryin_shift): Likewise.
12611 (rsbsi3_carryin_shift): Likewise.
12612 (arm_subsi3_insn): Likewise.
12613 (subsi3_compare0): Likewise.
12614 (subsi3_compare): Likewise.
12615 (arm_andsi3_insn): Likewise.
12616 (thumb1_andsi3_insn): Likewise.
12617 (andsi3_compare0): Likewise.
12618 (andsi3_compare0_scratch): Likewise.
12619 (zeroextractsi_compare0_scratch
12620 (andsi_not_shiftsi_si): Likewise.
12621 (iorsi3_insn): Likewise.
12622 (iorsi3_compare0): Likewise.
12623 (iorsi3_compare0_scratch): Likewise.
12624 (arm_xorsi3): Likewise.
12625 (thumb1_xorsi3_insn): Likewise.
12626 (xorsi3_compare0): Likewise.
12627 (xorsi3_compare0_scratch): Likewise.
12628 (satsi_<SAT:code>_shift): Likewise.
12630 (arm_shiftsi3): Likewise.
12631 (shiftsi3_compare0): Likewise.
12632 (not_shiftsi): Likewise.
12633 (not_shiftsi_compare0): Likewise.
12634 (not_shiftsi_compare0_scratch): Likewise.
12635 (arm_one_cmplsi2): Likewise.
12636 (thumb_one_complsi2): Likewise.
12637 (notsi_compare0): Likewise.
12638 (notsi_compare0_scratch): Likewise.
12639 (thumb1_zero_extendhisi2): Likewise.
12640 (arm_zero_extendhisi2): Likewise.
12641 (arm_zero_extendhisi2_v6): Likewise.
12642 (arm_zero_extendhisi2addsi): Likewise.
12643 (thumb1_zero_extendqisi2): Likewise.
12644 (thumb1_zero_extendqisi2_v6): Likewise.
12645 (arm_zero_extendqisi2): Likewise.
12646 (arm_zero_extendqisi2_v6): Likewise.
12647 (arm_zero_extendqisi2addsi): Likewise.
12648 (thumb1_extendhisi2): Likewise.
12649 (arm_extendhisi2): Likewise.
12650 (arm_extendhisi2_v6): Likewise.
12651 (arm_extendqisi): Likewise.
12652 (arm_extendqisi_v6): Likewise.
12653 (arm_extendqisi2addsi): Likewise.
12654 (thumb1_extendqisi2): Likewise.
12655 (thumb1_movdi_insn): Likewise.
12656 (arm_movsi_insn): Likewise.
12657 (movsi_compare0): Likewise.
12658 (movhi_insn_arch4): Likewise.
12659 (movhi_bytes): Likewise.
12660 (arm_movqi_insn): Likewise.
12661 (thumb1_movqi_insn): Likewise.
12662 (arm32_movhf): Likewise.
12663 (thumb1_movhf): Likewise.
12664 (arm_movsf_soft_insn): Likewise.
12665 (thumb1_movsf_insn): Likewise.
12666 (movdf_soft_insn): Likewise.
12667 (thumb_movdf_insn): Likewise.
12668 (arm_cmpsi_insn): Likewise.
12669 (cmpsi_shiftsi): Likewise.
12670 (cmpsi_shiftsi_swp): Likewise.
12671 (arm_cmpsi_negshiftsi_si): Likewise.
12672 (movsicc_insn): Likewise.
12673 (movsfcc_soft_insn): Likewise.
12674 (arith_shiftsi): Likewise.
12675 (arith_shiftsi_compare0
12676 (arith_shiftsi_compare0_scratch
12677 (sub_shiftsi): Likewise.
12678 (sub_shiftsi_compare0
12679 (sub_shiftsi_compare0_scratch
12680 (and_scc): Likewise.
12681 (cond_move): Likewise.
12682 (if_plus_move): Likewise.
12683 (if_move_plus): Likewise.
12684 (if_move_not): Likewise.
12685 (if_not_move): Likewise.
12686 (if_shift_move): Likewise.
12687 (if_move_shift): Likewise.
12688 (if_shift_shift): Likewise.
12689 (if_not_arith): Likewise.
12690 (if_arith_not): Likewise.
12691 (cond_move_not): Likewise.
12692 (thumb1_ashlsi3): Set type attribute.
12693 (thumb1_ashrsi3): Likewise.
12694 (thumb1_lshrsi3): Likewise.
12695 (thumb1_rotrsi3): Likewise.
12696 (shiftsi3_compare0_scratch): Likewise.
12697 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12698 (neon_mov<mode>): Likewise.
12699 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12701 (thumb2_movsi_insn): Likewise.
12702 (thumb2_cmpsi_neg_shiftsi): Likewise.
12703 (thumb2_extendqisi_v6): Likewise.
12704 (thumb2_zero_extendhisi2_v6): Likewise.
12705 (thumb2_zero_extendqisi2_v6): Likewise.
12706 (thumb2_shiftsi3_short): Likewise.
12707 (thumb2_addsi3_compare0_scratch): Likewise.
12708 (orsi_not_shiftsi_si): Likewise.
12709 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12710 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12712 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12713 (1020alu_shift_op): Likewise.
12714 (1020alu_shift_reg_op): Likewise.
12715 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12716 (alu_shift_op): Likewise.
12717 (alu_shift_reg_op): Likewise.
12718 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12719 (11_alu_shift_op): Likewise.
12720 (11_alu_shift_reg_op): Likewise.
12721 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12722 (9_alu_shift_reg_op): Likewise.
12723 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12725 (cortex_a15_alu_shift): Likewise.
12726 (cortex_a15_alu_shift_reg): Likewise.
12727 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12729 (cortex_a5_alu_shift): Likewise.
12730 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12732 (cortex_a53_alu_shift): Likewise.
12733 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12735 (cortex_a7_alu_reg): Likewise.
12736 (cortex_a7_alu_shift): Likewise.
12737 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12739 (cortex_a8_alu_shift): Likewise.
12740 (cortex_a8_alu_shift_reg): Likewise.
12741 (cortex_a8_mov): Likewise.
12742 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12743 (cortex_a9_dp_shift): Likewise.
12744 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12746 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12748 (cortex_r4_mov): Likewise.
12749 (cortex_r4_alu_shift): Likewise.
12750 (cortex_r4_alu_shift_reg): Likewise.
12751 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12752 (526_alu_shift_op): Likewise.
12753 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12754 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12755 (626te_alu_shift_op): Likewise.
12756 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12757 (726te_alu_op): Likewise.
12758 (726te_alu_shift_op): Likewise.
12759 (726te_alu_shift_reg_op): Likewise.
12760 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12761 (mp626_alu_shift_op): Likewise.
12762 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12763 (pj4_alu_e1_conds): Likewise.
12764 (pj4_alu): Likewise.
12765 (pj4_alu_conds): Likewise.
12766 (pj4_shift): Likewise.
12767 (pj4_shift_conds): Likewise.
12768 (pj4_alu_shift): Likewise.
12769 (pj4_alu_shift_conds): Likewise.
12770 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12772 (cortexa7_older_only): Likewise.
12773 (cortexa7_younger): Likewise.
12775 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12777 * ipa-pure-const.c (generate_summary): Rename to...
12778 (pure_const_generate_summary): ... this.
12780 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12782 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12784 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12787 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12790 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12792 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12793 enabled without -march=zEC12.
12794 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12797 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12799 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12800 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12801 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12802 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12803 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12804 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12805 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12807 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12808 and ISA_HAS_NMADD3_NMSUB3.
12809 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12810 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12811 (nmsub4<mode>, nmsub3<mode>): Likewise.
12812 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12814 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12816 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12817 TARGET_MIPS5400 checking.
12819 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12820 Peter Bergner <bergner@vnet.ibm.com>
12822 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12823 registers in the comment.
12824 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12825 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12826 rather than FIRST_PSEUDO_REGISTERS.
12828 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12830 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12831 enable extra ISA flags with TARGET_HTM.
12833 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12835 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12838 2013-07-15 Cong Hou <congh@google.com>
12840 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12841 in compare function for sorting.
12843 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12845 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12846 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12847 * config/rs6000/rs6000.opt: Add -mhtm option.
12848 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12849 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12850 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12851 __HTM__ if the HTM instructions are available.
12852 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12853 htm_spr_reg_operand): New define_predicates.
12854 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12855 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12857 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12858 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12859 HTM builtin functions.
12860 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12861 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12862 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12863 (rs6000_builtin_mask_calculate): Likewise.
12864 (rs6000_option_override_internal): Likewise.
12865 (bdesc_htm): Add new HTM builtin support.
12866 (htm_spr_num): New function.
12867 (htm_spr_regno): Likewise.
12868 (rs6000_htm_spr_icode): Likewise.
12869 (htm_expand_builtin): Likewise.
12870 (htm_init_builtins): Likewise.
12871 (rs6000_expand_builtin): Add support for HTM builtin functions.
12872 (rs6000_init_builtins): Likewise.
12873 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12875 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12876 (TARGET_HTM, MASK_HTM): Define macros.
12877 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12878 (FIXED_REGISTERS): Likewise.
12879 (CALL_USED_REGISTERS): Likewise.
12880 (CALL_REALLY_USED_REGISTERS): Likewise.
12881 (REG_ALLOC_ORDER): Likewise.
12882 (enum reg_class): Likewise.
12883 (REG_CLASS_NAMES): Likewise.
12884 (REG_CLASS_CONTENTS): Likewise.
12885 (REGISTER_NAMES): Likewise.
12886 (ADDITIONAL_REGISTER_NAMES): Likewise.
12887 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12888 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12889 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12890 * config/rs6000/htm.md: New file.
12891 * config/rs6000/htmintrin.h: New file.
12892 * config/rs6000/htmxlintrin.h: New file.
12894 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12896 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12897 Define SYMBOL_TINY_GOT, update comment.
12898 * config/aarch64/aarch64.c
12899 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12900 (aarch64_expand_mov_immediate): Likewise.
12901 (aarch64_print_operand): Likewise.
12902 (aarch64_classify_symbol): Likewise.
12903 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12904 (ldr_got_tiny): Define.
12906 2013-07-13 Tobias Grosser <tobias@grosser.es>
12908 PR tree-optimization/54094
12909 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12910 scheduling dimension for the parallelism check from the polyhedral
12911 information in the AST.
12912 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12915 2013-07-13 Jason Merrill <jason@redhat.com>
12917 * print-tree.c (debug_vec_tree): Use debug_raw.
12918 (debug_raw (vec<tree, va_gc> &)): New.
12919 (debug_raw (vec<tree, va_gc> *)): New.
12920 * tree.h: Declare them.
12922 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12924 * ifcvt.c (ifcvt_after_combine): New static variable.
12925 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12927 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12928 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12929 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12931 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12933 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12935 2013-07-12 Michael Matz <matz@suse.de>
12937 PR middle-end/55771
12938 * convert.c (convert_to_real): Reject non-float inner types.
12940 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12942 * config/aarch64/aarch64-protos.h
12943 (aarch64_simd_immediate_valid_for_move): Remove.
12944 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12945 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12947 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12949 2013-07-11 Steve Ellcey <sellcey@mips.com>
12951 * config/mips/mips.c (mips_conditional_register_usage): Do not
12952 use t[0-7] registers in MIPS16 mode when optimizing for size.
12954 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12956 * config/i386/i386.c (dispatch_function_versions): Fix array
12957 indexing of function_version_info to match actual_versions.
12959 2013-07-11 Teresa Johnson <tejohnson@google.com>
12961 * vec.h (struct va_gc): Move release out-of-line.
12962 (va_gc::release): Call ggc_free on released vec.
12964 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12966 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12967 Require GOT register as additional operand in UNSPEC.
12968 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12969 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12970 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12971 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12972 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12973 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12974 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12976 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12979 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12980 name seen by assembler/linker rather if available.
12982 2013-07-11 Andreas Schwab <schwab@suse.de>
12984 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12986 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12988 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12990 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12992 * doc/tm.texi.in: Move hook documentation to ....
12993 * target.def: ... here.
12995 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12996 text on @hook line.
12997 * doc/tm.texi: Regenerate.
12999 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13002 * doc/invoke.texi: Document -Wconditionally-supported.
13004 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13007 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13010 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13013 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13014 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13015 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13016 Remove duplicate devices.
13017 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13018 * config/avr/t-multilib: Regenerate.
13019 * config/avr/avr-tables.opt: Regenerate.
13020 * doc/avr-mmcu.texi: Regenerate.
13022 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13025 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13027 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13029 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13030 the cost of MULT when optimizing for size.
13032 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13034 * config/cr16/cr16-protos.h: Don't include target.h.
13036 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13038 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13039 adjust register size for TDmode and TFmode for VSX registers.
13041 2013-07-08 Kai Tietz <ktietz@redhat.com>
13044 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13045 hook_bool_const_tree_true.
13047 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13049 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13050 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13051 * config/s390/s390.md: Define FPR*_REGNUM constants.
13052 Fix FPR2_REGNUM constant (18 -> 17).
13053 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13054 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13055 ("trunc<BFP:mode><DFP_ALL:mode>2")
13056 ("trunc<DFP_ALL:mode><BFP:mode>2")
13057 ("*extend<BFP:mode><DFP_ALL:mode>2")
13058 ("*extend<DFP_ALL:mode><BFP:mode>2")
13059 ("extend<BFP:mode><DFP_ALL:mode>2")
13060 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13063 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13065 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13067 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13069 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13070 and cfun_fpr_bit_p to cfun_fpr_save_p.
13071 (s390_frame_area, s390_register_info, s390_frame_info)
13072 (s390_emit_prologue, s390_emit_epilogue)
13073 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13075 * config/s390/s390.h: Define *_REGNUM macros for floating point
13078 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13080 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13082 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13084 PR rtl-optimization/57786
13085 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13086 and break out of the loop when it is set to false.
13088 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13091 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13092 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13093 (const_int 63)) 0)).
13094 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13095 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13096 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13098 PR rtl-optimization/57829
13099 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13100 mask bits outside of mode are just sign-extension from mode to HWI.
13102 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13104 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13105 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13106 adjust_address instead of change_address to keep info about alignment.
13107 (emit_strmov): Remove.
13108 (emit_memmov): New function.
13109 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13110 (expand_movmem_epilogue): Likewise and return updated rtx for
13112 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13113 destination and source.
13114 (decide_alignment): Refactor, handle vector_loop.
13115 (ix86_expand_movmem): Likewise.
13116 (ix86_expand_setmem): Likewise.
13117 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13119 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13121 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13122 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13124 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13126 * config/i386/sse.md (sse_movlhps): Change alternative 3
13127 of operand 2 to "m".
13129 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13132 * config/i386/sse.md (iptr): New mode attribute.
13133 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13134 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13135 (<sse>_vmmul<mode>3): Ditto.
13136 (<sse>_vmdiv<mode>3): Ditto.
13137 (sse_vmrcpv4sf2): Ditto.
13138 (<sse>_vmsqrt<mode>2): Ditto.
13139 (sse_vmrsqrtv4sf2): Ditto.
13140 (<sse>_vm<code><mode>3): Ditto.
13141 (avx_vmcmp<mode>3): Ditto.
13142 (<sse>_vmmaskcmp<mode>3): Ditto.
13143 (<sse>_comi): Ditto.
13144 (<sse>_ucomi): Ditto.
13145 (*xop_vmfrcz_<mode>): Ditto.
13146 (*fmai_fmadd_<mode>): Ditto.
13147 (*fmai_fmsub_<mode>): Ditto.
13148 (*fmai_fnmadd_<mode>): Ditto.
13149 (*fmai_fnmsub_<mode>): Ditto.
13150 (*fma4i_vmfmadd_<mode>): Ditto.
13151 (*fma4i_vmfmsub_<mode>): Ditto.
13152 (*fma4i_vmfnmadd_<mode>): Ditto.
13153 (*fma4i_vmfnmsub_<mode>): Ditto.
13154 (*xop_vmfrcz_<mode>): Ditto.
13155 (sse_cvtps2pi): Ditto.
13156 (sse_cvttps2pi): Ditto.
13157 (sse_cvtss2si): Ditto.
13158 (sse_cvtss2si_2): Ditto.
13159 (sse_cvtss2siq_2): Ditto.
13160 (sse_cvttss2si): Ditto.
13161 (sse_cvttss2siq): Ditto.
13162 (sse_cvtsd2si): Ditto.
13163 (sse_cvtsd2si_2): Ditto.
13164 (sse_cvtsd2siq_2): Ditto.
13165 (sse_cvttsd2si): Ditto.
13166 (sse_cvttsd2siq): Ditto.
13167 (sse_cvtsd2ss): Ditto.
13168 (sse_cvtss2sd): Ditto.
13169 (avx2_pbroadcast<mode>): Ditto.
13170 (avx2_pbroadcast<mode>_1): Ditto.
13171 (*avx_vperm_broadcast_v4sf): Ditto.
13173 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13174 (sse_movlhps): Ditto.
13175 (sse_storehps): Ditto.
13176 (sse_loadhps): Ditto.
13177 (sse_storelps): Ditto.
13178 (sse_loadlps): Ditto.
13179 (*vec_concatv4sf): Ditto.
13180 (*vec_interleave_highv2df): Ditto.
13181 (*vec_interleave_lowv2df): Ditto.
13182 (*vec_extractv2df_1_sse): Ditto.
13183 (*vec_extractv2df_0_sse): Ditto.
13184 (sse2_storelpd): Ditto.
13185 (sse2_loadlpd): Ditto.
13186 (sse2_movsd): Ditto.
13187 (*vec_concatv4si): Ditto.
13188 (vec_concatv2di): Ditto.
13190 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13191 for Intel asm dialect.
13192 (mmx_punpcklwd): Ditto.
13193 (mmx_punpckldq): Ditto.
13195 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13196 for intel assembler dialect.
13198 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13201 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13202 for BUILT_IN_C{LZ,LRSB}*.
13203 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13204 * tree-vrp.c (extract_range_basic): Handle
13205 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13206 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13207 fall thru to code calling set_value*.
13208 * builtins.c (expand_builtin): Remove *IMAX cases.
13209 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13210 if width is bigger than 2*HWI.
13212 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13214 PR rtl-optimization/55342
13215 * lra-int.h (lra_subreg_reload_pseudos): New.
13216 * lra.c: Add undoing optional reloads to the block diagram.
13217 (lra_subreg_reload_pseudos): New.
13218 (lra_optional_reload_pseudos): Change comments.
13219 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13220 lra_optional_reload_pseudos after undo transformations.
13221 * lra-assigns.c (pseudo_prefix_title): New.
13222 (lra_setup_reg_renumber): Use it.
13223 (spill_for): Ditto. Check subreg reload pseudos too.
13224 (assign_by_spills): Consider subreg reload pseudos too.
13225 * lra-constraints.c (simplify_operand_subreg): Use
13226 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13227 (curr_insn_transform): Recognize and do optional reloads.
13228 (undo_optional_reloads): New.
13229 (lra_undo_inheritance): Call undo_optional_reloads.
13231 2013-07-05 Thomas Quinot <quinot@adacore.com>
13233 * tree-complex.c (expand_complex_operations_1): Fix typo.
13235 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13237 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13238 (tune_params): New member 'const vec_costs'.
13239 * config/aarch64/aarch64.c (generic_vector_cost): New.
13240 (generic_tunings): New member 'generic_vector_cost'.
13241 (aarch64_builtin_vectorization_cost): New.
13242 (aarch64_add_stmt_cost): New.
13243 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13244 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13246 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13249 * config/i386/predicates.md (vsib_address_operand): Disallow
13250 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13252 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13254 PR middle-end/55030
13255 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13256 expand_builtin_setjmp_receiver.
13257 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13258 with NULL for the label parameter.
13259 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13260 the frame-pointer. Adjust comments.
13261 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13262 only if LABEL is non-NULL.
13264 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13266 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13267 (ARM_ABI_AAPCS64): Ditto.
13269 (ARM_DEFAULT_ABI): Ditto.
13271 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13273 * config/aarch64/aarch64-builtins.c
13274 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13275 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13277 * config/aarch64/aarch64-simd.md
13278 (aarch64_ld1<VALL:mode>): New.
13279 (aarch64_st1<VALL:mode>): Likewise.
13280 * config/aarch64/arm_neon.h
13281 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13283 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13285 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13287 (ix86_option_override_internal):Turn on all -mavx target flags by
13288 default as they are dependent on AVX anyway.
13290 2013-07-02 Cary Coutant <ccoutant@google.com>
13292 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13293 deterministic hash.
13294 (loc_checksum_ordered): Likewise.
13295 (hash_loc_operands): Remove inline keyword.
13297 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13299 PR tree-optimization/57741
13300 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13301 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13302 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13303 Allow REAL_CST step_exprs if flag_associative_math.
13304 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13306 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13308 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13310 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13312 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13314 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13316 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13318 (iorsi3_insn): Likewise.
13319 (arm_xorsi3): Likewise.
13321 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13323 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13325 (attribute "type"): ... here, and prefix with "wmmx_".
13326 (attribute "core_cycles"): Update for attribute changes.
13327 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13328 (tbcstv4hi): Likewise.
13329 (tbcstv2si): Likewise.
13330 (iwmmxt_iordi3): Likewise.
13331 (iwmmxt_xordi3): Likewise.
13332 (iwmmxt_anddi3): Likewise.
13333 (iwmmxt_nanddi3): Likewise.
13334 (iwmmxt_arm_movdi): Likewise.
13335 (iwmmxt_movsi_insn): Likewise.
13336 (mov<mode>_internal): Likewise.
13337 (and<mode>3_iwmmxt): Likewise.
13338 (ior<mode>3_iwmmxt): Likewise.
13339 (xor<mode>3_iwmmxt): Likewise.
13340 (add<mode>3_iwmmxt): Likewise.
13341 (ssaddv8qi3): Likewise.
13342 (ssaddv4hi3): Likewise.
13343 (ssaddv2si3): Likewise.
13344 (usaddv8qi3): Likewise.
13345 (usaddv4hi3): Likewise.
13346 (usaddv2si3): Likewise.
13347 (sub<mode>3_iwmmxt): Likewise.
13348 (sssubv8qi3): Likewise.
13349 (sssubv4hi3): Likewise.
13350 (sssubv2si3): Likewise.
13351 (ussubv8qi3): Likewise.
13352 (ussubv4hi3): Likewise.
13353 (ussubv2si3): Likewise.
13354 (mulv4hi3_iwmmxt): Likewise.
13355 (smulv4hi3_highpart): Likewise.
13356 (umulv4hi3_highpart): Likewise.
13357 (iwmmxt_wmacs): Likewise.
13358 (iwmmxt_wmacsz): Likewise.
13359 (iwmmxt_wmacu): Likewise.
13360 (iwmmxt_wmacuz): Likewise.
13361 (iwmmxt_clrdi): Likewise.
13362 (iwmmxt_clrv8qi): Likewise.
13363 (iwmmxt_clr4hi): Likewise.
13364 (iwmmxt_clr2si): Likewise.
13365 (iwmmxt_uavgrndv8qi3): Likewise.
13366 (iwmmxt_uavgrndv4hi3): Likewise.
13367 (iwmmxt_uavgv8qi3): Likewise.
13368 (iwmmxt_uavgv4hi3): Likewise.
13369 (iwmmxt_tinsrb): Likewise.
13370 (iwmmxt_tinsrh): Likewise.
13371 (iwmmxt_tinsrw): Likewise.
13372 (iwmmxt_textrmub): Likewise.
13373 (iwmmxt_textrmsb): Likewise.
13374 (iwmmxt_textrmuh): Likewise.
13375 (iwmmxt_textrmsh): Likewise.
13376 (iwmmxt_textrmw): Likewise.
13377 (iwmxxt_wshufh): Likewise.
13378 (eqv8qi3): Likewise.
13379 (eqv4hi3): Likewise.
13380 (eqv2si3): Likewise.
13381 (gtuv8qi3): Likewise.
13382 (gtuv4hi3): Likewise.
13383 (gtuv2si3): Likewise.
13384 (gtv8qi3): Likewise.
13385 (gtv4hi3): Likewise.
13386 (gtv2si3): Likewise.
13387 (smax<mode>3_iwmmxt): Likewise.
13388 (umax<mode>3_iwmmxt): Likewise.
13389 (smin<mode>3_iwmmxt): Likewise.
13390 (umin<mode>3_iwmmxt): Likewise.
13391 (iwmmxt_wpackhss): Likewise.
13392 (iwmmxt_wpackwss): Likewise.
13393 (iwmmxt_wpackdss): Likewise.
13394 (iwmmxt_wpackhus): Likewise.
13395 (iwmmxt_wpackwus): Likewise.
13396 (iwmmxt_wpackdus): Likewise.
13397 (iwmmxt_wunpckihb): Likewise.
13398 (iwmmxt_wunpckihh): Likewise.
13399 (iwmmxt_wunpckihw): Likewise.
13400 (iwmmxt_wunpckilb): Likewise.
13401 (iwmmxt_wunpckilh): Likewise.
13402 (iwmmxt_wunpckilw): Likewise.
13403 (iwmmxt_wunpckehub): Likewise.
13404 (iwmmxt_wunpckehuh): Likewise.
13405 (iwmmxt_wunpckehuw): Likewise.
13406 (iwmmxt_wunpckehsb): Likewise.
13407 (iwmmxt_wunpckehsh): Likewise.
13408 (iwmmxt_wunpckehsw): Likewise.
13409 (iwmmxt_wunpckelub): Likewise.
13410 (iwmmxt_wunpckeluh): Likewise.
13411 (iwmmxt_wunpckeluw): Likewise.
13412 (iwmmxt_wunpckelsb): Likewise.
13413 (iwmmxt_wunpckelsh): Likewise.
13414 (iwmmxt_wunpckelsw): Likewise.
13415 (ror<mode>3): Likewise.
13416 (ashr<mode>3_iwmmxt): Likewise.
13417 (lshr<mode>3_iwmmxt): Likewise.
13418 (ashl<mode>3_iwmmxt): Likewise.
13419 (ror<mode>3_di): Likewise.
13420 (ashr<mode>3_di): Likewise.
13421 (lshr<mode>3_di): Likewise.
13422 (ashl<mode>3_di): Likewise.
13423 (iwmmxt_wmadds): Likewise.
13424 (iwmmxt_wmaddu): Likewise.
13425 (iwmmxt_tmia): Likewise.
13426 (iwmmxt_tmiaph): Likewise.
13427 (iwmmxt_tmiabb): Likewise.
13428 (iwmmxt_tmiatb): Likewise.
13429 (iwmmxt_tmiabt): Likewise.
13430 (iwmmxt_tmiatt): Likewise.
13431 (iwmmxt_tmovmskb): Likewise.
13432 (iwmmxt_tmovmskh): Likewise.
13433 (iwmmxt_tmovmskw): Likewise.
13434 (iwmmxt_waccb): Likewise.
13435 (iwmmxt_wacch): Likewise.
13436 (iwmmxt_waccw): Likewise.
13437 (iwmmxt_waligni): Likewise.
13438 (iwmmxt_walignr): Likewise.
13439 (iwmmxt_walignr0): Likewise.
13440 (iwmmxt_walignr1): Likewise.
13441 (iwmmxt_walignr2): Likewise.
13442 (iwmmxt_walignr3): Likewise.
13443 (iwmmxt_wsadb): Likewise.
13444 (iwmmxt_wsadh): Likewise.
13445 (iwmmxt_wsadbz): Likewise.
13446 (iwmmxt_wsadhz): Likewise.
13447 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13448 (iwmmxt_wabsdiffb): Likewise.
13449 (iwmmxt_wabsdiffh): Likewise.
13450 (iwmmxt_wabsdiffw): Likewise.
13451 (iwmmxt_waddsubhx): Likewise
13452 (iwmmxt_wsubaddhx): Likewise.
13453 (addc<mode>3): Likewise.
13454 (iwmmxt_avg4): Likewise.
13455 (iwmmxt_avg4r): Likewise.
13456 (iwmmxt_wmaddsx): Likewise.
13457 (iwmmxt_wmaddux): Likewise.
13458 (iwmmxt_wmaddsn): Likewise.
13459 (iwmmxt_wmaddun): Likewise.
13460 (iwmmxt_wmulwsm): Likewise.
13461 (iwmmxt_wmulwum): Likewise.
13462 (iwmmxt_wmulsmr): Likewise.
13463 (iwmmxt_wmulumr): Likewise.
13464 (iwmmxt_wmulwsmr): Likewise.
13465 (iwmmxt_wmulwumr): Likewise.
13466 (iwmmxt_wmulwl): Likewise.
13467 (iwmmxt_wqmulm): Likewise.
13468 (iwmmxt_wqmulwm): Likewise.
13469 (iwmmxt_wqmulmr): Likewise.
13470 (iwmmxt_wqmulwmr): Likewise.
13471 (iwmmxt_waddbhusm): Likewise.
13472 (iwmmxt_waddbhusl): Likewise.
13473 (iwmmxt_wqmiabb): Likewise.
13474 (iwmmxt_wqmiabt): Likewise.
13475 (iwmmxt_wqmiatb): Likewise.
13476 (iwmmxt_wqmiatt): Likewise.
13477 (iwmmxt_wqmiabbn): Likewise.
13478 (iwmmxt_wqmiabtn): Likewise.
13479 (iwmmxt_wqmiatbn): Likewise.
13480 (iwmmxt_wqmiattn): Likewise.
13481 (iwmmxt_wmiabb): Likewise.
13482 (iwmmxt_wmiabt): Likewise.
13483 (iwmmxt_wmiatb): Likewise.
13484 (iwmmxt_wmiatt): Likewise.
13485 (iwmmxt_wmiabbn): Likewise.
13486 (iwmmxt_wmiabtn): Likewise.
13487 (iwmmxt_wmiatbn): Likewise.
13488 (iwmmxt_wmiattn): Likewise.
13489 (iwmmxt_wmiawbb): Likewise.
13490 (iwmmxt_wmiawbt): Likewise.
13491 (iwmmxt_wmiawtb): Likewise.
13492 (iwmmxt_wmiawtt): Likewise.
13493 (iwmmxt_wmiawbbn): Likewise.
13494 (iwmmxt_wmiawbtn): Likewise.
13495 (iwmmxt_wmiawtbn): Likewise.
13496 (iwmmxt_wmiawttn): Likewise.
13497 (iwmmxt_wmerge): Likewise.
13498 (iwmmxt_tandc<mode>3): Likewise.
13499 (iwmmxt_torc<mode>3): Likewise.
13500 (iwmmxt_torvsc<mode>3): Likewise.
13501 (iwmmxt_textrc<mode>3): Likewise.
13502 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13503 (wmmxt_pack): Likewise.
13504 (wmmxt_mult_c1): Likewise.
13505 (wmmxt_mult_c2): Likewise.
13506 (wmmxt_alu_c1): Likewise.
13507 (wmmxt_alu_c2): Likewise.
13508 (wmmxt_alu_c3): Likewise.
13509 (wmmxt_transfer_c1): Likewise.
13510 (wmmxt_transfer_c2): Likewise.
13511 (wmmxt_transfer_c3): Likewise.
13512 (marvell_f_iwmmxt_wstr): Likewise.
13513 (marvell_f_iwmmxt_wldr): Likewise.
13515 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13517 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13519 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13522 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13523 * lra-constraints.c (need_for_split_p): Check call used hard regs
13524 living through calls.
13526 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13527 call used regs for call insn.
13529 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13532 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13533 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13535 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13537 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13539 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13541 * lra-constraints.c (need_for_split_p): Check call used hard regs
13542 living through calls.
13544 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13547 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13548 to tie with any other modes. Eliminate Altivec vector mode tests,
13549 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13550 code, to return 0 if testing MODE2 for a condition, if we've
13551 already tested MODE1 for the same condition.
13553 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13555 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13558 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13560 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13561 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13563 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13565 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13567 (aarch64_symbolic_constant_p): Remove.
13568 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13569 static. Fix line length and white space.
13570 (aarch64_symbolic_constant_p): Remove.
13571 * config/aarch64/predicates.md (aarch64_valid_symref):
13572 Use aarch64_classify_symbol_expression.
13574 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13576 * config/arm/constraints.md (Ts): New constraint.
13577 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13579 (compare_scc): Use "Ts" constraint for operand 0.
13580 (ior_scc_scc): Likewise.
13581 (and_scc_scc): Likewise.
13582 (and_scc_scc_nodom): Likewise.
13583 (ior_scc_scc_cmp): Likewise for operand 7.
13584 (and_scc_scc_cmp): Likewise.
13585 * config/arm/thumb2.md (thumb2_movsi_insn):
13586 Add alternatives for 16-bit encodings.
13587 (thumb2_movhi_insn): Likewise.
13588 (thumb2_movsicc_insn): Likewise.
13589 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13590 (thumb2_negscc): Use "Ts" constraint.
13591 Move mvn instruction outside cond_exec block.
13592 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13593 for 16-bit encodings.
13595 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13597 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13599 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13600 (mulsi3subsi): Likewise.
13601 (mulsidi3adddi): Likewise.
13602 (mulsidi3_v6): Likewise.
13603 (umulsidi3_v6): Likewise.
13604 (umulsidi3adddi_v6): Likewise.
13605 (smulsi3_highpart_v6): Likewise.
13606 (umulsi3_highpart_v6): Likewise.
13607 (mulhisi3tb): Likewise.
13608 (mulhisi3bt): Likewise.
13609 (mulhisi3tt): Likewise.
13610 (maddhisi4): Likewise.
13611 (maddhisi4tb): Likewise.
13612 (maddhisi4tt): Likewise.
13613 (maddhidi4): Likewise.
13614 (maddhidi4tb): Likewise.
13615 (maddhidi4tt): Likewise.
13616 (zeroextractsi_compare0_scratch): Likewise.
13617 (insv_zero): Likewise.
13618 (insv_t2): Likewise.
13619 (anddi_notzesidi_di): Likewise.
13620 (anddi_notsesidi_di): Likewise.
13621 (andsi_notsi_si): Likewise.
13622 (iordi_zesidi_di): Likewise.
13623 (xordi_zesidi_di): Likewise.
13624 (andsi_iorsi3_notsi): Likewise.
13625 (smax_0): Likewise.
13626 (smax_m1): Likewise.
13627 (smin_0): Likewise.
13628 (not_shiftsi): Likewise.
13629 (unaligned_loadsi): Likewise.
13630 (unaligned_loadhis): Likewise.
13631 (unaligned_loadhiu): Likewise.
13632 (unaligned_storesi): Likewise.
13633 (unaligned_storehi): Likewise.
13634 (extv_reg): Likewise.
13635 (extzv_t2): Likewise.
13636 (divsi3): Likewise.
13637 (udivsi3): Likewise.
13638 (arm_zero_extendhisi2addsi): Likewise.
13639 (arm_zero_extendqisi2addsi): Likewise.
13640 (compareqi_eq0): Likewise.
13641 (arm_extendhisi2_v6): Likewise.
13642 (arm_extendqisi2addsi): Likewise.
13643 (arm_movt): Likewise.
13644 (thumb2_ldrd): Likewise.
13645 (thumb2_ldrd_base): Likewise.
13646 (thumb2_ldrd_base_neg): Likewise.
13647 (thumb2_strd): Likewise.
13648 (thumb2_strd_base): Likewise.
13649 (thumb2_strd_base_neg): Likewise.
13650 (arm_negsi2): Add alternative for 16-bit encoding.
13651 (arm_one_cmplsi2): Likewise.
13653 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13655 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13656 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13657 (movdfcc): Likewise.
13658 * config/arm/vfp.md (*thumb2_movsf_vfp):
13659 Disable predication for arm_restrict_it.
13660 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13661 (*thumb2_movdfcc_vfp): Likewise.
13662 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13663 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13664 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13665 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13666 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13667 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13668 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13669 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13670 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13671 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13672 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13673 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13674 Disable predication for arm_restrict_it.
13676 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13678 * config/i386/bmiintrin.h (_bextr_u32): New.
13679 (_bextr_u64): Ditto.
13681 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13683 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13684 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13685 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13686 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13687 * config/mips/n32-elf.h: ...this new file.
13689 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13692 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13693 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13695 2013-06-27 Catherine Moore <clm@codesourcery.com>
13697 * config/mips/mips-tables.opt: Regenerate.
13698 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13699 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13700 * doc/invoke.texi: Add -m14kc.
13702 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13705 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13706 constraints of operand 1 and 2.
13709 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13710 to match RTL canonicalization. Swap predicates and
13711 constraints of operand 1 and 2.
13713 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13715 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13716 Process OP_INOUT regs for splitting too.
13718 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13720 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13721 decl before the loop, initialize to NULL.
13722 (vectorizable_load): Initialize ptr_incr to NULL.
13724 2013-06-27 Martin Jambor <mjambor@suse.cz>
13727 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13728 * ipa-ref.c (ipa_maybe_record_reference): New function.
13729 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13730 * ipa-cp.c (create_specialized_node): Record potential references from
13732 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13734 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13736 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13737 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13738 'mode' to force_reg.
13739 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13740 (aarch64_expand_mov_immediate): Likewise.
13742 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13744 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13745 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13747 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13749 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13750 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13751 (struct machine_function): Add tbegin_p.
13752 (s390_canonicalize_comparison): Fold CC mode compares to
13753 conditional jump if possible.
13754 (s390_emit_jump): Return the emitted jump.
13755 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13756 Handle CCRAWmode compares.
13757 (s390_option_override): Default to -mhtm if available.
13758 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13759 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13760 FPRs instead of df_regs_ever_live_p.
13761 (s390_optimize_nonescaping_tx): New function.
13762 (s390_init_frame_layout): Extend clobbered_regs array to cover
13764 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13765 (s390_expand_tbegin): New function.
13766 (enum s390_builtin): New enum definition.
13767 (code_for_builtin): New array definition.
13768 (s390_init_builtins): New function.
13769 (s390_expand_builtin): New function.
13770 (TARGET_INIT_BUILTINS): Define.
13771 (TARGET_EXPAND_BUILTIN): Define.
13772 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13773 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13774 (s390_alc_comparison): Likewise.
13775 * config/s390/s390-modes.def: Add CCRAWmode.
13776 * config/s390/s390.h (processor_flags): Add PF_TX.
13777 (TARGET_CPU_HTM): Define macro.
13778 (TARGET_HTM): Define macro.
13779 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13780 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13781 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13782 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13783 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13784 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13785 ("*cc_to_int"): Move up.
13786 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13787 constants other than 0.
13788 ("*ccraw_to_int"): New insn and splitter definition.
13789 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13790 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13791 ("tx_assist"): New expander.
13792 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13793 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13794 * config/s390/s390.opt: Add -mhtm option.
13795 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13796 * config/s390/htmxlintrin.h: New file.
13797 * config/s390/htmintrin.h: New file.
13798 * config/s390/s390intrin.h: New file.
13799 * doc/extend.texi: Document htm builtins.
13800 * config.gcc: Add the new header files to extra_headers.
13802 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13804 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13805 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13807 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13808 Pat Haugen <pthaugen@us.ibm.com>
13809 Peter Bergner <bergner@vnet.ibm.com>
13811 * config/rs6000/power8.md: New.
13812 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13813 setting for power8 entry.
13814 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13815 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13816 test for Power4/Power5 only.
13817 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13819 (force_new_group): Adjust comment.
13820 * config/rs6000/rs6000.md: Include power8.md.
13822 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13824 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13825 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13827 (tune_params): Update comment.
13828 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13829 (arm_max_conditional_execute): New function.
13830 (thumb2_final_prescan_insn): Use max_insn_skipped and
13831 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13833 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13835 PR tree-optimization/57705
13836 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13837 SSA_NAME step, provided that it is not defined inside the loop.
13838 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13839 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13841 2013-06-25 Martin Jambor <mjambor@suse.cz>
13843 PR middle-end/57670
13844 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13845 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13847 (ipa_note_param_call): Initialize member_ptr flag.
13848 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13849 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13850 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13851 (ipa_read_indirect_edge_info): Likewise.
13853 2013-06-25 Richard Biener <rguenther@suse.de>
13855 PR middle-end/56977
13856 * passes.c (init_optimization_passes): Move pass_fold_builtins
13857 and pass_dce earlier with -Og.
13859 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13861 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13862 <BIT_FIELD_REF>: Remove trailing TAB.
13863 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13866 2013-06-24 Martin Jambor <mjambor@suse.cz>
13868 PR tree-optimization/57358
13869 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13870 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13871 (ipa_analyze_params_uses): Generate pessimistic info when true.
13873 2013-06-24 Martin Jambor <mjambor@suse.cz>
13875 PR tree-optimization/57539
13876 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13877 global.inlined_to of the new node to it. All callers changed.
13878 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13879 inlining_into, pass it to cgraph_clone_node.
13880 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13881 ipa_free_edge_args_substructures.
13882 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13883 rdesc linked list. Do not assert rdesc edges have inlined caller.
13884 Assert we have found an rdesc in the rdesc list.
13886 2013-06-24 Richard Biener <rguenther@suse.de>
13888 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13889 (pointer_set_lookup): Declare.
13890 (class pointer_map): New template class implementing a
13891 generic pointer to T map.
13892 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13893 pointer_map<T>::contains, pointer_map<T>::insert,
13894 pointer_map<T>::traverse): New functions.
13895 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13896 (pointer_set_lookup): New function.
13897 (pointer_set_contains): Use pointer_set_lookup.
13898 (pointer_set_insert): Likewise.
13899 (insert_aux): Remove.
13900 (struct pointer_map_t): Embed a pointer_set_t.
13901 (pointer_map_create): Adjust.
13902 (pointer_map_destroy): Likewise.
13903 (pointer_map_contains): Likewise.
13904 (pointer_map_insert): Likewise.
13905 (pointer_map_traverse): Likewise.
13906 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13907 pointer_map<unsigned> instead of a pointer_map_t.
13908 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13909 (streamer_tree_cache_lookup): Likewise.
13910 (streamer_tree_cache_create): Likewise.
13911 (streamer_tree_cache_delete): Likewise.
13912 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13913 pointer_map<unsigned> instead of a pointer_map_t.
13914 (lto_init_tree_ref_encoder): Adjust.
13915 (lto_destroy_tree_ref_encoder): Likewise.
13916 * lto-section-out.c (lto_output_decl_index): Likewise.
13917 (lto_record_function_out_decl_state): Likewise.
13918 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13920 2013-06-24 Richard Biener <rguenther@suse.de>
13922 PR tree-optimization/57488
13923 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13925 2013-06-24 Alan Modra <amodra@gmail.com>
13927 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13928 (gen_easy_altivec_constant): Likewise.
13929 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13930 easy_vector_constant_msb): Likewise.
13932 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13935 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13936 add missing return true.
13938 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13941 * config/sh/predicates.md (general_extend_operand): Invoke
13942 general_movsrc_operand for memory operands.
13943 (general_movsrc_operand): Allow reg+reg addressing, do not use
13944 general_operand for memory operands.
13946 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13948 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13949 when current target options does not apply.
13950 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13951 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13952 * config/i386/bmiintrin.h: Pass appropriate target
13953 attributes to header.
13954 * config/i386/mmintrin.h: Ditto.
13955 * config/i386/nmmintrin.h: Ditto.
13956 * config/i386/avx2intrin.h: Ditto.
13957 * config/i386/fxsrintrin.h: Ditto.
13958 * config/i386/tbmintrin.h: Ditto.
13959 * config/i386/xsaveintrin.h: Ditto.
13960 * config/i386/f16cintrin.h: Ditto.
13961 * config/i386/xtestintrin.h: Ditto.
13962 * config/i386/xsaveoptintrin.h: Ditto.
13963 * config/i386/bmi2intrin.h: Ditto.
13964 * config/i386/lzcntintrin.h: Ditto.
13965 * config/i386/smmintrin.h: Ditto.
13966 * config/i386/wmmintrin.h: Ditto.
13967 * config/i386/x86intrin.h: Remove all header include guards.
13968 * config/i386/prfchwintrin.h: Ditto.
13969 * config/i386/pmmintrin.h: Ditto.
13970 * config/i386/tmmintrin.h: Ditto.
13971 * config/i386/xmmintrin.h: Ditto.
13972 * config/i386/popcntintrin.h: Ditto.
13973 * config/i386/rdseedintrin.h: Ditto.
13974 * config/i386/ammintrin.h: Ditto.
13975 * config/i386/emmintrin.h: Ditto.
13976 * config/i386/immintrin.h: Remove all header include guards.
13977 * config/i386/fma4intrin.h: Ditto.
13978 * config/i386/lwpintrin.h: Ditto.
13979 * config/i386/xopintrin.h: Ditto.
13980 * config/i386/ia32intrin.h: Ditto.
13981 * config/i386/avxintrin.h: Ditto.
13982 * config/i386/rtmintrin.h: Ditto.
13983 * config/i386/fmaintrin.h: Ditto.
13984 * config/i386/mm3dnow.h: Ditto.
13986 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13988 * common/config/i386/i386-common.c: Handle LZCNT.
13990 2013-06-22 Andi Kleen <ak@linux.intel.com>
13992 * doc/extend.texi: Use __atomic_store_n instead of
13993 __atomic_store in HLE example.
13995 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13997 * config/sh/sh.c: Remove <cstdlib> workaround.
13999 2013-06-21 Andi Kleen <ak@linux.intel.com>
14001 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14003 2013-06-21 Andi Kleen <ak@linux.intel.com>
14005 * doc/extend.texi: Document that __atomic_clear and
14006 __atomic_test_and_set should only be used with bool.
14008 2013-06-20 Jan Hubicka <jh@suse.cz>
14010 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14011 types_same_for_odr.
14012 * tree.c (decls_same_for_odr): New function.
14013 (same_for_edr): New function.
14014 (types_same_for_odr): New function.
14015 (get_binfo_at_offset): Use it.
14016 * tree.h (types_same_for_odr): Declare.
14018 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14019 Jason Merrill <jason@redhat.com>
14021 * system.h: Include <cstdlib> as well as <stdlib.h>.
14023 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14026 * config/i386/i386.c (construct_container): Report error if
14027 long double is used with disabled x87 float returns.
14029 2013-06-20 Jan Hubicka <jh@suse.cz>
14031 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14033 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14035 PR rtl-optimization/57425
14036 PR rtl-optimization/57569
14037 * alias.c (write_dependence_p): Remove parameters mem_mode and
14038 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14039 Changed all callers.
14040 (canon_anti_dependence): Get comments and semantics in sync.
14041 Add parameter mem_canonicalized. Changed all callers.
14042 * rtl.h (canon_anti_dependence): Update prototype.
14044 2013-06-20 Richard Biener <rguenther@suse.de>
14046 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14047 byte case, inline streamer_read_uchar and defer section
14050 2013-06-20 Richard Biener <rguenther@suse.de>
14052 PR tree-optimization/57584
14053 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14054 SSA names into the expanded expression that take part in
14055 abnormal coalescing.
14057 2013-06-19 Sharad Singhai <singhai@google.com>
14059 * gcov.c (print_usage): Handle new option.
14060 (process_args): Ditto.
14061 (get_gcov_intermediate_filename): New function.
14062 (output_intermediate_file): New function.
14063 (output_gcov_file): New function
14064 (generate_results): Handle new option.
14065 (release_function): Relase demangled name.
14066 (read_graph_file): Handle demangled name.
14067 (output_lines): Ditto.
14068 * doc/gcov.texi: Document gcov intermediate format.
14070 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14073 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14074 (lra_emit_add): Use the functions. Add comment about Y as an
14077 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14080 * collect2.c (collect_atexit): New.
14081 (collect_exit): Delete.
14082 (main): Register collect_atexit with atexit.
14083 (collect_wait): Change collect_exit to exit.
14085 * collect2.h (collect_exit): Delete.
14086 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14088 2013-06-19 Wei Mi <wmi@google.com>
14090 PR rtl-optimization/57518
14091 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14092 if regno is used in paradoxical subreg.
14093 (update_equiv_regs): Check pdx_subregs[regno] before
14094 set a reg to be equivalent with a mem.
14096 2013-06-19 Matthias Klose <doko@ubuntu.com>
14099 * file-find.h (find_a_file): Add a mode parameter.
14100 * file-find.c (find_a_file): Likewise.
14101 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14102 with X_OK for the executables.
14103 * collect2.c (main): Call find_a_file with X_OK.
14105 2013-06-19 Steve Ellcey <sellcey@mips.com>
14108 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14109 Use NEXT_INSN instead of next_real_insn.
14111 2013-06-19 Jan Hubicka <jh@suse.cz>
14113 * cgraph.h (const_value_known_p): Replace by ...
14114 (ctor_for_folding): .. this one.
14115 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14116 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14117 * expr.c (expand_expr_real_1): Likewise.
14118 (string_constant): Likewise.
14119 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14120 * ipa.c (process_references): Likewise.
14121 (symtab_remove_unreachable_nodes): Likewise.
14122 * ipa-inline-analysis.c (param_change_prob): Likewise.
14123 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14124 (get_base_constructor): Likwise.
14125 * varpool.c (varpool_remove_node): Likewise.
14126 (varpool_remove_initializer): LIkewise.
14127 (dump_varpool_node): LIkwise.
14128 (const_value_known_p): Rewrite to ...
14129 (ctor_for_folding): ... this one.
14131 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14134 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14135 PERSONALITY in $PATH derived prefixes.
14137 2013-06-19 Jeff Law <law@redhat.com>
14139 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14142 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14143 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14146 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14148 * config/arm/vfp.md: Move VFP instruction classification documentation
14150 * config/arm/arm.md: ... here. Update instruction classification
14153 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14155 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14156 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14157 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14159 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14162 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14163 with NEXT_INSN. Use tablejump_p to check for jump table data
14166 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14169 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14170 that now in C++ the value is correct per the C++ standards.
14172 2013-06-19 Richard Biener <rguenther@suse.de>
14174 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14175 for global context.
14177 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14180 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14183 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14184 with next_active_insn.
14186 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14188 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14189 functions are inlined during failures to flag an error.
14190 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14191 in early inline pass.
14193 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14195 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14198 2013-06-18 Julian Brown <julian@codesourcery.com>
14200 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14201 Permit virtual register pre-reload if !strict.
14202 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14204 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14206 * config/arm/neon.md (movmisalign<mode>): Use
14207 neon_perm_struct_or_reg_operand instead of
14208 neon_struct_or_register_operand.
14209 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14210 neon_permissive_struct_operand instead of neon_struct_operand.
14211 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14212 neon_vector_mem_operand.
14213 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14214 neon_vector_mem_operand.
14215 (neon_permissive_struct_operand): New.
14216 (neon_struct_or_register_operand): Rename to...
14217 (neon_perm_struct_or_reg_operand): This. Adjust call to
14218 neon_vector_mem_operand.
14220 2013-06-18 Richard Biener <rguenther@suse.de>
14222 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14223 * lto-streamer.h: Include pointer-set.h.
14224 (struct lto_decl_slot): Remove.
14225 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14226 Remove next_index entry.
14227 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14228 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14229 (lto_init_tree_ref_encoder): Adjust.
14230 (lto_destroy_tree_ref_encoder): Likewise.
14231 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14232 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14233 (lto_output_decl_index): Adjust.
14234 (lto_new_out_decl_state): Likewise.
14235 (lto_record_function_out_decl_state): Likewise.
14236 * lto-streamer-out.c (copy_function): Likewise.
14238 2013-06-18 Richard Biener <rguenther@suse.de>
14240 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14241 * cgraphunit.c: Include cfgloop.h.
14242 (init_lowered_empty_function): Initialize the loop tree.
14243 (assemble_thunk): Insert new BBs into loops.
14245 2013-06-18 Richard Biener <rguenther@suse.de>
14247 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14248 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14249 the map from cache entry to cache index optional.
14250 (streamer_tree_cache_replace_tree): Adjust accordingly.
14251 (streamer_tree_cache_append): Likewise.
14252 (streamer_tree_cache_delete): Likewise.
14253 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14254 streamer cache map from cache entry to cache index.
14255 * lto-streamer-out.c (create_output_block): Adjust.
14257 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14259 * config/arm/arm.md (attribute "insn"): Move multiplication and
14260 division attributes to...
14261 (attribute "type"): ... here. Remove mult.
14262 (attribute "mul32"): New attribute.
14263 (attribute "mul64"): Add umaal.
14264 (*arm_mulsi3): Update attributes.
14265 (*arm_mulsi3_v6): Likewise.
14266 (*thumb_mulsi3): Likewise.
14267 (*thumb_mulsi3_v6): Likewise.
14268 (*mulsi3_compare0): Likewise.
14269 (*mulsi3_compare0_v6): Likewise.
14270 (*mulsi_compare0_scratch): Likewise.
14271 (*mulsi_compare0_scratch_v6): Likewise.
14272 (*mulsi3addsi): Likewise.
14273 (*mulsi3addsi_v6): Likewise.
14274 (*mulsi3addsi_compare0): Likewise.
14275 (*mulsi3addsi_compare0_v6): Likewise.
14276 (*mulsi3addsi_compare0_scratch): Likewise.
14277 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14278 (*mulsi3subsi): Likewise.
14279 (*mulsidi3adddi): Likewise.
14280 (*mulsi3addsi_v6): Likewise.
14281 (*mulsidi3adddi_v6): Likewise.
14282 (*mulsidi3_nov6): Likewise.
14283 (*mulsidi3_v6): Likewise.
14284 (*umulsidi3_nov6): Likewise.
14285 (*umulsidi3_v6): Likewise.
14286 (*umulsidi3adddi): Likewise.
14287 (*umulsidi3adddi_v6): Likewise.
14288 (*smulsi3_highpart_nov6): Likewise.
14289 (*smulsi3_highpart_v6): Likewise.
14290 (*umulsi3_highpart_nov6): Likewise.
14291 (*umulsi3_highpart_v6): Likewise.
14292 (mulhisi3): Likewise.
14293 (*mulhisi3tb): Likewise.
14294 (*mulhisi3bt): Likewise.
14295 (*mulhisi3tt): Likewise.
14296 (maddhisi4): Likewise.
14297 (*maddhisi4tb): Likewise.
14298 (*maddhisi4tt): Likewise.
14299 (maddhidi4): Likewise.
14300 (*maddhidi4tb): Likewise.
14301 (*maddhidi4tt): Likewise.
14302 (divsi3): Likewise.
14303 (udivsi3): Likewise.
14304 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14305 (thumb2_mulsi_short_compare0): Likewise.
14306 (thumb2_mulsi_short_compare0_scratch): Likewise.
14307 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14308 (1020mult2): Likewise.
14309 (1020mult3): Likewise.
14310 (1020mult4): Likewise.
14311 (1020mult5): Likewise.
14312 (1020mult6): Likewise.
14313 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14315 (cortex_a15_mult64): Likewise.
14316 (cortex_a15_sdiv): Likewise.
14317 (cortex_a15_udiv): Likewise.
14318 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14324 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14325 (pj4_ir_div): Likewise.
14326 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14327 (11_mult2): Likewise.
14328 (11_mult3): Likewise.
14329 (11_mult4): Likewise.
14330 (11_mult5): Likewise.
14331 (11_mult6): Likewise.
14332 (11_mult7): Likewise.
14333 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14334 (cortex_a8_mla): Likewise.
14335 (cortex_a8_mull): Likewise.
14336 (cortex_a8_smulwy): Likewise.
14337 (cortex_a8_smlald): Likewise.
14338 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14339 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14340 (cortex_r4_mul_3): Likewise.
14341 (cortex_r4_mla_4): Likewise.
14342 (cortex_r4_mla_3): Likewise.
14343 (cortex_r4_smlald): Likewise.
14344 (cortex_r4_mull): Likewise.
14345 (cortex_r4_sdiv): Likewise.
14346 (cortex_r4_udiv): Likewise.
14347 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14348 (cortex_a7_idiv): Likewise.
14349 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14350 (9_mult2): Likewise.
14351 (9_mult3): Likewise.
14352 (9_mult4): Likewise.
14353 (9_mult5): Likewise.
14354 (9_mult6): Likewise.
14355 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14356 (cortex_a53_sdiv): Likewise.
14357 (cortex_a53_udiv): Likewise.
14358 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14359 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14360 (mp626_mult2): Likewise.
14361 (mp626_mult3): Likewise.
14362 (mp626_mult4): Likewise.
14363 * config/arm/fa526.md (526_mult1): Update attribute change.
14364 (526_mult2): Likewise.
14365 * config/arm/arm-generic.md (mult): Update attribute change.
14366 (mult_ldsched_strongarm): Likewise.
14367 (mult_ldsched): Likewise.
14368 (multi_cycle): Likewise.
14369 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14370 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14371 (606te_mult2): Likewise.
14372 (606te_mult3): Likewise.
14373 (606te_mult4): Likewise.
14374 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14375 (cortex_a9_mac16): Likewise.
14376 (cortex_a9_multiply): Likewise.
14377 (cortex_a9_mac): Likewise.
14378 (cortex_a9_multiply_long): Likewise.
14379 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14380 (626te_mult2): Likewise.
14381 (626te_mult3): Likewise.
14382 (626te_mult4): Likewise.
14384 2013-06-18 Richard Biener <rguenther@suse.de>
14387 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14389 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14392 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14393 with next_active_insn.
14395 2013-06-18 Alan Modra <amodra@gmail.com>
14397 * config/rs6000/rs6000.h (enum data_align): New.
14398 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14399 (DATA_ABI_ALIGNMENT): Define.
14400 (CONSTANT_ALIGNMENT): Correct comment.
14401 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14402 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14404 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14406 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14407 ATTRIBUTE_UNUSED marking.
14409 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14411 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14412 alternative and update.
14413 (aarch64_dup_lanedi): Delete.
14414 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14415 * config/aarch64/aarch64-simd-builtins.def: Update.
14417 2013-06-17 Richard Biener <rguenther@suse.de>
14419 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14420 (lto_input_scc): Declare.
14421 (lto_input_tree_1): Likewise.
14422 (struct lto_stats_d): Add num_tree_bodies_output and
14423 num_pickle_refs_output.
14424 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14425 (lto_read_tree_1): Split out from ...
14426 (lto_read_tree): ... this.
14427 (lto_input_scc): New function.
14428 (lto_input_tree_1): Split out from ...
14429 (lto_input_tree): ... this. Handle LTO_tree_scc.
14430 (lto_data_in_create): Create the streamer cache without hashes.
14431 * lto-streamer-out.c (create_output_block): Create the streamer
14432 cache with hashes when not doing WPA.
14433 (lto_write_tree_1): Split out from ...
14434 (lto_write_tree): ... this.
14435 (get_symbol_initial_value): New function.
14436 (lto_output_tree_1): Split out from ...
14437 (lto_output_tree): ... this. Write trees as series of SCCs
14438 using a DFS walk via DFS_write_tree.
14439 (struct sccs, struct scc_entry): New types.
14440 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14441 (DFS_write_tree_body): New function.
14442 (DFS_write_tree): Likewise.
14443 (hash_tree): Likewise.
14444 (scc_entry_compare): Likewise.
14445 (hash_scc): Likewise.
14446 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14447 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14448 TREE_CHAIN as regular reference.
14449 (streamer_read_integer_cst): Remove.
14450 (streamer_get_pickled_tree): Adjust.
14451 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14452 of DECL_EXTERNALs in BLOCK_VARS for now.
14453 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14455 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14456 Add hash value argument and record that if hashes are recorded
14458 (streamer_tree_cache_insert_1): Adjust.
14459 (streamer_tree_cache_insert): Likewise.
14460 (streamer_tree_cache_insert_at): Rename to ...
14461 (streamer_tree_cache_replace_tree): ... this and adjust.
14462 (streamer_tree_cache_append): Adjust.
14463 (record_common_node): Likewise.
14464 (streamer_tree_cache_create): Add argument whether to
14465 record hash values together with trees.
14466 (streamer_tree_cache_delete): Adjust.
14467 * tree-streamer.h (struct streamer_tree_cache_d): Add
14469 (streamer_read_integer_cst): Remove.
14470 (streamer_tree_cache_insert): Adjust.
14471 (streamer_tree_cache_append): Likewise.
14472 (streamer_tree_cache_insert_at): Rename to ...
14473 (streamer_tree_cache_replace_tree): ... this and adjust.
14474 (streamer_tree_cache_create): Add argument whether to record hashes.
14475 (streamer_tree_cache_get): Rename to ...
14476 (streamer_tree_cache_get_tree): ... this.
14477 (streamer_tree_cache_get_hash): New function.
14478 * tree.c (cache_integer_cst): New function.
14479 * tree.h (cache_integer_cst): Declare.
14480 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14481 * lto-symtab.c (lto_varpool_replace_node): Only release
14482 DECL_INITIAL of non-prevailing decls.
14483 * varpool.c (varpool_remove_initializer): Do not release
14484 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14486 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14488 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14489 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14490 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14491 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14492 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14493 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14494 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14495 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14496 instead of TARGET_64BIT.
14497 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14498 Require ISA_HAS_<D>DIV.
14500 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14502 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14503 (mips*-*-linux*): Move default with_llsc setting to where other
14505 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14507 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14508 Likewise. Remove default with_tune setting. Move default float
14509 setting to its own block. Handle with_llsc in the same block as above.
14511 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14513 PR rtl-optimization/57425
14514 PR rtl-optimization/57569
14515 * alias.c (write_dependence_p): Add new parameters mem_mode,
14516 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14517 Changed all callers.
14518 (canon_anti_dependence): New function.
14519 * cse.c (check_dependence): Use canon_anti_dependence.
14520 * cselib.c (cselib_invalidate_mem): Likewise.
14521 * rtl.h (canon_anti_dependence): Declare.
14523 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14525 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14526 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14527 ".set mips3" for 64-bit targets.
14529 2013-06-15 Dehao Chen <dehao@google.com>
14531 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14532 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14533 (gimple_check_call_args): Likewise.
14534 * value-prof.c (check_ic_target): Likewise.
14535 * ipa-inline.c (early_inliner): Likewise.
14536 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14537 * cgraph.c (cgraph_create_edge_1): Likewise.
14538 (cgraph_make_edge_direct): Likewise.
14540 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14543 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14544 rs6000_output_move_128bit to handle emitting quad memory
14545 operations. Set attribute length to 8 bytes.
14547 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14549 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14551 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14552 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14553 (aarch64_<su>mlsl<mode>): Likewise.
14555 2013-06-14 Mike Stump <mikestump@comcast.net>
14557 * Makefile.in (TARGET_H): Add insn-codes.h.
14559 2013-06-14 Alan Modra <amodra@gmail.com>
14561 PR middle-end/57134
14562 PR middle-end/57586
14563 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14564 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14565 bitfield expansion when EXPAND_MEMORY.
14566 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14568 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14570 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14571 test for clearing quad memory on 32-bit later.
14573 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14575 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14576 (fold_negate_expr): Likewise.
14577 (fold_real_zero_addition_p): Handle vectors.
14578 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14580 2013-06-14 Alan Modra <amodra@gmail.com>
14582 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14584 2013-06-13 Jan Hubicka <jh@suse.cz>
14586 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14587 Local comdats are not externally visible.
14588 * symtab.c (dump_symtab_base): Dump externally visible.
14589 (verify_symtab_base): Verify back links in the symtab hash.
14591 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14593 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14594 CONVERT_EXPR as equal nodes.
14596 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14598 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14600 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14602 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14603 Generalize to complex and vector.
14604 * tree.c (build_all_ones_cst): New function.
14605 * tree.h (build_all_ones_cst): Declare it.
14607 2013-06-13 Alan Modra <amodra@gmail.com>
14609 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14610 * config/rs6000/rs6000.md (signbittf2): New insn.
14611 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14612 (abstf2_internal, cmptf_internal2): Likewise.
14613 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14615 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14616 Pat Haugen <pthaugen@us.ibm.com>
14617 Peter Bergner <bergner@vnet.ibm.com>
14619 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14620 power8 byte, half-word, and quad-word atomic instructions.
14621 (emit_store_conditional): Likewise.
14622 (rs6000_expand_atomic_compare_and_swap): Likewise.
14623 (rs6000_expand_atomic_op): Likewise.
14625 * config/rs6000/sync.md (larx): Add new modes for power8.
14627 (AINT): New mode iterator to include TImode as well as normal
14628 integer modes on power8.
14629 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14630 that VSX registers are not considered. Use AINT mode iterator
14631 instead of INT1 to allow inclusion of quad word atomic operations
14633 (load_locked<mode>): Likewise.
14634 (store_conditional<mode>): Likewise.
14635 (atomic_compare_and_swap<mode>): Likewise.
14636 (atomic_exchange<mode>): Likewise.
14637 (atomic_nand<mode>): Likewise.
14638 (atomic_fetch_<fetchop_name><mode>): Likewise.
14639 (atomic_nand_fetch<mode>): Likewise.
14640 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14642 (ATOMIC): On power8, add QImode, HImode modes.
14643 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14644 modes that promote to SImode.
14645 (load_lockedti): Convert TImode arguments to PTImode, so that we
14646 get a guaranteed even/odd register pair.
14647 (load_lockedpti): Likewise.
14648 (store_conditionalti): Likewise.
14649 (store_conditionalpti): Likewise.
14651 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14652 atomic load/store instructions.
14655 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14657 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14659 (insn_count): New attribute, with most cases extracted from...
14660 (length): ...here. Redefine most cases in terms of insn_count.
14661 (single_insn): Delete.
14662 (can_delay): Use insn_count to check for single instructions.
14663 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14664 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14665 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14666 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14667 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14668 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14669 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14670 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14671 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14672 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14673 rather than "length".
14674 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14675 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14676 Use "insn_count" rather than "length".
14677 * config/mips/mips-dsp.md
14678 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14679 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14682 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14684 PR tree-optimization/57361
14685 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14687 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14689 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14691 (aarch64_simd_combine<mode>): New instruction expansion.
14692 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14693 function prototype.
14694 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14695 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14697 2013-06-12 Jan Hubicka <jh@suse.cz>
14699 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14700 decl has when in streaming stage.
14701 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14702 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14704 2013-06-12 Roland Stigge <stigge@antcom.de>
14707 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14709 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14711 PR tree-optimization/57537
14712 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14713 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14715 2013-06-12 Richard Biener <rguenther@suse.de>
14717 * data-streamer.h (streamer_write_char_stream): CSE
14718 obs->current_pointer.
14719 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14720 streamer_write_char_stream manually and optimize the resulting loop.
14721 (streamer_write_hwi_stream): Likewise.
14723 2013-06-12 Jan Hubicka <jh@suse.cz>
14725 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14726 * cgraph.h (varpool_create_empty_node): Declare.
14727 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14729 * symtab.c (symtab_unregister_node): Be lax about missin entries
14731 (symtab_get_node): Update comment.
14732 * varpool.c (varpool_create_empty_node): Break out from ...
14733 (varpool_node_for_decl): ... here.
14734 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14736 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14738 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14739 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14740 part. Use straight-line flow at the end.
14741 <COMPONENT_REF>: Remove superfluous else.
14742 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14744 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14747 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14748 target hook even for !TREE_PUBLIC decls. If no resolution info
14749 is available, return false for common and external decls.
14751 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14753 * config/rl78/constraints.md (U): New constraint.
14754 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14757 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14760 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14761 to allow returning address to AT_PLATFORM name.
14763 2013-06-11 Jan Hubicka <jh@suse.cz>
14765 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14766 * cgraph.h (symtab_node_base): Add weakref flag.
14767 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14768 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14769 (output_weakrefs): Use weakref flag.
14770 * fold-const.c (simple_operand_p): Handle WEAK.
14771 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14772 * ipa.c (varpool_externally_visible_p): Drop weakref.
14773 (function_and_variable_visibility): Update comment; fix weakref
14774 sanity checks; do not clear DECL_WEAK on them.
14775 * lto-cgraph.c (lto_output_node): update.
14776 (lto_output_varpool_node): Update.
14777 (input_overwrite_node): Update.
14778 (input_node): Update.
14779 (input_varpool_node): Update.
14780 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14781 (lto_symtab_merge_symbols): Add sanity check.
14782 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14783 * passes.c (rest_of_decl_compilation): Set static flag, too.
14784 * symtab.c (dump_symtab_base): Dump weakref.
14785 (verify_symtab_base): Sanity check weakrefs.
14786 (symtab_make_decl_local): Remove duplicated code.
14787 (symtab_alias_ultimate_target): Simplify.
14788 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14790 2013-06-11 Tom de Vries <tom@codesourcery.com>
14792 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14793 sequence_vect == NULL.
14795 2013-06-11 DJ Delorie <dj@redhat.com>
14797 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14798 (rl78_unwind_word_mode): New.
14800 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14802 * final.c (debug_prefix_maps): Make static.
14804 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14806 * function.c (initial_trampoline): Remove stray copy.
14808 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14810 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14812 2013-06-11 Martin Jambor <mjambor@suse.cz>
14814 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14815 within bounds at the beginning of the function.
14817 2013-06-11 Alan Modra <amodra@gmail.com>
14819 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14821 (get_named_section): Don't NULL !DECL_P decl.
14823 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14825 * doc/invoke.texi (core-avx2): Document.
14827 (atom): Updated with MOVBE.
14829 2013-06-11 Richard Biener <rguenther@suse.de>
14831 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14833 2013-06-11 Anton Blanchard <anton@samba.org>
14835 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14836 correct shift value in little-endian mode.
14838 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14841 * varasm.c (get_variable_align): Move #endif to the right place.
14843 2013-06-10 Cary Coutant <ccoutant@google.com>
14845 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14846 for hash so that hash table traversal order is deterministic.
14848 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14849 Pat Haugen <pthaugen@us.ibm.com>
14850 Peter Bergner <bergner@vnet.ibm.com>
14852 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14853 of vectors in GPRS if they are direct moves or quad word load or
14856 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14858 (direct_move_p): Likewise.
14859 (quad_load_store_p): Likewise.
14861 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14862 classes into bins based on the physical register type.
14863 (reg_class_to_reg_type): Likewise.
14864 (IS_STD_REG_TYPE): Likewise.
14865 (IS_FP_VECT_REG_TYPE): Likewise.
14866 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14867 use direct move instructions.
14868 (reload_gpr_vsx): Likewise.
14869 (reload_vsx_gpr): Likewise.
14870 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14871 information that is a simplification of register classes. Also
14872 precalculate direct move reload helpers.
14873 (direct_move_p): New function to return true if the operation can
14874 be done as a direct move instruciton.
14875 (quad_load_store_p): New function to return true if the operation
14876 is a quad memory operation.
14877 (rs6000_legitimize_address): If quad memory, only allow register
14878 indirect for TImode addresses.
14879 (rs6000_legitimate_address_p): Likewise.
14880 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14881 (rs6000_reload_register_type): Likewise.
14882 (register_to_reg_type): Return register type.
14883 (rs6000_secondary_reload_simple_move): New helper function for
14884 secondary reload and secondary memory needed to identify anything
14885 that is a simple move, and does not need reloading.
14886 (rs6000_secondary_reload_direct_move): New helper function for
14887 secondary reload to identify cases that can be done with several
14888 instructions via the direct move instructions.
14889 (rs6000_secondary_reload_move): New helper function for secondary
14890 reload to identify moves between register types that can be done.
14891 (rs6000_secondary_reload): Add support for quad memory operations
14892 and for direct move.
14893 (rs6000_secondary_memory_needed): Likewise.
14894 (rs6000_debug_secondary_memory_needed): Change argument names.
14895 (rs6000_output_move_128bit): New function to return the move to
14896 use for 128-bit moves, including knowing about the various
14897 limitations of quad memory operations.
14899 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14900 memory operations. call rs6000_output_move_128bit for the actual
14901 instruciton(s) to generate.
14902 (vsx_movti_64bit): Likewise.
14904 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14905 (UNSPEC_P8V_MTVSRWZ): Likewise.
14906 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14907 (UNSPEC_P8V_MTVSRD): Likewise.
14908 (UNSPEC_P8V_XXPERMDI): Likewise.
14909 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14910 (UNSPEC_FUSION_GPR): Likewise.
14911 (FMOVE128_GPR): New iterator for direct move.
14912 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14913 (f32_sv): Likewise.
14914 (f32_dm): Likewise.
14915 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14916 loads and direct move instructions.
14917 (zero_extendsidi2_lfiwzx): Likewise.
14918 (extendsidi2_lfiwax): Likewise.
14919 (extendsidi2_nocell): Likewise.
14920 (floatsi<mode>2_lfiwax): Likewise.
14921 (lfiwax): Likewise.
14922 (floatunssi<mode>2_lfiwzx): Likewise.
14923 (lfiwzx): Likewise.
14924 (fix_trunc<mode>_stfiwx): Likewise.
14925 (fixuns_trunc<mode>_stfiwx): Likewise.
14926 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14927 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14928 (parity<mode>2_cmpb): Set length/type attr.
14929 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14930 for 'mr.' to fast_compare.
14931 (bpermd_<mode>): Change type attr to popcnt.
14932 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14933 (p8_mtvsrwz_1): Likewise.
14934 (p8_mtvsrwz_2): Likewise.
14935 (reload_fpr_from_gpr<mode>): Likewise.
14936 (p8_mtvsrd_1): Likewise.
14937 (p8_mtvsrd_2): Likewise.
14938 (p8_xxpermdi_<mode>): Likewise.
14939 (reload_vsx_from_gpr<mode>): Likewise.
14940 (reload_vsx_from_gprsf): Likewise.
14941 (p8_mfvsrd_3_<mode>): LIkewise.
14942 (reload_gpr_from_vsx<mode>): Likewise.
14943 (reload_gpr_from_vsxsf): Likewise.
14944 (p8_mfvsrd_4_disf): Likewise.
14945 (multi-word GPR splits): Do not split direct moves or quad memory
14948 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14950 * tree-into-ssa.c (interesting_blocks): Make static.
14952 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14955 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14956 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14957 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14958 (get_variable_align): New function.
14959 (get_variable_section, emit_bss, emit_common,
14960 assemble_variable_contents, place_block_symbol): Use
14961 get_variable_align instead of DECL_ALIGN.
14962 (assemble_noswitch_variable): Add align argument, use it
14963 instead of DECL_ALIGN.
14964 (assemble_variable): Adjust caller. Use get_variable_align
14965 instead of DECL_ALIGN.
14966 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14968 (DATA_ABI_ALIGNMENT): Define.
14969 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14970 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14971 opt is false, only return the psABI mandated alignment increase.
14972 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14973 (DATA_ABI_ALIGNMENT): ... this.
14974 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14975 (DATA_ABI_ALIGNMENT): ... this.
14976 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14977 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14978 (DATA_ABI_ALIGNMENT): ... this.
14979 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14980 * doc/tm.texi: Regenerated.
14982 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14984 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14985 cmp_code to construct REG_EQUAL note.
14987 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14990 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14991 that operands[2] doesn't overlap with operands[0].
14993 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14994 Jan Hubicka <jh@suse.cz>
14996 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14997 hack to mark symbols as used.
14999 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15001 PR rtl-optimization/57559
15002 * lra-constraints.c (process_alt_operands): Don't discourage
15003 memory with known offset for offsetable memory constraint.
15004 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15006 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15008 * varasm.c (struct oc_local_state): Reorder fields.
15009 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15010 and adjust accordingly.
15011 (output_constructor): Reorder initialization code and adjust call to
15012 output_constructor_bitfield.
15014 2013-06-07 Jan Hubicka <jh@suse.cz>
15016 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15018 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15020 * tree-object-size.c (unknown): Make const.
15022 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15024 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15025 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15026 for last alternative in the cpu_facility attribute.
15028 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15031 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15032 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15033 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15034 * config/arm/constraints.md (Dg): New constraint.
15035 * config/arm/neon.md (xordi3_neon): Remove.
15036 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15037 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15039 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15041 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15042 Clean up alternatives.
15044 2013-06-07 Alan Modra <amodra@gmail.com>
15046 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15049 2013-06-07 Alan Modra <amodra@gmail.com>
15051 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15053 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15055 * config/arm/constraints.md (Df): New constraint.
15056 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15057 Correct length attribute for last two alternatives.
15059 2013-06-07 Alan Modra <amodra@gmail.com>
15061 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15062 override user -mfp-in-toc.
15063 (offsettable_ok_by_alignment): Consider just the current access
15064 rather than the whole object, unless BLKmode. Handle
15065 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15066 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15067 for -mcmodel=medium.
15068 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15069 override user -mfp-in-toc or -msum-in-toc. Default to
15070 -mno-fp-in-toc for -mcmodel=medium.
15072 2013-06-06 DJ Delorie <dj@redhat.com>
15074 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15075 TARGET_VALID_POINTER_MODE.
15077 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15078 Pat Haugen <pthaugen@us.ibm.com>
15079 Peter Bergner <bergner@vnet.ibm.com>
15081 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15082 Document new power8 builtins.
15084 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15085 condition code register, to allow 128-bit logical operations to be
15086 done in the VSX or GPR registers.
15087 (nor<mode>3): Use the canonical form for nor.
15088 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15089 vclz*, and vpopcnt* vector instructions.
15090 (nand<mode>3): Likewise.
15091 (orc<mode>3): Likewise.
15092 (clz<mode>2): LIkewise.
15093 (popcount<mode>2): Likewise.
15095 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15096 that only the GPRs are recognized.
15098 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15099 support for new power8 builtins.
15101 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15103 (xscvdpspn): Likewise.
15109 (vpopcnt): Likewise.
15110 (vpopcntb): Likewise.
15111 (vpopcnth): Likewise.
15112 (vpopcntw): Likewise.
15113 (vpopcntd): Likewise.
15115 (vmrgew): Likewise.
15116 (vmrgow): Likewise.
15118 (eqv_v16qi3): Likewise.
15119 (eqv_v8hi3): Likewise.
15120 (eqv_v4si3): Likewise.
15121 (eqv_v2di3): Likewise.
15122 (eqv_v4sf3): Likewise.
15123 (eqv_v2df3): Likewise.
15125 (nand_v16qi3): Likewise.
15126 (nand_v8hi3): Likewise.
15127 (nand_v4si3): Likewise.
15128 (nand_v2di3): Likewise.
15129 (nand_v4sf3): Likewise.
15130 (nand_v2df3): Likewise.
15132 (orc_v16qi3): Likewise.
15133 (orc_v8hi3): Likewise.
15134 (orc_v4si3): Likewise.
15135 (orc_v2di3): Likewise.
15136 (orc_v4sf3): Likewise.
15137 (orc_v2df3): Likewise.
15139 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15140 allow power8 quad mode in 64-bit.
15141 (rs6000_builtin_vectorized_function): Add support to vectorize
15142 ISA 2.07 count leading zeros, population count builtins.
15143 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15144 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15145 (builtin_function_type): Add vgbbd builtin function which takes an
15147 (altivec_expand_vec_perm_const): Add support for new power8 merge
15150 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15151 that does not include TImdoe for use with 32-bit.
15152 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15154 (UNSPEC_VSX_CVDPSPN): Likewise.
15155 (vsx_xscvdpspn): Likewise.
15156 (vsx_xscvspdpn): Likewise.
15157 (vsx_xscvdpspn_scalar): Likewise.
15158 (vsx_xscvspdpn_directmove): Likewise.
15159 (vsx_and<mode>3): Split logical operations into 32-bit and
15160 64-bit. Add support to do logical operations on TImode as well as
15161 VSX vector types. Allow logical operations to be done in either
15162 VSX registers or in general purpose registers in 64-bit mode. Add
15163 splitters if GPRs were used. For AND, add clobber of CCmode to
15164 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15166 (vsx_and<mode>3_32bit): Likewise.
15167 (vsx_and<mode>3_64bit): Likewise.
15168 (vsx_ior<mode>3): Likewise.
15169 (vsx_ior<mode>3_32bit): Likewise.
15170 (vsx_ior<mode>3_64bit): Likewise.
15171 (vsx_xor<mode>3): Likewise.
15172 (vsx_xor<mode>3_32bit): Likewise.
15173 (vsx_xor<mode>3_64bit): Likewise.
15174 (vsx_one_cmpl<mode>2): Likewise.
15175 (vsx_one_cmpl<mode>2_32bit): Likewise.
15176 (vsx_one_cmpl<mode>2_64bit): Likewise.
15177 (vsx_nor<mode>3): Likewise.
15178 (vsx_nor<mode>3_32bit): Likewise.
15179 (vsx_nor<mode>3_64bit): Likewise.
15180 (vsx_andc<mode>3): Likewise.
15181 (vsx_andc<mode>3_32bit): Likewise.
15182 (vsx_andc<mode>3_64bit): Likewise.
15183 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15184 and xxlorc instructions.
15185 (vsx_eqv<mode>3_64bit): Likewise.
15186 (vsx_nand<mode>3_32bit): Likewise.
15187 (vsx_nand<mode>3_64bit): Likewise.
15188 (vsx_orc<mode>3_32bit): Likewise.
15189 (vsx_orc<mode>3_64bit): Likewise.
15191 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15193 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15195 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15196 (p8_vmrgow): Likewise.
15197 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15198 GPRs to be split under VSX.
15199 (p8v_clz<mode>2): Add power8 count leading zero support.
15200 (p8v_popcount<mode>2): Add power8 population count support.
15201 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15204 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15207 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15209 (vec_nand): Likewise.
15210 (vec_vclz): Likewise.
15211 (vec_vclzb): Likewise.
15212 (vec_vclzd): Likewise.
15213 (vec_vclzh): Likewise.
15214 (vec_vclzw): Likewise.
15215 (vec_vgbbd): Likewise.
15216 (vec_vmrgew): Likewise.
15217 (vec_vmrgow): Likewise.
15218 (vec_vpopcnt): Likewise.
15219 (vec_vpopcntb): Likewise.
15220 (vec_vpopcntd): Likewise.
15221 (vec_vpopcnth): Likewise.
15222 (vec_vpopcntw): Likewise.
15224 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15226 PR rtl-optimization/57468
15227 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15230 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15232 PR rtl-optimization/57459
15233 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15234 type when setting live regs.
15236 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15238 * config/s390/s390.opt (mlra): New option.
15239 * config/s390/s390.c (s390_decompose_address): Check displacement
15240 for all registers for LRA.
15241 (s390_secondary_reload): Don't used secondary reloads for LRA.
15242 (s390_lra_p): New function.
15243 (TARGET_LRA_P): Define.
15244 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15245 of attribute cpu_facility to zarch for the last alternative.
15246 (*cmpmem_short): Ditto.
15248 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15250 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15251 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15252 (arm_expand_prologue): Likewise.
15254 2013-06-06 Teresa Johnson <tejohnson@google.com>
15257 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15258 as this is now done by redirect_edge_and_branch_force.
15259 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15260 barriers, and fix interaction with splitting.
15261 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15262 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15263 reflect changes made in the routine.
15264 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15265 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15266 since this is called in cfglayout mode, and replace partition fixup
15267 with assert as that is now done by force_nonfallthru_and_redirect.
15268 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15269 already be marked with region crossing note.
15270 (insert_section_boundary_note): Make non-static, gate on flag
15271 has_bb_partition, rewrite to also check for multiple partitions.
15272 (rest_of_handle_reorder_blocks): Remove call to
15273 insert_section_boundary_note, now done later during free_cfg.
15274 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15275 * bb-reorder.h (insert_section_boundary_note): Declare.
15276 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15277 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15278 invoke insert_section_boundary_note.
15279 (try_redirect_by_replacing_jump): Remove unnecessary
15280 check for region crossing note.
15281 (fixup_partition_crossing): New function.
15282 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15283 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15284 in non-cfglayout mode.
15285 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15286 remove old code that tried to do this. Emit barrier correctly
15287 when we are in cfglayout mode.
15288 (last_bb_in_partition): New function.
15289 (rtl_split_edge): Correctly fixup partition boundaries.
15290 (commit_one_edge_insertion): Remove old code that tried to
15291 fixup region crossing edge since this is now handled in
15292 split_block, and set up insertion point correctly since
15293 block may now end in a jump.
15294 (verify_hot_cold_block_grouping): Guard against checking when not in
15295 linearized RTL mode.
15296 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15298 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15299 rtl_verify_flow_info, so not called in cfglayout mode.
15300 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15301 (fixup_reorder_chain): Remove old code that attempted to fixup region
15302 crossing note as this is now handled in force_nonfallthru_and_redirect.
15303 (duplicate_insn_chain): Don't duplicate switch section notes.
15304 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15306 * basic-block.h (emit_barrier_after_bb): Declare.
15308 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15310 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15311 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15312 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15314 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15316 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15318 * config/arm/ldmstm.md: Regenerate.
15320 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15322 * config/arm/sync.md (atomic_loaddi_1):
15323 Disable predication for arm_restrict_it.
15324 (arm_load_exclusive<mode>): Likewise.
15325 (arm_load_exclusivesi): Likewise.
15326 (arm_load_exclusivedi): Likewise.
15327 (arm_load_acquire_exclusive<mode>): Likewise.
15328 (arm_load_acquire_exclusivesi): Likewise.
15329 (arm_load_acquire_exclusivedi): Likewise.
15330 (arm_store_exclusive<mode>): Likewise.
15331 (arm_store_exclusive<mode>): Likewise.
15332 (arm_store_release_exclusivedi): Likewise.
15333 (arm_store_release_exclusive<mode>): Likewise.
15335 2013-06-06 Richard Biener <rguenther@suse.de>
15337 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15339 (lto_tag_is_tree_code_p): Adjust.
15340 (lto_tag_is_gimple_code_p): Likewise.
15341 (lto_gimple_code_to_tag): Likewise.
15342 (lto_tag_to_gimple_code): Likewise.
15343 (lto_tree_code_to_tag): Likewise.
15344 (lto_tag_to_tree_code): Likewise.
15345 * data-streamer.h (streamer_write_hwi_in_range): Use
15346 uhwi streaming to stream the normalized range.
15347 (streamer_read_hwi_in_range): Likewise.
15349 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15351 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15352 (predicable_short_it): Likewise.
15353 (predicated): Likewise.
15354 (enabled): Handle above.
15355 (define_cond_exec): Set predicated attribute to yes.
15357 2013-06-05 Mike Stump <mikestump@comcast.net>
15359 * gdbinit.in (__FUNCTION__): Add.
15361 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15363 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15364 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15366 2013-06-05 Jan Hubicka <jh@suse.cz>
15368 * varasm.c (mark_decl_referenced): Revert the removal until targets
15371 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15373 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15374 instead of mark_decl_referenced.
15376 2013-06-05 Jan Hubicka <jh@suse.cz>
15378 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15379 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15380 and symtab_used_from_object_file_p.
15381 (cgraph_make_node_local_1): Clear forced_by_abi.
15382 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15383 * cgraph.h (symtab_node_base): Add forced_by_abi.
15384 (decide_is_variable_needed): Remove.
15385 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15386 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15387 (decide_is_symbol_needed): ... this one; handle symbols in general;
15388 always analyze virtuals; honnor forced_by_abi.
15389 (cgraph_finalize_function): Update.
15390 (varpool_finalize_decl): Update.
15391 (symbol_defined_and_needed): Remove.
15392 (analyze_functions): Update.
15393 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15394 output_refs, input_overwrite_node): Handle forced_by_abi.
15395 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15396 (address_taken_from_non_vtable_p): ... this one.
15397 (comdat_can_be_unshared_p_1): New function.
15398 (cgraph_comdat_can_be_unshared_p): Rename to ...
15399 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15400 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15401 (function_and_variable_visibility): Clear forced_by_abi as needed.
15402 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15403 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15404 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15405 * varpool.c (decide_is_variable_needed): Remove.
15407 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15409 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15410 (arm_option_override): Override arm_restrict_it where appropriate.
15411 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15412 * config/arm/arm.opt (mrestrict-it): New command-line option.
15413 * doc/invoke.texi: Document -mrestrict-it.
15415 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15417 * tsan.c (tsan_atomic_table): Make const.
15419 2013-06-05 Richard Biener <rguenther@suse.de>
15421 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15422 index associated with the tree we are supposed to replace.
15423 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15424 TREE_ASM_WRITTEN as zero for everything but SSA names.
15426 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15428 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15430 2013-06-04 Jan Hubicka <jh@suse.cz>
15432 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15433 (input_node, input_varpool_node): Handle correctly external same
15435 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15436 nodes at ltrans stage.
15438 2013-06-04 Jan Hubicka <jh@suse.cz>
15440 * ipa-inline.c (update_caller_keys): Fix availability test.
15441 (update_callee_keys): Likewise.
15442 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15443 to follow ELF standard.
15445 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15447 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15448 (mips64r5900el-*-elf*): New configurations.
15449 * config/mips/mips-cpus.def (r5900): New processor.
15450 * config/mips/mips-tables.opt: Regenerate.
15451 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15452 (mips_issue_rate): Handle PROCESSOR_R5900.
15453 (mips_reorg_process_insns): Force reorder mode for the R5900.
15454 * config/mips/mips.h (TARGET_MIPS5900): Define.
15455 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15457 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15459 * config/mips/mips.md (processor): Add r5900.
15460 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15462 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15464 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15465 into function to generate MOVI instruction.
15466 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15467 (aarch64_preferred_simd_mode): Turn into wrapper.
15468 (aarch64_output_scalar_simd_mov_immediate): New function.
15469 * config/aarch64/aarch64-protos.h: Add prototype for above.
15471 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15473 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15474 element_char member.
15475 (sizetochar): Return signed char.
15476 (aarch64_simd_valid_immediate): Remove elchar and other
15477 unnecessary variables.
15478 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15479 Calculate element_char as required.
15480 * config/aarch64/aarch64-protos.h: Update and move prototype
15481 for aarch64_output_simd_mov_immediate.
15482 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15485 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15487 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15488 information completed by aarch64_simd_valid_immediate.
15489 (aarch64_legitimate_constant_p): Update arguments.
15490 (aarch64_simd_valid_immediate): Work with struct rather than many
15492 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15493 (aarch64_simd_make_constant): Update arguments.
15494 (aarch64_output_simd_mov_immediate): Work with struct rather than
15495 many pointers. Output immediate directly rather than as operand.
15496 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15498 * config/aarch64/constraints.md (Dn): Update arguments.
15500 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15502 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15504 (aarch64_simd_immediate_valid_for_move): Remove.
15505 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15506 (aarch64_simd_make_constant): Update call.
15507 (aarch64_output_simd_mov_immediate): Update call.
15508 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15510 * config/aarch64/constraints.md (Dn): Update call.
15512 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15514 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15515 return type to bool for prototype.
15516 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15517 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15518 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15520 2013-06-04 Catherine Moore <clm@codesourcery.com>
15522 * config/mips/mips.opt (meva): New.
15523 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15524 (ASM_SPEC): Handle -meva.
15525 * doc/invoke.texi (meva): Document.
15527 2013-06-04 Alan Modra <amodra@gmail.com>
15529 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15532 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15534 * rtl.def: Add extra fourth optional field to define_cond_exec.
15535 * gensupport.c (process_one_cond_exec): Process attributes from
15537 * doc/md.texi: Document fourth field in define_cond_exec.
15539 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15541 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15542 out the processing order as in store_bit_field_1.
15544 2013-06-04 Jan Hubicka <jh@suse.cz>
15546 PR middle-end/57500
15547 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15548 non-VAR_DECL node if it does not exist yet.
15550 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15552 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15553 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15554 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15555 target_cpu_default setting.
15557 2013-06-03 Teresa Johnson <tejohnson@google.com>
15559 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15560 default to -fopt-info=optimized instead of all.
15561 * doc/invoke.texi: Ditto.
15562 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15563 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15564 (execute_vect_slp): Emit BB vectorization success under
15565 MSG_OPTIMIZED_LOCATIONS.
15566 * tree-vect-slp.c (vect_slp_transform_bb): Change
15567 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15568 * tree-vect-loop.c (vect_transform_loop): Ditto.
15570 2013-06-03 Jason Merrill <jason@redhat.com>
15573 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15574 Use TARGET_EXPR for C++.
15576 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15578 PR rtl-optimization/57268
15579 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15580 if DEBUG_INSN_P (insn).
15583 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15585 PR rtl-optimization/57268
15586 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15587 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15589 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15591 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15592 (ix86_avoid_lea_for_addr): Likewise.
15593 (exact_dependency_1): Likewise.
15594 (ix86_adjust_cost): Likewise.
15595 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15597 (do_reorder_for_imul): Fix typo, formatting and
15598 !reload_completed check removed.
15599 (ix86_sched_reorder): Fix typo and formatting.
15600 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15603 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15605 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15607 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15609 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15610 <STRING_CST>: Likewise.
15611 <VECTOR_CST>: Likewise.
15613 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15614 Mikael Morin <mikael@gcc.gnu.org>
15616 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15617 * config.in: Regenerated.
15618 * configure: Regenerated.
15620 2013-06-01 Jan Hubicka <jh@suse.cz>
15622 PR middle-end/57366
15623 * cgraphunit.c (compile): When weakref is not supported,
15624 set up transparent aliases before final output pass.
15625 * varasm.c (assemble_alias): Do not try to do it here.
15627 2013-06-01 Jan Hubicka <jh@suse.cz>
15629 PR middle-end/57467
15630 * passes.c (for_per_function): Skip unanalyzed functions.
15632 2013-06-01 Jan Hubicka <jh@suse.cz>
15634 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15635 (lto_symtab_merge_symbols_1): ... this one.
15636 (lto_symtab_merge_cgraph_nodes): Rename to ...
15637 (lto_symtab_merge_symbols): ... this one; simplify.
15638 * cgraph.c (same_body_aliases_done): Rename to ...
15639 (cpp_implicit_aliases_done): ... this one.
15640 (cgraph_create_function_alias): Update.
15641 (cgraph_same_body_alias): Update.
15642 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15643 (verify_edge_corresponds_to_fndecl): Simplify.
15644 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15645 (cgraph_node): Remove same_body_alias.
15646 (varpool_node): Remove alias_of and extra_name_alias.
15647 (same_body_aliases_done): Rename to ..
15648 (cpp_implicit_aliases_done): ... this one.
15649 (symtab_alias_ultimate_target): Add default parameter.
15650 (symtab_resolve_alias): New function.
15651 (fixup_same_cpp_alias_visibility): Declare.
15652 (cgraph_function_node): Add default parameter.
15653 (cgraph_node_asm_name): Likewise.
15654 (cgraph_function_or_thunk_node): Add default parameter; do
15655 not ICE when it is NULL.
15656 (varpool_variable_node): Likewise.
15657 * tree-emutls.c (create_emultls_var): Update.
15658 (ipa_lower_emutls): Update.
15659 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15660 (cgraph_reset_node): Reset alias info.
15661 (cgraph_finalize_function): Update.
15662 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15663 (analyze_function): Simplify.
15664 (cgraph_process_same_body_aliases): Simplify.
15665 (analyze_functions): Fixup same body aliases.
15666 (handle_alias_pairs): Simplify.
15667 (assemble_thunk): Update.
15668 (assemble_thunks_and_aliases): Update.
15669 (output_weakrefs): Rewrite.
15670 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15671 (lto_output_varpool_node): Likewise.
15672 (compute_ltrans_boundary): Remve assert.
15673 (get_alias_symbol): New functoin.
15674 (input_node): Rewrite alias handling.
15675 (input_varpool_node): Likewise.
15676 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15677 * ipa.c (process_references): Handle weakrefs correctly.
15678 (symtab_remove_unreachable_nodes): Likewise.
15679 * trans-mem.c (get_cg_data): Update.
15680 (ipa_tm_create_version_alias): Update.
15681 (ipa_tm_execute): Update.
15682 * symtab.c (dump_symtab_base): Dump aliases.
15683 (verify_symtab_base): Verify aliases.
15684 (symtab_node_availability): New function.
15685 (symtab_alias_ultimate_target): Simplify.
15686 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15687 handle all the fixup cases.
15688 (symtab_resolve_alias): New function.
15689 * passes.c (ipa_write_summaries): Handle weakrefs.
15690 * varpool.c (varpool_analyze_node): Simplify.
15691 (assemble_aliases): Update.
15692 (varpool_create_variable_alias): Simplify.
15693 (varpool_extra_name_alias): Simplify.
15694 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15695 (lto_symtab_merge_symbols): ... this one.
15697 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15700 PR rtl-optimization/57268
15701 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15702 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15704 2013-06-01 Tobias Burnus <burnus@net-b.de>
15706 Partially reverted:
15707 2013-05-31 Tobias Burnus <burnus@net-b.de>
15709 PR middle-end/57073
15710 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15713 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15715 PR rtl-optimization/57268
15716 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15717 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15719 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15721 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15722 unordered comparison operators when -fno-trapping-math is in effect
15724 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15725 and implement unordered comparison operators properly on the e500.
15727 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15729 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15730 for constant scalar integers.
15731 (simplify_relational_operation_1): Likewise.
15733 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15735 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15736 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15739 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15740 Igor Zamyatin <igor.zamyatin@intel.com>
15742 Silvermont (SLM) architecture performance tuning.
15743 * config/i386/i386.h (enum ix86_tune_indices): Add
15744 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15745 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15747 * config/i386/i386.c (initial_ix86_tune_features)
15748 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15749 (ix86_lea_outperforms): Handle Silvermont tuning.
15750 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15752 (ix86_use_lea_for_mov): Likewise.
15753 (ix86_avoid_lea_for_addr): Likewise.
15754 (ix86_lea_for_add_ok): Likewise.
15755 (exact_dependency_1): New function.
15756 (exact_store_load_dependency): Likewise.
15757 (ix86_adjust_cost): Handle Silvermont tuning.
15758 (do_reoder_for_imul): Likewise.
15759 (swap_top_of_ready_list): New function.
15760 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15762 * config/i386/i386.md (peepholes that split memory operand in fp
15765 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15767 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15768 Remove un-necessary braces.
15770 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15772 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15773 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15775 2013-05-31 Tobias Burnus <burnus@net-b.de>
15777 PR middle-end/57073
15778 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15780 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15783 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15784 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15785 * config/arm/neon.md (iordi3_neon): Remove.
15786 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15787 * config/arm/predicates.md (imm_for_neon_logic_operand):
15788 Move to earlier in the file.
15789 (neon_logic_op2): Likewise.
15790 (arm_iordi_operand_neon): New predicate.
15792 2013-05-31 Richard Biener <rguenther@suse.de>
15794 PR tree-optimization/57478
15795 PR tree-optimization/57453
15796 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15799 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15801 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15802 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15804 2013-05-30 Tobias Burnus <burnus@net-b.de>
15805 Thomas Koenig <tkoenig@gcc.gnu.org>
15807 PR middle-end/57073
15808 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15809 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15811 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15813 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15815 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15817 * target.def (register_usage_leveling_p): New hook.
15818 * targhooks.c (default_register_usage_leveling_p): New.
15819 * targhooks.h (default_register_usage_leveling_p): New prototype.
15820 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15821 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15822 * doc/tm.texi: Update.
15823 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15825 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15827 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15828 (*insv_reg<mode>): New define_insn.
15830 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15832 PR rtl-optimization/57439
15833 * postreload.c (move2add_valid_value_p): Check that we have
15834 a zero subreg_regno_offset when accessing the register in
15835 the requested mode.
15837 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15838 Igor Zamyatin <igor.zamyatin@intel.com>
15840 Silvermont (SLM) architecture pipeline model, tuning and
15842 * config.gcc: Add slm config options and target.
15844 * config/i386/slm.md: New.
15846 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15848 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15850 (ix86_target_macros_internal): Likewise.
15852 * config/i386/i386.c (slm_cost): New cost.
15853 (m_SLM): New macro flag.
15854 (initial_ix86_tune_features): Set m_SLM.
15855 (x86_accumulate_outgoing_args): Likewise.
15856 (x86_arch_always_fancy_math_387): Likewise.
15857 (processor_target_table): Add slm cost.
15858 (cpu_names): Add slm cpu name.
15859 (x86_option_override_internal): Set SLM ISA.
15860 (ix86_issue_rate): New case PROCESSOR_SLM.
15861 (ia32_multipass_dfa_lookahead): Likewise.
15862 (fold_builtin_cpu): Add slm.
15864 * config/i386/i386.h (TARGET_SLM): New target macro.
15865 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15866 (processor_type): Add PROCESSOR_SLM.
15868 * config/i386/i386.md (cpu): Add new value "slm".
15869 (slm.md): Include slm.md.
15871 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15872 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15874 * config/arm/arm-protos.h: Add and update function protos.
15875 * config/arm/arm.c (use_simple_return_p): New added.
15876 (thumb2_expand_return): Check simple_return flag.
15877 * config/arm/arm.md: Add simple_return and conditional simple_return.
15878 * config/arm/iterators.md: Add iterator for return and simple_return.
15880 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15882 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15883 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15884 (arm_emit_vfp_multi_reg_pop): Likewise.
15885 (thumb2_emit_ldrd_pop): Likewise.
15886 (arm_expand_epilogue): Add misc REG_CFA notes.
15887 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15889 2013-05-29 Lawrence Crowl <crowl@google.com>
15891 * config/arm/t-arm: Update for below.
15893 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15894 Change type to hash_table. Update dependent calls and types.
15896 * config/i386/t-cygming: Update for below.
15898 * config/i386/t-interix: Update for below.
15900 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15901 Change type to hash_table. Update dependent calls and types.
15902 (i386_find_on_wrapper_list::wrappers): Likewise.
15904 * config/ia64/t-ia64: Update for below.
15906 * config/ia64/ia64.c (bundle_state_table):
15907 Change type to hash_table. Update dependent calls and types.
15909 * config/mips/mips.c (mips_reorg_process_insns::htab):
15910 Change type to hash_table. Update dependent calls and types.
15912 * config/sol2.c (solaris_comdat_htab):
15913 Change type to hash_table. Update dependent calls and types.
15915 * config/t-sol2: Update for above.
15917 2013-05-29 Teresa Johnson <tejohnson@google.com>
15919 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15920 functions are not yet marked as defined.
15922 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15923 Pat Haugen <pthaugen@us.ibm.com>
15924 Peter Bergner <bergner@vnet.ibm.com>
15926 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15930 (vrotl<mode>3): Likewise.
15931 (vashl<mode>3): Likewise.
15932 (vlshr<mode>3): Likewise.
15933 (vashr<mode>3): Likewise.
15935 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15936 support for power8 V2DI builtins.
15938 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15939 power8 V2DI builtins.
15940 (vupkhsw): Likewise.
15941 (vupklsw): Likewise.
15942 (vaddudm): Likewise.
15943 (vminsd): Likewise.
15944 (vmaxsd): Likewise.
15945 (vminud): Likewise.
15946 (vmaxud): Likewise.
15947 (vpkudum): Likewise.
15948 (vpksdss): Likewise.
15949 (vpkudus): Likewise.
15950 (vpksdus): Likewise.
15955 (vsubudm): Likewise.
15956 (vcmpequd): Likewise.
15957 (vcmpgtsd): Likewise.
15958 (vcmpgtud): Likewise.
15959 (vcmpequd_p): Likewise.
15960 (vcmpgtsd_p): Likewise.
15961 (vcmpgtud_p): Likewise.
15962 (vupkhsw): Likewise.
15963 (vupklsw): Likewise.
15964 (vaddudm): Likewise.
15965 (vmaxsd): Likewise.
15966 (vmaxud): Likewise.
15967 (vminsd): Likewise.
15968 (vminud): Likewise.
15969 (vpksdss): Likewise.
15970 (vpksdus): Likewise.
15971 (vpkudum): Likewise.
15972 (vpkudus): Likewise.
15977 (vsubudm): Likewise.
15979 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15980 support for power8 V2DI instructions.
15982 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15983 power8 V2DI instructions. Combine pack and unpack insns to use an
15984 iterator for each mode. Check whether a particular mode supports
15985 Altivec instructions instead of just checking TARGET_ALTIVEC.
15986 (UNSPEC_VPKUWUM): Likewise.
15987 (UNSPEC_VPKSHSS): Likewise.
15988 (UNSPEC_VPKSWSS): Likewise.
15989 (UNSPEC_VPKUHUS): Likewise.
15990 (UNSPEC_VPKSHUS): Likewise.
15991 (UNSPEC_VPKUWUS): Likewise.
15992 (UNSPEC_VPKSWUS): Likewise.
15993 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15994 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15995 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15996 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15997 (UNSPEC_VUPKHSB): Likewise.
15998 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15999 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16000 (UNSPEC_VUPKHSH): Likewise.
16001 (UNSPEC_VUPKLSB): Likewise.
16002 (UNSPEC_VUPKLSH): Likewise.
16004 (VI_char): Likewise.
16005 (VI_scalar): Likewise.
16006 (VI_unit): Likewise.
16008 (VP_small): Likewise.
16009 (VP_small_lc): Likewise.
16010 (VU_char): Likewise.
16011 (add<mode>3): Likewise.
16012 (altivec_vaddcuw): Likewise.
16013 (altivec_vaddu<VI_char>s): Likewise.
16014 (altivec_vadds<VI_char>s): Likewise.
16015 (sub<mode>3): Likewise.
16016 (altivec_vsubcuw): Likewise.
16017 (altivec_vsubu<VI_char>s): Likewise.
16018 (altivec_vsubs<VI_char>s): Likewise.
16019 (altivec_vavgs<VI_char>): Likewise.
16020 (altivec_vcmpbfp): Likewise.
16021 (altivec_eq<mode>): Likewise.
16022 (altivec_gt<mode>): Likewise.
16023 (altivec_gtu<mode>): Likewise.
16024 (umax<mode>3): Likewise.
16025 (smax<mode>3): Likewise.
16026 (umin<mode>3): Likewise.
16027 (smin<mode>3): Likewise.
16028 (altivec_vpkuhum): Likewise.
16029 (altivec_vpkuwum): Likewise.
16030 (altivec_vpkshss): Likewise.
16031 (altivec_vpkswss): Likewise.
16032 (altivec_vpkuhus): Likewise.
16033 (altivec_vpkshus): Likewise.
16034 (altivec_vpkuwus): Likewise.
16035 (altivec_vpkswus): Likewise.
16036 (altivec_vpks<VI_char>ss): Likewise.
16037 (altivec_vpks<VI_char>us): Likewise.
16038 (altivec_vpku<VI_char>us): Likewise.
16039 (altivec_vpku<VI_char>um): Likewise.
16040 (altivec_vrl<VI_char>): Likewise.
16041 (altivec_vsl<VI_char>): Likewise.
16042 (altivec_vsr<VI_char>): Likewise.
16043 (altivec_vsra<VI_char>): Likewise.
16044 (altivec_vsldoi_<mode>): Likewise.
16045 (altivec_vupkhsb): Likewise.
16046 (altivec_vupkhs<VU_char>): Likewise.
16047 (altivec_vupkls<VU_char>): Likewise.
16048 (altivec_vupkhsh): Likewise.
16049 (altivec_vupklsb): Likewise.
16050 (altivec_vupklsh): Likewise.
16051 (altivec_vcmpequ<VI_char>_p): Likewise.
16052 (altivec_vcmpgts<VI_char>_p): Likewise.
16053 (altivec_vcmpgtu<VI_char>_p): Likewise.
16054 (abs<mode>2): Likewise.
16055 (vec_unpacks_hi_v16qi): Likewise.
16056 (vec_unpacks_hi_v8hi): Likewise.
16057 (vec_unpacks_lo_v16qi): Likewise.
16058 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16059 (vec_unpacks_lo_v8hi): Likewise.
16060 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16061 (vec_pack_trunc_v8h): Likewise.
16062 (vec_pack_trunc_v4si): Likewise.
16063 (vec_pack_trunc_<mode>): Likewise.
16065 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16067 (vec_vmaxsd): Likewise.
16068 (vec_vmaxud): Likewise.
16069 (vec_vminsd): Likewise.
16070 (vec_vminud): Likewise.
16071 (vec_vpksdss): Likewise.
16072 (vec_vpksdus): Likewise.
16073 (vec_vpkudum): Likewise.
16074 (vec_vpkudus): Likewise.
16075 (vec_vrld): Likewise.
16076 (vec_vsld): Likewise.
16077 (vec_vsrad): Likewise.
16078 (vec_vsrd): Likewise.
16079 (vec_vsubudm): Likewise.
16080 (vec_vupkhsw): Likewise.
16081 (vec_vupklsw): Likewise.
16083 2013-05-29 Jan Hubicka <jh@suse.cz>
16085 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16086 flags; reorder rest of fields in more consistent way.
16087 (varpool_node): Remove analyzed, finalized and alias.
16088 (cgraph_ndoe): Likewise.
16089 (symtab_alias_ultimate_target): New function.
16090 (cgraph_function_node): Move offline.
16091 (cgraph_reset_node): Declare.
16092 (cgraph_comdat_can_be_unshared_p): Remove.
16093 (varpool_remove_initializer): Declare.
16094 (varpool_first_defined_variable, varpool_next_defined_variable
16095 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16096 (cgraph_function_with_gimple_body_p): Update.
16097 (varpool_all_refs_explicit_p): Update.
16098 (symtab_alias_target): New function.
16099 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16100 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16101 (cgraph_function_or_thunk_node): Simplify using
16102 symtab_alias_ultimate_target.
16103 (varpool_variable_node): Likewise.
16104 * cgraph.c (cgraph_create_function_alias): Update.
16105 (cgraph_add_thunk): Update.
16106 (cgraph_remove_node): Update.
16107 (dump_cgraph_node): Do not dump removed flags.
16108 (cgraph_function_body_availability): Update.
16109 (cgraph_propagate_frequency): Update.
16110 (verify_cgraph_node): Check sanity of local flag.
16111 (cgraph_function_node): Move here from cgraph.h; revamp for
16112 cgraph_function_or_thunk_node.
16113 * lto-symtab.c (lto_varpool_replace_node): Update.
16114 (lto_symtab_resolve_can_prevail_p): Update.
16115 (lto_symtab_merge_cgraph_nodes): Update.
16116 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16117 propagate_constants_accross_call, devirtualization_time_bonus,
16118 ipcp_propagate_stage): Update.
16119 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16120 * ipa-inline-transform.c (clone_inlined_nodes,
16121 preserve_function_body_p): Update.
16122 * ipa-reference.c (propagate): Update.
16123 (write_node_summary_p): Update.
16124 * toplev.c (wrapup_global_declaration_2): Update.
16125 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16126 (analyze_function) ... this one.
16127 (cgraph_process_new_functions): Update.
16128 (cgraph_reset_node): Export.
16129 (cgraph_finalize_function): Update.
16130 (cgraph_add_new_function): Update.
16131 (process_function_and_variable_attributes): Update.
16132 (varpool_finalize_decl): Update.
16133 (symbol_finalized): Remove.
16134 (symbol_finalized_and_needed): Rename to ...
16135 (symbol_defined_and_needed): ... update.
16136 (cgraph_analyze_functions): Update.
16137 (handle_alias_pairs): Update.
16138 (mark_functions_to_output): Update.
16139 (assemble_thunk): Update.
16140 (output_in_order): Update.
16141 (output_weakrefs): Update.
16142 (finalize_compilation_unit): Update.
16143 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16144 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16145 input_node, input_varpool_node): Update.
16146 * dbxout.c (dbxout_expand_expr): Update.
16147 * cgraphclones.c (cgraph_clone_node): Update.
16148 (cgraph_copy_node_for_versioning): Update.
16149 (cgraph_materialize_clone): Update.
16150 (cgraph_materialize_all_clones): Update.
16151 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16152 propagate_pure_const, propagate_nothrow): Update.
16153 * lto-streamer-out.c (lto_output, write_symbol): Update.
16154 * ipa-utils.c (ipa_reverse_postorder): Update.
16155 * ipa-inline.c (can_inline_edge_p): Update.
16156 (update_caller_keys, ipa_inline): Update.
16157 * dwarf2out.c (reference_to_unused,
16158 premark_types_used_by_global_vars_helper): Update.
16159 * tree-eh.c (tree_could_trap_p): Update.
16160 * ipa-split.c (consider_split, execute_split_functions): Update.
16161 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16162 has_addr_references_p): Update; move ahead in file for better
16164 (process_references): Simplify.
16165 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16166 bodies are removed.
16167 (cgraph_comdat_can_be_unshared_p): Make static.
16168 (cgraph_externally_visible_p): Update.
16169 (varpool_externally_visible_p): Update.
16170 (function_and_variable_visibility): Update.
16171 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16172 ipa_tm_mark_force_output_node): Update.
16173 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16174 estimate_edge_devirt_benefit, inline_generate_summary,
16175 inline_write_summary): Update.
16176 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16177 * ipa-prop.c (ipa_compute_jump_functions): Update.
16178 (ipa_print_node_params, ipa_prop_read_section,
16179 ipa_update_after_lto_read, read_replacements_section): Update.
16180 * varasm.c (mark_decl_referenced): Update.
16181 (assemble_alias, dump_tm_clone_pairs): Update.
16182 * tree-inline.c (copy_bb): Update.
16183 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16185 * symtab.c (dump_symtab_base): Print new flags.
16186 (verify_symtab_base): Verify new flags.
16187 (symtab_alias_ultimate_target): New function.
16188 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16189 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16191 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16193 * i386.c (ix86_get_function_versions_dispatcher,
16194 ix86_generate_version_dispatcher_body): Update.
16195 (fold_builtin_cpu): Use varpool_add_new_variable.
16196 * varpool.c (varpool_remove_initializer): Break out from ...
16197 (varpool_remove_node): ... this one.
16198 (dump_varpool_node, varpool_node_for_asm,
16199 cgraph_variable_initializer_availability, varpool_analyze_node,
16200 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16201 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16204 2013-05-29 Jan Hubicka <jh@suse.cz>
16206 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16208 2013-05-29 Easwaran Raman <eraman@google.com>
16210 PR tree-optimization/57442
16211 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16212 when control exits the main loop.
16214 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16216 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16218 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16219 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16220 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16223 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16225 PR tree-optimization/57441
16226 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16227 Don't limit size of incr_vec to number of candidates.
16229 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16231 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16232 and mips16 directories.
16233 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16234 (MULTILIB_DIRNAMES): Ditto.
16235 (MULTILIB_EXCEPTIONS): Add new exceptions.
16236 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16237 (MULTILIB_DIRNAMES): Ditto.
16238 (MULTILIB_EXCEPTIONS): Add new exceptions.
16240 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16241 Marcus Shawcroft <marcus.shawcroft@arm.com>
16243 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16244 SYMBOL_TINY_ABSOLUTE.
16245 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16246 SYMBOL_TINY_ABSOLUTE.
16247 (aarch64_expand_mov_immediate): Likewise.
16248 (aarch64_classify_symbol): Likewise.
16249 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16250 Permit SYMBOL_TINY_ABSOLUTE.
16251 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16253 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16254 Marcus Shawcroft <marcus.shawcroft@arm.com>
16256 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16257 Refactor if/switch. Replace gcc_assert with if.
16259 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16261 * config/i386/i386.c (initial_ix86_tune_features): Enable
16262 FP Reassociation for AMD bdver1 and bdver2.
16264 2013-05-29 Martin Jambor <mjambor@suse.cz>
16266 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16267 and IMAGPART_EXPR do not occur within other handled_components.
16269 2013-05-29 Richard Biener <rguenther@suse.de>
16271 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16272 access on whether the use is in the BB we currently try to
16274 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16275 vectorize to vect_bb_slp_scalar_cost.
16277 2013-05-29 Richard Biener <rguenther@suse.de>
16279 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16280 computing scalar cost offsetted by stmts that are kept live
16282 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16283 for computation of scalar cost.
16285 2013-05-28 Steve Ellcey <sellcey@mips.com>
16287 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16289 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16291 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16292 array notation built-in reduction functions.
16293 * doc/passes.texi (Passes): Added documentation about changes done
16295 * doc/invoke.texi (C Dialect Options): Added documentation about
16296 the -fcilkplus flag.
16297 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16298 (BUILTINS_DEF): Depend on cilkplus.def.
16299 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16300 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16301 * cilkplus.def: New file.
16303 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16305 PR rtl-optimization/57439
16306 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16308 2013-05-28 Easwaran Raman <eraman@google.com>
16310 PR tree-optimization/57337
16311 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16312 (find_insert_point): Correctly identify the insertion point
16313 when two statements with the same UID is compared.
16315 2013-05-28 Richard Biener <rguenther@suse.de>
16317 PR tree-optimization/56787
16318 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16319 from the list of data references.
16320 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16322 (vect_analyze_loop_operations): Likewise.
16323 (vect_transform_loop): Remove clobbers.
16325 2013-05-28 Martin Jambor <mjambor@suse.cz>
16327 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16328 and REALPART_EXPRs have scalar type.
16330 2013-05-28 Richard Biener <rguenther@suse.de>
16332 PR tree-optimization/57411
16333 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16335 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16336 virtual operand propagation.
16338 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16340 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16341 destination register for bmasksi_vis.
16342 (vector_init_bshuffle): Likewise.
16343 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16345 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16347 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16348 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16349 mode if the instruction isn't available in the original mode.
16350 * config/sparc/sparc.opt (mfix-ut699): New option.
16351 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16352 (divdf3): Turn into expander.
16353 (divdf3_nofix): New insn.
16354 (divdf3_fix): Likewise.
16355 (divsf3): Disable if -mfix-ut699.
16356 (sqrtdf2): Turn into expander.
16357 (sqrtdf2_nofix): New insn.
16358 (sqrtdf2_fix): Likewise.
16359 (sqrtsf2): Disable if -mfix-ut699.
16361 2013-05-27 Richard Biener <rguenther@suse.de>
16363 PR middle-end/57412
16364 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16365 block for the new loop.
16367 2013-05-27 Richard Biener <rguenther@suse.de>
16369 PR tree-optimization/57343
16370 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16371 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16372 (number_of_iterations_cond): Do not build the folded tree.
16374 2013-05-27 Richard Biener <rguenther@suse.de>
16377 PR middle-end/57381
16378 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16379 OEP_CONSTANT_ADDRESS_OF retained.
16381 PR tree-optimization/57417
16382 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16383 for unchanged base.
16384 (set_ssa_val_to): Compare addresses using
16385 get_addr_base_and_unit_offset.
16387 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16389 PR rtl-optimization/56833
16390 * postreload.c (move2add_record_mode): New function.
16391 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16392 (move2add_use_add2_insn): Use move2add_record_sym_value.
16393 (move2add_use_add3_insn): Likewise.
16394 (reload_cse_move2add): Use move2add_valid_value_p and
16395 move2add_record_mode. Invalidate call-clobbered and REG_INC
16396 affected regs by setting reg_mode to VOIDmode.
16397 (move2add_note_store): Don't pretend the inside of a SUBREG is
16398 the actual destination. Invalidate single/leading registers by
16399 setting reg_mode to VOIDmode.
16400 Use move2add_record_sym_value, move2add_valid_value_p and
16401 move2add_record_mode.
16403 2013-05-27 Richard Biener <rguenther@suse.de>
16405 PR tree-optimization/57396
16406 * tree-affine.c (double_int_constant_multiple_p): Properly
16407 return false for val == 0 and div != 0.
16409 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16411 * config/mips/mips.h: Use #elif in preprocessor conditions.
16413 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16416 * config/mips/constraints.md (kl): New constraint.
16417 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16418 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16419 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16420 from LO for MIPS16.
16421 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16423 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16426 * config/mips/mips.c (mips_can_inline_p): New function.
16427 (TARGET_CAN_INLINE_P): Define.
16429 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16431 * sched-int.h (ds_t, dw_t): Make unsigned int.
16432 Fix documentation that describes how all the ds_t bits are used.
16433 Reserve the last bit for delayed-branch scheduling.
16434 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16435 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16436 (gen_dep_weak_1): Remove prototype.
16437 * sched-deps.c (get_dep_weak_1): Make static.
16438 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16439 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16440 * doc/tm.texi: Regenerate.
16441 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16443 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16446 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16448 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16449 Sandra Loosemore <sandra@codesourcery.com>
16451 * config.gcc (powerpc-*): Allow native for with-cpu.
16453 2013-05-24 Jeff Law <law@redhat.com>
16455 PR tree-optimization/57124
16456 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16457 conversion feeding a condition if the range has an overflow
16458 if -fstrict-overflow. Add warnings for when we do make the
16461 2013-05-24 Dehao Chen <dehao@google.com>
16463 * tree-cfg.c (locus_discrim_map): Fix the typo.
16464 (locus_discrim_hasher): Likewise.
16465 (locus_discrim_hasher::hash): Likewise.
16466 (locus_discrim_hasher::equal): Likewise.
16468 2013-05-24 Martin Jambor <mjambor@suse.cz>
16470 PR tree-optimization/57294
16471 * cgraph.h (ipa_record_stmt_references): Declare.
16472 * cgraphbuild.c (ipa_record_stmt_references): New function.
16473 (build_cgraph_edges): Use ipa_record_stmt_references.
16474 (rebuild_cgraph_edges): Likewise.
16475 (cgraph_rebuild_references): Likewise.
16476 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16477 associated with the old statement and build references from the
16478 newly built statements.
16479 * ipa-ref.c (ipa_remove_stmt_references): New function.
16480 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16482 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16484 * lra-constraints.c (emit_spill_move): Use smaller mode for
16486 (check_and_process_move): Consider mem-reg moves for secondary
16488 (curr_insn_transform): Don't lose insns emitted before for
16489 secondary memory moves.
16490 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16491 reg set up in the current insn.
16493 2013-05-24 Dehao Chen <dehao@google.com>
16495 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16497 (locus_descrim_hasher::equal): Likewise.
16498 (build_gimple_cfg): New discriminator assignment algorithm.
16499 (make_edges): Likewise.
16500 (next_discriminator_for_locus): Likewise.
16501 (same_line_p): Likewise.
16502 (assign_discriminators): Likewise.
16503 (make_cond_expr_edges): Likewise.
16504 (make_gimple_switch_edges): Likewise.
16505 (make_goto_expr_edges): Likewise.
16506 (make_gimple_asm_edges): Likewise.
16508 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16510 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16511 X format specifier to only display bottom 16 bits.
16512 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16513 immediate to match for operand 2, since it will be masked.
16515 2013-05-24 Richard Biener <rguenther@suse.de>
16517 PR tree-optimization/57287
16518 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16519 all SSA names that occur in abnormal PHIs.
16521 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16524 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16525 that index is not negative.
16527 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16529 PR rtl-optimization/55177
16530 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16531 (simplify_byte_swapping_operation): New.
16532 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16533 (simplify_relational_operation_1): Deal with BSWAP.
16535 2013-05-23 Richard Henderson <rth@redhat.com>
16538 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16539 (ix86_reorg): Call it.
16541 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16544 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16545 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16546 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16548 2013-05-23 Christian Bruel <christian.bruel@st.com>
16551 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16553 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16554 Marcus Shawcroft <marcus.shawcroft@arm.com>
16556 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16557 * config/aarch64/constraints.md (Usa): Remove.
16558 * doc/md.texi (AArch64 Usa): Remove.
16560 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16561 Marcus Shawcroft <marcus.shawcroft@arm.com>
16563 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16564 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16565 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16566 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16568 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16570 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16571 instruction (AdvSIMD).
16572 * config/aarch64/aarch64-builtins.c
16573 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16574 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16576 2013-05-23 Martin Jambor <mjambor@suse.cz>
16578 PR middle-end/57347
16579 * tree.h (contains_bitfld_component_ref_p): Declare.
16580 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16581 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16583 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16584 not access a bit-field. Assert all final offsets are byte-aligned.
16586 2013-05-23 Richard Biener <rguenther@suse.de>
16588 PR tree-optimization/57380
16589 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16590 least one invariant or re-used load.
16591 * passes.c (init_optimization_passes): Move pass_phiprop before
16594 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16596 * config/aarch64/aarch64-simd.md
16597 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16599 2013-05-23 Richard Biener <rguenther@suse.de>
16601 PR middle-end/57381
16602 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16603 OEP_CONSTANT_ADDRESS_OF retained.
16605 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16607 PR middle-end/57344
16608 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16609 don't lower unit. Handle unit not being always BITS_PER_WORD.
16611 2013-05-23 Richard Biener <rguenther@suse.de>
16613 PR rtl-optimization/57341
16614 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16615 instead of true_dependence.
16617 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16619 * bb-reorder.c (branch_threshold): Make const.
16620 (exec_threshold): Ditto.
16622 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16623 Pat Haugen <pthaugen@us.ibm.com>
16624 Peter Bergner <bergner@vnet.ibm.com>
16626 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16627 documentation for the power8 crypto builtins.
16629 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16631 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16632 macros for defining power8 builtin functions.
16633 (BU_P8V_AV_2): Likewise.
16634 (BU_P8V_AV_P): Likewise.
16635 (BU_P8V_VSX_1): Likewise.
16636 (BU_P8V_OVERLOAD_1): Likewise.
16637 (BU_P8V_OVERLOAD_2): Likewise.
16638 (BU_CRYPTO_1): Likewise.
16639 (BU_CRYPTO_2): Likewise.
16640 (BU_CRYPTO_3): Likewise.
16641 (BU_CRYPTO_OVERLOAD_1): Likewise.
16642 (BU_CRYPTO_OVERLOAD_2): Likewise.
16643 (XSCVSPDP): Fix typo, point to the correct instruction.
16644 (VCIPHER): Add power8 crypto builtins.
16645 (VCIPHERLAST): Likewise.
16646 (VNCIPHER): Likewise.
16647 (VNCIPHERLAST): Likewise.
16648 (VPMSUMB): Likewise.
16649 (VPMSUMH): Likewise.
16650 (VPMSUMW): Likewise.
16651 (VPERMXOR_V2DI): Likewise.
16652 (VPERMXOR_V4SI: Likewise.
16653 (VPERMXOR_V8HI: Likewise.
16654 (VPERMXOR_V16QI: Likewise.
16655 (VSHASIGMAW): Likewise.
16656 (VSHASIGMAD): Likewise.
16657 (VPMSUM): Likewise.
16658 (VPERMXOR): Likewise.
16659 (VSHASIGMA): Likewise.
16661 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16662 __CRYPTO__ if the crypto instructions are available.
16663 (altivec_overloaded_builtins): Add support for overloaded power8
16666 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16667 support for power8 crypto builtins.
16668 (builtin_function_type): Likewise.
16669 (altivec_init_builtins): Add support for builtins that take vector
16670 long long (V2DI) arguments.
16672 * config/rs6000/crypto.md: New file, define power8 crypto
16675 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16676 Pat Haugen <pthaugen@us.ibm.com>
16677 Peter Bergner <bergner@vnet.ibm.com>
16679 * doc/invoke.texi (Option Summary): Add power8 options.
16680 (RS/6000 and PowerPC Options): Likewise.
16682 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16683 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16684 wm, wn, wr documentation.
16686 * config/rs6000/constraints.md (wm): New constraint for VSX
16687 registers if direct move instructions are enabled.
16688 (wn): New constraint for no registers.
16689 (wq): New constraint for quad word even GPR registers.
16690 (wr): New constraint if 64-bit instructions are enabled.
16691 (wv): New constraint if power8 vector instructions are enabled.
16692 (wQ): New constraint for quad word memory locations.
16694 * config/rs6000/predicates.md (const_0_to_15_operand): New
16695 constraint for 0..15 for crypto instructions.
16696 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16697 as GPR and floating point registers.
16698 (int_reg_operand): New predicate to match only GPR registers.
16699 (base_reg_operand): New predicate to match base registers.
16700 (quad_int_reg_operand): New predicate to match even GPR registers
16701 for quad memory operations.
16702 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16703 operations in both GPR and VSX registers.
16704 (quad_memory_operand): New predicate for quad memory operations.
16705 (reg_or_indexed_operand): New predicate for direct move support.
16707 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16708 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16709 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16710 (POWERPC_MASKS): Add power8 options.
16711 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16714 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16715 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16717 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16718 (-mpower8-fusion): New power8 options.
16719 (-mpower8-fusion-sign): Likewise.
16720 (-mpower8-vector): Likewise.
16721 (-mcrypto): Likewise.
16722 (-mdirect-move): Likewise.
16723 (-mquad-memory): Likewise.
16725 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16726 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16728 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16729 (rs6000_debug_vector_unit): Add p8_vector.
16730 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16731 definitions. Also print fusion state.
16732 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16733 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16734 (rs6000_option_override_internal): Add support for power8.
16735 (rs6000_common_init_builtins): Add debugging for skipped builtins
16736 if -mdebug=builtin.
16737 (rs6000_adjust_cost): Add power8 support.
16738 (rs6000_issue_rate): Likewise.
16739 (insn_must_be_first_in_group): Likewise.
16740 (insn_must_be_last_in_group): Likewise.
16741 (force_new_group): Likewise.
16742 (rs6000_register_move_cost): Likewise.
16743 (rs6000_opt_masks): Likewise.
16745 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16746 power8 capable assembler, default to power7 options.
16747 (TARGET_DIRECT_MOVE): Likewise.
16748 (TARGET_CRYPTO): Likewise.
16749 (TARGET_P8_VECTOR): Likewise.
16750 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16751 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16752 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16753 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16754 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16755 (TARGET_XSCVDPSPN): Likewise.
16756 (TARGET_XSCVSPDPN): Likewsie.
16757 (TARGET_SYNC_HI_QI): Likewise.
16758 (TARGET_SYNC_TI): Likewise.
16759 (MASK_CRYPTO): Likewise.
16760 (MASK_DIRECT_MOVE): Likewise.
16761 (MASK_P8_FUSION): Likewise.
16762 (MASK_P8_VECTOR): Likewise.
16763 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16764 temporary used by some of the direct move instructions to get two FP
16765 temporary registers does not force creation of a stack frame.
16766 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16767 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16768 that any VSX registers are tieable, even if they are also an
16769 Altivec vector mode.
16770 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16771 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16772 (RS6000_BTM_CRYPTO): Likewise.
16773 (RS6000_BTM_COMMON): Likewise.
16775 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16776 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16777 (enum rs6000_vector): Add power8 vector support.
16779 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16783 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16784 (any_sibcall_could_use_r3): this and handle indirect calls.
16785 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16787 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16789 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16791 2013-05-22 Richard Biener <rguenther@suse.de>
16793 PR middle-end/57349
16794 * profile.c (branch_prob): Do not split blocks that are
16795 abnormally receiving from ECF_RETURNS_TWICE functions.
16797 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16799 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16800 address mode. Move pointer mode initialization to the same place.
16802 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16804 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16805 while it has any effect.
16807 2013-05-21 Easwaran Raman <eraman@google.com>
16809 PR tree-optimization/57322
16810 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16811 UID of the statement added to the BB to be 1.
16813 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16815 PR tree-optimization/57331
16816 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16817 of conversion from pointer type to integral type with integer.
16819 2013-05-21 Martin Jambor <mjambor@suse.cz>
16822 * ipa-prop.c (ipa_read_node_info): Process param_used and
16823 controlled_uses in the same order as when writing.
16825 2013-05-21 Magnus Granberg <baldrick@free.fr>
16828 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16830 2013-05-21 Richard Biener <rguenther@suse.de>
16832 PR tree-optimization/57318
16833 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16834 estimate stmts with side-effects as likely eliminated.
16836 2013-05-21 Richard Biener <rguenther@suse.de>
16838 PR tree-optimization/57330
16839 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16840 preserve the call stmts fntype.
16842 2013-05-21 Richard Biener <rguenther@suse.de>
16844 PR tree-optimization/57303
16845 * tree-ssa-sink.c (statement_sink_location): Improve killing
16846 stmt detection and properly handle self-assignments.
16848 2013-05-21 Christian Bruel <christian.bruel@st.com>
16850 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16851 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16852 registers. Set register size out of the PARALLEL loop.
16854 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16857 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16858 (*fmasf4, *fmasf4_media): New insns.
16860 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16862 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16863 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16864 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16865 (mips_idiv_insns): Update the comments to say that the returned
16866 instruction counts are in units of BASE_INSN_LENGTH.
16867 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16868 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16869 using 2 rather than 4 as the length of indirect MIPS16 and
16870 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16871 length of a NOP. Don't divide MIPS16 lengths by 2.
16872 (mips16_split_long_branches): Assume a branch is long if the
16873 length is greater than 4 rather than 8.
16874 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16875 rather than multiplying them by 2. Multiply instruction counts
16876 by BASE_INSN_LENGTH rather than 4.
16877 (*jump_mips16, tls_get_tp_mips16_<mode>)
16878 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16880 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16882 * config/mips/mips.md (extended_mips16): Remove branch case.
16883 (length): Remove duplicated extended_mips16 test.
16885 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16887 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16889 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16891 * recog.h (Recog_data): Rename to...
16892 (recog_data_d): ...this.
16893 (recog_data): Update accordingly.
16894 * recog.c (recog_data): Likewise.
16895 * reload.c (save_recog_data): Likewise.
16896 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16897 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16899 2013-05-17 Julian Brown <julian@codesourcery.com>
16901 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16902 found in a REG_EQUAL note, invalidate it.
16904 2013-05-17 Easwaran Raman <eraman@google.com>
16906 * tree-ssa-reassoc.c (find_insert_point): New function.
16907 (insert_stmt_after): Likewise.
16908 (get_def_stmt): Likewise.
16909 (ensure_ops_are_available): Likewise.
16910 (not_dominated_by): Likewise.
16911 (rewrite_expr_tree): Do not move statements beyond what is
16912 necessary. Remove call to swap_ops_for_binary_stmt...
16913 (reassociate_bb): ... and move it here.
16914 (build_and_add_sum): Assign UIDs for new statements.
16915 (linearize_expr): Likewise.
16916 (do_reassoc): Renumber gimple statement UIDs.
16918 2013-05-17 Jan Hubicka <jh@suse.cz>
16920 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16922 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16923 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16925 (output_weakrefs): Update.
16927 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16928 Martin Jambor <mjambor@suse.cz>
16930 PR middle-end/57276
16931 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16932 value that corresponds to the given aggval is found in values vector.
16934 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16936 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16937 sse, sse2, sse3, ssse3 and sse4a flags to options.
16939 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16941 * gengtype-state.c: (s_expr_writer): New class, to handle
16942 prettifying of output layout of s-expressions.
16943 (state_writer): New class, to write out gtype.state.
16944 (state_written_type_count): Move this variable into member data of
16946 (s_expr_writer::s_expr_writer): New code: constructor for new class
16947 (state_writer::state_writer(): ditto
16948 (s_expr_writer::write_new_line): New function
16949 (s_expr_writer::write_any_indent): ditto
16950 (s_expr_writer::begin_s_expr): ditto
16951 (s_expr_writer::end_s_expr): ditto
16952 (write_state_fileloc): convert to method of state_writer...
16953 (state_writer:: write_state_fileloc): ...and use methods of
16954 s_expr_writer to write indentation into the gtype.state output file
16955 to visually represent the hierarchical structure of the list
16957 (write_state_fields): ditto, renaming to...
16958 (state_writer::write_state_fields)
16959 (write_state_a_string): ditto, renaming to...
16960 (state_writer::write_state_a_string)
16961 (write_state_string_option): ditto, renaming to...
16962 (state_writer::write_state_string_option)
16963 (write_state_type_option): ditto, renaming to...
16964 (state_writer::write_state_type_option)
16965 (write_state_nested_option): ditto, renaming to...
16966 (state_writer::write_state_nested_option)
16967 (write_state_option): ditto, renaming to...
16968 (state_writer::write_state_option)
16969 (write_state_options): ditto, renaming to...
16970 (state_writer::write_state_options)
16971 (write_state_lang_bitmap): ditto, renaming to...
16972 (state_writer::write_state_lang_bitmap)
16973 (write_state_version): ditto, renaming to...
16974 (state_writer::write_state_version)
16975 (write_state_scalar_type): ditto, renaming to...
16976 (state_writer::write_state_scalar_type)
16977 (write_state_string_type): ditto, renaming to...
16978 (state_writer::write_state_string_type)
16979 (write_state_undefined_type): ditto, renaming to...
16980 (state_writer::write_state_undefined_type)
16981 (write_state_struct_union_type): ditto, renaming to...
16982 (state_writer::write_state_struct_union_type)
16983 (write_state_struct_type): ditto, renaming to...
16984 (state_writer::write_state_struct_type)
16985 (write_state_user_struct_type): ditto, renaming to...
16986 (state_writer::write_state_user_struct_type)
16987 (write_state_lang_struct_type): ditto, renaming to...
16988 (state_writer::write_state_lang_struct_type)
16989 (write_state_param_struct_type): ditto, renaming to...
16990 (state_writer::write_state_param_struct_type)
16991 (write_state_pointer_type): ditto, renaming to...
16992 (state_writer::write_state_pointer_type)
16993 (write_state_array_type): ditto, renaming to...
16994 (state_writer::write_state_array_type)
16995 (write_state_gc_used): ditto, renaming to...
16996 (state_writer::write_state_gc_used)
16997 (write_state_common_type_content): ditto, renaming to...
16998 (state_writer::write_state_common_type_content)
16999 (write_state_type): ditto, renaming to...
17000 (state_writer::write_state_type)
17001 (write_state_pair_list): ditto, renaming to...
17002 (state_writer::write_state_pair_list)
17003 (write_state_pair): ditto, renaming to...
17004 (state_writer::write_state_pair)
17005 (write_state_typedefs): ditto, renaming to...
17006 (state_writer::write_state_typedefs)
17007 (write_state_structures): ditto, renaming to...
17008 (state_writer::write_state_structures)
17009 (write_state_param_structs): ditto, renaming to...
17010 (state_writer::write_state_param_structs)
17011 (write_state_variables): ditto, renaming to...
17012 (state_writer::write_state_variables)
17013 (write_state_srcdir): ditto, renaming to...
17014 (state_writer::write_state_srcdir)
17015 (write_state_files_list): ditto, renaming to...
17016 (state_writer::write_state_files_list)
17017 (write_state_languages): ditto, renaming to...
17018 (state_writer::write_state_languages)
17019 (write_state): create a state_writer instance and use it when
17020 writing out the state file
17022 2013-05-17 Mike Stump <mikestump@comcast.net>
17024 PR rtl-optimization/57304
17025 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17026 accessing DF_REF_REAL_LOC.
17028 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17030 PR rtl-optimization/57281
17031 PR rtl-optimization/57300
17032 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17033 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17034 what the other splitter did if the registers are dead.
17036 2013-05-17 Richard Biener <rguenther@suse.de>
17038 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17041 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17043 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17046 2013-05-17 Marek Polacek <polacek@redhat.com>
17048 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17049 length when doing non-zero store of storing '\0' to '\0'.
17051 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17053 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17054 vect_external_def oprnd1 with loop_vinfo, try to emit
17055 optional cast, negation and and stmts on the loop preheader
17056 edge instead of into the pattern def seq.
17058 PR tree-optimization/57051
17059 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17060 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17062 2013-05-16 Nick Clifton <nickc@redhat.com>
17064 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17065 (rl78_is_naked_func): New function.
17066 (rl78_expand_prologue): Skip prologue generation for naked functions.
17067 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17068 * doc/extend.texi (naked): Add RL78 to the list of processors
17069 that supports this attribute.
17071 2013-05-16 Jeff Law <law@redhat.com>
17073 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17075 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17077 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17078 cache parameters using detect_caches_amd also for CYRIX,
17079 NSC and TM2 signatures.
17081 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17082 Dzianis Kahanovich <mahatma@eu.by>
17086 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17087 VIA/Centaur processors and determine their cache parameters
17088 using detect_caches_amd.
17090 2013-05-16 Teresa Johnson <tejohnson@google.com>
17092 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17093 (rtl_verify_edges): New function.
17094 (rtl_verify_bb_insns): Ditto.
17095 (rtl_verify_bb_pointers): Ditto.
17096 (rtl_verify_bb_insn_chain): Ditto.
17097 (rtl_verify_fallthru): Ditto.
17098 (rtl_verify_bb_layout): Ditto.
17099 (rtl_verify_flow_info_1): Outline checks into new functions.
17100 (rtl_verify_flow_info): Ditto.
17102 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17104 * cfghooks.c (copy_bbs): Add update_dominance argument.
17105 * cfghooks.h (copy_bbs): Update prototype.
17106 * tree-cfg.c (gimple_duplicate_sese_region):
17107 Add update_dominance argument.
17108 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17109 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17110 gimple_duplicate_sese_region call.
17111 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17112 Update copy_bbs call.
17113 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17114 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17116 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17118 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17119 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17120 vect_recog_rotate_pattern.
17121 (vect_recog_rotate_pattern): New function.
17123 2013-05-16 Jason Merrill <jason@redhat.com>
17125 * Makefile.in (LLINKER): New variable.
17126 (mostlyclean): Remove link mutex.
17127 * configure.ac: Handle --enable-link-mutex.
17128 * lock-and-run.sh: New script.
17130 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17133 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17136 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17138 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17140 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17142 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17143 * config/arm/arm.c (next_consecutive_mem): New function.
17144 (gen_movmem_ldrd_strd): Likewise.
17145 * config/arm/arm.md (movmemqi): Update condition and code.
17146 (unaligned_loaddi, unaligned_storedi): New patterns.
17148 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17150 * config.gcc: Obsolete *-*-solaris2.9*.
17151 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17153 2013-05-16 Richard Biener <rguenther@suse.de>
17155 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17156 earlier, after GRAPHITE transforms and IV canonicalization.
17158 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17160 * omp-low.c (extract_omp_for_data): For collapsed loops,
17161 if at least one of the loops is known at compile time to
17162 iterate zero times, set count to 0.
17163 (expand_omp_regimplify_p): New function.
17164 (expand_omp_for_generic): For collapsed loops, if at least
17165 one of the loops isn't known to iterate at least once,
17166 add runtime check with setting count to 0.
17167 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17168 For unsigned types if it isn't known at compile time that
17169 the loop will iterate at least once, add runtime check to bypass
17170 the whole loop if initial condition isn't true.
17172 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17174 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17176 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17178 PR middle-end/57286
17179 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17180 transformations to avoid an infinite loop.
17182 2013-05-16 Marek Polacek <polacek@redhat.com>
17184 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17186 2013-05-15 Leif Ekblad <leif@rdos.net>
17188 * config/i386/i386.c (ix86_decompose_address): Use
17189 DEFAULT_TLS_SEG_REG to access TLS segment register.
17190 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17191 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17192 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17194 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17197 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17198 sibling calls to functions that would normally be lazily bound,
17199 unless $gp is call-clobbered.
17201 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17203 * config/i386/i386.c (ix86_option_override_internal): Update
17204 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17205 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17206 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17207 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17209 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17211 2013-05-15 Andreas Schwab <schwab@suse.de>
17213 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17214 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17217 2013-05-15 Teresa Johnson <tejohnson@google.com>
17219 * loop-unroll.c (report_unroll_peel): Check decision before
17220 emitting unroll/peel message.
17222 2013-05-15 Teresa Johnson <tejohnson@google.com>
17224 * function.h (has_bb_partition): New rtl_data flag.
17225 (bb_reorder_complete): Ditto.
17226 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17227 instead of flag_reorder_blocks_and_partition.
17228 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17229 with some enhancements.
17230 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17231 * bb-reorder.c (connect_traces): Check for has_bb_partition
17232 instead of flag_reorder_blocks_and_partition.
17233 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17234 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17235 verify_hot_cold_block_grouping.
17236 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17238 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17241 * config/arm/predicates.md (call_insn_operand): New predicate.
17242 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17243 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17244 if insn is not a tail call.
17245 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17247 * config/arm/arm.h (enum reg_class): New caller save register class.
17248 (REG_CLASS_NAMES): Likewise.
17249 (REG_CLASS_CONTENTS): Likewise.
17250 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17253 2013-05-15 Richard Biener <rguenther@suse.de>
17255 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17256 of MSG_OPTIMIZED_LOCATIONS.
17257 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17258 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17260 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17261 of MSG_OPTIMIZED_LOCATIONS.
17262 (execute_vect_slp): Likewise.
17263 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17264 (vect_create_cond_for_alias_checks): Likewise.
17265 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17266 (vect_recog_widen_mult_pattern): Likewise.
17267 (vect_recog_widen_sum_pattern): Likewise.
17268 (vect_recog_over_widening_pattern): Likewise.
17269 (vect_recog_widen_shift_pattern): Likewise.
17270 (vect_recog_vector_vector_shift_pattern): Likewise.
17271 (vect_recog_divmod_pattern): Likewise.
17272 (vect_recog_mixed_size_cond_pattern): Likewise.
17273 (vect_recog_bool_pattern): Likewise.
17274 (vect_pattern_recog_1): Likewise.
17276 2013-05-15 Martin Jambor <mjambor@suse.cz>
17278 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17279 non-functions to builtin_unreachable.
17280 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17281 correct when new direct edges were discovered.
17283 2013-05-15 Martin Jambor <mjambor@suse.cz>
17285 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17286 header, print symbol order instead of node uid, print more information
17287 about indirect edge targets.
17288 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17290 (ipa_make_edge_direct_to_target): Likewise.
17291 (remove_described_reference): Likewise.
17292 (propagate_controlled_uses): Likewise.
17293 (ipa_print_node_params): Also print symbol order.
17294 (ipcp_transform_function): Print symbol order instead of node uids.
17295 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17296 (cgraph_get_create_real_symbol_node): Likewise.
17297 * ipa-cp.c (print_lattice): Likewise.
17298 (print_all_lattices): Likewise.
17299 (determine_versionability): Likewise.
17300 (initialize_node_lattices): Likewise.
17301 (estimate_local_effects): Likewise.
17302 (update_profiling_info): Likewise.
17303 (create_specialized_node): Likewise.
17304 (perhaps_add_new_callers): Likewise.
17305 (decide_about_value): Likewise.
17306 (decide_whether_version_node): Likewise.
17307 (identify_dead_nodes): Likewise.
17308 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17309 (dump_inline_summary): Likewise.
17310 (estimate_node_size_and_time): Likewise.
17311 (inline_analyze_function): Likewise.
17312 * ipa-inline.c (report_inline_failed_reason): Likewise.
17313 (want_early_inline_function_p): Likewise.
17314 (edge_badness): Likewise.
17315 (update_edge_key): Likewise.
17316 (inline_small_functions): Likewise. Add dumping of order to two other
17318 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17319 instead of node uids.
17320 (propagate_pure_const): Likewise.
17321 (propagate_pure_const): Likewise.
17322 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17323 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17324 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17326 * tree-pretty-print.c (dump_function_header): Likewise.
17327 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17328 Print symbol order instead of node uids.
17330 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17332 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17333 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17335 2013-05-15 Richard Biener <rguenther@suse.de>
17337 PR tree-optimization/57275
17338 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17339 return value for fail to do runtime alias checks for gather loads.
17341 2013-05-15 Jan Hubicka <jh@suse.cz>
17345 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17346 weakrefs are not external.
17347 (lto_symtab_merge_decls): Fix thinko when dealing with
17348 non-lto_symtab decls.
17349 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17350 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17351 * varpool.c (dump_varpool_node): Dump more flags.
17353 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17355 * config/i386/i386.c (processor_alias_table): Add instruction
17356 FSGSBASE for AMD bdver3 architecture.
17358 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17360 * tree.c (warn_deprecated_use): Print file:line using locus color.
17361 * diagnostic.c (diagnostic_report_current_module): Print file:line
17362 and file:line:column using locus color.
17364 2013-05-14 Mike Stump <mikestump@comcast.net>
17366 * gdbinit.in: Add __null.
17368 2013-05-14 Mike Stump <mikestump@comcast.net>
17370 * recog.h: Rename struct recog_data to Recog_data.
17371 * recog.c: Likewise.
17372 * reload.c (can_reload_into): Likewise.
17373 * config/picochip/picochip.c: Likewise.
17375 2013-05-14 Mike Stump <mikestump@comcast.net>
17377 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17379 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17381 * resource.h (struct resources): Remove unch_memory member.
17382 (CLEAR_RESOURCE): Don't clear unch_memory.
17383 * resource.c (mark_referenced_resources): Don't set it.
17384 (mark_set_resources): Likewise.
17385 (mark_target_live_regs): Don't clear it.
17386 (init_resource_info): Likewise.
17387 * reorg.c (resource_conflicts_p): Don't compare it.
17388 (redundant_insn): Don't set it.
17390 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17392 * emit-rtl.c (next_label): Remove unused function.
17393 (skip_consecutive_labels, link_cc0_insns): Move to ...
17394 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17395 only place where these functions are used, and make them static.
17397 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17399 * fold-const.c (fold_negate_expr): Handle vectors.
17400 (fold_truth_not_expr): Make it static.
17401 (fold_invert_truthvalue): New static function.
17402 (invert_truthvalue_loc): Handle vectors. Do not call
17403 fold_truth_not_expr directly.
17404 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17405 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17406 (fold_comparison): Handle vector constants.
17407 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17408 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17409 * tree.h (fold_truth_not_expr): Remove declaration.
17411 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17413 * config/aarch64/aarch64-simd.md
17414 (aarch64_vcond_internal<mode>): Rename to...
17415 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17416 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17417 float modes. Clarify all iterator modes.
17418 (vcond<mode><mode>): Use new name for vcond expanders.
17419 (vcond<v_cmp_result><mode>): Likewise.
17420 (vcondu<mode><mode>: Likewise.
17421 * config/aarch64/iterators.md (VDQF_COND): New.
17423 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17426 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17427 variable for the shift amount. Check that we shift by non-negative
17430 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17433 * config/arm/arm.h (EPILOGUE_USES): Only return true
17434 for LR_REGNUM after epilogue_completed.
17436 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17438 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17439 is error_mark_node.
17441 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17444 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17445 and Solaris 11+/x86 with gld.
17446 * configure: Regenerate.
17448 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17450 * expmed.c (expand_shift_1): Canonicalize rotates by
17451 constant bitsize / 2 to bitsize - 1.
17452 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17453 case ROTATERT>: Likewise.
17456 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17458 * config/i386/i386.md (rotateinv): New code attr.
17459 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17460 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17461 roll $31, %eax, etc.
17463 2013-05-14 Richard Biener <rguenther@suse.de>
17465 PR middle-end/57235
17466 * tree-eh.c (sink_clobbers): Give up for successors with
17467 multiple predecessors and no virtual uses.
17469 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17471 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17472 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17474 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17476 PR middle-end/57251
17477 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17478 the case when both op0 and op1 have VOIDmode.
17480 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17482 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17483 in multiply-accumulate mode.
17485 2013-05-13 Guozhi Wei <carrot@google.com>
17487 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17489 2013-05-13 Kai Tietz <ktietz@redhat.com>
17492 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17493 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17494 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17495 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17496 (ix86_expand_prologue): Likewise.
17497 (ix86_expand_split_stack_prologue): Likewise.
17498 (legitimate_pic_address_disp_p): Likewise.
17499 (legitimize_pic_address): Likewise.
17500 (legitimize_tls_address): Likewise.
17501 (legitimize_pe_coff_symbol): Likewise.
17502 (output_pic_addr_const): Likewise.
17503 (construct_plt_address): Likewise.
17504 (ix86_expand_call): Likewise.
17505 (x86_output_mi_thunk): Likewise.
17506 (x86_function_profiler): Likewise.
17508 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17510 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17511 similar switch cases.
17512 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17513 (aarch64_simd_mov_to_<mode>low): Delete.
17514 (aarch64_simd_mov_to_<mode>high): Delete.
17515 (move_lo_quad_<mode>): Add w<-r alternative.
17516 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17517 (aarch64_simd_mov_from_*): Update type attribute.
17518 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17521 2013-05-13 Jan Hubicka <jh@suse.cz>
17523 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17524 * config/i386/i386.c (ix86_compute_frame_layout,
17525 ix86_expand_epilogue, emit_i387_cw_initialization,
17526 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17527 ix86_local_alignment): Fix use of size/speed predicates.
17529 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17531 PR tree-optimization/45216
17532 PR tree-optimization/57157
17533 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17534 the (-Y) & (B - 1) variant if OP is |.
17535 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17536 return shifted. Use (-op1) & (prec - 1) as other_amount
17537 instead of prec - op1.
17539 2013-05-13 Martin Jambor <mjambor@suse.cz>
17541 PR middle-end/42371
17542 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17543 (ipa_constant_data): New type.
17544 (ipa_jump_func): Use ipa_constant_data to hold information about
17545 constant jump functions.
17546 (ipa_get_jf_constant): Adjust to jump function type changes.
17547 (ipa_get_jf_constant_rdesc): New function.
17548 (ipa_param_descriptor): New field controlled_uses.
17549 (ipa_get_controlled_uses): New function.
17550 (ipa_set_controlled_uses): Likewise.
17551 * ipa-ref.h (ipa_find_reference): Declare.
17552 * ipa-prop.c (ipa_cst_ref_desc): New type.
17553 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17555 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17556 New parameter cs. Adjust all callers.
17557 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17558 (remove_described_reference): New function.
17559 (jfunc_rdesc_usable): Likewise.
17560 (try_make_edge_direct_simple_call): Decrement controlled use count,
17561 attempt to remove reference if it hits zero.
17562 (combine_controlled_uses_counters): New function.
17563 (propagate_controlled_uses): Likewise.
17564 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17565 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17566 (ipa_print_node_params): Print described use counter.
17567 (ipa_write_jump_function): Adjust to jump function type changes.
17568 (ipa_read_jump_function): New parameter CS, pass it to
17569 ipa_set_jf_constant. Adjust caller.
17570 (ipa_write_node_info): Stream controlled use count
17571 (ipa_read_node_info): Likewise.
17572 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17574 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17575 count. Remove cloning-added reference if it reaches zero.
17576 * ipa-ref.c (ipa_find_reference): New function.
17578 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17580 * config/i386/i386.c (processor_target_table): Modified default
17581 alignment values for AMD BD and BT architectures.
17583 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17585 * tree-vect-generic.c (uniform_vector_p): Move ...
17586 * tree.c (uniform_vector_p): ... here.
17587 * tree.h (uniform_vector_p): Declare it.
17588 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17591 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17593 PR tree-optimization/57230
17594 * tree-ssa-strlen.c (handle_char_store): Record length for
17595 array store from STRING_CST.
17597 PR tree-optimization/57230
17598 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17601 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17603 * config/epiphany/epiphany.c (epiphany_init): Check size of
17604 NUM_MODES_FOR_MODE_SWITCHING.
17605 (epiphany_expand_prologue):
17606 Remove CONFIG_REGNUM initial value handling code.
17607 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17608 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17609 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17610 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17611 Don't return 1 for FP_MODE_NONE.
17612 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17613 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17614 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17615 * config/epiphany/epiphany.md (save_config): New pattern.
17617 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17619 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17621 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17623 * config/i386/i386.md (memory): Handle sseishft1.
17624 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17625 (*vec_extractv2di_1): Ditto.
17627 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17629 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17632 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17634 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17635 Add mthumb/march=armv7-a multilib.
17636 Add mthumb/march=armv7-r multilib.
17637 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17639 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17641 * config/v850/t-rtems: Add more multilibs.
17643 2013-05-10 Richard Biener <rguenther@suse.de>
17645 PR tree-optimization/57214
17646 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17647 not propagate from SSA names that occur in abnormal PHI nodes.
17649 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17651 * stor-layout.c (element_precision): New function.
17652 * machmode.h (element_precision): Declare it.
17653 * tree.c (build_minus_one_cst): New function.
17654 (element_precision): Likewise.
17655 * tree.h (build_minus_one_cst): Declare new function.
17656 (element_precision): Likewise.
17657 * fold-const.c (operand_equal_p): Use element_precision.
17658 (fold_binary_loc): Handle vector types.
17659 * convert.c (convert_to_integer): Use element_precision.
17660 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17663 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17665 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17666 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17667 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17668 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17669 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17670 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17671 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17672 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17673 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17674 (Uuw8): New constraints.
17675 (Usb4): Move into alphabetical order.
17676 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17677 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17678 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17679 previously unnamed patterns.
17680 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17681 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17682 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17683 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17684 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17685 of specific lengths.
17687 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17689 * config/i386/i386.md (rotateinv): New code attr.
17690 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17691 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17692 roll $31, %eax, etc.
17694 PR tree-optimization/45216
17695 PR tree-optimization/57157
17696 * tree-ssa-forwprop.c (simplify_rotate): New function.
17697 (ssa_forward_propagate_and_combine): Call it.
17699 2013-05-10 Richard Biener <rguenther@suse.de>
17701 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17702 disable peeling when we version for aliasing.
17703 (vector_alignment_reachable_p): Honor explicit user alignment.
17704 (vect_supportable_dr_alignment): Likewise.
17705 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17706 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17707 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17708 then peeling to arrange for the cost-model check to come first.
17710 2013-05-10 Alan Modra <amodra@gmail.com>
17712 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17713 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17714 * configure: Regenerate.
17716 2013-05-10 Alan Modra <amodra@gmail.com>
17719 * varasm.c (default_elf_select_section): Move !DECL_P check..
17720 (get_named_section): ..to here before calling get_section_name.
17722 (default_section_type_flags): Add DECL_P check.
17723 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17724 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17726 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17728 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17729 When using gen_stack_adjust_str with a register offset, add a
17730 REG_FRAME_RELATED_EXPR note.
17732 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17734 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17735 (*vec_extractv4si_zext_mem): Ditto.
17736 (*vec_extractv2di): Add 0->x and x->x alternatives.
17737 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17738 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17740 2013-05-09 Jason Merrill <jason@redhat.com>
17742 N3639 C++1y VLA support
17743 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17746 * tree.c (build_constructor_va): New.
17747 * tree.h: Declare it.
17749 2013-05-09 Martin Jambor <mjambor@suse.cz>
17752 * gimple-fold.c (canonicalize_constructor_val): Call
17753 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17755 2013-05-09 Jan Hubicka <jh@suse.cz>
17756 Richard Biener <rguenther@suse.de>
17759 * symtab.c (symtab_make_decl_local): Do not add private names.
17761 2013-05-09 Jan Hubicka <jh@suse.cz>
17764 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17765 (unlink_from_assembler_name_hash): Likewise.
17766 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17767 symtab_unregister_node, symtab_initialize_asm_name_hash,
17768 change_decl_assembler_name): Update.
17770 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17772 * config/aarch64/aarch64.md: New movtf split.
17773 (*movtf_aarch64): Update.
17774 (aarch64_movdi_tilow): Handle TF modes and rename to
17775 aarch64_movdi_<mode>low.
17776 (aarch64_movdi_tihigh): Handle TF modes and rename to
17777 aarch64_movdi_<mode>high
17778 (aarch64_movtihigh_di): Handle TF modes and rename to
17779 aarch64_mov<mode>high_di
17780 (aarch64_movtilow_di): Handle TF modes and rename to
17781 aarch64_mov<mode>low_di
17782 (aarch64_movtilow_tilow): Remove spurious whitespace.
17783 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17785 (aarch64_print_operand): Update.
17787 2013-05-09 Alan Modra <amodra@gmail.com>
17789 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17791 * configure: Regenerate.
17793 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17795 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17796 splitter preparation statements.
17797 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17798 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17801 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17803 * gimple-ssa-strength-reduction.c (count_candidates): Change
17804 return value to int.
17805 (analyze_candidates_and_replace): Change type of length to int.
17807 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17809 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17810 (*vec_extract<mode>): Use VI12_128 mode iterator.
17811 (*vec_extract<mode>_mem): Ditto.
17812 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17815 2013-05-08 Diego Novillo <dnovillo@google.com>
17820 2012-08-17 Diego Novillo <dnovillo@google.com>
17823 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17824 * config.in: Regenerate.
17825 * configure: Regenerate.
17826 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17828 2013-05-08 Jan Hubicka <jh@suse.cz>
17831 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17832 * cgraph.h (symtab_node_base): Add unique_name.
17833 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17834 input_overwrite_node, input_varpool_node): Stream unique_name.
17835 * cgraphclones.c (cgraph_create_virtual_clone,
17836 cgraph_function_versioning): Set unique_name.
17837 * ipa.c (function_and_variable_visibility): Set unique_name.
17839 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17841 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17842 (alloc_cand_and_find_basis): Restrict conditional candidate
17843 processing to CAND_MULTs.
17845 2013-05-08 Jan Hubicka <jh@suse.cz>
17848 lto-symtab.c (lto_symtab_symbol_p): New function.
17849 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17850 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17851 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17852 Skip static symbols.
17854 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17856 PR tree-optimization/57200
17857 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17858 Only call inform if the preceding warning_at returns true.
17860 2013-05-07 Han Shen <shenhan@google.com>
17862 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17863 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17864 * common.opt (fstack-protector-strong): New option.
17865 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17866 * doc/invoke.texi (Optimization Options): Document
17867 "-fstack-protector-strong".
17868 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17870 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17872 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17874 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17876 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17877 it is not lhs of the insn.
17878 (setup_reg_equiv): Remove insn having equiv memory which it is not
17880 * lra-constraints.c (process_address): Try to improve generation
17881 code for address base + disp.
17882 (lra_constraints): Make correct the code for checking insn setting
17883 up backward equivalence. Remove insn only if it is in the init
17885 * lra-eliminations.c (update_reg_eliminate): Change return value.
17886 (lra_eliminate): Use the result.
17888 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17890 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17891 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17892 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17893 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17894 register target operands.
17895 (*vec_extractv8hi_sse2): New pattern.
17896 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17897 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17898 (*vec_extract<mode>_mem): New insn and split pattern.
17900 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17902 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17903 (TARGET_ASAN_SHADOW_OFFSET): Define.
17904 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17905 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17907 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17909 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17910 (incr_vec_index): Return -1 if increment not found.
17911 (create_add_on_incoming_edge): Assert if increment not found.
17912 (record_increment): Limit number of increments recorded.
17913 (all_phi_incrs_profitable): Return false if an increment not found.
17914 (replace_profitable_candidates): Don't process increments that were
17916 (analyze_candidates_and_replace): Limit size of incr_vec.
17918 2013-05-07 Richard Biener <rguenther@suse.de>
17920 * calls.c (special_function_p): setjmp-like functions are leaf.
17921 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17922 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17924 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17926 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17927 (aarch64_simd_mov<mode>): New expander.
17928 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17929 (aarch64_simd_mov_to_<mode>high): Likewise.
17930 (aarch64_simd_mov_from_<mode>low): Likewise.
17931 (aarch64_simd_mov_from_<mode>high): Likewise.
17932 (aarch64_dup_lane<mode>): Update.
17933 (aarch64_dup_lanedi): New instruction pattern.
17934 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17935 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17937 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17939 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17940 (replace_mult_candidate): Remove unnecessary argument; remove
17941 unnecessary parameter from call to introduce_cast_before_cand.
17942 (replace_unconditional_candidate): Remove unnecessary parameter
17943 from call to replace_mult_candidate.
17944 (replace_conditional_candidate): Likewise.
17945 (insert_initializers): Use make_temp_ssa_name.
17946 (introduce_cast_before_cand): Remove unnecessary argument; use
17947 make_temp_ssa_name.
17948 (replace_one_candidate): Remove unnecessary argument; remove
17949 unnecessary parameter from calls to introduce_cast_before_cand.
17950 (replace_profitable_candidates): Remove unnecessary parameters
17951 from calls to replace_one_candidate.
17953 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17955 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17956 phi def as possibly hiding a basis for a CAND_ADD whose operands
17957 have been commuted in the analysis.
17958 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17960 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17962 * config/aarch64/aarch64.md
17963 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17964 shift value between 0-4.
17966 2013-05-07 Richard Biener <rguenther@suse.de>
17968 * double-int.h (rshift): New overload.
17969 * double-int.c (rshift): New function.
17970 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17971 (create_reference_ops_from_ref): Remove.
17972 (vn_reference_insert): Use shared ops for constructing the
17973 reference and copy it.
17975 2013-05-07 Richard Biener <rguenther@suse.de>
17977 PR middle-end/57190
17978 * tree-eh.c (sink_clobbers): Properly propagate
17979 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17981 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17983 PR tree-optimization/57149
17984 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17985 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17986 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17987 uninit_undefined_value_p instead of ssa_undefined_value_p.
17990 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17991 for modifier == EXPAND_INITIALIZER.
17993 2013-05-07 Anton Blanchard <anton@samba.org>
17995 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17996 for powerpc64 little endian.
17997 * configure: Regenerate.
17999 2013-05-06 Graham Stott <grahams@btinternet.com>
18001 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18002 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18003 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18004 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18006 2013-05-06 Graham Stott <grahams@btinternet.com>
18008 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18009 codes which allow non-lvalues.
18011 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18013 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18014 components are all 1s.
18015 (integer_minus_onep): New function.
18016 * tree.h (integer_minus_onep): Declare it.
18017 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18018 integer_minus_onep instead of integer_all_onesp.
18020 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18023 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18024 variations of these patterns.
18026 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18028 * config/i386/i386.md (isa): Add x64_sse4 member.
18029 (enabled): Handle x64_sse4.
18030 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18031 instruction for 64bit SSE4_1 targets. Update insn attributes.
18032 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18033 instruction for SSE4_1 targets. Update insn attributes.
18034 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18035 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18036 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18038 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18039 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18040 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18042 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18045 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18047 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18049 * final.c (do_assembler_dialects): Don't handle curly braces and
18050 vertical bar escaped by % as dialect delimiters.
18051 (output_asm_insn): Print curly braces and vertical bar if escaped
18052 by % and ASSEMBLER_DIALECT defined.
18053 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18054 * doc/tm.texi: Regenerated.
18056 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18058 * config/mips/mips.c: Include tree-pass.h.
18059 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18060 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18061 (pass_mips_machine_reorg2): New machine specific pass.
18062 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18063 (mips_option_override): Register the new pass.
18064 * rtl.h (cleanup_barriers): Remove prototype.
18065 (dbr_schedule): Likewise.
18066 * jump.c (cleanup_barriers): Make static.
18067 * reorg.c (dbr_schedule): Likewise.
18069 2013-05-06 Richard Biener <rguenther@suse.de>
18071 PR tree-optimization/57185
18072 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18073 SSA names properly.
18075 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18078 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18079 between "first_arg" and "insn", not "last" and "insn".
18081 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18083 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18084 (find_candidates_in_block): Re-enable slsr_process_phi.
18085 (create_phi_basis): Fix double counting of candidate adjustment.
18087 2013-05-06 Richard Biener <rguenther@suse.de>
18089 PR middle-end/57147
18090 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18091 the edge is also fallthru, preserve it and just clear the
18093 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18094 also complex, preserve that and just clear the fallthru flag.
18095 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18096 update virtual operands.
18098 2013-05-06 Alan Modra <amodra@gmail.com>
18100 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18101 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18102 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18103 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18104 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18105 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18107 2013-05-06 Alan Modra <amodra@gmail.com>
18109 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18110 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18111 (DEFAULT_ASM_ENDIAN): Define.
18112 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18113 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18114 Update -K PIC clause from sysv4.h.
18115 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18116 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18118 2013-05-06 Alan Modra <amodra@gmail.com>
18120 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18121 twice for little-endian.
18122 (ashrdi3_no_power, ashrdi3): Support little-endian.
18124 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18127 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18128 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18130 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18131 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18132 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18135 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18136 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18138 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18139 * config/pa/pa.opt: Make mbig-switch a no-op.
18140 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18141 (CASE_VECTOR_MODE): Always return SImode.
18142 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18143 for the !TARGET_BIG_SWITCH case.
18144 * config/pa/pa-linux.h: Likewise.
18145 * config/pa/pa-openbsd.h: Likewise.
18146 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18147 * config/pa/pa.md (short_jump): Remove define_insn.
18148 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18149 (casesi0): Remove define_insn.
18150 (type): Remove btable_branch.
18151 (pa_combine_type): Likewise.
18152 (in_nullified_branch_delay): Likewise.
18153 (in_call_delay): Likewise.
18154 (define_delay): Likewise.
18155 (define_insn_reservation "Z3"): Likewise.
18156 (define_insn_reservation "Z4"): Likewise.
18157 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18158 (pa_adjust_insn_length): Remove adjustment for btable branches.
18159 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18162 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18164 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18165 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18166 Add m->r,x alternatives.
18167 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18168 splitters using SWI48x mode iterator.
18169 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18170 TARGET_64BIT. Add m->x alternative.
18171 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18172 Add o->x alternative. Enable for TARGET_SSE.
18173 (sse_storeq): Remove expander.
18174 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18175 with memory input operand.
18176 (*vec_extractv2di_1 splitter): New.
18177 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18178 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18180 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18182 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18183 (INT_LOWPART): Delete.
18184 (extract_MB): Adjust.
18185 (extract_ME): Adjust.
18186 (print_operand): Adjust.
18188 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18190 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18191 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18192 (reg_or_logical_cint_operand, easy_fp_constant,
18193 logical_const_operand): Delete "CONST_DOUBLE" case.
18194 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18195 "HOST_BITS_PER_WIDE_INT == 64" test.
18196 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18197 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18198 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18199 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18201 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18202 CONST_DOUBLE DImode/VOIDmode case.
18203 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18204 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18205 CONST_DOUBLE VOIDmode case.
18206 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18207 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18208 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18209 Delete CONST_DOUBLE case.
18210 (splitters for mov FMOVE64 const_double): Delete
18211 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18212 "HOST_BITS_PER_WIDE_INT >= 64" test.
18213 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18215 (mov DI const_double): Delete.
18217 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18219 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18220 on op shows all bits zero in mode of a lowpart subreg, return zero.
18222 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18225 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18226 to save TFmode registers and DImode to save TImode registers for
18227 caller save operations.
18228 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18229 mark being partially clobbered since they only use the first
18232 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18233 and TDmode only use the upper 64-bits of each VSX register.
18235 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18237 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18238 (find_candidates_in_block): Disable slsr_process_phi.
18240 2013-05-03 Guozhi Wei <carrot@google.com>
18242 * coverage.c (coverage_obj_init): Move the construction of gcov
18244 (build_init_ctor): ... here.
18246 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18248 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18249 (slsr_cand_d): Redefine def_phi.
18250 (stride_status, phi_adjust_status, count_phis_status): New enums.
18251 (find_phi_def): New.
18252 (find_basis_for_base_expr): New.
18253 (find_basis_for_candidate): Handle hidden bases.
18254 (alloc_cand_and_find_basis): Handle phi candidates.
18255 (slsr_process_phi): New.
18256 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18257 (create_mul_imm_cand): Likewise.
18258 (create_add_ssa_cand): Exclude phi base candidates.
18259 (create_add_imm_cand): Likewise.
18260 (slsr_process_cast): Likewise.
18261 (slsr_process_copy): Likewise.
18262 (find_candidates_in_block): Handle phi candidates.
18263 (dump_candidate): Likewise.
18264 (unconditional_cands): Delete.
18265 (unconditional_cands_with_known_stride_p): Delete.
18266 (phi_dependent_cand_p): New.
18267 (cand_increment): Handle phi-dependent candidates.
18268 (replace_dependent): Delete.
18269 (replace_mult_candidate): New.
18270 (replace_unconditional_candidate): New.
18271 (incr_vec_index): Move to avoid forward reference.
18272 (create_add_on_incoming_edge): New.
18273 (create_phi_basis): New.
18274 (replace_dependents): Delete.
18275 (replace_conditional_candidate): New.
18276 (phi_add_costs): New.
18277 (replace_uncond_cands_and_profitable_phis): New.
18278 (record_increment): Handle phi adjustments.
18279 (record_phi_increments): New.
18280 (record_increments): Handle phi adjustments.
18281 (phi_incr_cost): New.
18282 (lowest_cost_path): Handle phis.
18283 (total_savings): Likewise.
18284 (analyze_increments): Likewise.
18285 (ncd_with_phi): New.
18286 (ncd_of_cand_and_phis): New.
18287 (nearest_common_dominator_for_cands): Handle phi increments.
18288 (all_phi_incrs_profitable): New.
18289 (replace_profitable_candidates): Handle phi-dependent candidates.
18290 (analyze_candidates_and_replace): Likewise.
18292 2013-05-03 Teresa Johnson <tejohnson@google.com>
18295 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18296 do not exceed REG_BR_PROB_BASE.
18298 2013-05-03 Jeff Law <law@redhat.com>
18300 PR tree-optimization/57144
18301 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18302 operand of the condition will bit into the new type when eliminating
18303 a cast feeding a condition.
18305 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18307 PR rtl-optimization/57130
18308 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18309 of COMPARE as in_code to the recursive call if needed.
18311 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18313 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18314 (enabled): Handle new members.
18315 * config/i386/sse.md (*vec_concatv2si): Merge from
18316 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18317 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18319 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18321 PR tree-optimization/57027
18322 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18323 for fnms opportunity, check we got the prerequisite kind
18324 of tree / gimple before using accessor functions.
18326 2013-05-03 Richard Biener <rguenther@suse.de>
18328 * double-int.h (lshift): New overload without precision
18329 and arith argument.
18330 (operator *=, operator +=, operator -=): Move ...
18331 * double-int.c (operator *=, operator +=, operator -=): ... here
18332 and implement more efficiently.
18333 (mul_double_with_sign): Remove.
18334 (lshift_double): Adjust to take unsinged shift argument, push
18335 dispatching code to callers.
18336 (mul_double_wide_with_sign): Add early out for callers that
18337 are not interested in high parts or overflow.
18338 (lshift): New function.
18339 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18340 dispatch code here.
18341 (lrotate, rrotate): Use logical shifts.
18342 * expr.c (get_inner_reference): Use lshift.
18343 * fixed-value.c (do_fixed_divide): Likewise.
18344 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18345 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18346 (indirect_refs_may_alias_p): Likewise.
18347 (stmt_kills_ref_p_1): Likewise.
18349 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18351 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18353 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18355 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18356 scalar form of FABD instruction.
18358 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18360 * lra-constraints.c (process_alt_operands): Add checking alt
18361 number to choose the best alternative.
18363 2013-05-02 Richard Biener <rguenther@suse.de>
18365 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18366 bitmap and its handling.
18367 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18369 2013-05-02 Richard Biener <rguenther@suse.de>
18371 PR middle-end/57140
18372 * tree-inline.c (copy_loops): Properly handle removed loops.
18373 (copy_cfg_body): Mark destination loops for fixup if source
18374 loops needed fixup.
18376 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18379 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18380 generating simple_return for naked functions.
18382 2013-05-02 Martin Jambor <mjambor@suse.cz>
18384 PR middle-end/56988
18385 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18386 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18388 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18389 ipa_agg_replacement_value structures.
18390 (known_aggs_to_agg_replacement_list): Likewise.
18391 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18392 (read_agg_replacement_chain): Likewise.
18393 (ipcp_transform_function): Also check that by_ref flags match.
18395 2013-05-02 Richard Biener <rguenther@suse.de>
18397 * graphds.h (struct graph): Add obstack member.
18398 * graphds.c (new_graph): Initialize obstack and allocate
18400 (add_edge): Allocate edge from the obstack.
18401 (free_graph): Free the obstack instead of all edges and vertices.
18403 2013-05-02 Teresa Johnson <tejohnson@google.com>
18405 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18407 * cfg.c (update_bb_profile_for_threading): Ditto.
18408 * tree-inline.c (copy_bb): Ditto.
18409 (copy_edges_for_bb): Ditto.
18410 (initialize_cfun): Ditto.
18411 (copy_cfg_body): Ditto.
18412 (expand_call_inline): Ditto.
18413 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18414 (estimate_node_size_and_time): Ditto.
18415 (inline_merge_summary): Ditto.
18416 * cgraphclones.c (cgraph_clone_edge): Ditto.
18417 (cgraph_clone_node): Ditto.
18418 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18419 (compute_trg_info): Ditto.
18421 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18423 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18424 S reg when fp attribute set.
18425 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18427 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18429 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18431 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18432 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18433 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18435 2013-05-02 Richard Biener <rguenther@suse.de>
18437 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18438 (struct instantiate_cache_entry): New type.
18439 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18440 (struct instantiate_cache_type): New type.
18441 (set_instantiated_value, get_instantiated_value): Remove.
18442 (get_instantiated_value_entry): New function.
18443 (instantiate_scev_name): Use the new cache and adjust.
18444 (instantiate_scev_poly): Adjust.
18445 (instantiate_scev_binary): Likewise.
18446 (instantiate_array_ref): Likewise.
18447 (instantiate_scev_convert): Likewise.
18448 (instantiate_scev_not): Likewise.
18449 (instantiate_scev_3): Likewise.
18450 (instantiate_scev_2): Likewise.
18451 (instantiate_scev_r): Likewise.
18452 (instantiate_scev): Likewise.
18453 (resolve_mixers): Likewise.
18455 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18458 * lra-constraints.c (best_small_class_operands_num): Remove.
18459 (process_alt_operands): Remove small_class_operands_num. Take
18460 small classes operands into losers and only if the operand is not
18461 matched. Modify debugging output.
18462 (curr_insn_transform): Remove best_small_class_operands_num.
18465 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18467 * config/aarch64/aarch64-builtins.c
18468 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18469 * config/aarch64/aarch64-simd-builtins.def
18470 (reduc_splus_): Add new modes.
18471 (reduc_uplus_): New.
18472 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18473 (reduc_uplus_v4sf): Likewise.
18474 (reduc_splus_v4sf): Likewise.
18475 (aarch64_addv<mode>): Likewise.
18476 (reduc_uplus_<mode>): Likewise.
18477 (reduc_splus_<mode>): Likewise.
18478 (aarch64_addvv2di): Likewise.
18479 (reduc_uplus_v2di): Likewise.
18480 (reduc_splus_v2di): Likewise.
18481 (aarch64_addvv2si): Likewise.
18482 (reduc_uplus_v2si): Likewise.
18483 (reduc_splus_v2si): Likewise.
18484 (reduc_<sur>plus_<mode>): New.
18485 (reduc_<sur>plus_v2di): Likewise.
18486 (reduc_<sur>plus_v2si): Likewise.
18487 (reduc_<sur>plus_v4sf): Likewise.
18488 (aarch64_addpv4sf): Likewise.
18489 * config/aarch64/arm_neon.h
18490 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18491 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18492 add UNSPEC_SADDV, UNSPEC_UADDV.
18494 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18496 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18498 * config/aarch64/arm_neon.h
18499 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18501 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18503 * config/aarch64/aarch64-builtins
18504 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18506 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18508 * config/aarch64/aarch64-simd-builtins.def
18509 (reduc_smax_): New.
18510 (reduc_smin_): Likewise.
18511 (reduc_umax_): Likewise.
18512 (reduc_umin_): Likewise.
18513 (reduc_smax_nan_): Likewise.
18514 (reduc_smin_nan_): Likewise.
18517 (smax): Update for V2SF, V4SF and V2DF modes.
18520 (smin_nan): Likewise.
18521 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18522 (<su><maxmin><mode>3): ...This, refactor.
18523 (s<maxmin><mode>3): New.
18524 (<maxmin_uns><mode>3): Likewise.
18525 (reduc_<maxmin_uns>_<mode>): Refactor.
18526 (reduc_<maxmin_uns>_v4sf): Likewise.
18527 (reduc_<maxmin_uns>_v2si): Likewise.
18528 (aarch64_<fmaxmin><mode>: Remove.
18529 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18531 (vmin<q>_f<32,64>): Likewise.
18532 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18534 (su): Add mappings for smax, smin, umax, umin.
18536 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18537 (FMAXMIN): Rename as...
18538 (FMAXMIN_UNS): ...This.
18540 (fmaxminv): Likewise.
18541 (fmaxmin): Likewise.
18543 (maxmin_uns_op): Likewise.
18545 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18547 * config/aarch64/arm_neon.h
18548 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18549 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18550 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18552 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18554 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18555 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18557 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18559 * config/aarch64/aarch64-simd.md
18560 (vcond<mode>_internal): Handle special cases for constant masks.
18561 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18562 (vcondu<mode><mode>): Likewise.
18563 (vcond<v_cmp_result><mode>): New.
18565 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18567 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18568 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18569 * config/aarch64/aarch64-simd-builtins.def
18570 (cmeq): Update to BUILTIN_VALLDI.
18575 * config/aarch64/arm_neon.h
18576 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18577 to builtins or C as appropriate.
18579 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18581 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18583 (cmhi): Rename to...
18585 * config/aarch64/aarch64-simd.md
18586 (simd_mode): Add SF.
18587 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18588 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18589 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18590 (cstore<mode>_neg): ...This.
18591 * config/aarch64/iterators.md
18593 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18594 (COMPARISONS): New.
18595 (UCOMPARISONS): Likewise.
18596 (optab): Add missing comparisons.
18602 (VCMP_S): Likewise.
18603 (VCMP_U): Likewise.
18604 (V_cmp_result): Add DF, SF modes.
18605 (v_cmp_result): Likewise.
18607 (vmtype): Likewise.
18608 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18610 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18612 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18613 define_insn to define_insn_and_split.
18614 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18615 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18616 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18617 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18618 (thumb2_negscc): Likewise.
18620 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18622 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18624 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18626 * config/arm/thumb2.md: Remove trailing whitespaces.
18628 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18630 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18631 Use gen_int_mode rather than GEN_INT.
18633 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18635 * value-prof.c (stream_in_histogram_value): Remove the strayed
18638 2013-04-30 Richard Biener <rguenther@suse.de>
18640 PR middle-end/57122
18641 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18643 2013-04-30 Richard Biener <rguenther@suse.de>
18645 PR middle-end/57107
18646 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18648 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18650 PR rtl-optimization/56957
18651 PR rtl-optimization/57105
18652 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18653 variable. Use just INSN_UID for determining whether an insn
18654 should be only disconnected from the insn stream.
18655 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18657 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18659 PR tree-optimization/57104
18660 * tsan.c (instrument_expr): Don't instrument accesses to
18661 DECL_HARD_REGISTER VAR_DECLs.
18663 2013-04-30 Richard Biener <rguenther@suse.de>
18665 * function.h (loops_for_fn): New inline function.
18666 (set_loops_for_fn): Likewise.
18667 * cfgloop.h (place_new_loop): Add struct function parameter.
18668 (get_loop): Likewise.
18669 (get_loops): Likewise.
18670 (number_of_loops): Likewise.
18671 (fel_next): Adjust.
18672 (fel_init): Likewise.
18673 * cfg.c (get_loop_copy): Adjust.
18674 * cfgloop.c (flow_loops_dump): Likewise.
18675 (record_loop_exits): Likewise.
18676 (verify_loop_structure): Likewise.
18677 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18678 (estimate_reg_pressure_cost): Likewise.
18679 (mark_loop_exit_edges): Likewise.
18680 * cfgloopmanip.c (place_new_loop): Likewise.
18681 (add_loop): Likewise.
18682 (duplicate_loop): Likewise.
18683 * graph.c (draw_cfg_nodes): Likewise.
18684 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18685 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18686 (extract_affine_chrec): Likewise.
18687 (build_scop_iteration_domain): Likewise.
18688 * graphite.c (graphite_initialize): Likewise.
18689 * ira-build.c (create_loop_tree_nodes): Likewise.
18690 (more_one_region_p): Likewise.
18691 (rebuild_regno_allocno_maps): Likewise.
18692 (mark_loops_for_removal): Likewise.
18693 (mark_all_loops_for_removal): Likewise.
18694 (remove_unnecessary_regions): Likewise.
18695 (ira_build): Likewise.
18696 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18697 * loop-init.c (fix_loop_structure): Likewise.
18698 (gate_rtl_move_loop_invariants): Likewise.
18699 (gate_rtl_unswitch): Likewise.
18700 (gate_rtl_unroll_and_peel_loops): Likewise.
18701 (rtl_doloop): Likewise.
18702 * lto-streamer-in.c (input_cfg): Likewise.
18703 * lto-streamer-out.c (output_cfg): Likewise.
18704 * modulo-sched.c (sms_schedule): Likewise.
18705 * predict.c (tree_estimate_probability): Likewise.
18706 (tree_estimate_probability_driver): Likewise.
18707 (estimate_loops): Likewise.
18708 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18709 (move_sese_region_to_fn): Likewise.
18710 (debug_loop_num): Likewise.
18711 * tree-chrec.c (chrec_evaluate): Likewise.
18712 (hide_evolution_in_other_loops_than_loop): Likewise.
18713 (chrec_component_in_loop_num): Likewise.
18714 (reset_evolution_in_loop): Likewise.
18715 (evolution_function_is_invariant_rec_p): Likewise.
18716 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18717 * tree-inline.c (copy_loops): Likewise.
18718 (copy_cfg_body): Likewise.
18719 (tree_function_versioning): Likewise.
18720 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18721 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18723 (add_to_evolution_1): Likewise.
18724 (scev_const_prop): Likewise.
18725 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18726 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18727 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18728 (tree_ssa_lim_initialize): Likewise.
18729 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18730 (verify_loop_closed_ssa): Likewise.
18731 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18732 (tree_ssa_loop_im): Likewise.
18733 (tree_ssa_loop_unswitch): Likewise.
18734 (tree_vectorize): Likewise.
18735 (check_data_deps): Likewise.
18736 (tree_ssa_loop_ivcanon): Likewise.
18737 (tree_ssa_loop_bounds): Likewise.
18738 (tree_complete_unroll): Likewise.
18739 (tree_complete_unroll_inner): Likewise.
18740 (tree_parallelize_loops): Likewise.
18741 (tree_ssa_loop_prefetch): Likewise.
18742 (tree_ssa_loop_ivopts): Likewise.
18743 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18744 * tree-vectorizer.c (vectorize_loops): Likewise.
18746 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18748 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18749 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18750 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18751 with EABI_LINK_SPEC.
18753 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18756 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18759 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18762 * lra-constraints.c (process_alt_operands): Discourage a bit more
18763 using memory for pseudos. Print cost dump for alternatives.
18764 Modify cost values for conflicts with early clobbers.
18765 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18767 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18770 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18772 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18774 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18775 from/to S register.
18776 (movdi_aarch64): Support LDR/STR from/to D register.
18778 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18780 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18781 or higher by default.
18783 2013-04-29 Richard Biener <rguenther@suse.de>
18785 PR middle-end/57075
18786 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18787 even if not adding abnormal edges for calls that can make
18790 2013-04-29 Richard Biener <rguenther@suse.de>
18792 PR middle-end/57103
18793 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18795 (move_stmt_r): Remove redundant checking.
18797 2013-04-29 Teresa Johnson <tejohnson@google.com>
18800 * basic-block.h (apply_scale): New function.
18801 (apply_probability): Use apply_scale.
18802 * gimple-streamer-in.c (input_bb): Ditto.
18803 * lto-streamer-in.c (input_cfg): Ditto.
18804 * lto-cgraph.c (merge_profile_summaries): Ditto.
18805 * tree-optimize.c (execute_fixup_cfg): Ditto.
18806 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18807 (copy_edges_for_bb): Ditto.
18808 (copy_cfg_body): Ditto.
18810 2013-04-29 Tom de Vries <tom@codesourcery.com>
18812 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18813 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18814 (tail_merge_optimize): Handle current_loops == NULL.
18816 2013-04-26 Jeff Law <law@redhat.com>
18818 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18819 (simplify_cond_using_ranges): Generalize code to simplify
18820 COND_EXPRs where one argument is a constant and the other
18821 is an SSA_NAME created by an integral type conversion.
18823 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18825 * config/arm/arm.md (store_minmaxsi): Use only when
18826 optimize_insn_for_size_p.
18828 2013-04-29 Christian Bruel <christian.bruel@st.com>
18831 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18833 2013-04-29 Richard Biener <rguenther@suse.de>
18835 PR middle-end/57089
18836 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18837 loop tree make sure to schedule a fixup for the child as well.
18838 (expand_omp_for_generic): Properly add loops.
18839 (expand_omp_for_static_nochunk): Likewise.
18840 (expand_omp_for_static_chunk): Likewise.
18841 (expand_omp_for): For the degenerate case fixup loops.
18842 (expand_omp_sections): Fix default bb placement in loops.
18843 (expand_omp_atomic_pipeline): Properly add loops.
18845 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18847 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18849 2013-04-29 Tom de Vries <tom@codesourcery.com>
18851 * tree-ssa-tail-merge.c: Update header comment.
18853 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18855 * config/aarch64/arm_neon.h
18856 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18857 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18858 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18859 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18860 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18861 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18862 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18863 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18865 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18867 * config/aarch64/aarch64-simd.md
18868 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18869 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18870 fix_trunc, fixuns_trunc.
18871 (ftrunc<VDQF:mode>2): New.
18872 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18873 (fix_trunc_optab): New.
18875 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18877 * config/aarch64/aarch64-builtins.c
18878 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18879 iceilf, lround, iroundf.
18881 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18884 * config/i386/i386.h (enum ix86_tune_indices)
18885 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18886 New, split from X86_TUNE_INTER_UNIT_MOVES.
18887 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18888 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18889 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18890 (TARGET_INTER_UNIT_MOVES): Remove.
18891 * config/i386/i386.c (initial_ix86_tune_features): Update.
18892 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18893 (ix86_expand_convert_uns_didf_sse): Use
18894 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18895 (ix86_expand_vector_init_one_nonzero): Ditto.
18896 (ix86_expand_vector_init_interleave): Ditto.
18897 (inline_secondary_memory_needed): Return true for moves from SSE class
18898 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18899 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18900 * config/i386/constraints.md (Yi, Ym): Depend on
18901 TARGET_INTER_UNIT_MOVES_TO_VEC.
18902 (Yj, Yn): New constraints.
18903 * config/i386/i386.md (*movdi_internal): Change constraints of
18904 operand 1 from Yi to Yj and from Ym to Yn.
18905 (*movsi_internal): Ditto.
18906 (*movdf_internal): Ditto.
18907 (*movsf_internal): Ditto.
18908 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18909 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18910 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18911 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18912 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18913 * config/i386/sse.md (movdi_to_sse): Ditto.
18914 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18915 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18916 TARGET_INTER_UNIT_MOVES.
18917 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18918 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18919 instead of TARGET_INTER_UNIT_MOVES.
18920 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18921 operand 1 from Yi to Yj and from Ym to Yn.
18923 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18925 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18926 (float_truncate_hi_): Likewise.
18927 (float_extend_lo_): Likewise.
18928 (float_truncate_lo_): Likewise.
18929 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18930 (aarch64_float_extend_lo_v2df): Likewise.
18931 (vec_unpacks_hi_v4sf): Likewise.
18932 (aarch64_float_truncate_lo_v2sf): Likewise.
18933 (aarch64_float_truncate_hi_v4sf): Likewise.
18934 (vec_pack_trunc_v2df): Likewise.
18935 (vec_pack_trunc_df): Likewise.
18937 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18939 * config/aarch64/aarch64-builtins.c
18940 (aarch64_fold_builtin): Fold float conversions.
18941 * config/aarch64/aarch64-simd-builtins.def
18942 (floatv2si, floatv4si, floatv2di): New.
18943 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18944 * config/aarch64/aarch64-simd.md
18945 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18946 * config/aarch64/iterators.md (FLOATUORS): New.
18947 (optab): Add float, floatuns.
18948 (su_optab): Likewise.
18950 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18952 * config/aarch64/aarch64-builtins.c
18953 (aarch64_builtin_vectorized_function): Use new names for
18955 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18956 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18957 (fcvtzu): Split as...
18958 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18959 (fcvtas): Split as...
18960 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18961 (fcvtau): Split as...
18962 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18963 (fcvtps): Split as...
18964 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18965 (fcvtpu): Split as...
18966 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18967 (fcvtms): Split as...
18968 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18969 (fcvtmu): Split as...
18970 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18971 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18972 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18973 (lfrintnusf, lfrintnudf): Likewise.
18974 * config/aarch64/aarch64-simd.md
18975 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18977 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18978 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18979 (fcvt_pattern): Likewise.
18981 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18983 * config/aarch64/aarch64-simd.md
18984 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18985 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18987 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18989 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18990 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18991 (vrnd<a,m,n,p>_f32): Implement using builtins.
18992 (vrnd<i,x><q>_f<32, 64>): New.
18994 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18996 * config/aarch64/aarch64-builtins.c
18997 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18998 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18999 (frintz): Rename to...
19001 (frintp): Rename to...
19003 (frintm): Rename to...
19005 (frinti): Rename to...
19006 (nearbyint): ...this.
19007 (frintx): Rename to...
19009 (frinta): Rename to...
19011 * config/aarch64/aarch64-simd.md
19012 (aarch64_frint<frint_suffix><mode>): Delete.
19013 (<frint_pattern><mode>2): Convert to insn.
19014 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19015 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19016 (frint_pattern): Likewise.
19017 (frint_suffix): Likewise.
19019 2013-04-29 Richard Biener <rguenther@suse.de>
19021 PR tree-optimization/57081
19022 * loop-init.c: Include tree-flow.h.
19023 (loop_optimizer_finalize): Free number of iteration estimates.
19024 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19026 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19028 PR tree-optimization/57083
19029 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19030 non-singleton shift count range, zero extend low_bound for uns case.
19032 * config/i386/predicates.md (general_vector_operand): New predicate.
19033 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19034 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19035 if they aren't nonimmediate operands. If their original values
19036 satisfy const_vector_equal_evenodd_p, don't shift them.
19037 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19038 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19039 if not nonimmediate_operand.
19040 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19041 instead of register_operand.
19042 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19044 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19046 * stor-layout.c (finalize_size_functions): Allocate a structure and
19047 reset cfun before dumping the functions.
19049 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19051 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19054 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19055 use xop_pmacsdqh if uns_p.
19056 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19057 the immediate rotate count.
19059 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19061 * rtl.h (struct rtx_def): Add comment for field jump.
19062 (LRA_SUBREG_P): New macro.
19063 * recog.c (register_operand): Check LRA_SUBREG_P.
19064 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19066 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19067 (lra_final_code_change): Skip subreg change for operators.
19068 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19069 if there are no operand changes.
19070 * lra-constraints.c (curr_insn_set): New.
19071 (match_reload): Set LRA_SUBREG_P.
19072 (emit_spill_move): Ditto.
19073 (check_and_process_move): Use curr_insn_set. Process only single
19074 set insns. Don't initialize sec_mem_p and change_p.
19075 (simplify_operand_subreg): Use LRA_SUBREG_P.
19076 (reg_in_class_p): New function.
19077 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19078 of #ifdef. Add code to remove cycling.
19079 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19080 non-null disp. Reload inner instead of disp when base and index
19081 are null. Try to put lo_sum into register.
19082 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19083 (check_and_process_move): Move code for move cost check to
19084 simple_move_p. Remove equiv_substitution.
19085 (simple_move_p): New function.
19086 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19087 curr_insn_set. Call check_and_process_move only for single set
19088 insns. Use the new function. Move call of check_and_process_move
19089 after operand equiv substitution and address process.
19091 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19094 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19095 with nonlocal goto receivers or returns twice calls, ignore
19096 unininitialized values from abnormal edges to nl goto receiver
19097 or returns twice call.
19099 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19101 PR tree-optimization/57051
19102 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19103 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19106 2013-04-26 Richard Biener <rguenther@suse.de>
19108 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19109 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19110 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19111 (fixup_loop_arrays_after_move): New function.
19112 (move_sese_region_to_fn): Properly outline the loop tree parts
19113 of the SESE region.
19115 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19117 * config/i386/i386.md (type, unit): Fix long lines.
19119 2013-04-26 Richard Biener <rguenther@suse.de>
19121 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19122 (lto-streamer-out.o): Likewise.
19123 * cfgloop.c (init_loops_structure): Export, add struct function
19124 argument and adjust.
19125 (flow_loops_find): Adjust.
19126 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19127 (init_loops_structure): Declare.
19128 * lto-streamer-in.c: Include cfgloop.h.
19129 (input_cfg): Input the loop tree.
19130 * lto-streamer-out.c: Include cfgloop.h.
19131 (output_cfg): Output the loop tree.
19132 (output_struct_function_base): Do not drop PROP_loops.
19134 2013-03-26 Richard Biener <rguenther@suse.de>
19136 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19137 (pass_build_cfg): Provide PROP_loops.
19138 (move_sese_region_to_fn): Remove loops that are outlined into fn
19140 * tree-inline.c: Include cfgloop.h.
19141 (initialize_cfun): Do not drop PROP_loops.
19142 (copy_loops): New function.
19143 (copy_cfg_body): Copy loop structure.
19144 (tree_function_versioning): Initialize destination loop tree.
19145 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19146 (pass_parallelize_loops): Do IL verification.
19147 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19148 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19149 the CFG make sure we fixup loops as well.
19150 * tree-ssa-tail-merge.c: Include cfgloop.h.
19151 (replace_block_by): When merging loop latches mark loops for fixup.
19152 * lto-streamer-out.c (output_struct_function_base): Drop
19153 PROP_loops for now.
19154 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19155 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19156 * ipa-split.c: Include cfgloop.h.
19157 (split_function): Add the new return block to the loop tree root.
19158 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19159 whether we have removed the forwarder block.
19160 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19161 * cfgloop.h (place_new_loop): Declare.
19162 * cfgloopmanip.c (place_new_loop): Export.
19163 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19164 (tree-switch-conversion.o): Likewise.
19165 (tree-complex.o): Likewise.
19166 (tree-inline.o): Likewise.
19167 (tree-ssa-tailmerge.o): Likewise.
19168 (ipa-split.o): Likewise.
19169 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19170 (tree-ssa-copy.o): Likewise.
19171 * tree-switch-conversion.c: Include cfgloop.h
19172 (process_switch): If we emit a bit-test cascade, schedule loops
19174 * tree-complex.c: Include cfgloop.h.
19175 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19176 * asan.c: Include cfgloop.h.
19177 (create_cond_insert_point): Properly add new basic-blocks to
19178 loops, schedule loop fixup.
19179 * cfgloop.c (verify_loop_structure): Check that looks are not
19181 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19183 (expand_omp_for_generic): Likewise.
19184 (expand_omp_sections): Likewise.
19185 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19186 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19187 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19188 is initialized, not when loops are present.
19189 * tree-parloops.c (parallelize_loops): Remove checking here.
19190 * passes.c (init_optimization_passes): Schedule a copy-propagation
19191 pass before complete unrolling of inner loops.
19193 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19195 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19196 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19197 (colorize_init): Add argument to _WIN32 version.
19198 * toplev.c: Include diagnostic-color.h.
19199 (process_options): Default to -fdiagnostics-color=auto if
19200 GCC_COLORS env var is in the environment.
19201 * common.opt (fdiagnostics-color=): Add Var and Init.
19202 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19203 env var is in the environment, the default is auto rather than never.
19205 * diagnostic.h (file_name_as_prefix): Add context argument.
19206 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19207 the string as locus.
19208 * langhooks.c (lhd_print_error_function): Adjust caller.
19210 2013-04-25 Lawrence Crowl <crowl@google.com>
19212 * var-tracking.c (shared_hash_def::htab):
19213 Change type to hash_table. Update dependent calls and types.
19215 2013-04-25 Lawrence Crowl <crowl@google.com>
19217 * Makefile.in: Update as needed below.
19219 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19220 Move declaration to after the type's method definitons.
19222 * attribs.c (htab_t scoped_attributes::attribute_hash):
19223 Change type to hash_table. Update dependent calls and types.
19225 * bitmap.c (htab_t bitmap_desc_hash):
19226 Change type to hash_table. Update dependent calls and types.
19228 * cselib.c (htab_t cselib_hash_table):
19229 Change type to hash_table. Update dependent calls and types.
19231 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19232 (hash_string_slot_node): Move implementation into lto-streamer.h
19233 struct string_slot_hasher.
19234 (eq_string_slot_node): Likewise.
19236 * data-streamer-out.c: Update output_block::string_hash_table
19237 dependent calls and types.
19239 * dwarf2cfi.c (htab_t trace_index):
19240 Change type to hash_table. Update dependent calls and types.
19242 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19243 Change type to hash_table. Update dependent calls and types.
19244 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19245 (htab_t optimize_external_refs::map): Likewise.
19246 (htab_t output_comp_unit::extern_map): Likewise.
19247 (htab_t output_comdat_type_unit::extern_map): Likewise.
19248 (htab_t output_macinfo::macinfo_htab): Likewise.
19249 (htab_t optimize_location_lists::htab): Likewise.
19250 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19252 * except.c (htab_t ehspec_hash_type):
19253 Change type to hash_table. Update dependent calls and types.
19254 (assign_filter_values::ttypes): Likewise.
19255 (assign_filter_values::ehspec): Likewise.
19256 (sjlj_assign_call_site_values::ar_hash): Likewise.
19257 (convert_to_eh_region_ranges::ar_hash): Likewise.
19259 * gcse.c (htab_t pre_ldst_table):
19260 Change type to hash_table. Update dependent calls and types.
19262 * ggc-common.c (htab_t saving_htab):
19263 Change type to hash_table. Update dependent calls and types.
19264 (htab_t loc_hash): Likewise.
19265 (htab_t ptr_hash): Likewise.
19266 (call_count): Rename ggc_call_count.
19267 (call_alloc): Rename ggc_call_alloc.
19268 (loc_descriptor): Rename make_loc_descriptor.
19269 (add_statistics): Rename ggc_add_statistics.
19271 * ggc-common.c (saving_htab):
19272 Change type to hash_table. Update dependent calls and types.
19274 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19275 (push_gimplify_context): Likewise.
19276 (pop_gimplify_context): Likewise.
19277 (struct gimple_temp_hash_elt): Added.
19278 (struct gimplify_hasher): Likewise.
19279 (struct gimplify_ctx.temp_htab):
19280 Change type to hash_table. Update dependent calls and types.
19282 * gimple-fold.c: Include gimplify-ctx.h.
19284 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19285 Change type to hash_table. Update dependent calls and types.
19286 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19287 avoid potential global name collision.
19289 * gimplify.c: Include gimplify-ctx.h.
19290 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19291 (htab_t gimplify_ctx::temp_htab):
19292 Update dependent calls and types for new type hash_table.
19293 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19294 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19296 * gimplify-ctx.h: New.
19297 (struct gimple_temp_hash_elt): Move from gimplify.c.
19298 (class gimplify_hasher): New.
19299 (struct gimplify_ctx): Move from gimple.h.
19300 (htab_t gimplify_ctx::temp_htab):
19301 Change type to hash_table. Update dependent calls and types.
19303 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19304 (htab_t ivs_params::newivs_index):
19305 Change type to hash_table. Update dependent calls and types.
19306 (htab_t ivs_params::params_index): Likewise.
19307 (htab_t print_generated_program::params_index): Likewise.
19308 (htab_t gloog::newivs_index): Likewise.
19309 (htab_t gloog::params_index): Likewise.
19311 * graphite.c: Include graphite-htab.h.
19312 4htab_t graphite_transform_loops::bb_pbb_mapping):
19313 Change type to hash_table. Update dependent calls and types.
19315 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19316 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19317 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19319 * graphite-dependences.c: Include graphite-htab.h.
19320 (loop_is_parallel_p): Change hash table type of parameter.
19322 * graphite-htab.h: New.
19323 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19324 (extern find_pbb_via_hash): Move from graphite-poly.h.
19325 (extern loop_is_parallel_p): Move from graphite-poly.h.
19326 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19328 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19329 (extern loop_is_parallel_p): Move to graphite-htab.h.
19330 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19332 * haifa-sched.c (htab_t delay_htab):
19333 Change type to hash_table. Update dependent calls and types.
19334 (htab_t delay_htab_i2): Likewise.
19336 * ira-color.c (htab_t allocno_hard_regs_htab):
19337 Change type to hash_table. Update dependent calls and types.
19339 * ira-costs.c (htab_t cost_classes_htab):
19340 Change type to hash_table. Update dependent calls and types.
19342 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19343 Change type to hash_table. Update dependent calls and types.
19345 * loop-iv.c (htab_t bivs):
19346 Change type to hash_table. Update dependent calls and types.
19348 * loop-unroll.c (htab_t opt_info::insns_to_split):
19349 Change type to hash_table. Update dependent calls and types.
19350 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19352 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19353 (struct string_slot_hasher): New.
19354 (htab_t output_block::string_hash_table):
19355 Change type to hash_table. Update dependent calls and types.
19357 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19358 (htab_t file_name_hash_table):
19359 Change type to hash_table. Update dependent calls and types.
19361 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19364 * lto-streamer.c (htab_t tree_htab):
19365 Change type to hash_table. Update dependent calls and types.
19367 * omp-low.c: Include gimplify-ctx.h.
19369 * passes.c (htab_t name_to_pass_map):
19370 Change type to hash_table. Update dependent calls and types.
19371 (pass_traverse): Rename to passes_pass_traverse.
19373 * plugin.c (htab_t event_tab):
19374 Change type to hash_table. Update dependent calls and types.
19376 * postreload-gcse.c (htab_t expr_table):
19377 Change type to hash_table. Update dependent calls and types.
19378 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19380 * sese.c (debug_rename_map_1): Make extern.
19381 (htab_t copy_bb_and_scalar_dependences::rename_map):
19382 Change type to hash_table. Update dependent calls and types.
19384 * sese.h (extern debug_rename_map): Move to .c file.
19386 * store-motion.c (htab_t store_motion_mems_table):
19387 Change type to hash_table. Update dependent calls and types.
19389 * trans-mem.c (htab_t tm_new_mem_hash):
19390 Change type to hash_table. Update dependent calls and types.
19392 * tree-browser.c (htab_t TB_up_ht):
19393 Change type to hash_table. Update dependent calls and types.
19395 * tree-cfg.c (htab_t discriminator_per_locus):
19396 Change type to hash_table. Update dependent calls and types.
19398 * tree-complex.c: Include tree-hasher.h
19399 (htab_t complex_variable_components):
19400 Change type to hash_table. Update dependent calls and types.
19402 * tree-eh.c (htab_t finally_tree):
19403 Change type to hash_table. Update dependent calls and types.
19405 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19406 struct int_tree_hasher.
19407 (extern int_tree_map_eq): Likewise.
19408 (uid_decl_map_hash): Removed.
19409 (extern decl_tree_map_eq): Likewise.
19411 * tree-hasher.h: New.
19412 (struct int_tree_hasher): New.
19413 (typedef int_tree_htab_type): New.
19415 * tree-inline.c: Include gimplify-ctx.h.
19417 * tree-mudflap.c: Include gimplify-ctx.h.
19419 * tree-parloops.c: Include tree-hasher.h.
19420 (htab_t eliminate_local_variables_stmt::decl_address):
19421 Change type to hash_table. Update dependent calls and types.
19422 (htab_t separate_decls_in_region::decl_copies): Likewise.
19424 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19425 Change type to hash_table. Update dependent calls and types.
19427 * tree-sra.c (candidates):
19428 Change type to hash_table. Update dependent calls and types.
19430 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19432 (int_tree_map_hash): Likewise.
19434 * tree-ssa-dom.c (htab_t avail_exprs):
19435 Change type to hash_table. Update dependent calls and types.
19437 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19438 Change type to hash_table. Update dependent calls and types.
19440 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19441 Change type to hash_table. Update dependent calls and types.
19443 * tree-ssa-phiopt.c (seen_ssa_names):
19444 Change type to hash_table. Update dependent calls and types.
19446 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19447 Change type to hash_table. Update dependent calls and types.
19449 * tree-ssa-uncprop.c (equiv):
19450 Change type to hash_table. Update dependent calls and types.
19452 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19454 PR rtl-optimization/57003
19455 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19456 call note_stores with kill_clobbered_value callback again after
19457 killing regs_invalidated_by_call.
19459 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19461 * config/aarch64/aarch64-simd.md
19462 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19463 (aarch64_simd_bsl<mode>): Likewise.
19464 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19466 2013-04-25 Marek Polacek <polacek@redhat.com>
19468 PR tree-optimization/57066
19469 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19471 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19473 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19475 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19477 * config/aarch64/aarch64-builtins.c
19478 (aarch64_fold_builtin): New.
19479 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19480 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19481 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19482 * config/aarch64/arm_neon.h
19483 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19485 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19486 Tejas Belagod <tejas.belagod@arm.com>
19488 * config/aarch64/aarch64-builtins.c
19489 (aarch64_gimple_fold_builtin): New.
19490 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19491 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19492 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19493 (addvv4sf): Update.
19494 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19496 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19498 * config/aarch64/aarch64.md
19499 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19501 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19503 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19504 (*ngcsi_uxtw): New pattern.
19506 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19507 Julian Brown <julian@codesourcery.com>
19509 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19510 (TB_DREG): Add T_V4HF.
19511 (v4hf_UP): New macro.
19512 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19513 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19514 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19516 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19517 (arm_vector_mode_supported_p): Handle V4HF.
19518 (arm_mangle_map): Handle V4HFmode.
19519 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19520 * config/arm/arm_neon_builtins.def: Add entries for
19521 vcvtv4hfv4sf, vcvtv4sfv4hf.
19522 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19523 (neon_vcvtv4hfv4sf): Likewise.
19524 * config/arm/neon-gen.ml: Handle half-precision floating point
19526 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19527 * config/arm/arm_neon.h: Regenerate.
19528 * config/arm/neon.ml (type elts): Add F16.
19529 (type vectype): Add T_float16x4, T_floatHF.
19530 (type vecmode): Add V4HF.
19531 (type features): Add Requires_FP_bit feature.
19532 (elt_width): Handle F16.
19533 (elt_class): Likewise.
19534 (elt_of_class_width): Likewise.
19535 (mode_of_elt): Refactor.
19536 (type_for_elt): Handle F16, fix error messages.
19537 (vectype_size): Handle T_float16x4.
19538 (vcvt_sh): New function.
19539 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19540 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19541 (string_of_mode): Handle V4HF.
19542 * doc/arm-neon-intrinsics.texi: Regenerate.
19544 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19546 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19547 format specifier in 'X' case.
19549 2013-04-25 Alan Modra <amodra@gmail.com>
19552 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19553 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19554 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19555 Repeat for many other rotate/shift and mask patterns using subregs.
19556 Name lshiftrt insns.
19557 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19558 on WORDS_BIG_ENDIAN.
19560 2013-04-25 Alan Modra <amodra@gmail.com>
19562 * config.gcc: Support little-endian powerpc-linux targets.
19563 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19564 (LINK_OS_LINUX_SPEC): Define.
19565 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19566 Preserve MASK_LITTLE_ENDIAN.
19567 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19568 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19569 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19570 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19571 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19572 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19573 Correct fp word order for little-endian. Don't shift toc entries
19574 smaller than a word for little-endian.
19575 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19576 (bswapdi2 splits): Correct low-part subreg for little-endian.
19577 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19578 low/high where such is correct only for be.
19579 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19580 little-endian for -mcall-aixdesc.
19582 2013-04-25 Alan Modra <amodra@gmail.com>
19584 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19585 replace_equiv_address_nv.
19587 2013-04-25 Alan Modra <amodra@gmail.com>
19589 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19591 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19594 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19595 * rtl.h (struct rtx_def): ...
19597 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19599 PR rtl-optimizations/57046
19600 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19601 for multi-reg splits.
19603 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19605 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19607 2013-04-24 Sterling Augustine <saugustine@google.com>
19609 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19610 (comp_dir_string, debug_str_dwo_section): New.
19611 (DEBUG_STR_DWO_SECTION): Rename to ...
19612 (DEBUG_DWO_STR_SECTION): ... this.
19613 (DEBUG_NORM_STR_SECTION): Delete.
19614 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19615 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19616 (find_AT_string): Move most logic to ...
19617 (find_AT_string_in_table): ... here. New.
19618 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19619 add_skeleton_AT_string. Delete logic.
19620 (output_skeleton_debug_sections): Remove call to
19621 add_top_level_skeleton_die_attrs.
19622 (add_comp_dir_attribute): Move logic to comp_dir_string.
19623 (dwarf2out_init): Initialize debug_str_dwo_section.
19624 (output_indirect_string): Call find_string_form.
19625 (output_indirect_strings): Rewrite.
19626 (prune_unused_types): Empty skeleton_debug_str_hash.
19627 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19628 (dwarf2out_finish): Call output_indirect_strings.
19630 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19632 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19634 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19636 * rtl.h (struct rtx_def): Add comment for field jump.
19637 (LRA_SUBREG_P): New macro.
19638 * recog.c (register_operand): Check LRA_SUBREG_P.
19639 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19641 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19642 (lra_final_code_change): Skip subreg change for operators.
19643 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19644 if there are no operand changes.
19645 * lra-constraints.c (curr_insn_set): New.
19646 (match_reload): Set LRA_SUBREG_P.
19647 (emit_spill_move): Ditto.
19648 (check_and_process_move): Use curr_insn_set. Process only single
19649 set insns. Don't initialize sec_mem_p and change_p.
19650 (simplify_operand_subreg): Use LRA_SUBREG_P.
19651 (reg_in_class_p): New function.
19652 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19653 of #ifdef. Add code to remove cycling.
19654 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19655 non-null disp. Reload inner instead of disp when base and index
19656 are null. Try to put lo_sum into register.
19657 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19658 (check_and_process_move): Move code for move cost check to
19659 simple_move_p. Remove equiv_substitution.
19660 (simple_move_p): New function.
19661 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19662 curr_insn_set. Call check_and_process_move only for single set
19663 insns. Use the new function. Move call of check_and_process_move
19664 after operand equiv substitution and address process.
19666 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19668 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19669 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19670 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19672 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19674 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19676 2013-04-24 Marek Polacek <polacek@redhat.com>
19678 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19679 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19680 (select_loops_exit_conditions): Likewise.
19681 (number_of_iterations_for_all_loops): Likewise.
19682 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19683 (scev_analysis): Likewise.
19685 2013-04-02 Catherine Moore <clm@codesourcery.com>
19686 Chao-ying Fu <fu@mips.com>
19688 * config/mips/micromips.md (jraddiusp): New pattern.
19689 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19690 instruction if possible.
19692 2013-04-24 Alan Modra <amodra@gmail.com>
19694 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19696 2013-04-24 Julian Brown <julian@codesourcery.com>
19697 Chung-Lin Tang <cltang@codesourcery.com>
19699 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19700 dependency behavior in enumeration type DIE generation. Add TODO note
19701 to comments about future DW_FORM_sdata/udata re-work of related code.
19703 2013-04-23 Lawrence Crowl <crowl@google.com>
19705 * Makefile.in: Update as needed below.
19707 * hash-table.h (class hash_table):
19708 Correct many methods with parameter types compare_type to the correct
19709 value_type. (Correct code was unlikely to notice the change.)
19710 (hash_table::elements_with_deleted) New.
19711 (class hashtable::iterator): New.
19712 (hashtable::begin()): New.
19713 (hashtable::end()): New.
19714 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19716 * statistics.c (statistics_hashes):
19717 Change type to hash_table. Update dependent calls and types.
19719 * tree-into-ssa.c (var_infos):
19720 Change type to hash_table. Update dependent calls and types.
19722 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19723 Change type to hash_table. Update dependent calls and types.
19725 * tree-ssa-loop-im.c (struct mem_ref.refs):
19726 Change type to hash_table. Update dependent calls and types.
19728 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19729 Change type to hash_table. Update dependent calls and types.
19731 * tree-ssa-sccvn.c (vn_tables_s::nary):
19732 Change type to hash_table. Update dependent calls and types.
19733 (vn_tables_s::phis): Likewise.
19734 (vn_tables_s::references): Likewise.
19736 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19737 (vn_reference_eq): Update parameter and return types.
19739 * tree-ssa-structalias.c (pointer_equiv_class_table):
19740 Change type to hash_table. Update dependent calls and types.
19741 (location_equiv_class_table): Likewise.
19743 * tree-vect-data-refs.c: Consequential changes for making
19744 peeling a hash_table.
19746 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19747 (destroy_loop_vec_info): Dependent hash_table update.
19749 * tree-vectorizer.h (peeling_htab):
19750 Change type to hash_table. Update dependent calls and types.
19752 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19754 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19755 to check the register content is equal or not.
19756 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19757 to assign register content record.
19758 * lra-eliminations.c (update_reg_eliminate): Use
19759 lra_update_reg_val_offset to update register content offset.
19760 * lra-int.h (struct lra_reg): Add offset member.
19761 (lra_reg_val_equal_p): New static inline function.
19762 (lra_update_reg_val_offset): New static inline function.
19763 (lra_assign_reg_val): New static inline function.
19764 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19765 to assign register content record.
19766 (initialize_lra_reg_info_element): Initial offset to zero.
19768 2013-04-23 Catherine Moore <clm@codesourcery.com>
19770 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19771 operands. Record compression.
19773 2013-04-23 Xinliang David Li <davidxl@google.com>
19775 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19777 2013-04-23 Richard Biener <rguenther@suse.de>
19779 PR middle-end/57036
19780 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19781 parameter, only add abnormal goto edges from the copied body
19782 if the call could perform abnormal gotos.
19783 (copy_cfg_body): Adjust.
19785 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19787 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19789 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19791 * coretypes.h (gimple_stmt_iterator): Add struct to make
19794 2013-04-23 Richard Biener <rguenther@suse.de>
19796 PR tree-optimization/57026
19797 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19798 from SSA names occuring in abnormal PHI nodes.
19800 2013-04-22 Andi Kleen <ak@linux.intel.com>
19802 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19804 2013-04-22 Andi Kleen <ak@linux.intel.com>
19806 * lto/lto.c (print_lto_report_1): Declare early.
19807 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19809 2013-04-22 Andi Kleen <ak@linux.intel.com>
19811 * common.opt (-flto-report-wpa): Add.
19812 * doc/invoke.texi (-flto-report-wpa): Add.
19813 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19816 2013-04-22 Xinliang David Li <davidxl@google.com>
19818 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19819 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19820 * Makefile.in: New dependency
19822 David Daney <ddaney.cavm@gmail.com>
19824 * configure.ac (gcc_cv_as_micromips_support): Use the
19825 --fatal-warnings option.
19826 * configure: Regenerate.
19828 2013-04-22 Marek Polacek <polacek@redhat.com>
19831 * tsan.c (instrument_expr): Don't instrument expression
19832 in case its size is zero.
19834 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19838 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19840 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19842 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19844 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19845 (gimple_stmt_iterator): New typedef.
19846 * gimple.h (gimple_stmt_iterator): Rename to...
19847 (gimple_stmt_iterator_d): ... This.
19848 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19849 trees be valid for GIMPLE and GENERIC.
19850 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19851 * gimple-fold.c (gimple_fold_call): Call target hook
19852 gimple_fold_builtin.
19853 * hooks.c (hook_bool_gsiptr_false): New.
19854 * hooks.h (hook_bool_gsiptr_false): New.
19855 * target.def (fold_stmt): New.
19856 * doc/tm.texi: Regenerate.
19858 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19861 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19862 a set sp if no stack realignment.
19864 2013-04-22 Nick Clifton <nickc@redhat.com>
19866 * config.gcc (tilegx-linux): Extend extra_objs rather than
19868 (tilepro-linux): Likewise.
19870 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19872 * config/aarch64/aarch64-builtins.c
19874 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19875 (VAR<1-12>): Add MAP parameter.
19876 (BUILTIN_*): Likewise.
19877 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19878 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19879 (aarch64_ushl_n<mode>): Likewise.
19880 (aarch64_sshr_n<mode>): Likewise.
19881 (aarch64_ushr_n<mode>): Likewise.
19882 (aarch64_<maxmin><mode>): Likewise.
19883 (aarch64_sqrt<mode>): Likewise.
19884 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19885 (vshr<q>_n_*): Likewise.
19887 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19889 * config/aarch64/aarch64-builtins.c
19890 (aarch64_simd_builtin_type_mode): Handle SF types.
19892 (BUILTIN_GPF): Define.
19893 (aarch64_init_simd_builtins): Handle SF types.
19894 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19895 (frecps): Likewise.
19896 (frecpx): Likewise.
19897 * config/aarch64/aarch64-simd.md
19898 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19899 (aarch64_frecpe<mode>): New.
19900 (aarch64_frecps<mode>): Likewise.
19901 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19902 (v8type): Add frecp<esx>.
19903 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19904 (aarch64_frecps<mode>): Likewise.
19905 * config/aarch64/iterators.md (FRECP): New.
19906 (frecp_suffix): Likewise.
19907 * config/aarch64/arm_neon.h
19908 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19910 2013-04-22 Christian Bruel <christian.bruel@st.com>
19913 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19914 (REG_CLASS_NAMES): Idem.
19915 (REG_CLASS_CONTENTS): Idem.
19916 (REGCLASS_HAS_FP_REG): Idem.
19917 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19918 (sh_conditional_register_usage): Idem.
19920 2013-04-21 Jeff Law <law@redhat.com>
19922 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19923 (ssa_forward_propagate_and_combine): Use it.
19925 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19927 * lra.c: Update the flow chart diagram.
19929 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19931 PR rtl-optimization/56847
19932 * lra-constraints.c (process_alt_operands): Discourage alternative
19933 with non-matche doffsettable memory constraint fro memory with
19936 2013-04-19 Richard Biener <rguenther@suse.de>
19938 PR tree-optimization/56982
19939 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19941 * gimplify.c (gimplify_call_expr): Notice special calls.
19942 (gimplify_modify_expr): Likewise.
19943 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19944 abnormal control flow receivers.
19945 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19946 in the same way as cfun->has_nonlocal_labels.
19947 (gimple_purge_dead_abnormal_call_edges): Likewise.
19948 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19949 receivers start a basic-block.
19951 2013-04-19 Richard Biener <rguenther@suse.de>
19953 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19955 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19956 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19957 (SLP_TREE_LOAD_PERMUTATION): Add.
19958 (vect_transform_slp_perm_load): Adjust prototype.
19959 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19960 (vect_free_slp_instance): Likewise.
19961 (vect_create_new_slp_node): Likewise.
19962 (vect_supported_slp_permutation_p): Remove.
19963 (vect_slp_rearrange_stmts): Adjust.
19964 (vect_supported_load_permutation_p): Likewise. Inline
19965 vect_supported_slp_permutation_p here.
19966 (vect_analyze_slp_instance): Compute load permutations per
19967 slp node instead of per instance.
19968 (vect_get_slp_defs): Adjust.
19969 (vect_transform_slp_perm_load): Likewise.
19970 (vect_schedule_slp_instance): Remove redundant code.
19971 (vect_schedule_slp): Remove hack for PR56270, add it ...
19972 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19973 CSE loads for SLP. Adjust.
19975 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19977 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19978 spelling in two comments.
19980 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19983 * config/arm/arm.c (load_multiple_sequence): Require SP
19984 as base register for loads if SP is in the register list.
19986 2013-04-19 Martin Jambor <mjambor@suse.cz>
19988 PR tree-optimization/56718
19989 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19990 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19991 and made public. Adjusted all callers.
19992 (ipa_intraprocedural_devirtualization): New function.
19993 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19994 (ipa_intraprocedural_devirtualization): Likewise.
19995 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19997 2013-04-19 Richard Biener <rguenther@suse.de>
19999 PR tree-optimization/57000
20000 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20002 2013-04-19 Terry Guo <terry.guo@arm.com>
20004 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20006 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20007 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20008 (cortex_m4_fmacs): Use new reservations.
20009 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20011 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20013 PR rtl-optimization/56999
20014 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20016 (lra_coalesce): Remove split_origin_bitmap and related code.
20017 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20018 ranges if necessary.
20020 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20022 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20024 (ix86_expand_call): Remove clobbered_registers array and use
20025 x86_64_ms_sysv_extra_clobbered_registers instead.
20026 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20028 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20030 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20031 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20032 (*call_value_rex64_ms_sysv): Ditto.
20034 2013-04-18 Cary Coutant <ccoutant@google.com>
20036 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20037 parent instead of die_mark.
20039 2013-04-18 Diego Novillo <dnovillo@google.com>
20041 * gimple.c (create_gimple_tmp): New.
20042 (get_expr_type): New.
20043 (build_assign): New.
20044 (build_type_cast): New.
20045 * gimple.h (enum ssa_mode): Define.
20046 (gimple_seq_set_location): New.
20047 * asan.c (build_check_stmt): Change some gimple_build_* calls
20048 to use build_assign and build_type_cast.
20050 2013-04-18 Richard Biener <rguenther@suse.de>
20052 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20053 handle negative step. Remove redundant checks.
20054 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20055 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20056 for negative step and grouped loads fail to vectorize.
20058 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20060 * emit-rtl.c (reset_insn_used_flags): New function.
20061 (reset_all_used_flags): Use it.
20062 (verify_insn_sharing): New function.
20063 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20065 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20067 PR tree-optimization/56984
20068 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20069 and (x >> M) >= N don't register any assertion if N << M is the
20072 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20074 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20075 do not delete_insn insns that have not yet been emitted, only
20076 unlink them with remove_insn.
20077 * df-scan.c (df_insn_delete): Revert r197492.
20079 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20081 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20082 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20084 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20086 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20087 define_insn_and_split.
20088 (and_scc,ior_scc,negscc): Likewise.
20089 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20091 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20093 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20094 can benefit from using a sequence of LDRD instructions in epilogue
20095 instead of a single LDM instruction.
20097 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20100 * doc/extend.texi: Fix typo.
20102 2013-04-17 Richard Biener <rguenther@suse.de>
20104 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20105 (vect_build_slp_tree): ... here.
20106 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20107 match. Remove special-casing of mismatched complex loads.
20108 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20109 re-try the match with swapped commutative operands.
20110 (vect_supported_load_permutation_p): Remove special-casing of
20111 mismatched complex loads.
20112 (vect_analyze_slp_instance): Adjust.
20114 2013-04-17 Richard Biener <rguenther@suse.de>
20116 PR rtl-optimization/56921
20117 * cfgloop.h (struct loop): Add simple_loop_desc member.
20118 (struct niter_desc): Mark with GTY(()).
20119 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20120 * loop-iv.c (get_simple_loop_desc): Likewise.
20121 (free_simple_loop_desc): Likewise.
20124 2013-04-16 Richard Biener <rguenther@suse.de>
20126 PR rtl-optimization/56921
20127 * loop-init.c (pass_rtl_move_loop_invariants): Add
20128 TODO_do_not_ggc_collect to todo_flags_finish.
20129 (pass_rtl_unswitch): Same.
20130 (pass_rtl_unroll_and_peel_loops): Same.
20131 (pass_rtl_doloop): Same.
20133 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20135 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20136 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20137 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20139 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20140 * tree-streamer.c (record_common_node): Adjust reference in comment.
20142 2013-04-17 Terry Guo <terry.guo@arm.com>
20144 * config/arm/cortex-m4.md: Add a new bypass.
20146 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20148 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20150 (*subs_<optab><mode>_multp2): New pattern.
20151 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20152 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20154 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20156 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20157 (*subs_mul_imm_<mode>): New pattern.
20159 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20162 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20163 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20164 (vsx_movti_32bit): Change j->wa to O->wa.
20166 2013-04-16 Richard Biener <rguenther@suse.de>
20168 PR rtl-optimization/56921
20169 * loop-init.c (pass_rtl_move_loop_invariants): Add
20170 TODO_do_not_ggc_collect to todo_flags_finish.
20171 (pass_rtl_unswitch): Same.
20172 (pass_rtl_unroll_and_peel_loops): Same.
20173 (pass_rtl_doloop): Same.
20175 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20177 * config/arm/arm.c (emit_multi_reg_push): New declaration
20178 for an existing function.
20179 (arm_emit_strd_push): New function.
20180 (arm_expand_prologue): Used here.
20181 (arm_emit_ldrd_pop): New function.
20182 (arm_expand_epilogue): Used here.
20183 (arm_get_frame_offsets): Update condition.
20184 (arm_emit_multi_reg_pop): Add a special case for load of a single
20185 register with writeback.
20187 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20189 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20192 2013-04-16 Richard Biener <rguenther@suse.de>
20194 PR tree-optimization/56756
20195 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20196 (first_mem_ref_loc): New.
20197 (execute_sm): Place the load temporarily before a previous
20198 access instead of in the latch edge to ensure its SSA dependencies
20199 are defined at points dominating the load.
20201 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20203 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20204 correct fix by moving header and footer insn to the footer of
20205 the merged basic block. Clear BB_END of the merged-away block.
20207 PR middle-end/43631
20208 * emit-rtl.c (make_note_raw): New function.
20209 (link_insn_into_chain): New static inline function.
20210 (add_insn): Use it.
20211 (add_insn_before, add_insn_after): Factor insn chain linking code...
20212 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20213 using link_insn_into_chain.
20214 (note_outside_basic_block_p): New helper function for emit_note_after
20215 and emit_note_before.
20216 (emit_note_after): Use nobb variant of add_insn_after if the note
20217 should not be contained in a basic block.
20218 (emit_note_before): Use nobb variant of add_insn_before if the note
20219 should not be contained in a basic block.
20220 (emit_note_copy): Use make_note_raw.
20221 (emit_note): Likewise.
20222 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20223 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20224 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20225 the moved barrier the tail of the basic block it follows.
20226 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20228 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20230 PR tree-optimization/56962
20231 * gimple-ssa-strength-reduction.c (record_increment): Only set
20232 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20233 either rhs1 or rhs2 is equal to c->base_expr.
20235 2013-04-15 Richard Biener <rguenther@suse.de>
20237 PR tree-optimization/56933
20238 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20240 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20241 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20242 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20243 dependence check ...
20244 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20246 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20247 GROUP_READ_WRITE_DEPENDENCE.
20249 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20251 * emit-rtl.c (reset_all_used_flags): New function.
20252 (verify_rtl_sharing): Call reset_all_used_flags before and after
20253 performing the checks.
20255 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20257 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20258 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20259 * config/arm/constraints.md (De): New constraint.
20260 * config/arm/neon.md (anddi3_neon): Delete.
20261 (neon_vand<mode>): Expand to standard anddi3 pattern.
20262 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20263 Move earlier in the file.
20264 (neon_inv_logic_op2): Likewise.
20265 (arm_anddi_operand_neon): New predicate.
20267 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20269 * configure.ac (gcc_cv_ld_as_needed): Set
20270 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20271 Use -z ignore, -z record on *-*-solaris2*.
20272 (HAVE_LD_AS_NEEDED): Update comment.
20273 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20274 * configure: Regenerate.
20275 * config.in: Regenerate.
20276 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20277 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20278 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20279 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20280 equivalents. Fix markup.
20281 * doc/tm.texi: Regenerate.
20283 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20285 * config/i386/i386.opt: New option mstack-protector-guard=.
20286 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20287 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20288 TARGET_SSP_TLS_GUARD.
20289 * config/i386/i386.c (ix86_option_override_internal): Set
20290 ix86_stack_protector_guard.
20291 * config/i386/i386.md (stack_protect_set): Enable for
20292 TARGET_SSP_TLS_GUARD only.
20293 (stack_protect_set_<mode>): Ditto.
20294 (stack_protect_test): Ditto.
20295 (stack_protect_test_<mode>): Ditto.
20296 * doc/invoke.texi (i386 Option): Document.
20298 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20301 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20302 (S_MODES): Set H_MODE bit.
20303 (SF_MODES): Set only S_MODE and SF_MODE bits.
20304 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20305 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20306 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20307 <MODE_FLOAT>: Likewise.
20309 2013-04-15 Joey Ye <joey.ye@arm.com>
20311 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20313 2013-04-15 Joey Ye <joey.ye@arm.com>
20315 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20317 (thumb_far_jump_used_p): Count instruction size and set
20320 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20322 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20323 * resource.c (mark_target_live_regs): Reformat conditional block.
20325 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20327 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20328 notes, they are emitted only just before final.
20329 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20331 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20333 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20334 * cfgrtl.c (delete_insn): Call it here instead.
20335 * lra-spills.c (lra_final_code_change): Use delete_insn.
20336 * haifa-sched.c (sched_remove_insn): Likewise.
20337 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20338 returning to the nop pool.
20339 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20340 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20342 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20344 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20345 * doc/tm.texi: Regenerated.
20347 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20349 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20352 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20354 * df-core.c (df_find_def): Compare register numbers.
20355 (df_find_use): Likewise.
20357 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20360 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20361 lra_in_progress for return.
20363 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20365 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20366 define_insn into define_insn_and_split and emit movsicc patterns.
20368 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20370 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20372 2013-04-12 Richard Biener <rguenther@suse.de>
20374 * tree-pass.h (TODO_do_not_ggc_collect): New.
20375 * passes.c (execute_one_ipa_transform_pass): Honor
20376 TODO_do_not_ggc_collect.
20377 (execute_one_pass): Likewise.
20380 2013-04-10 Richard Biener <rguenther@suse.de>
20382 * passes.c (init_optimization_passes): Remove reload pass.
20383 * ira.c (do_reload): Merge into ...
20385 (rest_of_handle_reload): Remove.
20386 (pass_reload): Likewise.
20387 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20388 of reload for vzeroupper pass placement.
20390 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20392 PR tree-optimization/56918
20393 PR tree-optimization/56920
20394 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20395 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20396 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20397 use wide_mul_with_sign method.
20399 2013-04-12 Richard Biener <rguenther@suse.de>
20401 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20402 not be considered a gimple constant.
20404 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20406 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20407 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20410 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20411 Jakub Jelinek <jakub@redhat.com>
20413 * opts.c: Include diagnostic-color.h.
20414 (common_handle_option): Handle OPT_fdiagnostics_color_.
20415 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20416 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20417 (diagnostic-color.o): New.
20418 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20419 (diagnostic_color_rule): New enum.
20420 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20421 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20422 the location string.
20423 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20424 either NULL, or color kind.
20425 * diagnostic-color.c: New file.
20426 * diagnostic-color.h: New file.
20427 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20429 * doc/invoke.texi (-fdiagnostics-color): Document.
20430 * pretty-print.h (pp_show_color): Define.
20431 (struct pretty_print_info): Add show_color field.
20432 * diagnostic.c: Include diagnostic-color.h.
20433 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20434 macros. Colorize error:, warning: etc. strings and also the location
20436 (diagnostic_show_locus): Colorize the caret line.
20437 * pretty-print.c: Include diagnostic-color.h.
20438 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20439 inside of %< %> quotes or quoted through q format modifier.
20441 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20443 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20445 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20447 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20448 code in CC_NZ mode.
20449 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20452 2013-04-11 Marek Polacek <polacek@redhat.com>
20454 PR tree-optimization/48184
20455 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20457 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20459 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20460 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20461 (skip_simple_arithmetic): Tidy up.
20462 * tree.h (skip_simple_constant_arithmetic): Declare.
20464 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20466 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20468 2013-04-11 Richard Biener <rguenther@suse.de>
20470 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20471 generate vector constants.
20473 2013-04-11 Richard Biener <rguenther@suse.de>
20475 PR tree-optimization/56878
20476 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20477 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20479 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20480 Prefer to align the DR with the most invariant base address.
20482 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20484 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20487 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20489 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20490 floating-point vector comparisons against 0.
20492 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20494 PR tree-optimization/56899
20495 * fold-const.c (extract_muldiv_1): Apply distributive law
20496 only if TYPE_OVERFLOW_WRAPS (ctype).
20498 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20501 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20502 loading has side effect.
20504 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20506 * config/sparc/sparc.c: Include tree-pass.h.
20507 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20508 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20509 head of file. Change return type. Split off gate function.
20510 (sparc_gate_work_around_errata): New function.
20511 (pass_work_around_errata): New pass definition.
20512 (insert_pass_work_around_errata) New pass insert definition to
20513 insert pass_work_around_errata just after delayed-branch scheduling.
20514 (sparc_option_override): Insert the pass.
20515 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20517 2013-04-10 David S. Miller <davem@davemloft.net>
20519 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20520 or -mcpu=hypersparc.
20522 * target.def (cstore_mode): New hook.
20523 * target.h: Include insn-codes.h
20524 * targhooks.c: Likewise.
20525 (default_cstore_mode): New function.
20526 * targhooks.h: Declare it.
20527 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20528 * doc/tm.texi: Rebuild.
20529 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20530 target hook, rather than inspecting the insn_data.
20531 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20532 (TARGET_CSTORE_MODE): Redefine.
20533 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20535 * config/sparc/predicates.md (cstore_result_operand): New special
20537 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20538 Use it for operand 0.
20539 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20540 (*snesi_special): Likewise.
20541 (*snesi_zero): Likewise.
20542 (*seqsi_zero): Likewise.
20543 (*sltu_insn): Likewise.
20544 (*sgeu_insn): Likewise.
20545 (*seqdi_special): Make operand 0 and comparison operation be of
20547 (*snedi_special): Likewise.
20548 (*snedi_special_vis3): Likewise.
20549 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20550 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20551 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20552 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20553 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20554 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20555 (*sltu_extend_sp64): Likewise.
20556 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20557 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20558 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20559 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20560 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20562 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20564 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20565 (aarch64_start_file): Use the new function.
20567 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20568 Jason Merrill <jason@redhat.com>
20570 * common.opt: Add -gdwarf.
20571 * opts.c (common_handle_option): Handle it.
20572 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20574 2013-04-10 Richard Biener <rguenther@suse.de>
20576 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20577 (execute_one_ipa_transform_pass): But unconditionally here.
20578 (execute_one_pass): And here.
20579 (init_optimization_passes): Remove reload pass.
20580 * tree-pass.h (TODO_ggc_collect): Remove.
20581 (pass_reload): Likewise.
20582 * ira.c (do_reload): Merge into ...
20584 (rest_of_handle_reload): Remove.
20585 (pass_reload): Likewise.
20586 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20587 of reload for vzeroupper pass placement.
20588 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20589 and todo_flags_finish of all passes.
20591 2013-04-10 Richard Biener <rguenther@suse.de>
20593 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20594 first_const_oprnd field, rename first_def_type to first_op_type.
20595 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20596 (vect_get_and_check_slp_defs): Always use the type of the
20597 operand. Allow mixed vect_external_def, vect_constant_def types.
20598 (vect_get_constant_vectors): Handle mixed vect_external_def,
20599 vect_constant_def types.
20601 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20603 PR tree-optimization/55524
20604 * tree-ssa-math-opts.c
20605 (convert_mult_to_fma): Don't use an fms construct
20606 when we don't have an fms operation, but fnma, and it looks
20607 likely that we'll be able to use the latter.
20609 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20611 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20613 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20614 inline fail caused by overwritable functions.
20616 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20618 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20619 unnecessary bits in the constant power of two case.
20621 2013-04-10 Richard Biener <rguenther@suse.de>
20623 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20624 broken code swapping operands.
20625 (vect_build_slp_tree): Do not compute load permutations here.
20626 (vect_analyze_slp_instance): Compute load permutations here,
20627 after building the SLP tree.
20629 2013-04-09 Christian Bruel <christian.bruel@st.com>
20631 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20632 of next/prev_real_insn.
20634 2013-04-09 Jan Hubicka <jh@suse.cz>
20636 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20637 Drop aliased parameter.
20638 (function_and_variable_visibility): Do not handle alias pairs.
20639 * cgraph.c (varpool_externally_visible_p): Update prototype.
20640 * varpool.c (varpool_add_new_variable): Update.
20642 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20644 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20646 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20648 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20650 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20652 2013-04-09 Marek Polacek <polacek@redhat.com>
20654 PR tree-optimization/48762
20655 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20657 2013-04-09 Richard Biener <rguenther@suse.de>
20659 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20661 (vect_build_slp_tree): Likewise.
20662 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20663 calculating the cost of a SLP instance.
20664 (vect_analyze_slp_instance): Use it from here, after building
20667 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20669 PR middle-end/56883
20670 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20671 expand_omp_for_static_chunk): Use simple_p = true in
20672 force_gimple_operand_gsi calls when assigning to addressable decls.
20674 2013-04-09 Jeff Law <law@redhat.com>
20676 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20677 when the boolean was created by converting a wider object which
20678 had a boolean range.
20680 2013-04-09 Richard Biener <rguenther@suse.de>
20682 * tree-vectorizer.h (slp_void_p): Remove.
20683 (slp_tree): Typedef before _slp_tree declaration.
20684 (struct _slp_tree): Use a vector of slp_tree as children.
20685 (vect_get_place_in_interleaving_chain): Remove.
20686 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20688 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20690 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20691 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20692 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20693 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20694 Use slp_node instead of slp_void_p and adjust.
20696 2013-04-09 Richard Biener <rguenther@suse.de>
20698 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20699 work that is not necessary.
20701 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20703 PR tree-optimization/56854
20704 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20705 forward into clobber stmts if it would change MEM_REF lhs into
20708 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20710 * tree.c (type_hash_lookup, type_hash_add): Make static.
20711 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20713 2013-04-09 Richard Biener <rguenther@suse.de>
20715 * tree.h (unsave_expr_now): Remove.
20716 * tree-inline.c (mark_local_for_remap_r): Remove.
20717 (unsave_expr_1): Likewise.
20718 (unsave_r): Likewise.
20719 (unsave_expr_now): Likewise.
20720 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20721 (propagate_tree_value): Likewise.
20723 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20725 * doc/rtl.texi (sequence): Rewrite documentation to match the
20726 current use of SEQUENCE rtl objects.
20727 * rtl.def (SEQUENCE): Likewise.
20729 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20730 Update documentation.
20731 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20732 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20734 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20736 2013-04-08 Teresa Johnson <tejohnson@google.com>
20738 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20739 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20741 (estimate_edge_size_and_time): Add comment to suggest using rounding
20743 (estimate_node_size_and_time): Ditto.
20744 (remap_edge_change_prob): Use helper rounding divide methods.
20745 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20746 (gimple_mod_pow2_value_transform): Ditto.
20747 (gimple_mod_subtract_transform): Ditto.
20748 (gimple_ic_transform): Ditto.
20749 (gimple_stringops_transform): Ditto.
20750 * stmt.c (conditional_probability): Ditto.
20751 (emit_case_dispatch_table): Ditto.
20752 * lto-cgraph.c (merge_profile_summaries): Ditto.
20753 * tree-optimize.c (execute_fixup_cfg): Ditto.
20754 * cfgcleanup.c (try_forward_edges): Ditto.
20755 * cfgloopmanip.c (scale_loop_profile): Ditto.
20757 (duplicate_loop_to_header_edge): Ditto.
20758 (lv_adjust_loop_entry_edge): Ditto.
20759 * tree-vect-loop.c (vect_transform_loop): Ditto.
20760 * profile.c (compute_branch_probabilities): Ditto.
20761 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20762 * lto-streamer-in.c (input_cfg): Ditto.
20763 * gimple-streamer-in.c (input_bb): Ditto.
20764 * ipa-cp.c (update_profiling_info): Ditto.
20765 (update_specialized_profile): Ditto.
20766 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20767 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20769 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20770 (compute_trg_info): Ditto.
20771 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20772 (purge_dead_edges): Ditto.
20773 * loop-unswitch.c (unswitch_loop): Ditto.
20774 * cgraphclones.c (cgraph_clone_edge): Ditto.
20775 (cgraph_clone_node): Ditto.
20776 * tree-inline.c (copy_bb): Ditto.
20777 (copy_edges_for_bb): Ditto.
20778 (initialize_cfun): Ditto.
20779 (copy_cfg_body): Ditto.
20780 (expand_call_inline): Ditto.
20782 2013-04-08 Kai Tietz <ktietz@redhat.com>
20784 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20785 TARGET_CYGWIN64 by TARGET_64BIT.
20787 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20789 * config/epiphany/epiphany.md (GPR_1): New constant.
20790 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20791 * config/epiphany/epiphany.c (gen_compare_reg):
20792 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20793 is already in place.
20794 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20795 Don't require being called during rtl expansion; If y operlaps r0,
20797 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20798 (epiphany_expand_epilogue): Likewise.
20800 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20801 Don't use CC_FPmode for ORDERED / UNORDERED.
20802 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20804 * config/epiphany/constraints.md (CnL): New constraint.
20805 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20806 * config/epiphany/predicates.md (add_operand): Allow 1024.
20808 * config/epiphany/epiphany.md (logical_op): New code iterator.
20809 (op_mnc): New code attribute.
20810 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20811 (mov_f+1, mov_f+2): New peephole2 patterns.
20813 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20814 (cstoresi4): Also allow re-use of zero result when doing a NE
20815 comparison to a non-zero operand.
20816 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20818 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20819 Use gen_addsi3_i / gen_subsi3_i.
20821 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20825 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20826 contain anything but clobbers, at most one __builtin_stack_restore,
20827 optionally debug stmts and final resx, and if it has at least one
20828 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20829 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20830 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20831 which isn't defaut definition, remove them.
20832 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20833 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20834 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20835 with MEM_REF LHS with SSA_NAME address.
20837 2013-04-08 Jeff Law <law@redhat.com>
20839 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20841 2013-04-08 Richard Biener <rguenther@suse.de>
20843 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20845 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20846 determined vector type.
20847 (vect_analyze_data_refs): Likewise.
20848 (vect_get_new_vect_var): Adjust.
20849 (vect_create_destination_var): Preserve SSA name versions.
20850 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20851 not dump anything here.
20853 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20855 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20856 Add member lr_slot_known.
20857 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20859 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20860 Remove code that sets lr_slot_offset according to what a previous
20861 version of epiphany_emit_save_restore used to do.
20862 (epiphany_emit_save_restore): When doing an lr save or restore,
20863 set/verify lr_slot_known and lr_slot_offset.
20865 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20868 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20871 2013-04-08 Richard Biener <rguenther@suse.de>
20873 * alias.c (find_base_term): Fix thinko in previous change.
20875 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20877 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20878 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20879 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20880 if possible to compute val.
20881 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20882 For QImode integers don't require anything about precision. Use
20883 const_with_all_bytes_same to find out if the constant doesn't have
20884 repeated bytes in it.
20886 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20888 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20891 2013-04-08 Marek Polacek <polacek@redhat.com>
20893 PR rtl-optimization/48182
20894 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20897 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20900 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20901 nuses, make sure we have a label.
20903 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20906 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20907 (rs6000_emit_swdiv_low_precision): Remove.
20908 (rs6000_emit_swdiv): Rewrite to handle between one and four
20909 iterations of Newton-Raphson generally; modify required number of
20910 iterations for some cases.
20911 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20913 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20915 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20916 set-but-unused variable.
20918 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20919 basic blocks of released function bodies garbage-collectable.
20921 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20922 (struct rtl_opt_pass): Add TODO_df_finish.
20924 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20926 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20928 * config/arm/constraints.md (q): New constraint.
20929 * config/arm/ldrdstrd.md: New file.
20930 * config/arm/arm.md (ldrdstrd.md) New include.
20931 (arm_movdi): Use "q" instead of "r" constraint
20932 for double-word memory access.
20933 (movdf_soft_insn): Likewise.
20934 * config/arm/vfp.md (movdi_vfp): Likewise.
20935 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20936 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20937 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20938 (mem_ok_for_ldrd_strd): Likewise.
20939 (output_move_double): Update assertion.
20941 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20943 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20945 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20947 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20948 define_insn_and_split.
20949 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20951 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20953 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20954 define_insn_and_split.
20955 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20956 (shiftsi3_compare): New pattern.
20957 (rrx): New pattern.
20958 * config/arm/unspecs.md (UNSPEC_RRX): New.
20960 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20962 * config/arm/arm.md (negdi_extendsidi): New pattern.
20963 (negdi_zero_extendsidi): Likewise.
20965 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20967 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20968 define_insn_and_split.
20969 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20970 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20972 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20974 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20975 define_insn_and_split.
20976 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20977 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20979 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20981 * config/arm/arm.md (subsi3_carryin): New pattern.
20982 (subsi3_carryin_const): Likewise.
20983 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20984 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20986 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20988 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20990 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20992 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20993 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20995 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20997 * config/arm/arm.c (arm_expand_builtin): Change fcode
20998 type to unsigned int.
21000 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21002 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21004 2013-04-04 Ian Lance Taylor <iant@google.com>
21006 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21009 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21011 PR middle-end/56729
21012 * df-scan.c (df_insn_delete): Disable failing assert.
21014 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21016 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21017 New function prototype.
21018 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21019 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21020 (arm_builtin_vectorized_function): New function.
21022 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21024 * config/arm/arm_neon_builtins.def: New file.
21025 * config/arm/arm.c (neon_builtin_data): Move contents to
21026 arm_neon_builtins.def.
21027 (enum arm_builtins): Include neon builtin definitions.
21028 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21029 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21031 2013-04-04 Marek Polacek <polacek@redhat.com>
21033 PR tree-optimization/48186
21034 * predict.c (maybe_hot_frequency_p): Return false if
21035 HOT_BB_FREQUENCY_FRACTION is 0.
21036 (cgraph_maybe_hot_edge_p): Likewise.
21038 2013-04-04 Richard Biener <rguenther@suse.de>
21040 PR tree-optimization/56826
21041 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21044 2013-04-04 Richard Biener <rguenther@suse.de>
21046 PR tree-optimization/56213
21047 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21048 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21050 2013-04-04 Richard Biener <rguenther@suse.de>
21052 PR tree-optimization/56837
21053 * tree-loop-distribution.c (classify_partition): For non-zero
21054 values require that the value has the same precision as its
21055 mode to be useful as memset value.
21057 2013-04-03 Nick Clifton <nickc@redhat.com>
21059 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21060 (fmssf4): Use fmsf.s on E3V5 architectures.
21061 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21062 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21064 2013-04-03 Jeff Law <law@redhat.com>
21066 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21067 (lra-eliminations.o): Likewise.
21069 2013-04-03 Teresa Johnson <tejohnson@google.com>
21071 * gcov-io.c (compute_working_sets): Moved most of body of old
21072 compute_working_sets here from profile.c.
21073 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21074 (gcov_working_set_t): Moved typedef here from basic-block.h
21075 (compute_working_set): Declare.
21076 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21077 (get_working_sets): Renamed from compute_working_set,
21078 replace most of body with call to new compute_working_sets.
21079 (get_exec_counts): Replace call to compute_working_sets
21080 to get_working_sets.
21081 * profile.h (get_working_sets): Renamed from compute_working_set.
21082 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21083 to get_working_sets.
21084 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21085 * gcov-dump.c (dump_working_sets): New function.
21087 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21089 * hwint.c (sext_hwi, zext_hwi): New functions.
21090 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21091 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21092 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21093 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21094 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21095 (sext_hwi, zext_hwi): New functions.
21097 2013-04-03 Jeff Law <law@redhat.com>
21099 PR tree-optimization/56799
21100 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21101 back test for widening conversion erroneously dropped in prior change.
21103 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21106 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21107 instead of next_real_insn.
21109 2013-04-03 Marek Polacek <polacek@redhat.com>
21112 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21114 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21117 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21119 (thumb1_output_casesi): Likewise.
21120 (thumb2_output_casesi): Likewise.
21122 2013-04-03 Richard Biener <rguenther@suse.de>
21124 PR tree-optimization/56817
21125 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21127 (tree_unroll_loops_completely_1): ... new function to manually
21128 walk the loop tree, properly defering outer loops of unrolled
21129 loops to later iterations.
21131 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21133 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21134 (vectorizable_load): Likewise.
21135 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21136 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21138 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21140 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21143 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21145 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21147 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21149 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21151 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21153 PR tree-optimization/56790
21154 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21157 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21159 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21161 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21162 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21165 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21168 * tree.h (force_folding_builtin_constant_p): New decl.
21169 * builtins.c (force_folding_builtin_constant_p): New variable.
21170 (fold_builtin_constant_p): Fold immediately also if
21171 force_folding_builtin_constant_p.
21173 2013-04-03 Richard Biener <rguenther@suse.de>
21175 PR tree-optimization/56812
21176 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21177 DRs of the same interleaving chain are independent.
21179 2013-04-02 Jason Merrill <jason@redhat.com>
21181 * gdbinit.in (pbb): Use debug fn.
21183 2013-04-02 Lawrence Crowl <crowl@google.com>
21185 * sese.h (struct ivtype_map_elt_s): Remove unused.
21186 (extern debug_ivtype_map): Remove unused.
21187 (extern eq_ivtype_map_elts): Remove unused.
21188 * sese.c (debug_ivtype_map): Removed unused.
21189 (debug_ivtype_map_1): Removed unused.
21190 (debug_ivtype_elt): Remove unused.
21191 (eq_ivtype_map_elts): Remove unused.
21194 2013-04-02 Kai Tietz <ktietz@redhat.com>
21197 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21198 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21199 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21201 (legitimize_pe_coff_symbol): Likewise.
21202 (is_imported_p): New helper-function.
21203 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21204 for Windows x64 targets.
21205 (ix86_expand_prologue): Optimize for pe-coff targets.
21206 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21207 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21208 medium/large code-model.
21209 (legitimize_pic_address): Likewise.
21210 (legitimize_tls_address): Likewise.
21211 (ix86_expand_call): Likewise.
21212 (x86_output_mi_thunk): Likewise.
21213 (get_dllimport_decl): Add new beimport argument.
21214 (construct_plt_address): Don't assert for x64 pe-coff targets.
21215 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21217 (SYMBOL_FLAG_STUBVAR): New macro.
21218 (SYMBOL_REF_STUBVAR_P): Likewise.
21219 * config/i386/winnt.c (stub_list): New structure.
21220 (stub_head): New local variable.
21221 (i386_pe_record_stub): New function.
21222 (i386_pe_file_end): Emit refptr-stubs.
21224 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21226 PR rtl-optimization/56745
21227 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21228 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21231 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21232 and both of them are MEM_REFs, just compare first argument for
21233 equality and attempt to deal even with differing offsets.
21236 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21237 of gimple_clobber_p to be MEM_REF.
21238 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21239 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21240 after gimplification.
21241 * asan.c (get_mem_ref_of_assignment): Don't instrument
21242 gimple_clobber_p stmts.
21243 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21244 gimple_clobber_p stmt if they have MEM_REF lhs and
21245 are dead because of another gimple_clobber_p stmt.
21246 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21247 gimple_clobber_p stmts like debug stmts.
21248 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21249 that refer to unused VAR_DECLs or uninitialized values.
21250 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21251 gimple_clobber_p stmts if they refer to removed parameters.
21252 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21255 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21257 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21258 using SWI48 mode attribute.
21260 2013-04-02 Wei Mi <wmi@google.com>
21262 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21263 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21264 *<rotate_insn><mode>3_mask in i386.md.
21266 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21268 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21270 2013-04-02 Richard Biener <rguenther@suse.de>
21272 PR tree-optimization/56778
21273 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21274 Runtime alias tests are not supported for gather loads.
21275 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21276 stmts referenced from SSA operands before updating SSA form.
21278 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21279 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21281 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21282 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21283 * config/arm/cortex-a53.md: New file.
21284 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21285 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21286 * config/arm/arm.c (arm_issue_rate): Likewise.
21287 * config/arm/arm-tune.md: Regenerate
21288 * config/arm/arm-tables.opt: Regenerate.
21289 * config/arm/arm-cores.def: Add cortex-a53.
21291 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21293 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21296 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21298 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21299 scalar load/store operations using B/H registers.
21300 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21302 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21304 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21306 * config/aarch64/aarch64.c
21307 (aarch64_simd_scalar_immediate_valid_for_move): New.
21308 * config/aarch64/aarch64-protos.h
21309 (aarch64_simd_scalar_immediate_valid_for_move): New.
21310 * config/aarch64/constraints.md (Dh, Dq): New.
21311 * config/aarch64/iterators.md (hq): New.
21313 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21315 * reorg.c (get_branch_condition): Deal with conditional returns.
21316 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21318 2013-04-01 Wei Mi <wmi@google.com>
21320 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21321 Truncate operand 2 using %b asm operand modifier.
21322 (*<shift_insn><mode>3_mask): Ditto.
21323 (*<rotate_insn><mode>3_mask): Ditto.
21325 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21327 PR middle-end/56798
21328 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21330 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21332 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21334 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21336 2013-03-30 Lawrence Crowl <crowl@google.com>
21338 * dse.c (clear_alias_sets): Remove never set.
21339 (disqualified_clear_alias_sets): Remove never set.
21340 (clear_alias_mode_pool): Remove never set.
21341 (dse_step0): Remove condition that is never true.
21342 (canon_address): Remove condition that is never true.
21343 (dse_step7): Remove condition that is never true.
21344 (rest_of_handle_dse): Remove condition that is never true.
21345 (rest_of_handle_dse::did_global): Remove never read from above.
21346 (dse_step2_spill): Remove never called from above.
21347 (dse_step5_spill): Remove never called from above.
21349 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21351 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21352 JUMP_TABLE_DATA changes.
21353 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21354 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21355 (Insns) <jump_table_data>: New entry.
21356 * doc/tm.texi: Regenerate.
21358 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21360 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21361 for table jump at the end of a basic block using tablejump_p.
21362 * targhooks.c (default_invalid_within_doloop): Likewise.
21363 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21364 target hook implementation that is identical to the default hook.
21365 (rs6000_invalid_within_doloop): Remove.
21367 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21368 unused variable from tablejump_p call.
21370 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21371 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21372 (INSN_DELETED_P): Likewise.
21373 (emit_jump_table_data): New prototype.
21374 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21375 after 4th as unused.
21376 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21377 * sched-vis.c (print_insn): Likewise.
21378 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21379 insn for compatibility with back ends that use next_active_insn to
21380 identify jump table data.
21381 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21382 (remove_insn): Likewise.
21383 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21385 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21386 (emit_jump_table_data): New function.
21388 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21389 basic block, a JUMP_TABLE_DATA never is.
21390 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21391 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21392 off from code handling real insns.
21393 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21394 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21395 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21396 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21397 is not a NONDEBUG_INSN_P.
21398 * ira-costs.c (scan_one_insn): Likewise.
21399 * jump.c (mark_all_labels): Likewise.
21400 (mark_jump_label_1): Likewise.
21401 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21402 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21404 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21405 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21406 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21407 JUMP_TABLE_DATA_P insns.
21408 (calculate_elim_costs_all_insns): Likewise.
21409 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21410 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21411 (delete_output_reload): Code style fixups.
21412 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21413 insn flags on this non-insn.
21414 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21415 as scheduling barriers, for pre-change compatibility.
21416 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21417 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21419 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21420 redundant JUMP_TABLE_DATA_P test.
21421 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21422 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21423 (frv_for_each_packet): Likewise.
21424 * config/i386/i386.c (min_insn_size): Likewise.
21425 (ix86_avoid_jump_mispredicts): Likewise.
21426 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21427 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21428 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21429 (mips16_insn_length): Robustify.
21430 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21431 (mips16_split_long_branches): Likewise.
21432 * config/pa/pa.c (pa_combine_instructions): Likewise.
21433 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21434 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21435 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21436 as contributing to pool range lengths.
21437 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21438 Remove redundant JUMP_TABLE_DATA_P test.
21439 (sh_loop_align): Likewise.
21440 (split_branches): Likewise.
21441 (sh_insn_length_adjustment): Likewise.
21442 * config/spu/spu.c (get_branch_target): Likewise.
21444 2013-03-29 Jan Hubicka <jh@suse.cz>
21446 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21447 gcov streaming; stream hot bb threshold to ltrans.
21448 * predict.c (get_hot_bb_threshold): Break out from ....
21449 (maybe_hot_count_p): ... here.
21450 (set_hot_bb_threshold): New function.
21451 * lto-section-in.c (lto_section_name): Add profile.
21452 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21453 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21454 and data-streamer.h
21455 (histogram_entry): New structure.
21456 (histogram, histogram_pool): New global vars.
21457 (histogram_hash): New structure.
21458 (histogram_hash::hash): New method.
21459 (histogram_hash::equal): Likewise.
21460 (account_time_size): New function.
21461 (cmp_counts): New function.
21462 (dump_histogram): New function.
21463 (ipa_profile_generate_summary): New function.
21464 (ipa_profile_write_summary): New function.
21465 (ipa_profile_read_summary): New function.
21466 (ipa_profile): Decide on threshold.
21467 (pass_ipa_profile): Add ipa_profile_write_summary and
21468 ipa_profile_read_summary.
21469 * Makefile.in (ipa.o): Update dependencies.
21470 * lto-streamer.h (LTO_section_ipa_profile): New section.
21472 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21474 * tree.h (VAR_P): New.
21476 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21479 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21481 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21483 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21484 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21485 (control_flow_insn_p): Likewise.
21486 * cfgrtl.c (duplicate_insn_chain): Likewise.
21487 * final.c (get_attr_length_1): Likewise.
21488 (shorten_branches): Likewise.
21489 (final_scan_insn): Likewise.
21490 * function.c (instantiate_virtual_regs): Likewise.
21491 * gcse.c (insert_insn_end_basic_block): Likewise.
21492 * ira-costs.c (scan_one_insn): Likewise.
21493 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21494 * lra.c (check_rtl): Likewise.
21495 * reload1.c (elimination_costs_in_insn): Likewise.
21496 * reorg.c (follow_jumps): Likewise.
21498 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21499 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21500 (thumb_far_jump_used_p): Likewise.
21501 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21502 (workaround_speculation): Likewise.
21503 (add_sched_insns_for_speculation): Likewise.
21504 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21505 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21506 (frv_for_each_packet): Likewise.
21507 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21508 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21509 (final_emit_insn_group_barriers): Likewise.
21510 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21511 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21512 (mips16_insn_length): Likewise.
21513 * config/pa/pa.c (pa_reorg): Likewise.
21514 (pa_combine_instructions): Likewise.
21515 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21516 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21517 (sh_reorg): Likewise.
21518 (split_branches): Likewise.
21519 * config/spu/spu.c (get_branch_target): Likewise.
21521 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21524 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21526 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21527 Fix declaration name.
21529 2013-03-28 Lawrence Crowl <crowl@google.com>
21531 * graphds.h (struct graph.indicies): Remove unused.
21532 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21533 (SCOP_ORIGINAL_PDDRS): Remove unused.
21534 * sese.h (extern insert_loop_close_phis): Removed unused.
21535 (extern insert_guard_phis): Removed unused.
21536 (extern ivtype_map_elt_info): Removed unused.
21537 (new_ivtype_map_elt): Removed unused.
21538 * sese.c (ivtype_map_elt_info): Removed unused.
21540 2013-03-28 Lawrence Crowl <crowl@google.com>
21542 * Makefile.in: Add several missing include dependences.
21544 (test-dump.o): New. This object is not added to any executable,
21545 but is present for ad-hoc testing.
21547 (debug (const bitmap_head_def &)): New.
21548 (debug (const bitmap_head_def *)): New.
21550 (extern debug (const bitmap_head_def &)): New.
21551 (extern debug (const bitmap_head_def *)): New.
21553 (debug (edge_def &)): New.
21554 (debug (edge_def *)): New.
21556 (debug (basic_block_def &)): New.
21557 (debug (basic_block_def *)): New.
21559 (dump_node (const_tree, int, FILE *)): Correct source file.
21561 (debug (die_struct &)): New.
21562 (debug (die_struct *)): New.
21564 (extern debug (die_struct &)): New.
21565 (extern debug (die_struct *)): New.
21566 * gimple-pretty-print.c
21567 (debug (gimple_statement_d &)): New.
21568 (debug (gimple_statement_d *)): New.
21569 * gimple-pretty-print.h
21570 (extern debug (gimple_statement_d &)): New.
21571 (extern debug (gimple_statement_d *)): New.
21573 (debug (ira_allocno_copy &)): New.
21574 (debug (ira_allocno_copy *)): New.
21575 (debug (ira_allocno &)): New.
21576 (debug (ira_allocno *)): New.
21578 (extern debug (ira_allocno_copy &)): New.
21579 (extern debug (ira_allocno_copy *)): New.
21580 (extern debug (ira_allocno &)): New.
21581 (extern debug (ira_allocno *)): New.
21583 (debug (live_range &)): New.
21584 (debug (live_range *)): New.
21586 (debug (lra_live_range &)): New.
21587 (debug (lra_live_range *)): New.
21589 (debug (lra_live_range &)): New.
21590 (debug (lra_live_range *)): New.
21592 (debug (omega_pb_d &)): New.
21593 (debug (omega_pb_d *)): New.
21595 (extern debug (omega_pb_d &)): New.
21596 (extern debug (omega_pb_d *)): New.
21598 (debug (const rtx_def &)): New.
21599 (debug (const rtx_def *)): New.
21601 (debug_tree (tree): Move within file.
21602 (debug_raw (const tree_node &)): New.
21603 (debug_raw (const tree_node *)): New.
21604 (dump_tree_via_hooks (const tree_node *, int)): New.
21605 (debug (const tree_node &)): New.
21606 (debug (const tree_node *)): New.
21607 (debug_verbose (const tree_node &)): New.
21608 (debug_verbose (const tree_node *)): New.
21609 (debug_head (const tree_node &)): New.
21610 (debug_head (const tree_node *)): New.
21611 (debug_body (const tree_node &)): New.
21612 (debug_body (const tree_node *)): New.
21613 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21614 (debug (vec<tree, va_gc> &)): New.
21615 (debug (vec<tree, va_gc> *)): New.
21617 (extern debug (const rtx_def &)): New.
21618 (extern debug (const rtx_def *)): New.
21620 (debug_raw (simple_bitmap_def &)): New.
21621 (debug_raw (simple_bitmap_def *)): New.
21622 (debug (simple_bitmap_def &)): New.
21623 (debug (simple_bitmap_def *)): New.
21625 (extern debug (simple_bitmap_def &)): New.
21626 (extern debug (simple_bitmap_def *)): New.
21627 (extern debug_raw (simple_bitmap_def &)): New.
21628 (extern debug_raw (simple_bitmap_def *)): New.
21630 (debug (vinsn_def &)): New.
21631 (debug (vinsn_def *)): New.
21632 (debug_verbose (vinsn_def &)): New.
21633 (debug_verbose (vinsn_def *)): New.
21634 (debug (expr_def &)): New.
21635 (debug (expr_def *)): New.
21636 (debug_verbose (expr_def &)): New.
21637 (debug_verbose (expr_def *)): New.
21638 (debug (vec<rtx> &)): New.
21639 (debug (vec<rtx> *)): New.
21641 (extern debug (vinsn_def &)): New.
21642 (extern debug (vinsn_def *)): New.
21643 (extern debug_verbose (vinsn_def &)): New.
21644 (extern debug_verbose (vinsn_def *)): New.
21645 (extern debug (expr_def &)): New.
21646 (extern debug (expr_def *)): New.
21647 (extern debug_verbose (expr_def &)): New.
21648 (extern debug_verbose (expr_def *)): New.
21649 (extern debug (vec<rtx> &)): New.
21650 (extern debug (vec<rtx> *)): New.
21652 (_list_iter_cond_expr): Make inline instead of static.
21654 (debug (sreal &)): New.
21655 (debug (sreal *)): New.
21657 (extern debug (sreal &)): New.
21658 (extern debug (sreal *)): New.
21660 (extern debug_raw (const tree_node &)): New.
21661 (extern debug_raw (const tree_node *)): New.
21662 (extern debug (const tree_node &)): New.
21663 (extern debug (const tree_node *)): New.
21664 (extern debug_verbose (const tree_node &)): New.
21665 (extern debug_verbose (const tree_node *)): New.
21666 (extern debug_head (const tree_node &)): New.
21667 (extern debug_head (const tree_node *)): New.
21668 (extern debug_body (const tree_node &)): New.
21669 (extern debug_body (const tree_node *)): New.
21670 (extern debug (vec<tree, va_gc> &)): New.
21671 (extern debug (vec<tree, va_gc> *)): New.
21673 (debug (struct loop &)): New.
21674 (debug (struct loop *)): New.
21675 (debug_verbose (struct loop &)): New.
21676 (debug_verbose (struct loop *)): New.
21677 * tree-dump.c: Add header dependence.
21679 (extern debug (struct loop &)): New.
21680 (extern debug (struct loop *)): New.
21681 (extern debug_verbose (struct loop &)): New.
21682 (extern debug_verbose (struct loop *)): New.
21684 (debug (data_reference &)): New.
21685 (debug (data_reference *)): New.
21686 (debug (vec<data_reference_p> &)): New.
21687 (debug (vec<data_reference_p> *)): New.
21688 (debug (vec<ddr_p> &)): New.
21689 (debug (vec<ddr_p> *)): New.
21691 (extern debug (data_reference &)): New.
21692 (extern debug (data_reference *)): New.
21693 (extern debug (vec<data_reference_p> &)): New.
21694 (extern debug (vec<data_reference_p> *)): New.
21695 (extern debug (vec<ddr_p> &)): New.
21696 (extern debug (vec<ddr_p> *)): New.
21698 (debug (pt_solution &)): New.
21699 (debug (pt_solution *)): New.
21701 (extern debug (pt_solution &)): New.
21702 (extern debug (pt_solution *)): New.
21704 (debug (_var_map &)): New.
21705 (debug (_var_map *)): New.
21706 (debug (tree_live_info_d &)): New.
21707 (debug (tree_live_info_d *)): New.
21709 (extern debug (_var_map &)): New.
21710 (extern debug (_var_map *)): New.
21711 (extern debug (tree_live_info_d &)): New.
21712 (extern debug (tree_live_info_d *)): New.
21714 2013-03-28 Jan Hubicka <jh@suse.cz>
21716 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21718 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21720 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21721 record only when desired or required.
21723 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21725 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21726 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21728 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21730 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21731 (*andsi3_compare0_uxtw): New pattern.
21732 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21733 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21735 2013-03-28 Jan Hubicka <jh@suse.cz>
21737 * data-streamer-in.c (streamer_read_gcov_count): New function.
21738 * gimple-streamer-out.c: Include value-prof.h.
21739 (output_gimple_stmt): Output histogram.
21740 (output_bb): Use streamer_write_gcov_count.
21741 * value-prof.c: Include data-streamer.h
21742 (dump_histogram_value): Add HIST_TYPE_MAX.
21743 (stream_out_histogram_value): New function.
21744 (stream_in_histogram_value): New function.
21745 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21746 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21747 * data-streamer-out.c (streamer_write_gcov_count): New function.
21748 (streamer_write_gcov_count_stream): New function.
21749 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21750 (lto_output_node): Likewise.
21751 (input_node, input_edge): Likewise.
21752 * lto-streamer-out.c (output_cfg): Update streaming.
21753 * lto-streamer-in.c (input_cfg): Likewise.
21754 * data-streamer.h (streamer_write_gcov_count,
21755 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21756 * gimple-streamer-in.c: Include value-prof.h
21757 (input_gimple_stmt): Input histograms.
21758 (input_bb): Update profile streaming.
21760 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21762 * genmodes.c (emit_max_int): New function.
21763 (emit_insn_modes_h): Added call to emit_max_function.
21764 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21766 * machmode.def: Fixed comment.
21768 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21770 * combine.c (try_combine): Removed useless assert.
21771 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21773 2013-03-28 Marek Polacek <polacek@redhat.com>
21774 Richard Biener <rguenther@suse.de>
21776 PR tree-optimization/56695
21777 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21778 build signed result of a vector comparison.
21779 * tree-cfg.c (verify_gimple_comparison): Check that a result
21780 of a vector comparison has signed type.
21782 2013-03-28 Richard Biener <rguenther@suse.de>
21784 PR tree-optimization/37021
21785 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21786 do not restrict gaps between groups.
21787 * tree-vect-stmts.c (vectorizable_load): Properly account for
21788 a gap between groups.
21790 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21792 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21793 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21796 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21798 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21799 * doc/extend.texi (Named Address Spaces): Ditto.
21800 (Variable Attributes): Ditto.
21802 2013-03-27 Kai Tietz <ktietz@redhat.com>
21804 * config.build: Add support for cygwin x64 target.
21805 * config.gcc: Likewise.
21806 * config.host: Likewise.
21807 * configure.ac: Likewise
21808 * configure: Regenerated.
21810 2013-03-27 Kai Tietz <ktietz@redhat.com>
21812 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21813 * config/i386/t-cygwin-w64: New file.
21814 * config/i386/cygwin-w64.h: New file.
21815 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21816 and add support for x64-cygwin target.
21817 (CPP_SPEC): Likewise.
21818 (CXX_WRAP_SPEC_LIST): Undefine before define.
21819 (LIBGCJ_SONAME): Use 15 as version.
21821 2013-03-27 Richard Biener <rguenther@suse.de>
21823 PR tree-optimization/56716
21824 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21825 dumping for ref nodes.
21827 2013-03-27 Martin Jambor <mjambor@suse.cz>
21829 PR tree-optimization/55334
21830 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21831 restricted pointers to arrays.
21833 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21835 * Makefile.in (.SUFFIXES): Add .cc.
21836 (.c.o): Apply same recipe for implicit rule .cc.o.
21838 2013-03-27 Richard Biener <rguenther@suse.de>
21840 PR tree-optimization/37021
21841 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21842 REALPART/IMAGPART_EXPRs around the supported refs.
21843 * tree-ssa-structalias.c (find_func_aliases): Assume that
21844 floating-point values are not used to transfer pointers.
21846 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21848 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21849 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21850 * doc/tm.texi: Regenerate.
21851 * targhooks.h (default_has_ifunc_p): New.
21852 * targhooks.c (default_has_ifunc_p): Ditto.
21853 * config/linux-protos.h: New file.
21854 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21855 hook for linux which disables support of indirect functions in android.
21856 * config/linux-android.c: New file.
21857 * config/t-linux-android.c: Ditto.
21858 * config.gcc: Added new object file linux-android.o.
21859 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21860 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21861 * varasm.c (do_assemble_alias): Likewise.
21862 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21863 doesn't support indirect functions.
21864 * configure: Regenerate.
21866 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21869 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21870 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21873 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21875 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21877 2013-03-27 Terry Guo <terry.guo@arm.com>
21879 * config/arm/arm-cores.def: Added core cortex-r7.
21880 * config/arm/arm-tune.md: Regenerated.
21881 * config/arm/arm-tables.opt: Regenerated.
21882 * doc/invoke.texi: Added entry for core cortex-r7.
21884 2013-03-27 Walter Lee <walt@tilera.com>
21886 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21887 double-decrement of next_scratch_regno.
21889 2013-03-27 Walter Lee <walt@tilera.com>
21891 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21893 (insn_v1mulus): Ditto.
21894 (insn_v2muls): Ditto.
21896 2013-03-27 Walter Lee <walt@tilera.com>
21898 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21899 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21901 2013-03-27 Walter Lee <walt@tilera.com>
21903 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21904 (*sibcall_value): Ditto.
21906 2013-03-27 Walter Lee <walt@tilera.com>
21908 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21909 (insn_mnz_v8qi): ... this ...
21910 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21912 (insn_v<n>mnz): Replaced by ...
21913 (insn_v1mnz): ... this ...
21914 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21916 (insn_mz_<mode>): Replaced by ...
21917 (insn_mz_v8qi): ... this ...
21918 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21920 (insn_v<n>mz): Replaced by ...
21921 (insn_v1mz): ... this ...
21922 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21925 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21927 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21929 2013-03-26 Roland McGrath <mcgrathr@google.com>
21931 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21932 than fprintf with a non-constant, non-format string.
21934 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21936 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21937 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21938 operand 0 predicate.
21939 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21940 attribute. Use general_x64nomem_operand as operand 1 predicate.
21941 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21942 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21943 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21944 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21945 (mov<mode>_insv_1): Remove expander. Merge insn with
21946 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21947 Use general_x64nomem_operand as operand 1 predicate.
21948 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21949 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21950 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21951 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21952 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21953 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21954 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21955 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21956 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21957 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21958 (general_x64nomem_operand): Ditto.
21960 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21962 * config/rtems.opt: Add -pthread option.
21964 2013-03-26 Richard Biener <rguenther@suse.de>
21966 * alias.c (find_base_term): Avoid redundant and not used recursion.
21967 (base_alias_check): Get the initial base term from the caller.
21968 (true_dependence_1): Compute and pass base terms to base_alias_check.
21969 (write_dependence_p): Likewise.
21970 (may_alias_p): Likewise.
21972 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21974 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21975 PC-relative load in SI modes and above only.
21977 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21979 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21980 * config/arm/iwmmxt.md (WCGR0): Update.
21981 (WCGR1, WCGR2, WCGR3): Likewise.
21983 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21985 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21986 Use x64 and nox64 isa attributes.
21988 2013-03-26 Richard Biener <rguenther@suse.de>
21990 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21991 alignment computations and rely on get_object_alignment_1
21992 for the !TYPE_P case.
21993 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21995 2013-03-26 Walter Lee <walt@tilera.com>
21997 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21998 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22000 2013-03-25 Jeff Law <law@redhat.com>
22002 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22003 check for INTEGRAL_TYPE_P that was missing due to checking in
22004 wrong version of prior patch.
22006 2013-03-25 Walter Lee <walt@tilera.com>
22008 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22009 TILEGX_INSN_SHUFFLEBYTES1.
22010 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22012 (tilegx_builtins): Ditto.
22013 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22015 2013-03-25 Walter Lee <walt@tilera.com>
22017 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22018 (floatunssisf2): New pattern.
22019 (floatsidf2): New pattern.
22020 (floatunssidf2): New pattern.
22022 2013-03-25 Walter Lee <walt@tilera.com>
22024 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22025 tests for constraint J, K, N, P.
22027 2013-03-25 Walter Lee <walt@tilera.com>
22029 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22030 Use indirect/pcrel encoding.
22031 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22034 2013-03-25 Steve Ellcey <sellcey@mips.com>
22036 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22037 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22038 * config/mips/mips.c (mips_option_override): Set IMADD default.
22039 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22040 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22041 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22042 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22043 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22045 2013-03-25 Jeff Law <law@redhat.com>
22047 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22048 slightly to avoid creating and folding useless trees. Simplify
22049 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22051 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22053 * config/i386/i386.md (*zero_extendsidi2): Merge with
22054 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22055 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22056 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22057 targets. Clarify comment.
22059 2013-03-25 Martin Jambor <mjambor@suse.cz>
22061 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22062 pass-through jump functions differently.
22063 (ipa_read_jump_function): Likewise. Also use setter functions to set
22066 2013-03-25 Martin Jambor <mjambor@suse.cz>
22068 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22069 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22071 (ipa_get_indirect_edge_target): New function.
22072 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22073 ipa_get_indirect_edge_target. Update all callers.
22074 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22075 ipa_get_indirect_edge_target_1 instead of calling
22076 ipa_get_indirect_edge_target.
22077 (create_specialized_node): Pass aggvlas to
22078 ipcp_discover_new_direct_edges.
22080 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22082 * config/arm/arm.md (f_sels, f_seld): New types.
22083 (*cmov<mode>): New pattern.
22084 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22087 2013-03-25 Kai Tietz <ktietz@redhat.com>
22089 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22090 POSIX-printf for mingw-hosted builds.
22092 2013-03-25 Richard Biener <rguenther@suse.de>
22094 PR middle-end/56694
22095 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22096 must-not-throw stmt location.
22098 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22100 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22101 Emit load-acquire versions when acq is true.
22102 (arm_emit_store_exclusive): Add rel parameter.
22103 Emit store-release versions when rel is true.
22104 (arm_split_compare_and_swap): Use acquire-release instructions
22106 of barriers when appropriate.
22107 (arm_split_atomic_op): Likewise.
22108 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22109 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22110 (VUNSPEC_SLX): Likewise.
22111 (VUNSPEC_LDA): Likewise.
22112 (VUNSPEC_STL): Likewise.
22113 * config/arm/sync.md (atomic_load<mode>): New pattern.
22114 (atomic_store<mode>): Likewise.
22115 (arm_load_acquire_exclusive<mode>): Likewise.
22116 (arm_load_acquire_exclusivesi): Likewise.
22117 (arm_load_acquire_exclusivedi): Likewise.
22118 (arm_store_release_exclusive<mode>): Likewise.
22120 2013-03-25 Catherine Moore <clm@codesourcery.com>
22122 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22123 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22124 * config/mip/predicates.md (lwsp_swsp_operand,
22125 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22126 sb16_operand, db4_operand, db7_operand, ib3_operand,
22127 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22128 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22129 andi16_operand): New predicates.
22130 * config/mips/mips.md (compression): New attribute.
22131 (enabled): New attribute.
22132 (length): Consider compression in computing length.
22133 (shift_compression): New code attribute.
22134 (*add<mode>3): New operands. Record compression.
22135 (sub<mode>3): Likewise.
22136 (one_cmpl<mode>2): Likewise.
22137 (*and<mode>3): Likewise.
22138 (*ior<mode>3): Likewise.
22139 (unnamed pattern for xor): Likewise.
22140 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22141 (*<optab><mode>3): Likewise.
22142 (*mov<mode>_internal: Likewise.
22143 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22144 (mips_unsigned_immediate_p): New.
22145 (umips_lwsp_swsp_address_p): New.
22146 (m16_based_address_p): New.
22147 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22148 (mips_unsigned_immediate_p): New prototype.
22149 (lwsp_swsp_address_p): New prototype.
22150 (m16_based_address_p): New prototype.
22151 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22152 (mips_signed_immediate_p): New function.
22153 (m16_based_address_p): New function.
22154 (lwsp_swsp_address_p): New function.
22155 (mips_print_operand_punctuation): Recognize short delay slot insns
22156 for microMIPS.add<mode>3"
22158 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22161 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22162 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22164 2013-03-25 Richard Biener <rguenther@suse.de>
22166 PR tree-optimization/56689
22167 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22170 2013-03-25 Richard Biener <rguenther@suse.de>
22172 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22174 (memory_references): Likewise.
22175 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22176 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22177 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22178 (gather_mem_refs_in_loops): Fold into ...
22179 (analyze_memory_references): ... this. Move initialization
22180 to tree_ssa_lim_initialize.
22181 (fill_always_executed_in): Rename to ...
22182 (fill_always_executed_in_1): ... this.
22183 (fill_always_executed_in): Move contains_call computation to
22184 this new function from ...
22185 (tree_ssa_lim_initialize): ... here.
22186 (tree_ssa_lim): Call fill_always_executed_in.
22188 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22190 * postreload.c (reload_combine): Fix code detecting returns.
22192 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22194 * function.c (emit_use_return_register_into_block): On cc0 targets,
22195 do not emit the sequence between cc0 setter and user.
22197 2013-03-25 Kai Tietz <ktietz@redhat.com>
22199 * config/i386/predicates.md (local_symbolic_operand): Interpret
22200 dll-imported symbols as none-local.
22202 2013-03-25 Richard Biener <rguenther@suse.de>
22204 * tree-ssa-loop-im.c (struct depend): Remove.
22205 (struct lim_aux_data): Make depends a vec of gimples.
22206 (free_lim_aux_data): Adjust.
22207 (add_dependency): Likewise.
22208 (set_level): Likewise.
22210 2013-03-25 Richard Biener <rguenther@suse.de>
22212 PR middle-end/56434
22213 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22214 the pointer returned by calls with ECF_MALLOC set.
22216 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22218 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22220 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22222 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22223 using MMXMODE mode iterator.
22224 (*move<mode>_internal): Merge with *movv2sf_internal and
22225 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22227 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22229 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22230 (record_last_mem_set_info): Likewise.
22232 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22233 of XNEWVEC followed by memset.
22234 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22236 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22238 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22239 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22240 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22241 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22242 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22243 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22244 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22245 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22246 BARRIER_P instead of GET_CODE.
22248 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22250 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22251 inaccuracy in the probing code.
22253 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22254 (ctrapdi4): Likewise.
22256 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22258 * calls.c (expand_call): Add missing guard to code handling return
22259 of non-BLKmode structures in MSB.
22260 * function.c (expand_function_end): Likewise.
22262 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22264 * combine.c (try_combine): Adjust comment. Do not add the set of
22265 insn #0 if the destination indirectly is set or dies in insn #2.
22266 Tidy up code to distribute a new note.
22268 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22270 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22271 also for alternatives 16 and 17.
22273 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22275 * config/i386/sse.md (*mov<mode>_internal): Merge with
22276 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22277 Emit insn template depending on type attribute. Use
22278 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22279 movd instead of movq mnemonic for interunit moves. Rewrite mode
22280 attribute calculation. Remove unit attribute calculation.
22281 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22282 Set prefix_data16 attribute for DImode ssemov types.
22283 Use Ym instead of y for SSE-MMX conversion alternatives.
22284 Reorder operand constraints.
22286 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22288 * df.h (df_insn_delete): Adjust prototype.
22289 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22290 and let it decide whether mark the basic block dirty.
22291 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22292 * df-scan.c (df_insn_info_delete): New helper function, split
22293 off from df_insn_delete.
22294 (df_scan_free_bb_info): Use it.
22295 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22297 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22298 that the insn is actually an insn and it has a non-NULL basic block.
22299 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22301 2013-03-22 Richard Biener <rguenther@suse.de>
22303 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22305 (mem_ref_alloc): Do not allocate them.
22306 (refs_independent_p): Do not query or maintain a cache.
22308 2013-03-22 Richard Biener <rguenther@suse.de>
22310 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22311 (gather_mem_refs_in_loops): Do not compute it.
22312 (analyze_memory_references): Do not allocate it.
22313 (tree_ssa_lim_finalize): Do not free it.
22314 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22316 2013-03-22 Richard Biener <rguenther@suse.de>
22318 * is-a.h (as_a): Use gcc_checking_assert.
22320 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22322 * config/aarch64/aarch64.c (aarch64_print_operand): New
22323 format specifier for printing a constant in hex.
22324 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22325 format specifier for printing second operand.
22327 2013-03-22 Richard Biener <rguenther@suse.de>
22329 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22331 (gather_mem_refs_in_loops): Perform store accumulation here.
22332 (create_vop_ref_mapping_loop): Remove.
22333 (create_vop_ref_mapping): Likewise.
22334 (analyze_memory_references): Initialize refs_stored_in_loop.
22335 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22336 (record_indep_loop): Remove.
22337 (record_dep_loop): New function.
22338 (ref_indep_loop_p_1): Adjust to only walk over references
22339 in the loop, not its subloops.
22340 (ref_indep_loop_p): Rename to ...
22341 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22342 maintaining a more fine-grained cache.
22343 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22344 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22346 2013-03-22 Richard Biener <rguenther@suse.de>
22348 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22349 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22350 aggregate mem_ref_loc.
22351 (free_mem_ref_locs): Inline into ...
22352 (memref_free): ... this and adjust.
22353 (mem_ref_alloc): Adjust.
22354 (mem_ref_locs_alloc): Remove.
22355 (record_mem_ref_loc): Adjust.
22356 (get_all_locs_in_loop): Rewrite into ...
22357 (for_all_locs_in_loop): ... this iterator.
22358 (rewrite_mem_ref_loc): New functor.
22359 (rewrite_mem_refs): Use for_all_locs_in_loop.
22360 (sm_set_flag_if_changed): New functor.
22361 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22362 (ref_always_accessed): New functor.
22363 (ref_always_accessed_p): Use for_all_locs_in_loop.
22365 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22367 * tree-pass.h (PROP_gimple_lvec): New.
22368 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22369 (init_optimization_passes): Move pass_lower_vector.
22370 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22372 (pass_lower_vector): Provide PROP_gimple_lvec.
22373 (pass_lower_vector_ssa): Likewise.
22374 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22376 2013-03-21 Mark Wielaard <mjw@redhat.com>
22378 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22380 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22382 * config/i386/i386.md (*movdi_internal): Disparage slightly
22383 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22384 conversion alternatives.
22386 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22388 PR middle-end/48087
22389 * diagnostic.def (DK_WERROR): New kind.
22390 * diagnostic.h (werrorcount): Define.
22391 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22392 promoted to DK_ERROR, increment DK_WERROR counter instead of
22394 * toplev.c (toplev_main): Call print_ignored_options even if
22395 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22396 even if just werrorcount is non-zero.
22399 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22401 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22402 (string_cst_pool_decl): New function.
22403 (optimize_one_addr_into_implicit_ptr): New function.
22404 (resolve_addr_in_expr): Optimize DWARF location expression
22405 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22406 which doesn't live in memory, but has DW_AT_location or
22407 DW_AT_const_value, or refers to a string literal, into
22408 DW_OP_GNU_implicit_pointer.
22409 (optimize_location_into_implicit_ptr): New function.
22410 (resolve_addr): If removing DW_AT_location of a variable because
22411 it was DW_OP_addr of address of the variable, but the variable doesn't
22412 live in memory, try to emit const value attribute for the initializer.
22414 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22416 * tree.h (VECTOR_TYPE_P): New macro.
22417 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22418 TYPE_MODE): Use it.
22419 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22420 VEC_COND_EXPR cannot be lvalues.
22421 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22423 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22425 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22426 Restrict the transformation to equal modes.
22428 2013-03-21 Richard Biener <rguenther@suse.de>
22430 PR tree-optimization/39326
22431 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22432 (MEM_ANALYZABLE): Adjust.
22433 (record_mem_ref_loc): Move bitmap ops ...
22434 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22435 unanalyzable refs, do not record locations for it.
22436 (analyze_memory_references): Allocate ref zero as shared
22438 (refs_independent_p): Do not test for unanalyzed mems here.
22439 (ref_indep_loop_p_1): Special-case disambiguation against
22440 the unanalyzed ref.
22441 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22443 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22445 * config/arm/arm-protos.h (tune_params): Add
22446 prefer_neon_for_64bits field.
22447 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22448 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22449 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22450 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22451 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22452 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22453 (arm_option_override): Handle -mneon-for-64bits new option.
22454 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22455 (prefer_neon_for_64bits): Declare new variable.
22456 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22457 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22458 (arch_enabled): Handle new arch types. Remove support for onlya8
22460 (one_cmpldi2): Use new arch names.
22461 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22462 * config/arm/arm.opt (mneon-for-64bits): Add option.
22463 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22464 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22465 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22467 * doc/invoke.texi (-mneon-for-64bits): Document.
22469 2013-03-21 Richard Biener <rguenther@suse.de>
22471 PR tree-optimization/39326
22472 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22473 (sort_bbs_in_loop_postorder_cmp): New function.
22474 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22476 2013-03-21 Richard Biener <rguenther@suse.de>
22478 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22479 (vect_insert_into_interleaving_chain): Likewise.
22480 (vect_drs_dependent_in_basic_block): Inline ...
22481 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22483 (vect_analyze_data_ref_dependence): ... here. Simplify.
22484 (vect_check_interleaving): Simplify.
22485 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22486 (vect_slp_analyze_data_ref_dependences): ... this new function.
22487 (dr_group_sort_cmp): New function.
22488 (vect_analyze_data_ref_accesses): Compute data-reference groups
22489 here instead of in vect_analyze_data_ref_dependence. Use
22490 a more efficient algorithm.
22491 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22492 vect_slp_analyze_data_ref_dependences. Call
22493 vect_analyze_data_ref_accesses earlier.
22494 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22495 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22496 (vect_slp_analyze_data_ref_dependences): New prototype.
22498 2013-03-21 Richard Biener <rguenther@suse.de>
22500 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22501 ref is stored in the loop.
22502 (find_refs_for_sm): Walk only over all stores.
22503 (store_motion_loop): Allocate from lim_bitmap_obstack.
22504 (store_motion): Likewise.
22506 2013-03-21 Richard Biener <rguenther@suse.de>
22508 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22509 Update virtual SSA form.
22511 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22513 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22514 * configure: Regenerate.
22515 * config.in: Regenerate.
22516 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22517 if !HAVE_LD_EH_FRAME_CIEV3.
22519 2013-03-21 Richard Biener <rguenther@suse.de>
22521 * tree-cfg.c (verify_expr_no_block): New function.
22522 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22523 nor DECL_VALUE_EXPR have locations with associated blocks.
22524 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22525 (clear_unused_block_pointer): Remove code dealing with
22526 blocks in DECL_DEBUG_EXPR locations.
22528 2013-03-21 Richard Biener <rguenther@suse.de>
22530 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22531 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22532 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22533 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22534 instead of DECL_DEBUG_EXPR_IS_FROM.
22535 * gimplify.c (gimplify_modify_expr): Likewise.
22536 * tree-cfg.c (verify_expr_location_1): Likewise.
22537 * tree-complex.c (create_one_component_var): Likewise.
22538 * tree-sra.c (create_access_replacement): Likewise.
22539 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22540 (clear_unused_block_pointer): Likewise.
22541 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22542 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22543 * var-tracking.c (var_debug_decl): Likewise.
22544 (track_expr_p): Likewise.
22545 * tree-inline.c (add_local_variables): Likewise. Set
22546 DECL_HAS_DEBUG_EXPR_P after copying it.
22547 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22548 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22550 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22553 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22554 * configure: Regenerate.
22555 * config.in: Regenerate.
22556 * config/i386/i386.md (*movdf_internal): Use
22557 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22558 movd instead of movq mnemonic for interunit moves.
22559 (*movdi_internal): Ditto.
22561 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22563 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22564 (abd<mode>_3): New pattern.
22565 (aba<mode>_3): New pattern.
22566 (fabd<mode>_3): New pattern.
22568 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22570 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22571 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22572 occurrence of REGISTER_PREFIX as its empty string.
22574 2013-03-20 Jeff Law <law@redhat.com>
22576 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22577 addititional equivalences for equality comparisons between an SSA_NAME
22578 and a constant where the SSA_NAME was set from a widening conversion.
22580 2013-03-20 Walter Lee <walt@tilera.com>
22582 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22584 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22586 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22587 depending on type attribute.
22588 (*movti_internal): Ditto.
22589 (*movtf_internal): Ditto.
22590 (*movxf_internal): Ditto.
22591 (*movdf_internal): Ditto.
22592 (*movsf_internal): Ditto.
22594 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22596 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22597 maybe_vex for sselog1 and ssemov types.
22598 (*movdi_internal): Reorder operand constraints.
22599 (*movsi_internal): Ditto. Set prefix attribute to
22600 maybe_vex for sselog1 and ssemov types.
22601 (*movtf_internal): Set prefix attribute to maybe_vex
22602 for sselog1 and ssemov types.
22603 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22604 DImode ssemov types. Reorder operand constraints.
22605 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22606 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22607 attribute for SImode ssemov types. Reorder operand constraints.
22609 2013-03-20 Martin Jambor <mjambor@suse.cz>
22611 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22612 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22614 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22616 * config/rs6000/predicates.md (indexed_address, update_address_mem
22617 update_indexed_address_mem): New predicates.
22618 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22619 attribute for load/store instructions.
22620 * config/rs6000/dfp.md (movsd_store): Likewise.
22621 (movsd_load): Likewise.
22622 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22623 (unnamed HI->DI extend define_insn): Likewise.
22624 (unnamed SI->DI extend define_insn): Likewise.
22625 (unnamed QI->SI extend define_insn): Likewise.
22626 (unnamed QI->HI extend define_insn): Likewise.
22627 (unnamed HI->SI extend define_insn): Likewise.
22628 (unnamed HI->SI extend define_insn): Likewise.
22629 (extendsfdf2_fpr): Likewise.
22630 (movsi_internal1): Likewise.
22631 (movsi_internal1_single): Likewise.
22632 (movhi_internal): Likewise.
22633 (movqi_internal): Likewise.
22634 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22635 attribute for load/store instructions.
22636 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22638 (mov<mode>_softfloat): Likewise.
22639 (mov<mode>_hardfloat32): Likewise.
22640 (mov<mode>_hardfloat64): Likewise.
22641 (mov<mode>_softfloat64): Likewise.
22642 (movdi_internal32): Likewise.
22643 (movdi_internal64): Likewise.
22644 (probe_stack_<mode>): Likewise.
22646 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22648 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22649 floating point, and decimal floating point to reload iterator.
22651 * config/rs6000/constraints.md (wl constraint): New constraints to
22652 return FLOAT_REGS if certain options are used to reduce the number
22653 of separate patterns that exist in the file.
22654 (wx constraint): Likewise.
22655 (wz constraint): Likewise.
22657 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22658 -mdebug=reg, print wg, wl, wx, and wz constraints.
22659 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22660 Initialize the reload functions for 64-bit binary/decimal floating
22662 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22663 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22664 create the buffer on the stack to overcome not having a 32-bit
22666 (rs6000_emit_move): Likewise.
22667 (rs6000_secondary_memory_needed_rtx): Likewise.
22668 (rs6000_alloc_sdmode_stack_slot): Likewise.
22669 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22670 via xxlxor, just like DFmode 0.0.
22672 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22673 define as 1 if we are running on a power7 or newer.
22674 (enum r6000_reg_class_enum): Add new constraints.
22676 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22677 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22678 with other moves by using conditional constraits (wg). Use LFIWZX
22679 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22680 (movsd splitter): Likewise.
22681 (movsd_hardfloat): Likewise.
22682 (movsd_softfloat): Likewise.
22684 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22685 binary and decimal floating point moves.
22686 (fmove_ok): New attributes to combine binary and decimal floating
22687 point moves, and to combine power6x (mfpgpr) moves along normal
22689 (real_value_to_target): Likewise.
22690 (f32_lr): Likewise.
22691 (f32_lm): Likewise.
22692 (f32_li): Likewise.
22693 (f32_sr): Likewise.
22694 (f32_sm): Likewise.
22695 (f32_si): Likewise.
22696 (movsf): Combine binary and decimal floating point moves. Combine
22697 power6x (mfpgpr) moves with other moves by using conditional
22698 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22699 (mov<mode> for SFmode/SDmode); Likewise.
22700 (SFmode/SDmode splitters): Likewise.
22701 (movsf_hardfloat): Likewise.
22702 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22703 (movsf_softfloat): Likewise.
22704 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22706 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22707 wx and wz constraints.
22709 * config/rs6000/constraints.md (wg constraint): New constraint to
22710 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22712 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22715 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22716 -mdebug=reg, print wg, wl, wx, and wz constraints.
22717 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22718 Initialize the reload functions for 64-bit binary/decimal floating
22720 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22721 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22722 create the buffer on the stack to overcome not having a 32-bit
22724 (rs6000_emit_move): Likewise.
22725 (rs6000_secondary_memory_needed_rtx): Likewise.
22726 (rs6000_alloc_sdmode_stack_slot): Likewise.
22727 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22728 via xxlxor, just like DFmode 0.0.
22730 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22731 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22732 with other moves by using conditional constraits (wg). Use LFIWZX
22733 and STFIWX for loading SDmode on power7.
22734 (movdd splitters): Likewise.
22735 (movdd_hardfloat32): Likewise.
22736 (movdd_softfloat32): Likewise.
22737 (movdd_hardfloat64_mfpgpr): Likewise.
22738 (movdd_hardfloat64): Likewise.
22739 (movdd_softfloat64): Likewise.
22741 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22742 64-bit binary and decimal floating point moves.
22743 (FMOVE64X): Likewise.
22744 (movdf): Combine 64-bit binary and decimal floating point moves.
22745 Combine power6x (mfpgpr) moves with other moves by using
22746 conditional constraits (wg).
22747 (mov<mode> for DFmode/DDmode): Likewise.
22748 (DFmode/DDmode splitters): Likewise.
22749 (movdf_hardfloat32): Likewise.
22750 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22751 (movdf_softfloat32): Likewise.
22752 (movdf_hardfloat64_mfpgpr): Likewise.
22753 (movdf_hardfloat64): Likewise.
22754 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22755 (movdf_softfloat64): Likewise.
22756 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22757 (reload_<mode>_load): Move to later in the file so they aren't in
22758 the middle of the floating point move insns.
22759 (reload_<mode>_store): Likewise.
22761 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22764 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22765 constraint if -mdebug=reg.
22766 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22767 Enable using dd reload support if needed.
22769 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22770 binary and decimal floating point moves in rs6000.md.
22771 (movtd_internal): Likewise.
22773 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22774 decimal floating point moves.
22776 (movtf_internal): Likewise.
22777 (mov<mode>_internal, TDmode/TFmode): Likewise.
22778 (movtf_softfloat): Likewise.
22779 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22781 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22782 movdi_internal64, using wg constraint for move direct operations.
22783 (movdi_internal64): Likewise.
22785 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22786 MODES_TIEABLE_P for selected modes. Print the numerical value of
22787 the various virtual registers. Use GPR/FPR first/last values,
22788 instead of hard coding the register numbers. Print which modes
22789 have reload functions registered.
22790 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22791 settings before/after setting cpu, target and subtarget settings.
22792 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22793 and for secondary reload failures in rs6000_secondary_reload_inner.
22794 (rs6000_secondary_reload_fail): Likewise.
22795 (rs6000_secondary_reload_inner): Likewise.
22797 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22798 macros for first/last GPR and FPR registers.
22799 (LAST_GPR_REGNO): Likewise.
22800 (FIRST_FPR_REGNO): Likewise.
22801 (LAST_FPR_REGNO): Likewise.
22803 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22804 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22805 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22806 (vcond<mode><mode>): Likewise.
22807 (vcondu<mode><mode>): Likewise.
22808 (vector_gtu<mode>): Likewise.
22809 (vector_gte<mode>): Likewise.
22810 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22811 to prevent the compiler from converting DImode operations to TImode.
22812 (ior<mode>3): Likewise.
22813 (and<mode>3): Likewise.
22814 (one_cmpl<mode>2): Likewise.
22815 (nor<mode>3): Likewise.
22816 (andc<mode>3): Likewise.
22818 * config/rs6000/constraints.md (wt constraint): New constraint
22819 that returns VSX_REGS if TImode is allowed in VSX registers.
22821 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22822 constant under VSX.
22824 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22825 similar to TImode, but it is restricted to being in the GPRs.
22827 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22828 TImode to occupy a single VSX register.
22830 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22831 -mvsx-timode for power7/power8.
22832 (power7 cpu): Likewise.
22833 (power8 cpu): Likewise.
22835 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22836 sure that TFmode/TDmode take up two registers if they are ever
22837 allowed in the upper VSX registers.
22838 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22840 (rs6000_init_hard_regno_mode_ok): Likewise.
22841 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22842 constraint. Print if LRA is turned on.
22843 (rs6000_option_override_internal): Give an error if -mvsx-timode
22844 and VSX is not enabled.
22845 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22846 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22847 to reg+offset addressing. Use PTImode when checking offset
22848 addresses for validity.
22849 (reg_offset_addressing_ok_p): Likewise.
22850 (rs6000_legitimate_offset_address_p): Likewise.
22851 (rs6000_legitimize_address): Likewise.
22852 (rs6000_legitimize_reload_address): Likewise.
22853 (rs6000_legitimate_address_p): Likewise.
22854 (rs6000_eliminate_indexed_memrefs): Likewise.
22855 (rs6000_emit_move): Likewise.
22856 (rs6000_secondary_reload): Likewise.
22857 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22858 reloads to fpr registers to continue to use reg+offset addressing,
22859 but 64-bit reloads to altivec registers need reg+reg addressing.
22860 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22861 it. Treat LO_SUM like a PLUS operation.
22862 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22863 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22864 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22865 registers to share a register with a smaller sized type, since VSX
22866 puts scalars in the upper 64-bits.
22867 (print_operand): Add support for PTImode.
22868 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22869 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22870 registers, but don't have arithmetic support.
22871 (rs6000_memory_move_cost): Add test for VSX.
22872 (rs6000_opt_masks): Add -mvsx-timode.
22874 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22877 (VSr): Use wt constraint for TImode.
22878 (VSv): Drop TImode support.
22879 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22880 (vsx_movti_64bit): Likewise.
22881 (vsx_movti_32bit): Likewise.
22882 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22883 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22884 one '?' on the appropriate output constraint. Do not allow TImode
22885 logical operations on 32-bit systems.
22886 (vsx_ior<mode>3): Likewise.
22887 (vsx_xor<mode>3): Likewise.
22888 (vsx_one_cmpl<mode>2): Likewise.
22889 (vsx_nor<mode>3): Likewise.
22890 (vsx_andc<mode>3): Likewise.
22891 (vsx_concat_<mode>): Likewise.
22892 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22894 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22895 OPTION_MASK_VSX_TIMODE.
22896 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22897 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22899 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22900 (TI2 iterator): New iterator for TImode, PTImode.
22901 (wd mode attribute): Add values for vector types.
22902 (movti_string): Replace TI move operations with operations for TImode
22903 and PTImode. Add support for TImode being allowed in VSX registers.
22904 (mov<mode>_string, TImode/PTImode): Likewise.
22905 (movti_ppc64): Likewise.
22906 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22907 (TI mode splitters): Likewise.
22909 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22912 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22914 PR tree-optimization/56355
22915 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22916 Also handle integers with undefined overflow.
22918 2013-03-20 Catherine Moore <clm@codesourcery.com>
22919 Maciej W. Rozycki <macro@codesourcery.com>
22920 Tom de Vries <tom@codesourcery.com>
22921 Nathan Sidwell <nathan@codesourcery.com>
22922 Iain Sandoe <iain@codesourcery.com>
22923 Nathan Froyd <froydnj@codesourcery.com>
22924 Chao-ying Fu <fu@mips.com>
22926 * doc/extend.texi: (micromips, nomicromips, nocompression):
22927 Document new function attributes.
22928 * doc/invoke.texi (minterlink-compressed, mmicromips,
22929 m14k, m14ke, m14kec): Document new options.
22930 (minterlink-mips16): Update documentation.
22931 * doc/md.texi (ZC, ZD): Document new constraints.
22932 * configure.ac (gcc_cv_as_micromips): Check if linker
22933 supports the .set micromips directive.
22934 * configure: Regenerate.
22935 * config.in: Regenerate.
22936 * config/mips/mips-tables.opt: Regenerate.
22937 * config/mips/micromips.md: New file.
22938 * constraints.md (ZC, ZD): New constraints.
22939 * config/mips/predicates.md (movep_src_register): New predicate.
22940 (movep_src_operand): New predicate.
22941 (non_volatile_mem_operand): New predicate.
22942 * config/mips/mips.md (multimem): New type.
22943 (length): Differentiate between 17-bit and 18-bit branch offsets.
22944 (MOVEP1, MOVEP2): New mode iterator.
22945 (mov_<load>l): Use ZC constraint.
22946 (mov_<load>r): Likewise.
22947 (mov_<store>l): Likewise.
22948 (mov_<store>r): Likewise.
22949 (*branch_equality<mode>_inverted): Add microMIPS support.
22950 (*branch_equality<mode>): Likewise.
22951 (*jump_absolute): Likewise.
22952 (indirect_jump_<mode>): Likewise.
22953 (tablejump_<mode>): Likewise.
22954 (<optab>_internal): Likewise.
22955 (sibcall_internal): Likewise.
22956 (sibcall_value_internal): Likewise.
22957 (prefetch): Use constraint ZD.
22958 * config/mips/mips.opt (minterlink-compressed): New option.
22959 (minterlink-mips16): Now an alias for minterlink-compressed.
22960 (mmicromips): New option.
22961 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22962 (compare_and_swap_12): Likewise.
22963 (sync_add<mode>): Likewise.
22964 (sync_<optab>_12): Likewise.
22965 (sync_old_<optab>_12): Likewise.
22966 (sync_new_<optab>_12): Likewise.
22967 (sync_nand_12): Likewise.
22968 (sync_old_nand_12): Likewise.
22969 (sync_new_nand_12): Likewise.
22970 (sync_sub<mode>): Likewise.
22971 (sync_old_add<mode>): Likewise.
22972 (sync_old_sub<mode>): Likewise.
22973 (sync_new_add<mode>): Likewise.
22974 (sync_new_sub<mode>): Likewise.
22975 (sync_<optab><mode>): Likewise.
22976 (sync_old_<optab><mode>): Likewise.
22977 (sync_new_<optab><mode>): Likewise.
22978 (sync_nand<mode>): Likewise.
22979 (sync_old_nand<mode>): Likewise.
22980 (sync_new_nand<mode>): Likewise.
22981 (sync_lock_test_and_set<mode>): Likewise.
22982 (test_and_set_12): Likewise.
22983 (atomic_compare_and_swap<mode>): Likewise.
22984 (atomic_exchange<mode>_llsc): Likewise.
22985 (atomic_fetch_add<mode>_llsc): Likewise.
22986 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22987 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22988 (umips_save_restore_pattern_p): Likewise.
22989 (umips_load_store_pair_p): Likewise.
22990 (umips_output_load_store_pair): Likewise.
22991 (umips_movep_target_p): Likewise.
22992 (umips_12bit_offset_address_p): Likewise.
22993 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22994 (mips_base_mips16): Rename this...
22995 (mips_base_compression_flags): ...to this. Update all uses.
22996 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22997 (mips_mips16_decl_p): Delete.
22998 (mips_nomips16_decl_p): Delete.
22999 (mips_get_compress_on_flags): New function.
23000 (mips_get_compress_off_flags): New function.
23001 (mips_get_compress_mode): New function.
23002 (mips_get_compress_on_name): New function.
23003 (mips_get_compress_off_name): New function.
23004 (mips_insert_attributes): Support multiple compression types.
23005 (mips_merge_decl_attributes): Likewise.
23006 (umips_12bit_offset_address_p): New function.
23007 (mips_start_function_definition): Emit .set micromips directive.
23008 (mips_call_may_need_jalx_p): New function.
23009 (mips_function_ok_for_sibcall): Add microMIPS support.
23010 (mips_print_operand_punctuation): Support short delay slots and
23012 (umips_swm_mask, umips_swm_encoding): New.
23013 (umips_build_save_restore): New function.
23014 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23015 (was_mips16_p): Remove.
23016 (old_compression_mode): New.
23017 (mips_set_compression_mode): New function.
23018 (mips_set_current_function): Add microMIPS support.
23019 (mips_option_override): Likewise.
23020 (umips_save_restore_pattern_p): New function.
23021 (umips_output_save_restore): New function.
23022 (umips_load_store_pair_p_1): New function.
23023 (umips_load_store_pair_p): New function.
23024 (umips_output_load_store_pair_1): New function.
23025 (umips_output_load_store_pair): New function.
23026 (umips_movep_target_p) New function.
23027 (mips_prepare_pch_save): Add microMIPS support.
23028 * config/mips/mips.h (TARGET_COMPRESSION): New.
23029 (TARGET_CPU_CPP_BUILTINS): Update macro
23030 to use new compression flags and to support microMIPS.
23031 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23032 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23033 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23034 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23035 (ASM_SPEC): Support mmicromips and mno-micromips.
23036 (M16STORE_REG_P): New macro.
23037 (MIPS_CALL): Support TARGET_MICROMIPS.
23038 (MICROMIPS_J): New macro.
23039 (mips_base_mips16): Rename this...
23040 (mips_base_compression_flags): ...to this.
23041 (UMIPS_12BIT_OFFSET_P): New macro.
23042 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23043 (MULTILIB_DIRNAMES): Likewise.
23044 2013-03-20 Richard Biener <rguenther@suse.de>
23046 PR tree-optimization/56661
23047 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23048 the result does not have to be distinct.
23050 2013-03-20 Richard Biener <rguenther@suse.de>
23052 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23055 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23056 Steven Bosscher <steven@gcc.gnu.org>
23058 PR rtl-optimization/56605
23059 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23061 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23064 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23065 that require movd instead of movq.
23067 2013-03-20 Richard Biener <rguenther@suse.de>
23069 * tree-ssa-structalias.c (struct variable_info): Add pointer
23070 to the first field of an aggregate with sub-vars. Make
23071 this and the pointer to the next subfield its ID.
23072 (vi_next): New function.
23073 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23074 storedanything_id, integer_id): Increment by one.
23075 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23076 get_call_clobber_vi): Adjust.
23077 (solution_set_expand): Simplify and speedup.
23078 (solution_set_add): Inline into ...
23079 (set_union_with_increment): ... this. Adjust accordingly.
23080 (do_sd_constraint): Likewise.
23081 (do_ds_constraint): Likewise.
23082 (do_complex_constraint): Simplify.
23083 (build_pred_graph): Adjust.
23084 (solve_graph): Likewise. Simplify and speedup.
23085 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23086 get_constraint_for_component_ref, get_constraint_for_1,
23087 first_vi_for_offset, first_or_preceding_vi_for_offset,
23088 create_function_info_for, create_variable_info_for_1,
23089 create_variable_info_for, intra_create_variable_infos): Adjust.
23090 (init_base_vars): Push NULL for ID zero.
23091 (compute_points_to_sets): Adjust.
23093 2013-03-20 Richard Biener <rguenther@suse.de>
23095 * cfgloop.c (verify_loop_structure): Streamline and avoid
23096 ICEing on corrupt loop tree.
23097 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23100 2013-03-20 Richard Biener <rguenther@suse.de>
23102 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23103 check whether an SSA update is needed.
23105 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23107 * config/mips/constraints.md (T): Rename to...
23111 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23112 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23114 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23116 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23117 (*subsi3_carryin_uxtw): Likewise.
23119 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23121 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23122 (*rorsi3_insn_uxtw): Likewise.
23124 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23126 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23127 (*extrsi5_insn_uxtw): Likewise.
23129 2013-03-19 Richard Biener <rguenther@suse.de>
23131 PR tree-optimization/56273
23132 * passes.c (init_optimization_passes): Move second VRP after DOM.
23134 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23136 * config/i386/i386.md (*movti_internal): Merge from
23137 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23138 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23139 nox64 isa attributes.
23141 2013-03-18 Richard Biener <rguenther@suse.de>
23143 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23145 (merge_node_constraints): Likewise.
23146 (build_succ_graph): Likewise.
23147 (valid_graph_edge): Inline into single caller.
23148 (unify_nodes): Likewise. Use bitmap_set_bit return value
23150 (scc_visit): Fix formatting and variable use.
23151 (do_sd_constraint): Use gcc_checking_assert.
23152 (do_ds_constraint): Likewise.
23153 (do_complex_constraint): Likewise.
23154 (condense_visit): Likewise. Cleanup.
23155 (dump_pred_graph): New function.
23156 (perform_var_substitution): Dump the pred-graph before
23157 variable substitution.
23158 (find_equivalent_node): Use gcc_checking_assert.
23159 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23161 2013-03-18 Richard Biener <rguenther@suse.de>
23163 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23164 Remove cond_expr_stmt_list argument and do not gimplify the
23166 (vect_loop_versioning): Adjust.
23167 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23168 Cleanup to use less temporaries.
23169 (vect_create_data_ref_ptr): Cleanup.
23171 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23173 PR tree-optimization/56635
23174 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23175 require types_compatible_p types.
23177 2013-03-18 Nick Clifton <nickc@redhat.com>
23179 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23180 spurious backslash.
23182 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23183 Add missing line to comment describing function.
23185 2013-03-18 Richard Biener <rguenther@suse.de>
23187 PR tree-optimization/56210
23188 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23189 Handle string / character search functions.
23190 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23192 2013-03-18 Richard Biener <rguenther@suse.de>
23194 PR middle-end/56483
23195 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23196 and implement properly.
23197 * gimple.h (gimple_cond_single_var_p): Remove.
23199 2013-03-18 Richard Biener <rguenther@suse.de>
23201 * tree-data-ref.h (find_data_references_in_loop): Declare.
23202 * tree-data-ref.c (get_references_in_stmt): Use a stack
23203 vector pre-allocated in the callers.
23204 (find_data_references_in_stmt): Adjust.
23205 (graphite_find_data_references_in_stmt): Likewise.
23206 (create_rdg_vertices): Likewise.
23207 (find_data_references_in_loop): Export.
23208 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23209 Compute dependences here...
23210 (vect_analyze_data_refs): ...not here. When we encounter
23211 a non-vectorizable data reference in basic-block vectorization
23212 truncate the data reference vector. Do not bother to
23213 fixup data-dependence information for gather loads.
23214 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23215 of data references, as reported.
23217 2013-03-18 Richard Biener <rguenther@suse.de>
23219 PR tree-optimization/3713
23220 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23221 has_constants and expr.
23222 (stmt_has_constants): Properly valueize SSA names when deciding
23223 whether the stmt has constants.
23225 2013-03-18 Richard Biener <rguenther@suse.de>
23227 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23228 whole function when there is nothing to do.
23229 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23230 * tree-vectorizer.c (vectorize_loops): Update virtual and
23231 loop-closed SSA once.
23232 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23234 2013-03-18 Richard Biener <rguenther@suse.de>
23236 PR middle-end/56113
23237 * domwalk.c (bb_postorder): New global static.
23238 (cmp_bb_postorder): New function.
23239 (walk_dominator_tree): Replace scheme imposing an order for
23240 visiting dominator sons by one sorting them at the time they
23241 are pushed on the stack.
23243 2013-03-18 Richard Biener <rguenther@suse.de>
23245 PR tree-optimization/39326
23246 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23247 (struct mem_ref): Replace mem member with ao_ref typed member.
23248 (MEM_ANALYZABLE): Adjust.
23249 (memref_eq): Likewise.
23250 (mem_ref_alloc): Likewise.
23251 (gather_mem_refs_stmt): Likewise.
23252 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23253 (execute_sm_if_changed_flag_set): Adjust.
23254 (execute_sm): Likewise.
23255 (ref_always_accessed_p): Likewise.
23256 (refs_independent_p): Likewise.
23257 (can_sm_ref_p): Likewise.
23259 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23262 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23263 return 1 even for !unsignedp.
23265 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23267 * config/i386/i386.md (isa): Add x64 and nox64.
23268 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23269 (*pushtf): Enable *roF alternative for x64 isa only.
23270 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23271 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23272 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23273 (*movtf_internal): Merge from *movtf_internal_rex64 and
23274 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23275 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23276 nox64 isa attributes.
23277 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23278 nox64 isa attributes.
23279 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23281 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23283 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23285 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23288 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23289 class names. Remove trailing comma after "ALL_REGS".
23291 2013-03-16 Jan Hubicka <jh@suse.cz>
23293 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23294 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23295 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23296 of cgraph_get_create_node.
23297 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23299 2013-03-16 Jason Merrill <jason@redhat.com>
23302 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23303 with DW_AT_default_value.
23305 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23307 * BASE-VER: Set to 4.9.0.
23309 2013-03-14 Andi Kleen <ak@linux.intel.com>
23312 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23313 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23314 Document _x* TSX intrinsics.
23316 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23317 David Holsgrove <david.holsgrove@xilinx.com>
23319 * configure.ac: Add MicroBlaze TLS support detection.
23320 * configure: Regenerate.
23321 * config/microblaze/microblaze-protos.h
23322 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23323 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23324 * config/microblaze/microblaze.c (microblaze_address_type): Add
23325 ADDRESS_TLS and tls_reloc address types.
23326 (microblaze_address_info): Add tls_reloc.
23327 (TARGET_HAVE_TLS): Define.
23328 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23329 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23330 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23331 load_tls_operand, microblaze_call_tls_get_addr,
23332 microblaze_legitimize_tls_address): New functions.
23333 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23334 (get_base_reg): Use microblaze_tls_symbol_p.
23335 (microblaze_classify_address): Handle TLS.
23336 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23337 label_mentioned_p and microblaze_tls_referenced_p.
23338 (microblaze_legitimize_address): Handle TLS.
23339 (microblaze_address_insns): Handle ADDRESS_TLS.
23340 (pic_address_needs_scratch): Handle TLS.
23341 (print_operand_address): Handle TLS.
23342 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23343 (microblaze_expand_move): Handle TLS.
23344 (microblaze_legitimate_constant_p): Check
23345 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23346 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23347 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23348 (PIC_OFFSET_TABLE_REGNUM): Set.
23349 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23350 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23351 (addsi3, movsi_internal2, movdf_internal): Update constraints
23352 * config/microblaze/predicates.md (arith_plus_operand): Define
23353 (move_operand): Redefine as move_src_operand,
23354 check microblaze_tls_referenced_p.
23356 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23358 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23359 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23361 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23363 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23366 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23368 PR tree-optimization/53265
23369 * common.opt (Waggressive-loop-optimizations): New option.
23370 * tree-ssa-loop-niter.c: Include tree-pass.h.
23371 (do_warn_aggressive_loop_optimizations): New function.
23372 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23373 if number_of_latch_executions returned constant.
23374 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23375 early. If number_of_latch_executions returned constant, set
23376 nb_iterations_upper_bound back to it.
23377 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23379 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23380 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23382 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23383 (MULTILIB_OSDIRNAMES): Set.
23384 * genmultilib: If defaultosdirname doesn't start with :: , set
23385 defaultosdirname2 instead, clear it and emit two . multilib_raw
23386 entries instead of just one.
23388 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23390 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23391 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23392 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23393 (SUBTARGET_OVERRIDE_OPTIONS): New.
23395 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23398 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23399 (musermode): Convert to Var(TARGET_USERMODE).
23400 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23401 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23402 * config/sh/sh.c (sh_option_override): Use
23403 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23404 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23406 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23408 (udivsi3_i4_single, divsi3_i4_single): Use
23409 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23411 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23413 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23416 2013-03-13 Richard Biener <rguenther@suse.de>
23418 PR tree-optimization/56608
23419 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23420 calls when vectorizing basic-blocks.
23422 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23425 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23428 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23430 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23432 2013-03-11 Jan Hubicka <jh@suse.cz>
23435 * lto-streamer-out.c (output_symbol_p): Skip references from
23436 constructors of external variables.
23438 2013-03-11 Jan Hubicka <jh@suse.cz>
23440 PR middle-end/56571
23441 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23443 * emit-rtl.c (verify_rtx_sharing): Likewise.
23444 (copy_insn_1): Likewise.
23445 * rtl.c (copy_rtx): Likewise.
23447 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23450 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23451 output_operand_lossage message.
23453 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23456 * arm.c (shift_op): Validate RTL pattern on the fly.
23457 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23460 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23463 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23464 new call patterns that clobber %fr12.
23465 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23466 split and postreload patterns.
23467 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23468 registers %fr12 and %fr12R as call used.
23470 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23472 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23473 (canon_address, record_store, replace_read, check_mem_read_rtx,
23474 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23475 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23476 rest_of_handle_dse): Likewise.
23478 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23480 PR middle-end/56524
23481 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23483 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23484 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23485 (save_optabs_if_changed): Replace with...
23486 (init_tree_optimization_optabs): ...this.
23487 * optabs.c (save_optabs_if_changed): Rename to...
23488 (init_tree_optimization_optabs): ...this. Take the optimization node
23489 as argument. Do nothing if the base optabs are already correct.
23490 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23491 to recompute optabs.
23492 * function.h (function): Remove optabs field.
23493 * function.c (invoke_set_current_function_hook): Call
23494 init_tree_optimization_optabs. Use the result to initialize
23497 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23499 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23500 if GTMA_HAS_NO_INSTRUMENTATION.
23501 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23502 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23503 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23504 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23505 GTMA_HAS_NO_INSTRUMENTATION.
23507 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23509 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23512 2013-03-08 Marek Polacek <polacek@redhat.com>
23513 Jakub Jelinek <jakub@redhat.com>
23515 PR tree-optimization/56478
23516 * predict.c (is_comparison_with_loop_invariant_p): Change the
23517 type of loop_step to tree.
23518 (predict_loops): Adjust.
23519 (predict_iv_comparison): Perform the computations on double_ints.
23521 2013-03-08 Richard Biener <rguenther@suse.de>
23523 PR tree-optimization/56570
23524 * tree-cfg.c (verify_expr_location_1): Verify locations for
23526 * tree-sra.c (create_access_replacement): Strip locations
23527 from DECL_DEBUG_EXPRs.
23529 2013-03-08 Richard Biener <rguenther@suse.de>
23531 * tree-inline.c (expand_call_inline): Do not associate
23532 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23533 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23535 2013-03-08 Richard Biener <rguenther@suse.de>
23537 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23538 or block changes with -Og. Fix for location / block encoding
23539 changes and PHI arguments with locations.
23541 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23543 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23545 (struct output_info): Likewise.
23546 (register_overhead): Remove bad gcc_assert.
23547 (bitmap_find_bit): If there is only a single bitmap element, do not
23548 count a miss as a search.
23549 (print_statistics): Update for counter type changes.
23550 (dump_bitmap_statistics): Likewise. Print headers such that they
23551 are properly lined up with the printed counters.
23553 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23555 PR tree-optimization/56559
23556 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23557 check that it has only a single use.
23559 2013-03-07 Richard Biener <rguenther@suse.de>
23561 * doc/invoke.texi (fwhole-program): Discourage use in combination
23564 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23566 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23568 PR tree-optimization/56539
23569 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23570 instead of GSI_CONTINUE_LINKING as last argument to
23571 force_gimple_operand_gsi. Adjust function comment.
23573 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23576 PR middle-end/56548
23577 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23578 promoted mode, convert the result back to the original mode.
23580 2013-03-06 Richard Biener <rguenther@suse.de>
23582 PR middle-end/56294
23583 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23584 (insert_updated_phi_nodes_compare_uids): New function.
23585 (update_ssa): Sort symbols_to_rename after UID before
23586 traversing it to insert PHI nodes.
23588 2013-03-06 Richard Biener <rguenther@suse.de>
23590 PR middle-end/50494
23591 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23592 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23595 2013-02-13 Richard Biener <rguenther@suse.de>
23598 * varasm.c (output_constant_def_1): Get the decl representing
23599 the constant as argument.
23600 (output_constant_def): Wrap output_constant_def_1.
23601 (make_decl_rtl): Use output_constant_def_1 with the decl
23602 representing the constant.
23603 (build_constant_desc): Optionally re-use a decl already
23604 representing the constant.
23605 (tree_output_constant_def): Adjust.
23607 2013-03-06 Joey Ye <joey.ye@arm.com>
23610 * gcc.c (convert_white_space): New function.
23611 (main): Handles white space in function name.
23613 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23616 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23617 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23618 to SH_DIV_CALL_TABLE for TARGET_SH2.
23619 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23621 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23622 call-table options.
23624 2013-03-05 Sterling Augustine <saugustine@google.com>
23625 Cary Coutant <ccoutant@google.com>
23628 * dwarf2out.c (resolve_addr): Don't call
23629 remove_loc_list_addr_table_entries a second time for the same
23632 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23635 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23636 (avoid_complex_debug_insns): New function.
23637 (expand_debug_locations): Call it.
23639 PR rtl-optimization/56484
23640 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23641 lifetimes of hard registers on small register class machines.
23643 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23645 * config/microblaze/microblaze-protos.h: Rename
23646 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23647 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23649 (microblaze_fast_interrupt_function_p): New function.
23650 (microblaze_is_interrupt_handler): Rename to
23651 microblaze_is_interrupt_variant and add fast_interrupt check.
23652 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23653 (save_restore_insns): Likewise.
23654 (compute_frame_size): Likewise.
23655 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23656 (microblaze_globalize_label): Likewise.
23657 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23658 * config/microblaze/microblaze.md: Use wrapper
23659 microblaze_is_interrupt_variant.
23661 2013-03-05 Kai Tietz <ktietz@redhat.com>
23663 * sdbout.c (sdbout_one_type): Switch to current function's section
23664 supporting cold/hot.
23666 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23668 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23671 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23673 PR middle-end/56461
23674 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23675 if VALGRIND_GET_VBITS is defined, temporarily make object
23676 memory all defined, and restore previous valgrind addressability
23677 and definability afterwards. Free this_object at the end.
23679 PR middle-end/56461
23680 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23681 right before calling lra_create_live_ranges, also call it
23682 when clearing live_p. Only call lra_clear_live_ranges
23683 at the end if live_p.
23685 PR middle-end/56461
23686 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23688 2013-03-05 Richard Biener <rguenther@suse.de>
23690 PR tree-optimization/56521
23691 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23694 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23697 * except.h (remove_unreachable_eh_regions): New prototype.
23698 * except.c (remove_eh_handler_splicer): New function, split out
23699 of remove_eh_handler.
23700 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23701 warning about running it on many EH regions one at a time.
23702 (remove_unreachable_eh_regions_worker): New function, walk the
23703 EH tree in depth-first order and remove non-marked regions.
23704 (remove_unreachable_eh_regions): New function.
23705 * tree-eh.c (mark_reachable_handlers): New function, split out
23706 from remove_unreachable_handlers.
23707 (remove_unreachable_handlers): Use mark_reachable_handlers and
23708 remove_unreachable_eh_regions.
23709 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23710 and remove_unreachable_eh_regions.
23712 2013-03-05 Richard Biener <rguenther@suse.de>
23714 PR middle-end/56525
23715 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23716 not freeing them until the end.
23718 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23720 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23722 2013-03-05 Richard Biener <rguenther@suse.de>
23724 PR tree-optimization/56270
23725 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23726 of loads after scheduling an SLP instance.
23728 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23730 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23732 (check_gcc_parallelize): Run guality.exp as a separate job from
23733 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23734 struct-layout-1.exp with stackalign.exp.
23736 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23738 PR middle-end/56461
23739 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23740 load_index sbitmap even if some bit in it isn't set.
23742 PR middle-end/56461
23743 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23744 (discover_iteration_bound_by_body_walk): Change queues to
23745 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23746 spelling in comment. Call safe_push on queues[bound_index] directly.
23747 Release queues[queue_index] in every iteration unconditionally.
23748 Release bounds vector.
23750 PR middle-end/56461
23751 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23752 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23753 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23756 2013-03-05 Richard Biener <rguenther@suse.de>
23759 * tree-inline.c (remap_blocks_to_null): New function.
23760 (expand_call_inline): When expanding a call stmt without
23761 an associated block inline remap all callee blocks to NULL.
23763 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23765 PR rtl-optimization/56494
23766 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23767 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23768 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23770 PR middle-end/56461
23771 * sel-sched-ir.c (free_sched_pools): Release
23772 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23773 if succs_info_pool.max_top isn't -1.
23776 * opts.c (opts_obstack, opts_concat): Moved to...
23777 * opts-common.c (opts_obstack, opts_concat): ... here.
23779 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23781 PR middle-end/56461
23782 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23784 2013-03-04 Martin Jambor <mjambor@suse.cz>
23786 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23787 all appropriate places.
23789 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23791 PR tree-optimization/56424
23792 * ipa-split.c (split_function): Do not set the RSO flag if result is
23793 not by reference and its type is a register type.
23795 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23797 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23798 (microblaze_legitimate_pic_operand): Likewise
23799 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23800 new function microblaze_legitimate_pic_operand
23801 * config/microblaze/microblaze-protos.h
23802 (microblaze_legitimate_pic_operand): Declare.
23804 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23806 * config/microblaze/predicates.md (call_insn_simple_operand):
23807 New predicate for supported rtx code types.
23808 * config/microblaze/microblaze.md (call_internal1): Use
23809 call_insn_simple_operand predicate.
23811 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23813 PR middle-end/56461
23814 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23815 partitions.ordered_remove.
23817 PR middle-end/56461
23818 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23819 vec_oprnds0.create (1) for modifier == NONE.
23821 PR middle-end/56461
23822 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23823 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23824 vec_oprnds1 right before pushing anything to it for
23827 PR middle-end/56461
23828 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23829 set nbbs to 0 instead of having separate code path.
23830 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23831 instead of false as last argument if returning NULL.
23833 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23835 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23836 the attribute is now called "target" instead of "option".
23837 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23838 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23839 attribute/pragma name for TARGET_OPTION_VALID_P and
23840 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23841 * doc/tm.texi: Regenerated.
23843 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23845 * config/microblaze/microblaze.c:
23846 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23847 * config/microblaze/microblaze.h: Add -mxl-reorder to
23849 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23850 instructions emitted if TARGET_REORDER.
23851 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23852 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23855 2013-03-01 Xinliang David Li <davidxl@google.com>
23857 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23860 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23862 PR middle-end/56461
23863 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23864 vector even when returning true. Fix up function comment formatting.
23866 PR middle-end/56461
23867 * ira-build.c (ira_loop_nodes_count): New variable.
23868 (create_loop_tree_nodes): Initialize it.
23869 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23871 PR middle-end/56461
23872 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23873 method on dr_chain and result_chain.
23874 * tree-vect-stmts.c (vectorizable_store): Only call
23875 result_chain.create if j == 0.
23877 PR middle-end/56461
23878 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23879 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23880 before overwriting it.
23882 2013-03-01 Tobias Burnus <burnus@net-b.de>
23884 * doc/extended.texi (C Extensions): Change order in @menu
23886 (Other MIPS Built-in Functions): Move last MIPS entry before
23887 "picoChip Built-in Functions".
23888 (SH Built-in Functions): Move after RX Built-in Functions.
23889 * doc/gcc.texi (Introduction): Change order in @menu
23891 * doc/md.texi (Constraints): Ditto.
23892 * gty.texi (Type Information): Ditto.
23893 (User-provided marking routines for template types): Make
23895 * doc/invoke.texi (AArch64 Options): Move before
23896 "Adapteva Epiphany Options".
23898 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23899 Jakub Jelinek <jakub@redhat.com>
23902 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23903 no_address_safety_analysis attribute.
23904 * doc/extend.texi (no_address_safety_attribute): Rename to
23905 no_sanitize_address attribute, mention no_address_safety_analysis
23906 attribute as deprecated alias.
23908 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23910 PR middle-end/56461
23911 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23912 type to vec<vec<tree> > *.
23913 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23914 to be vec<tree> instead of vec<tree> *, set vec_defs
23915 to vNULL and call vec_defs.create (number_of_vects), adjust other
23917 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23918 vectorizable_condition): Adjust vect_get_slp_defs callers.
23920 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23922 * config/aarch64/aarch64.c
23923 (aarch64_float_const_representable): Remove unused variable.
23925 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23927 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23929 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23931 * config/aarch64/aarch64-builtins.c
23932 (aarch64_init_simd_builtins): Make static.
23934 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23936 * config/aarch64/aarch64.c
23937 (aarch64_simd_make_constant): Make static.
23939 2013-02-28 Martin Jambor <mjambor@suse.cz>
23941 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23942 with no initialization to the RHS of debug statements.
23944 2013-02-28 Martin Jambor <mjambor@suse.cz>
23946 PR tree-optimization/56294
23947 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23949 (get_access_replacement): Do not call create_access_replacement.
23950 Assert a replacement exists.
23951 (get_repl_default_def_ssa_name): Create the replacement declaration
23954 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23956 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23957 final_end_function.
23959 2013-02-28 Marek Polacek <polacek@redhat.com>
23961 PR rtl-optimization/56466
23962 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23963 if we're changing a loop.
23964 (peel_loops_completely): Likewise.
23966 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23969 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23971 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23974 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23975 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23976 INTX_FTYPE_FX, FX_FTYPE_INTX.
23977 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23979 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23981 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23982 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23983 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23984 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23985 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23986 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23987 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23988 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23989 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23990 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23991 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23992 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23993 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23994 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23995 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23996 (avrxmega6): Increase max flash segments from 5 to 6.
23997 * config/avr/t-multilib: Regenerate.
23998 * config/avr/avr-tables.opt: Regenerate.
23999 * doc/avr-mmcu.texi: Regenerate.
24001 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24003 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24004 (avr_device_to_arch): Rename to avr_device_to_ld.
24005 (avr_device_to_as): New prototype.
24006 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24007 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24008 * config/avr/driver-avr.c (avr_device_to_as): New.
24009 (avr_device_to_arch): Rename to avr_device_to_ld.
24011 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24013 PR middle-end/56461
24014 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24015 method on dr_chain and result_chain.
24017 PR middle-end/56461
24018 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24019 pointer_set_destroy on not_executed_last_iteration.
24021 PR middle-end/56461
24022 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24024 PR middle-end/56461
24025 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24026 FOR_EACH_DEFINED_FUNCTION when freeing state.
24028 PR middle-end/56461
24029 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24031 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24034 PR middle-end/56461
24035 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24036 known_aggs[i].items and release known_aggs vector.
24038 PR middle-end/56461
24039 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24041 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24043 * config/microblaze/microblaze.c (microblaze_emit_compare):
24044 Use xor for EQ/NE comparisions.
24045 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24046 (cbranchsf4): Adjust operator to comparison_operator.
24048 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24050 PR middle-end/56461
24051 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24053 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24054 vec_safe_push, always update *slot.
24055 (redirect_edge_var_map_clear): Use vec_free.
24056 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24057 (free_var_map_entry): Use vec_free.
24058 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24059 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24061 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24063 PR middle-end/45472
24064 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24065 when the may_trap_p bit of the exprs being merged differs.
24066 Reorder tests for speculativeness in the logical and operator.
24068 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24070 * incpath.c (add_standard_paths): Use reconcat instead of concat
24071 where appropriate and avoid leaking memory.
24073 * opts.h: Include obstack.h.
24074 (opts_concat): New prototype.
24075 (opts_obstack): New declaration.
24076 * opts.c (opts_concat): New function.
24077 (opts_obstack): New variable.
24078 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24079 (finish_options): Use opts_concat instead of concat
24080 and XOBNEWVEC instead of XNEWVEC.
24081 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24082 generate_option): Likewise.
24083 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24084 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24087 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24088 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24090 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24092 PR middle-end/56461
24093 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24095 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24097 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24098 (arm_block_move_unaligned_straight): Likewise.
24099 (arm_adjust_block_mem): Likewise.
24101 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24104 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24105 temp, cond and label.
24106 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24109 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24110 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24113 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24114 of prologue/epilogue functions.
24117 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24118 Remove unused variable cfa_offset.
24119 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24122 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24123 type promotion to unsigned.
24126 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24127 for HOST_WIDE_INT of 32 bit / same size as int.
24128 (arm_block_move_unaligned_straight): Likewise.
24129 (arm_adjust_block_mem): Likewise.
24132 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24135 2013-02-26 Marek Polacek <polacek@redhat.com>
24137 PR tree-optimization/56426
24138 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24140 2013-02-26 Richard Biener <rguenther@suse.de>
24143 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24144 unused variable loops.
24146 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24148 PR tree-optimization/56448
24149 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24150 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24151 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24152 later operands of the references, or even first operand for
24153 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24155 PR tree-optimization/56443
24156 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24157 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24158 to type_for_mode langhook.
24160 2013-02-25 Matt Turner <mattst88@gmail.com>
24162 * doc/invoke.texi: Document r4700.
24164 2013-02-25 Richard Biener <rguenther@suse.de>
24166 PR tree-optimization/56175
24167 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24169 (simplify_bitwise_binary): ... here. Also guard the conversion
24170 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24172 2013-02-25 Catherine Moore <clm@codesourcery.com>
24175 2013-02-24 Catherine Moore <clm@codesourcery.com>
24176 Maciej W. Rozycki <macro@codesourcery.com>
24177 Tom de Vries <tom@codesourcery.com>
24178 Nathan Sidwell <nathan@codesourcery.com>
24179 Iain Sandoe <iain@codesourcery.com>
24180 Nathan Froyd <froydnj@codesourcery.com>
24181 Chao-ying Fu <fu@mips.com>
24183 * doc/extend.texi: (micromips, nomicromips, nocompression):
24184 Document new function attributes.
24185 * doc/invoke.texi (minterlink-compressed, mmicromips,
24186 m14k, m14ke, m14kec): Document new options.
24187 (minterlink-mips16): Update documentation.
24188 * doc/md.texi (ZC, ZD): Document new constraints.
24189 * configure.ac (gcc_cv_as_micromips): Check if linker
24190 supports the .set micromips directive.
24191 * configure: Regenerate.
24192 * config.in: Regenerate.
24193 * config/mips/mips-tables.opt: Regenerate.
24194 * config/mips/micromips.md: New file.
24195 * constraints.md (ZC, AD): New constraints.
24196 * config/mips/predicates.md (movep_src_register): New predicate.
24197 (movep_src_operand): New predicate.
24198 (non_volatile_mem_operand): New predicate.
24199 * config/mips/mips.md (multimem): New type.
24200 (length): Differentiate between 17-bit and 18-bit branch offsets.
24201 (MOVEP1, MOVEP2): New mode iterator.
24202 (mov_<load>l): Use ZC constraint.
24203 (mov_<load>r): Likewise.
24204 (mov_<store>l): Likewise.
24205 (mov_<store>r): Likewise.
24206 (*branch_equality<mode>_inverted): Add microMIPS support.
24207 (*branch_equality<mode>): Likewise.
24208 (*jump_absolute): Likewise.
24209 (indirect_jump_<mode>): Likewise.
24210 (tablejump_<mode>): Likewise.
24211 (<optab>_internal): Likewise.
24212 (sibcall_internal): Likewise.
24213 (sibcall_value_internal): Likewise.
24214 (prefetch): Use constraint ZD.
24215 * config/mips/mips.opt (minterlink-compressed): New option.
24216 (minterlink-mips16): Now an alias for minterlink-compressed.
24217 (mmicromips): New option.
24218 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24219 (compare_and_swap_12): Likewise.
24220 (sync_add<mode>): Likewise.
24221 (sync_<optab>_12): Likewise.
24222 (sync_old_<optab>_12): Likewise.
24223 (sync_new_<optab>_12): Likewise.
24224 (sync_nand_12): Likewise.
24225 (sync_old_nand_12): Likewise.
24226 (sync_new_nand_12): Likewise.
24227 (sync_sub<mode>): Likewise.
24228 (sync_old_add<mode>): Likewise.
24229 (sync_old_sub<mode>): Likewise.
24230 (sync_new_add<mode>): Likewise.
24231 (sync_new_sub<mode>): Likewise.
24232 (sync_<optab><mode>): Likewise.
24233 (sync_old_<optab><mode>): Likewise.
24234 (sync_new_<optab><mode>): Likewise.
24235 (sync_nand<mode>): Likewise.
24236 (sync_old_nand<mode>): Likewise.
24237 (sync_new_nand<mode>): Likewise.
24238 (sync_lock_test_and_set<mode>): Likewise.
24239 (test_and_set_12): Likewise.
24240 (atomic_compare_and_swap<mode>): Likewise.
24241 (atomic_exchange<mode>_llsc): Likewise.
24242 (atomic_fetch_add<mode>_llsc): Likewise.
24243 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24244 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24245 (umips_save_restore_pattern_p): Likewise.
24246 (umips_load_store_pair_p): Likewise.
24247 (umips_output_load_store_pair): Likewise.
24248 (umips_movep_target_p): Likewise.
24249 (umips_12bit_offset_address_p): Likewise.
24250 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24251 (mips_base_mips16): Rename this...
24252 (mips_base_compression_flags): ...to this. Update all uses.
24253 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24254 (mips_mips16_decl_p): Delete.
24255 (mips_nomips16_decl_p): Delete.
24256 (mips_get_compress_on_flags): New function.
24257 (mips_get_compress_off_flags): New function.
24258 (mips_get_compress_mode): New function.
24259 (mips_get_compress_on_name): New function.
24260 (mips_get_compress_off_name): New function.
24261 (mips_insert_attributes): Support multiple compression types.
24262 (mips_merge_decl_attributes): Likewise.
24263 (umips_12bit_offset_address_p): New function.
24264 (mips_start_function_definition): Emit .set micromips directive.
24265 (mips_call_may_need_jalx_p): New function.
24266 (mips_function_ok_for_sibcall): Add microMIPS support.
24267 (mips_print_operand_punctuation): Support short delay slots and
24269 (umips_swm_mask, umips_swm_encoding): New.
24270 (umips_build_save_restore): New function.
24271 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24272 (was_mips16_p): Remove.
24273 (old_compression_mode): New.
24274 (mips_set_compression_mode): New function.
24275 (mips_set_current_function): Add microMIPS support.
24276 (mips_option_override): Likewise.
24277 (umips_save_restore_pattern_p): New function.
24278 (umips_output_save_restore): New function.
24279 (umips_load_store_pair_p_1): New function.
24280 (umips_load_store_pair_p): New function.
24281 (umips_output_load_store_pair_1): New function.
24282 (umips_output_load_store_pair): New function.
24283 (umips_movep_target_p) New function.
24284 (mips_prepare_pch_save): Add microMIPS support.
24285 * config/mips/mips.h (TARGET_COMPRESSION): New.
24286 (TARGET_CPU_CPP_BUILTINS): Update macro
24287 to use new compression flags and to support microMIPS.
24288 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24289 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24290 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24291 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24292 (ASM_SPEC): Support mmicromips and mno-micromips.
24293 (M16STORE_REG_P): New macro.
24294 (MIPS_CALL): Support TARGET_MICROMIPS.
24295 (MICROMIPS_J): New macro.
24296 (mips_base_mips16): Rename this...
24297 (mips_base_compression_flags): ...to this.
24298 (UMIPS_12BIT_OFFSET_P): New macro.
24299 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24300 (MULTILIB_DIRNAMES): Likewise.
24302 2013-02-25 Tom de Vries <tom@codesourcery.com>
24304 PR rtl-optimization/56131
24305 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24306 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24307 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24309 2013-02-25 Tobias Burnus <burnus@net-b.de>
24311 * doc/invoke.texi (-fsanitize=): Move from optimization
24312 to debugging options.
24314 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24316 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24318 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24319 Alexander Monakov <amonakov@ispras.ru>
24321 PR middle-end/56077
24322 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24323 flush pending lists also on non-jumps. Adjust comment.
24325 2013-02-24 Catherine Moore <clm@codesourcery.com>
24326 Maciej W. Rozycki <macro@codesourcery.com>
24327 Tom de Vries <tom@codesourcery.com>
24328 Nathan Sidwell <nathan@codesourcery.com>
24329 Iain Sandoe <iain@codesourcery.com>
24330 Nathan Froyd <froydnj@codesourcery.com>
24331 Chao-ying Fu <fu@mips.com>
24333 * doc/extend.texi: (micromips, nomicromips, nocompression):
24334 Document new function attributes.
24335 * doc/invoke.texi (minterlink-compressed, mmicromips,
24336 m14k, m14ke, m14kec): Document new options.
24337 (minterlink-mips16): Update documentation.
24338 * doc/md.texi (ZC, ZD): Document new constraints.
24339 * configure.ac (gcc_cv_as_micromips): Check if linker
24340 supports the .set micromips directive.
24341 * configure: Regenerate.
24342 * config.in: Regenerate.
24343 * config/mips/mips-tables.opt: Regenerate.
24344 * config/mips/micromips.md: New file.
24345 * constraints.md (ZC, AD): New constraints.
24346 * config/mips/predicates.md (movep_src_register): New predicate.
24347 (movep_src_operand): New predicate.
24348 (non_volatile_mem_operand): New predicate.
24349 * config/mips/mips.md (multimem): New type.
24350 (length): Differentiate between 17-bit and 18-bit branch offsets.
24351 (MOVEP1, MOVEP2): New mode iterator.
24352 (mov_<load>l): Use ZC constraint.
24353 (mov_<load>r): Likewise.
24354 (mov_<store>l): Likewise.
24355 (mov_<store>r): Likewise.
24356 (*branch_equality<mode>_inverted): Add microMIPS support.
24357 (*branch_equality<mode>): Likewise.
24358 (*jump_absolute): Likewise.
24359 (indirect_jump_<mode>): Likewise.
24360 (tablejump_<mode>): Likewise.
24361 (<optab>_internal): Likewise.
24362 (sibcall_internal): Likewise.
24363 (sibcall_value_internal): Likewise.
24364 (prefetch): Use constraint ZD.
24365 * config/mips/mips.opt (minterlink-compressed): New option.
24366 (minterlink-mips16): Now an alias for minterlink-compressed.
24367 (mmicromips): New option.
24368 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24369 (compare_and_swap_12): Likewise.
24370 (sync_add<mode>): Likewise.
24371 (sync_<optab>_12): Likewise.
24372 (sync_old_<optab>_12): Likewise.
24373 (sync_new_<optab>_12): Likewise.
24374 (sync_nand_12): Likewise.
24375 (sync_old_nand_12): Likewise.
24376 (sync_new_nand_12): Likewise.
24377 (sync_sub<mode>): Likewise.
24378 (sync_old_add<mode>): Likewise.
24379 (sync_old_sub<mode>): Likewise.
24380 (sync_new_add<mode>): Likewise.
24381 (sync_new_sub<mode>): Likewise.
24382 (sync_<optab><mode>): Likewise.
24383 (sync_old_<optab><mode>): Likewise.
24384 (sync_new_<optab><mode>): Likewise.
24385 (sync_nand<mode>): Likewise.
24386 (sync_old_nand<mode>): Likewise.
24387 (sync_new_nand<mode>): Likewise.
24388 (sync_lock_test_and_set<mode>): Likewise.
24389 (test_and_set_12): Likewise.
24390 (atomic_compare_and_swap<mode>): Likewise.
24391 (atomic_exchange<mode>_llsc): Likewise.
24392 (atomic_fetch_add<mode>_llsc): Likewise.
24393 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24394 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24395 (umips_save_restore_pattern_p): Likewise.
24396 (umips_load_store_pair_p): Likewise.
24397 (umips_output_load_store_pair): Likewise.
24398 (umips_movep_target_p): Likewise.
24399 (umips_12bit_offset_address_p): Likewise.
24400 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24401 (mips_base_mips16): Rename this...
24402 (mips_base_compression_flags): ...to this. Update all uses.
24403 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24404 (mips_mips16_decl_p): Delete.
24405 (mips_nomips16_decl_p): Delete.
24406 (mips_get_compress_on_flags): New function.
24407 (mips_get_compress_off_flags): New function.
24408 (mips_get_compress_mode): New function.
24409 (mips_get_compress_on_name): New function.
24410 (mips_get_compress_off_name): New function.
24411 (mips_insert_attributes): Support multiple compression types.
24412 (mips_merge_decl_attributes): Likewise.
24413 (umips_12bit_offset_address_p): New function.
24414 (mips_start_function_definition): Emit .set micromips directive.
24415 (mips_call_may_need_jalx_p): New function.
24416 (mips_function_ok_for_sibcall): Add microMIPS support.
24417 (mips_print_operand_punctuation): Support short delay slots and
24419 (umips_swm_mask, umips_swm_encoding): New.
24420 (umips_build_save_restore): New function.
24421 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24422 (was_mips16_p): Remove.
24423 (old_compression_mode): New.
24424 (mips_set_compression_mode): New function.
24425 (mips_set_current_function): Add microMIPS support.
24426 (mips_option_override): Likewise.
24427 (umips_save_restore_pattern_p): New function.
24428 (umips_output_save_restore): New function.
24429 (umips_load_store_pair_p_1): New function.
24430 (umips_load_store_pair_p): New function.
24431 (umips_output_load_store_pair_1): New function.
24432 (umips_output_load_store_pair): New function.
24433 (umips_movep_target_p) New function.
24434 (mips_prepare_pch_save): Add microMIPS support.
24435 * config/mips/mips.h (TARGET_COMPRESSION): New.
24436 (TARGET_CPU_CPP_BUILTINS): Update macro
24437 to use new compression flags and to support microMIPS.
24438 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24439 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24440 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24441 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24442 (ASM_SPEC): Support mmicromips and mno-micromips.
24443 (M16STORE_REG_P): New macro.
24444 (MIPS_CALL): Support TARGET_MICROMIPS.
24445 (MICROMIPS_J): New macro.
24446 (mips_base_mips16): Rename this...
24447 (mips_base_compression_flags): ...to this.
24448 (UMIPS_12BIT_OFFSET_P): New macro.
24449 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24450 (MULTILIB_DIRNAMES): Likewise.
24452 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24455 * target-globals.c (save_target_globals): For init_reg_sets and
24456 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24458 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24460 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24461 * config/aarch64/t-aarch64
24462 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24464 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24466 PR inline-asm/56148
24467 * lra-constraints.c (process_alt_operands): Reload operand
24468 conflicting with earlier clobber only if no more other conflicting
24471 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24474 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24475 if not linking a shared library.
24477 2013-02-22 Seth LaForge <sethml@google.com>
24479 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24481 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24483 * config/arm/arm.md (split for extendsidi): Update condition.
24484 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24485 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24486 (qhs_zextenddi_cstr): Likewise.
24488 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24490 PR middle-end/56420
24491 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24492 avoid signed wrapping.
24493 (expand_mult): Handle properly multiplication by
24494 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24495 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24496 in the compiler if coeff is HOST_WIDE_INT_MIN.
24497 (expand_divmod): Don't make ext_op1 static, change it's type to
24498 uhwi. Avoid undefined behavior in -INTVAL (op1).
24500 PR rtl-optimization/50339
24501 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24503 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24504 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24505 into splitting_ashiftrt field.
24506 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24508 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24511 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24513 PR middle-end/56108
24514 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24515 are sure to go irrevocable.
24517 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24519 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24520 scalars are valid operands.
24522 2013-02-21 Martin Jambor <mjambor@suse.cz>
24524 PR tree-optimization/56310
24525 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24526 only matching indices and non-negative final offsets.
24527 (intersect_aggregates_with_edge): Pass src_idx to
24528 agg_replacements_to_vector. Pass src_idx insstead of index to
24529 intersect_with_agg_replacements.
24531 2013-02-21 Martin Jambor <mjambor@suse.cz>
24533 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24534 instead of hard-wired defaults.
24536 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24538 * doc/invoke.texi (MIPS Options): Update documentation of the
24539 floating-point multiply-accumulate instruction restrictions.
24541 2013-02-21 Kostya Serebryany <kcc@google.com>
24543 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24544 asan_shadow_offset on x86_64 linux.
24546 2013-02-21 Richard Biener <rguenther@suse.de>
24548 PR tree-optimization/56415
24550 2013-02-11 Richard Biener <rguenther@suse.de>
24552 PR tree-optimization/56273
24553 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24556 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24559 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24562 PR inline-asm/56405
24563 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24564 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24566 2013-02-20 Jan Hubicka <jh@suse.cz>
24568 PR tree-optimization/56265
24569 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24570 when target is referenced for first time.
24572 2013-02-20 Richard Biener <rguenther@suse.de>
24574 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24575 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24576 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24577 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24578 not return anything.
24579 (rename_ssa_copies): Do not remove unused locals.
24580 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24581 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24582 * passes.c (execute_function_todo): Do not schedule unused locals
24583 removal if cleanup_tree_cfg did something.
24584 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24585 about the number of removed locals.
24587 2013-02-20 Richard Biener <rguenther@suse.de>
24589 PR tree-optimization/56398
24590 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24592 2013-02-20 Martin Jambor <mjambor@suse.cz>
24594 PR tree-optimization/55334
24595 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24596 restricted pointers to arrays.
24598 2013-02-20 Richard Biener <rguenther@suse.de>
24599 Jakub Jelinek <jakub@redhat.com>
24601 PR tree-optimization/56396
24602 * tree-ssa-ccp.c (n_const_val): New static variable.
24603 (get_value): Return NULL for SSA names we don't have a lattice
24605 (ccp_initialize): Initialize n_const_val.
24606 * tree-ssa-copy.c (n_copy_of): New static variable.
24607 (init_copy_prop): Initialize n_copy_of.
24608 (get_value): Return NULL_TREE for SSA names we don't have a
24611 2013-02-20 Martin Jambor <mjambor@suse.cz>
24613 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24615 2013-02-20 Richard Biener <rguenther@suse.de>
24617 * genpreds.c (write_lookup_constraint): Do not compare first
24618 letter of the constraint again.
24620 2013-02-20 Richard Biener <rguenther@suse.de>
24622 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24624 (get_use_iv_cost): Terminate hashtable walk when coming across
24627 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24629 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24630 reassociation for avx2 targets.
24632 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24634 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24635 Add version check for v8.10.a to enable microblaze_has_clz
24636 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24637 version and TARGET_PATTERN_COMPARE check
24638 * config/microblaze/microblaze.md: New clzsi2 instruction
24640 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24642 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24643 function before branching.
24645 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24647 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24649 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24651 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24653 PR middle-end/55889
24654 * sel-sched.c: Include ira.h.
24655 (implicit_clobber_conflict_p): New function.
24656 (moveup_expr): Use it.
24657 * Makefile.in (sel-sched.o): Depend on ira.h.
24659 2013-02-19 Richard Biener <rguenther@suse.de>
24661 PR tree-optimization/56384
24662 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24663 (vn_hash_type): Split out from ...
24664 (vn_hash_constant_with_type): ... here.
24665 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24666 (vn_phi_eq): Compare types from vn_phi_s structure.
24667 (vn_phi_lookup): Populate vn_phi_s type.
24668 (vn_phi_insert): Likewise.
24670 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24672 PR tree-optimization/56350
24673 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24674 if haven't found reduction or nested cycle operand, rather than
24675 asserting we must find it.
24677 PR tree-optimization/56381
24678 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24681 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24682 Jakub Jelinek <jakub@redhat.com>
24685 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24686 (swap_optab_enable): Same.
24687 (init_all_optabs): Use argument instead of global.
24688 * tree.h (struct tree_optimization_option): New field target_optabs.
24689 * expr.h (init_all_optabs): Add argument to prototype.
24690 (TREE_OPTIMIZATION_OPTABS): New.
24691 (save_optabs_if_changed): Protoize.
24692 * optabs.h: Declare this_fn_optabs.
24693 * optabs.c (save_optabs_if_changed): New.
24694 Declare this_fn_optabs.
24695 (init_optabs): Add argument to init_all_optabs() call.
24696 * function.c (invoke_set_current_function_hook): Handle per
24698 * function.h (struct function): New field optabs.
24699 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24700 optimization_current_node has changed.
24701 * target-globals.h (save_target_globals_default_opts): Protoize.
24702 * target-globals.c (save_target_globals_default_opts): New.
24704 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24707 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24708 registers %fr12 and %fr12R as call used.
24711 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24712 and HImode, require all displacements to be an integer multiple of
24714 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24715 only allow QImode and HImode when reload is in progress and strict is
24716 true. Likewise for symbolic addresses. Use base14_operand to check
24717 displacements in REG+BASE addresses.
24719 2013-02-18 Richard Biener <rguenther@suse.de>
24721 PR tree-optimization/56366
24722 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24723 handle sign-conversion of outer-loop initial induction value.
24725 2013-02-18 Richard Biener <rguenther@suse.de>
24727 PR middle-end/56349
24728 * cfghooks.c (merge_blocks): If we merge a latch into another
24729 block adjust references to it.
24730 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24731 (verify_loop_structure): Verify that a recorded latch is in fact
24734 2013-02-18 Richard Biener <rguenther@suse.de>
24736 PR tree-optimization/56321
24737 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24738 order SSA name release and virtual operand unlinking.
24740 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24742 * config/microblaze/microblaze.md (save_stack_block): Define.
24743 (restore_stack_block): Likewise.
24745 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24747 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24748 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24749 * config/microblaze/microblaze.c (microblaze_option_override):
24750 Bail out early for PIC modes when target does not support PIC.
24752 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24754 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24755 Replace with a microblaze version.
24756 (microblaze_trampoline_init): Adapt for microblaze.
24757 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24760 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24761 Dodji Seketeli <dodji@redhat.com>
24764 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24765 (instrument_mem_region_access): Do not forget to always put
24766 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24767 0) statement, even for cases where either 'base' or 'base + len'
24768 are not instrumented -- because they have been previously
24769 instrumented. Simplify the logic by putting all the statements
24770 instrument 'base + len' inside a sequence, and then insert that
24771 sequence right before the current insertion point. Then, to
24772 instrument 'base + len', just get an iterator on that statement.
24773 And do not forget to update the pointer to iterator the function
24774 received as argument.
24776 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24778 PR rtl-optimization/56348
24779 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24781 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24783 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24784 (clean_graph_dump_file): Pass base to start_graph_dump.
24786 2013-02-14 Richard Henderson <rth@redhat.com>
24789 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24791 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24793 * collect2-aix.h: Define F_LOADONLY.
24795 2013-02-14 Richard Biener <rguenther@suse.de>
24798 * varasm.c (output_constant_def_1): Get the decl representing
24799 the constant as argument.
24800 (output_constant_def): Wrap output_constant_def_1.
24801 (make_decl_rtl): Use output_constant_def_1 with the decl
24802 representing the constant.
24803 (build_constant_desc): Optionally re-use a decl already
24804 representing the constant.
24805 (tree_output_constant_def): Adjust.
24807 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24810 * asan.c (instrument_builtin_call): Really put the length of the
24811 second source argument into src1_len.
24813 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24815 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24816 argument. If it is false, don't create edge from then_bb to
24818 (insert_if_then_before_iter): Pass true to it.
24819 (build_check_stmt): Pass false to it.
24820 (transform_statements): Flush hash table only on extended basic
24821 block boundaries, rather than at the beginning of every bb.
24822 Don't flush hash table on nonfreeing_call_p calls.
24823 * tree-flow.h (nonfreeing_call_p): New prototype.
24824 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24826 2013-02-13 David S. Miller <davem@davemloft.net>
24828 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24830 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24833 * ira.c (max_regno_before_ira): Move from ...
24835 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24838 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24840 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24842 2013-02-13 Richard Biener <rguenther@suse.de>
24845 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24846 globals in MEM_REFs.
24848 2013-02-13 Richard Biener <rguenther@suse.de>
24850 * loop-init.c (loop_optimizer_init): Clear loop state when
24851 re-initializing preserved loops.
24852 * loop-unswitch.c (unswitch_single_loop): Return whether
24853 we unswitched the loop. Do not verify loop state here.
24854 (unswitch_loops): When we unswitched a loop discover new loops.
24856 2013-02-13 Kostya Serebryany <kcc@google.com>
24858 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24860 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24862 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24864 Avoid instrumenting duplicated memory access in the same basic block
24865 * Makefile.in (asan.o): Add new dependency on hash-table.h
24866 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24867 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24868 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24869 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24870 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24871 (get_mem_ref_of_assignment): New functions.
24872 (get_mem_refs_of_builtin_call): Extract from
24873 instrument_builtin_call and tweak a little bit to make it fit with
24875 (instrument_builtin_call): Use the new
24876 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24877 of is_gimple_builtin_call.
24878 (instrument_derefs, instrument_mem_region_access): Insert the
24879 instrumented memory reference into the hash table.
24880 (maybe_instrument_assignment): Renamed instrument_assignment into
24881 this, and change it to advance the iterator when instrumentation
24882 actually happened and return true in that case. This makes it
24883 homogeneous with maybe_instrument_assignment, and thus give a
24884 chance to callers to be more 'regular'.
24885 (transform_statements): Clear the memory reference hash table
24886 whenever we enter a new BB, when we cross a function call, or when
24887 we are done transforming statements. Use
24888 maybe_instrument_assignment instead of instrumentation. No more
24889 need to special case maybe_instrument_assignment and advance the
24890 iterator after calling it; it's now handled just like
24891 maybe_instrument_call. Update comment.
24893 2013-02-13 Richard Biener <rguenther@suse.de>
24895 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24896 Fix loop discovery code.
24898 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24900 PR inline-asm/56148
24901 * lra-constraints.c (process_alt_operands): Match early clobber
24902 operand with itself. Check conflicts with earlyclobber only if
24903 the operand is not reloaded. Prefer to reload conflicting operand
24904 if earlyclobber and matching operands are the same.
24906 2013-02-12 Richard Biener <rguenther@suse.de>
24909 * lto-streamer-out.c (write_symbol): Do not output symbols
24910 for hard register variables.
24912 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24915 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24916 (umulsidi3_insn, mulsidi3_insn): New insns.
24918 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24920 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24921 (struct tune_params): Add vec_costs field.
24922 * config/arm/arm.c (arm_builtin_vectorization_cost)
24923 (arm_add_stmt_cost): New functions.
24924 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24925 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24926 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24927 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24928 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24929 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24930 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24932 2013-02-12 Richard Biener <rguenther@suse.de>
24935 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24936 decls again if possible.
24938 2013-02-12 Richard Biener <rguenther@suse.de>
24940 PR middle-end/56288
24941 * tree-ssa.c (verify_ssa_name): Fix check, move
24942 SSA_NAME_IN_FREE_LIST check up.
24944 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24945 Steven Bosscher <steven@gcc.gnu.org>
24947 PR rtl-optimization/56151
24948 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24949 equal to op0 or op1, and last_insn pattern is CODE operation
24950 with MEM dest and one of the operands matches that MEM.
24952 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24954 * doc/extend.texi: Document Function Multiversioning and "default"
24955 parameter string to target attribute.
24956 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24957 target attribute parameter is "default".
24958 (ix86_compare_version_priority): Remove checks for target attribute.
24959 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24960 Remove check for target attribute equal to NULL. Add assert.
24961 (ix86_generate_version_dispatcher_body): Change error to sorry.
24963 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24964 Jack Howarth <howarth@bromo.med.uc.edu>
24965 Patrick Marlier <patrick.marlier@gmail.com>
24968 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24969 define ENDFILE_SPEC as TM_DESTRUCTOR.
24970 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24972 2013-02-11 Alexander Potapenko <glider@google.com>
24973 Jack Howarth <howarth@bromo.med.uc.edu>
24974 Jakub Jelinek <jakub@redhat.com>
24977 * config/darwin.c (cdtor_record): Rename ctor_record.
24978 (sort_cdtor_records): Rename sort_ctor_records.
24979 (finalize_dtors): New routine to sort destructors by
24980 priority before use in assemble_integer.
24981 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24983 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24985 PR rtl-optimization/56275
24986 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24987 offset is non-negative and less than cmode size before
24988 calling simplify_subreg.
24990 2013-02-11 Richard Biener <rguenther@suse.de>
24992 PR tree-optimization/56264
24993 * cfgloop.h (fix_loop_structure): Adjust prototype.
24994 * loop-init.c (fix_loop_structure): Return the number of
24995 newly discovered loops.
24996 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24997 are discovered, do a full loop-closed SSA rewrite.
24999 2013-02-11 Richard Biener <rguenther@suse.de>
25001 PR tree-optimization/56273
25002 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25004 (check_array_ref): Fix missing newline in dumps.
25005 (search_for_addr_array): Likewise.
25007 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25009 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25011 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25014 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25016 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25018 PR rtl-optimization/56246
25019 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25021 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25022 constraints are satisfied.
25024 2013-02-08 Jeff Law <law@redhat.com>
25027 * emit-rtl.c (reg_is_parm_p): New function.
25028 * regs.h (reg_is_parm_p): New prototype.
25029 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25030 callee-clobbered registers.
25032 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25035 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25036 If there is no implicit builtin declaration, just return NULL.
25038 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25040 * config/i386/sse.md (FMAMODEM): New mode iterator.
25041 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25042 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25044 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25046 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25047 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25048 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25050 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25052 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25053 (microblaze*-*-elf): Likewise.
25054 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25056 * config/microblaze/microblaze-c.c: Add builtin defines for
25057 _LITTLE_ENDIAN and _BIG_ENDIAN.
25058 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25059 add to TARGET_DEFAULT flags.
25060 Expand ASM_SPEC and LINK_SPEC.
25061 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25062 * config/microblaze/microblaze.md: Update extendsidi2 and
25063 movdi_internal instructions to use low-order / high-order reg
25065 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25066 options and inversemask / mask of LITTLE_ENDIAN.
25067 * config/microblaze/t-microblaze: Expand multilib options to
25068 include mlittle-endian (le) and update exceptions patterns.
25070 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25072 PR rtl-optimization/56195
25073 * lra-constraints.c (get_reload_reg): Don't reuse regs
25074 if they have smaller mode than requested, if they have
25075 wider mode than requested, try to return a SUBREG.
25077 PR tree-optimization/56250
25078 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25079 if type is unsigned and code isn't MULT_EXPR.
25081 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25083 PR tree-optimization/56064
25084 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25085 bits according to mode.
25086 * fixed-value.h (fixed_from_double_int)
25087 (const_fixed_from_double_int): Adjust comments.
25089 2013-02-08 Richard Biener <rguenther@suse.de>
25092 * lto-streamer.h (struct data_in): Remove current_file, current_line
25093 and current_col members.
25094 * lto-streamer-out.c (lto_output_location): Stream changed bits
25095 en-block for efficiency.
25096 * lto-streamer-in.c (clear_line_info): Remove.
25097 (lto_input_location): Cache current file, line and column
25098 globally via local statics. Read changed bits en-block.
25099 (input_function): Do not call clear_line_info.
25100 (lto_read_body): Likewise.
25101 (lto_input_toplevel_asms): Likewise.
25103 2013-02-08 Michael Matz <matz@suse.de>
25105 PR tree-optimization/52448
25106 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25107 (nt_call_phase): New static.
25108 (add_or_mark_expr): Only mark accesses with newer phase than any
25110 (nonfreeing_call_p): New.
25111 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25112 (nt_fini_block): Keep blocks marked as visited.
25113 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25115 2013-02-08 Richard Biener <rguenther@suse.de>
25117 * ira.c (ira): Free broken dominator information.
25119 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25121 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25123 2013-02-08 Marek Polacek <polacek@redhat.com>
25125 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25127 2013-02-08 Richard Biener <rguenther@suse.de>
25129 PR middle-end/56181
25130 * cfgloop.h (flow_loops_find): Adjust.
25131 (bb_loop_header_p): Declare.
25132 * cfgloop.c (bb_loop_header_p): New function split out from ...
25133 (flow_loops_find): ... here. Adjust function signature,
25134 support incremental loop structure update.
25135 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25136 * cfgloopmanip.c (fix_loop_structure): Move ...
25137 * loop-init.c (fix_loop_structure): ... here.
25138 (apply_loop_flags): Split out from ...
25139 (loop_optimizer_init): ... here.
25140 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25141 in incremental mode, only remove dead loops here.
25143 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25146 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25147 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25148 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25149 (*round<mode>3.libgcc): New insns for fixed-modes.
25150 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25151 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25152 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25153 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25154 implementations. Define to __builtin_avr_absFX,
25155 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25156 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25157 __builtin_avr_countlsFX, respectively.
25158 * config/avr/avr-c.c (target.h): Include it.
25159 (enum avr_builtin_id): New enum.
25160 (avr_resolve_overloaded_builtin): New static function.
25161 (avr_register_target_pragmas): Use it to set
25162 targetm.resolve_overloaded_builtin.
25163 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25164 tree nodes used by DEF_BUILTIN.
25165 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25166 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25167 <AVR_BUILTIN_xxBITS>: Same.
25169 2013-02-08 Richard Biener <rguenther@suse.de>
25171 * cfgloop.c (verify_loop_structure): Properly handle
25172 a loop exiting to another loop header.
25173 * ira-int.h (ira_loops): Remove.
25174 * ira.c (ira_loops): Remove.
25175 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25176 (do_reload): Use loop_optimizer_finalize.
25177 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25178 number_of_loops to access the loop tree.
25179 (more_one_region_p): Likewise.
25180 (finish_loop_tree_nodes): Likewise.
25181 (rebuild_regno_allocno_maps): Likewise.
25182 (mark_loops_for_removal): Likewise.
25183 (mark_all_loops_for_removal): Likewise.
25184 (remove_unnecessary_regions): Likewise.
25185 (ira_build): Likewise.
25186 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25188 2013-02-08 Richard Biener <rguenther@suse.de>
25190 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25191 * ipa-pure-const.c (analyze_function): Avoid calling
25192 mark_irreducible_loops twice.
25193 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25195 2013-02-07 David S. Miller <davem@davemloft.net>
25197 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25199 * var-tracking.c (vt_add_function_parameter): Test the presence of
25200 HAVE_window_save properly and do not remap argument registers when
25201 we have a leaf function.
25203 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25206 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25208 * config/i386/i386.c (ix86_print_operand): Ditto.
25210 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25212 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25214 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25216 PR rtl-optimization/56225
25217 * lra-constraints.c (process_alt_operands): Check that reload hard
25218 reg can hold value for strict_low_part.
25220 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25223 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25224 dwarf2out_end_function.
25225 (in_first_function_p, maybe_at_text_label_p,
25226 first_loclabel_num_not_at_text_label): New variables.
25227 (dwarf2out_var_location): In the first function find out
25228 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25229 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25232 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25234 PR rtl-optimization/56178
25235 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25236 SUBREG of a register. Tidy up related block of code.
25237 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25238 note if the source is a register or a SUBREG of a register.
25240 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25243 * config/rs6000/rs6000.md (ptrm): New mode attr.
25244 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25245 call_value_indirect_aix<pttrsize>,
25246 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25249 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25251 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25252 if -bnortl. Convert to strcmp and strncmp.
25254 2013-02-07 Alan Modra <amodra@gmail.com>
25257 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25258 addresses won't wrap when offsetting.
25259 (rs6000_secondary_reload): Provide secondary reloads needed for
25260 wrapping LO_SUM addresses.
25262 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25264 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25265 MACH, just __MACH__.
25267 2013-02-06 Richard Biener <rguenther@suse.de>
25269 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25270 instead of calling fix_loop_structure.
25272 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25274 PR middle-end/56217
25275 * omp-low.c (use_pointer_for_field): Return false if
25276 lower_send_shared_vars doesn't generate any copy-out code.
25278 2013-02-06 Tom de Vries <tom@codesourcery.com>
25280 PR rtl-optimization/56131
25281 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25282 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25283 of the label is NULL. Add comment.
25285 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25287 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25290 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25291 (STATIC_LIBTSAN_LIBS): Likewise.
25292 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25293 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25294 is defined, don't add anything else beyond that.
25295 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25296 (LINK_COMMAND_SPEC): Use them.
25298 PR tree-optimization/56205
25299 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25300 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25301 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25303 2013-02-05 Richard Biener <rguenther@suse.de>
25305 PR tree-optimization/53342
25306 PR tree-optimization/53185
25307 * tree-vectorizer.h (vect_check_strided_load): Remove.
25308 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25309 not disallow peeling for vectorized strided loads.
25310 (vect_check_strided_load): Make static and simplify.
25311 (vect_analyze_data_refs): Adjust.
25312 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25313 correctly when vectorizing strided loads.
25315 2013-02-05 Richard Biener <rguenther@suse.de>
25317 * doc/install.texi: Refer to ISL, not PPL.
25319 2013-02-05 Jan Hubicka <jh@suse.cz>
25321 PR tree-optimization/55789
25322 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25324 2013-02-05 Jan Hubicka <jh@suse.cz>
25326 PR tree-optimization/55789
25327 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25328 the dead call anyway.
25330 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25333 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25335 2013-02-04 Alexander Potapenko <glider@google.com>
25336 Jack Howarth <howarth@bromo.med.uc.edu>
25337 Jakub Jelinek <jakub@redhat.com>
25340 * config/darwin.c (sort_ctor_records): Stabilized qsort
25341 on constructor priority by using original position.
25342 (finalize_ctors): New routine to sort constructors by
25343 priority before use in assemble_integer.
25344 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25346 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25349 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25350 about visibility on artificial decls.
25351 * config/sol2.c (solaris_assemble_visibility): Likewise.
25353 2013-02-04 Kai Tietz <ktietz@redhat.com>
25356 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25357 argument and improve checking of return-argument types for 16-byte
25359 (ix86_function_value_1): Add additional valtype argument on call
25360 of function_value_64.
25361 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25362 handling infunction_value_64 function.
25364 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25366 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25368 2013-02-04 Richard Biener <rguenther@suse.de>
25370 PR tree-optimization/56188
25371 * tree-ssa-structalias.c (label_visit): Consider case with
25372 initially non-empty points-to set.
25373 (perform_var_substitution): Dump node mapping and clean up.
25375 2013-02-04 Richard Guenther <rguenther@suse.de>
25378 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25379 node prevail as last resort.
25380 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25381 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25383 2013-02-04 Richard Biener <rguenther@suse.de>
25385 PR tree-optimization/56113
25386 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25388 (equiv_class_lookup_or_add): ... this.
25389 (label_visit): Adjust and fix error in previous patch.
25390 (perform_var_substitution): Adjust.
25392 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25394 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25395 * config/sh/sh4-300.md: Likewise.
25396 * config/sh/sh4a.md: Likewise.
25397 * config/sh/constraints.md: Likewise.
25398 * config/sh/sh.md: Likewise.
25399 * config/sh/netbsd-elf.h: Likewise.
25400 * config/sh/predicates.md: Likewise.
25401 * config/sh/sh-protos.h: Likewise.
25402 * config/sh/ushmedia.h: Likewise.
25403 * config/sh/linux.h: Likewise.
25404 * config/sh/sh.c: Likewise.
25405 * config/sh/superh.h: Likewise.
25406 * config/sh/elf.h: Likewise.
25407 * config/sh/sh4.md: Likewise.
25408 * config/sh/sh.h: Likewise.
25410 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25412 * config/pa/constraints.md: Adjust unused letters. Change "T"
25413 constraint to match_test floating_point_store_memory_operand().
25414 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25415 (base14_operand): New.
25416 (floating_point_store_memory_operand): New.
25417 (integer_store_memory_operand): Revise to use base14_operand and
25418 reg_plus_base_memory_operand.
25419 (move_dest_operand): Allow symbolic_memory_operands.
25420 (symbolic_memory_operand): Check for LO_SOM.
25421 (symbolic_operand): Change default case to break.
25422 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25423 CONST_DOUBLE values to be reloaded by putting them into memory when
25424 the destination is a floating point register.
25425 (movdf): Remove code to handle CONST_DOUBLE.
25427 (reload_indf_r1): New.
25428 (reload_insf_r1): New.
25429 Consistently use "Q" and "T" constraints with integer and floating
25430 point move instructions, respectively.
25431 (movdi): Remove FAIL.
25432 Change predicate for source operand unamed DImode move from
25433 general_operand to move_src_operand.
25434 (umulsidi3): Change predicate for destination operand to
25436 Likewise for similar unamed patterns.
25437 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25438 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25439 (hppa_legitimize_address): Simplify mask calculation.
25440 (pa_emit_move_sequence): Revised handling of secondary reloads from
25441 REG+D addresses for floating point loads and stores. Directly handle
25442 loading CONST0_RTX (mode) to a floating point register.
25443 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25444 to floating point registers. Don't restrict secondary reloads to
25445 floating point registers to integer modes. Revise some comments and
25447 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25448 (pa_legitimate_address_p): New.
25449 (pa_legitimize_reload_address): New.
25450 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25451 (STRICT_REG_OK_FOR_BASE_P): New.
25452 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25453 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25455 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25456 Andrew Dixie <andrewd@gentrack.com>
25458 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25461 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25463 * expmed.c (extract_bit_field_1): Pass the full width of the
25464 structure to get_best_reg_extraction_insn.
25466 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25469 * configure.ac (use_cxa_atexit): Add AIX.
25470 * configure: Regenerate.
25472 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25474 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25477 * final.c (need_profile_function): New variable.
25478 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25479 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25480 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25481 notes, targetm.asm_out.function_prologue doesn't emit anything,
25482 HAVE_prologue and profiler should be emitted before prologue,
25483 set need_profile_function instead of emitting it.
25484 (final_scan_insn): If need_profile_function, emit
25485 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25486 NOTE_INSN_FUNCTION_BEG note.
25488 2013-02-01 Richard Henderson <rth@redhat.com>
25490 * config/rs6000/rs6000.md (smulditi3): New.
25493 * config/alpha/alpha.md (umulditi3): New.
25495 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25497 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25498 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25500 2013-02-01 Richard Biener <rguenther@suse.de>
25502 PR tree-optimization/56113
25503 * tree-ssa-structalias.c (label_visit): Reduce work for
25504 single-predecessor nodes.
25506 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25508 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25509 range isn't testing for zero.
25511 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25513 PR middle-end/56113
25514 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25516 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25517 Nick Clifton <nickc@redhat.com>
25519 * config/v850/constraints.md (Q): Define as a memory constraint.
25520 * config/v850/predicates.md (label_ref_operand): New predicate.
25521 (e3v5_shift_operand): New predicate.
25522 (ior_operator): New predicate.
25523 * config/v850/t-v850: Add e3v5 multilib.
25524 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25525 (v850_gen_movdi): Prototype.
25526 * config/v850/v850.c: Add support for e3v5 architecture.
25527 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25529 (construct_save_jarl): Add e3v5 long JARL support.
25530 (v850_adjust_insn_length): New function. Adjust length of call
25531 insns when using e3v5 instructions.
25532 (v850_gen_movdi): New function: Generate instructions to move a
25534 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25535 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25536 (TARGET_USE_FPU): Enable for e3v5.
25537 (CONST_OK_FOR_W): New macro.
25538 (ADJUST_INSN_LENGTH): Define.
25539 * config/v850/v850.md (UNSPEC_LOOP): Define.
25540 (attr cpu): Add v850e3v5.
25541 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25542 (movdi): New pattern.
25543 (movdi_internal): New pattern.
25544 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25545 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25546 (cstoresf4): Likewise.
25547 (cstoredf4): Likewise.
25548 (insv): New pattern.
25549 (rotlso3_a): New pattern.
25550 (rotlsi3_b): New pattern
25551 (rotlsi3_v850e3v5): New pattern.
25552 (doloop_begin): New pattern.
25553 (fix_loop_counter): New pattern.
25554 (doloop_end): New pattern.
25555 (branch_normal): Add e3v5 long branch support.
25556 (branch_invert): Likewise.
25557 (branch_z_normal): Likewise.
25558 (branch_z_invert): Likewise.
25559 (branch_nz_normal): Likewise.
25560 (branch_nz_invert): Likewise.
25561 (call_internal_short): Add e3v5 register-indirect JARL support.
25562 (call_internal_long): Likewise.
25563 (call_value_internal_short): Likewise.
25564 (call_value_internal_long): Likewise.
25565 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25566 (mloop): New option.
25567 * config.gcc: Add support for configuring v840e3v5 target.
25568 * doc/invoke.texi: Document new v850 specific command line options.
25570 2013-01-31 Paul Koning <ni1d@arrl.net>
25574 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25575 children if parent is a class.
25576 (prune_unused_types_prune): Don't add DW_AT_declaration.
25578 2013-01-31 Richard Biener <rguenther@suse.de>
25580 PR tree-optimization/56157
25581 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25582 match up operand with SLP child.
25584 2013-01-31 Jason Merrill <jason@redhat.com>
25587 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25588 parameters the first time.
25589 (gen_scheduled_generic_parms_dies): Check completeness here.
25591 2013-01-31 Richard Biener <rguenther@suse.de>
25593 PR middle-end/53073
25594 * common.opt (faggressive-loop-optimizations): New flag,
25595 enabled by default.
25596 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25597 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25598 infer_loop_bounds_from_undefined by it.
25600 2013-01-31 Richard Biener <rguenther@suse.de>
25602 PR tree-optimization/56150
25603 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25604 visit virtual operands.
25605 (find_uses_to_rename_bb): Likewise.
25607 2013-01-31 Richard Biener <rguenther@suse.de>
25609 PR tree-optimization/56150
25610 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25611 mixed store non-store stmts.
25613 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25616 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25617 LIBASAN_EARLY_SPEC is defined.
25618 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25619 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25621 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25624 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25625 invalid args instead of ICEing on it.
25626 (ix86_valid_target_attribute_tree): Return error_mark_node if
25627 ix86_valid_target_attribute_inner_p failed.
25628 (ix86_valid_target_attribute_p): Return false only if
25629 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25630 target("default") attribute.
25631 (sorted_attr_string): Change argument from const char * to tree,
25632 merge in all target attribute arguments rather than just one.
25633 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25634 instead of free. Avoid using strcat.
25635 (ix86_mangle_function_version_assembler_name): Mangle
25636 target("default") as if no target attribute is present. Adjust
25637 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25638 instead of xmalloc and XDELETEVEC instead of free.
25639 (ix86_function_versions): Don't return true if one of the decls
25640 doesn't have target attribute. If they don't and one of the decls
25641 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25642 sorted_attr_string caller. Use XDELETEVEC instead of free.
25643 (ix86_supports_function_versions): Remove.
25644 (make_name): Fix up formatting.
25645 (make_dispatcher_decl): Remove resolver_name and its initialization.
25646 Avoid leaking memory.
25647 (is_function_default_version): Return true if there is
25648 target("default") attribute rather than no target attribute at all.
25649 (make_resolver_func): Avoid leaking memory.
25650 (ix86_generate_version_dispatcher_body): Likewise.
25651 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25652 * target.def (supports_function_versions): Remove.
25653 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25654 * doc/tm.texi: Regenerated.
25656 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25658 PR rtl-optimization/56144
25659 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25660 for values with side effects.
25662 2013-01-30 Richard Biener <rguenther@suse.de>
25664 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25665 (sparseset_pop): Likewise.
25666 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25667 to be able to use quick_push in the worker loop.
25669 2013-01-30 Marek Polacek <polacek@redhat.com>
25671 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25673 2013-01-30 Richard Biener <rguenther@suse.de>
25676 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25678 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25680 PR tree-optimization/56064
25681 * fixed-value.c (fixed_from_double_int): New function.
25682 * fixed-value.h (fixed_from_double_int): New prototype.
25683 (const_fixed_from_double_int): New static inline function.
25684 * fold-const.c (native_interpret_fixed): New static function.
25685 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25686 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25687 (native_encode_fixed): New static function.
25688 (native_encode_expr) <FIXED_CST>: Use it.
25689 (native_interpret_int): Move double_int worker code to...
25690 * double-int.c (double_int::from_buffer): ...this new static method.
25691 * double-int.h (double_int::from_buffer): Prototype it.
25693 2013-01-30 Richard Biener <rguenther@suse.de>
25695 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25696 New pointer-map and obstack.
25697 (init_alias_vars): Allocate pointer-map and obstack.
25698 (delete_points_to_sets): Free them.
25699 (find_what_var_points_to): Cache result.
25700 (find_what_p_points_to): Adjust for changed interface of
25701 find_what_var_points_to.
25702 (compute_points_to_sets): Likewise.
25703 (ipa_pta_execute): Likewise.
25705 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25707 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25708 * configure: Regenerate.
25709 * config.in: Regenerate.
25710 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25711 #nobits/#progbits if supported.
25713 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25716 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25717 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25718 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25720 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25722 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25723 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25725 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25727 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25728 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25730 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25732 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25734 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25735 * config/arm/cortex-a7.md: New bypasses using
25736 arm_mac_accumulator_is_result.
25738 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25740 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25741 (cortex_a7_neon_mla): Likewise.
25742 (cortex_a7_fpfmad): New reservation.
25743 (cortex_a7_fpmacs): Use ffmas and update required units.
25744 (cortex_a7_fpmuld): Update required units and latency.
25745 (cortex_a7_fpmacd): Likewise.
25746 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25747 (cortex_a7_neon). Likewise.
25748 (bypass) Update participating units.
25750 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25752 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25753 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25755 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25756 (vfp_fmul): Use ffmad.
25757 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25758 (cortex_r4_fmacd): Use ffmad.
25759 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25760 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25761 (cortex_a9_fmacd): Use ffmad.
25762 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25763 (cortex_a8_vfp_macd): Use ffmad.
25764 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25765 (cortex_a5_fpmacd): Use ffmad.
25766 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25767 (cortex_a15_vfp_macd): Use ffmad.
25768 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25770 2013-01-29 Jason Merrill <jason@redhat.com>
25773 * varasm.c (default_assemble_visibility): Don't warn about
25774 visibility on artificial decls.
25776 2013-01-29 Richard Biener <rguenther@suse.de>
25778 PR tree-optimization/56113
25779 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25781 (label_visit): Free duplicate bitmaps and record the leader instead.
25782 (perform_var_substitution): Adjust.
25784 2013-01-29 Richard Biener <rguenther@suse.de>
25786 PR tree-optimization/55270
25787 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25788 the CFG, schedule loops for fixup.
25790 2013-01-29 Nick Clifton <nickc@redhat.com>
25792 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25795 2013-01-28 Leif Ekblad <leif@rdos.net>
25797 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25798 * config/i386/i386.h (TARGET_RDOS): New macro.
25799 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25800 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25801 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25802 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25803 DEFAULT_LARGE_SECTION_THRESHOLD.
25804 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25805 * config/i386/rdos.h: New file.
25806 * config/i386/rdos64.h: New file.
25808 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25811 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25814 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25816 PR rtl-optimization/56117
25817 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25818 call cselib_lookup_from_insn on the MEM before calling
25819 add_insn_mem_dependence.
25821 2013-01-28 Richard Biener <rguenther@suse.de>
25823 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25824 to a stmt that didn't have one.
25825 (copy_phis_for_bb): Likewise for PHI arguments.
25826 (copy_debug_stmt): Likewise for debug stmts.
25828 2013-01-28 Richard Biener <rguenther@suse.de>
25830 PR tree-optimization/56034
25831 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25832 (partition_builtin_p): Adjust.
25833 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25834 it is the last partition.
25835 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25836 up the vertex for the definition.
25837 (classify_partition): Classify whether a partition is a
25838 PKIND_REDUCTION, thus has uses outside of the loop.
25839 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25840 Merge all PKIND_REDUCTION partitions into the last partition.
25841 (tree_loop_distribution): Seed partitions from reductions as well.
25843 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25845 PR tree-optimization/56125
25846 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25847 pow(x,c) into sqrt(x) * powi(x, n/2) or
25848 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25849 optimizing for size.
25850 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25851 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25854 PR tree-optimization/56094
25855 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25856 to UNKNOWN_LOCATION while gimplifying expr.
25858 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25861 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25862 operand 0 in movabs insn template for -masm=intel asm alternative.
25863 (*movabs<mode>_2): Ditto for operand 1.
25865 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25868 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25871 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25873 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25876 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25878 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25879 Undef to avoid warning.
25881 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25883 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25884 * configure: Regenerate.
25886 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25888 PR tree-optimization/56098
25889 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25890 for stmts with volatile ops.
25891 (cond_store_replacement): Don't optimize if assign has volatile ops.
25892 (cond_if_else_store_replacement_1): Don't optimize if either
25893 then_assign or else_assign have volatile ops.
25894 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25897 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25899 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25901 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25903 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25904 missing ':' in asm example.
25906 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25908 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25909 entries into lane and laneq entries.
25910 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25911 Remove AdvSIMD scalar modes.
25912 (aarch64_sq<r>dmulh_laneq<mode>): New.
25913 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25915 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25916 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25917 * config/aarch64/iterators.md (VCOND): New.
25920 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25923 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25924 Add NULL LIBNAME argument to existing definitions.
25925 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25926 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25927 * config/avr/avr.c (DEF_BUILTIN): Same.
25928 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25929 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25930 implementation is available (DECL_ASSEMBLER_NAME is set).
25931 (avr_fold_absfx): New static function.
25932 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25933 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25934 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25935 AVR_BUILTIN_ABSLLK.
25936 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25937 (abshk, absk, abslk, absllk): Provide as static inline functions.
25939 2013-01-25 Marek Polacek <polacek@redhat.com>
25941 PR tree-optimization/56035
25942 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25944 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25946 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25947 (*movtf_internal_rex64): Add (!o,C) alternative
25948 (*movxf_internal_rex64): Ditto.
25949 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25951 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25953 * doc/invoke.texi: fix typo.
25954 * doc/objc.texi: fix typo.
25956 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25958 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25959 for the first two alternatives.
25961 2013-01-24 Diego Novillo <dnovillo@google.com>
25963 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25964 (ggc-zone.o): Remove.
25965 * configure.ac: Remove option --with-gc.
25966 * configure: Re-generate.
25967 * doc/install.texi: Remove documentation for --with-gc.
25968 * gengtype.c (write_enum_defn): Remove. Update all users.
25969 (write_Types_process_field): Remove generation of gt_e_* argument.
25970 (output_type_enum): Remove. Update all users.
25971 (write_enum_defn): Remove. Update all users.
25972 (enum alloc_zone): Remove. Update all users.
25973 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25974 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25975 Update all callers.
25976 (struct ptr_data): Remove field TYPE. Update all users.
25977 (gt_pch_note_object): Remove argument TYPE. Update all users.
25978 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25980 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25981 (struct alloc_zone): Remove.
25982 (ggc_internal_alloc_zone_stat): Remove.
25983 (ggc_internal_cleared_alloc_zone_stat): Remove.
25984 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25985 (ggc_pch_count_object): Remove last argument. Update all users.
25986 (ggc_pch_alloc_object): Remove last argument. Update all users.
25987 (struct alloc_zone): Remove.
25988 * ggc-zone.c: Remove.
25989 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25990 (struct alloc_zone): Remove.
25991 (ggc_alloc_typed_stat): Remove.
25992 (ggc_alloc_typed): Remove.
25993 (ggc_splay_alloc): Remove first argument.
25994 (rtl_zone): Remove. Update all users.
25995 (tree_zone): Remove. Update all users.
25996 (tree_id_zone): Remove. Update all users.
25997 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25998 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25999 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26000 * tree-ssanames.c: Remove references to zone allocator in comments.
26002 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26004 * config/avr/avr.c (avr_out_fract): Make register numbers that
26005 might be outside of source operand signed.
26007 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26009 * config/i386/constraints.md (Yf): New constraint.
26010 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26011 of f constraint to conditionaly disable x87 register preferences.
26012 (*movdf_internal): Ditto.
26013 (*movsf_internal): Ditto.
26015 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26017 PR inline-asm/55934
26018 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26019 that have operands with impossible constraints.
26020 Add a FIXME for a speed-up opportunity.
26021 * lra-constraints.c (process_alt_operands): Verify that a class
26022 selected from constraints on asms is valid for the operand mode.
26023 (curr_insn_transform): Remove incorrect comment.
26025 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26027 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26028 TOC operand is a valid symbol ref in the constant pool.
26030 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26032 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26034 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26037 * config/avr/stdfix.h: New file.
26038 * t-avr (stdfix-gcc.h): New rule to build it.
26039 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26041 2013-01-23 Kostya Serebryany <kcc@google.com>
26043 * config/darwin.h: remove dependency on
26044 CoreFoundation (asan on Mac OS).
26046 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26049 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26050 instead of cmpdi_operand for first comparison operand.
26051 Don't assert that comparison operands aren't both constants.
26053 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26055 * doc/install.texi (Downloading the Source): Update references to
26056 downloading separate components.
26058 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26060 * doc/extend.texi (__int128): Improve grammar.
26062 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26065 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26066 alternative to (o,r).
26067 (*movdi_internal_rex64): Remove (!o,n) alternative.
26068 (DImode immediate->memory splitter): Remove.
26069 (DImode immediate->memory peephole2): Remove.
26070 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26071 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26072 alternative to (!o,*r).
26073 (*movtf_internal_sse): New pattern.
26074 (*movxf_internal_rex64): New pattern.
26075 (*movxf_internal): Disable for TARGET_64BIT.
26076 (*movdf_internal_rex64): Remove (!o,F) alternative.
26078 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26080 PR middle-end/56074
26081 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26082 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26083 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26084 stmt locations where LOCATION_LOCUS of the stmt location is
26085 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26088 * config/i386/i386.md (UNSPEC_STOS): New.
26089 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26090 *strsetqi_1): Add UNSPEC_STOS.
26092 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26095 * doc/invoke.texi: Remove left over -Wsynth example.
26097 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26099 PR tree-optimization/56051
26100 * fold-const.c (fold_binary_loc): Don't fold
26101 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26102 a narrowing conversion, or widening conversion from signed
26105 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26107 PR rtl-optimization/56023
26108 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26109 dependent on debug instruction.
26111 2013-01-21 Martin Jambor <mjambor@suse.cz>
26113 PR middle-end/56022
26114 * function.c (allocate_struct_function): Call
26115 invoke_set_current_function_hook earlier.
26117 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26119 * reload1.c (init_reload): Only initialize reload_obstack
26120 during the first call.
26122 2013-01-21 Marek Polacek <polacek@redhat.com>
26124 * cfgloop.c (verify_loop_structure): Fix up grammar.
26126 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26128 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26129 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26131 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26134 * config/arm/marvell-pj4.md: Update copyright year.
26135 Fix up use of alu to alu_reg and simple_alu_imm.
26137 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26139 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26141 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26144 * lra-constraints.c (curr_insn_transform): Don't reuse original
26145 insn for secondary memory move when memory mode should be different.
26147 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26149 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26150 atomic_storedi_1): New patterns.
26152 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26154 btver2 pipeline descriptions.
26155 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26157 * config/i386/i386.md (btver2_decode): New type attributes.
26158 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26160 * config/i386/btver2.md: New file describing btver2 pipelines.
26162 2013-01-19 Andrew Pinski <apinski@cavium.com>
26164 PR tree-optimization/52631
26165 * tree-ssa-sccvn (visit_use): Before looking up the original
26166 statement, try looking up the simplified expression.
26168 2013-01-19 Anthony Green <green@moxielogic.com>
26170 * config/moxie/moxie.c (moxie_expand_prologue): Set
26171 current_function_static_stack_size.
26173 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26175 PR tree-optimization/56029
26176 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26177 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26179 2013-01-18 Sharad Singhai <singhai@google.com>
26181 PR tree-optimization/55995
26182 * dumpfile.c (dump_loc): Print location only if available.
26183 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26185 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26188 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26189 secondary memory move.
26190 (inherit_reload_reg): Use rclass instead of cl for
26191 check_secondary_memory_needed_p.
26193 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26195 PR middle-end/56015
26196 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26197 the case where writing real complex part of target modifies op1.
26199 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26201 * config/aarch64/aarch64-simd.md
26202 (aarch64_vcond_internal<mode>): Handle unordered cases.
26203 * config/aarch64/iterators.md (v_cmp_result): New.
26205 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26206 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26208 * config/arm/marvell-pj4.md: New file.
26209 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26210 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26211 (generic_vfp): Likewise.
26212 * config/arm/arm-cores.def: Add marvell-pj4.
26213 * config/arm/arm-tune.md: Regenerate.
26214 * config/arm/arm-tables.opt: Regenerate.
26215 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26216 * doc/invoke.texi: Document marvell-pj4.
26218 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26220 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26222 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26227 * var-tracking.c (negative_power_of_two_p): New.
26228 (global_get_addr_cache, local_get_addr_cache): New.
26229 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26230 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26232 (vt_stack_offset_p): Remove.
26233 (vt_canon_true_dep): Always canonicalize loc's address.
26234 (clobber_overlapping_mems): Make sure we have a MEM.
26235 (local_get_addr_clear_given_value): New.
26236 (val_reset): Clear local cached entries.
26237 (compute_bb_dataflow): Create and release the local cache.
26238 Disable duplicate MEMs clobbering.
26239 (emit_notes_in_bb): Clobber MEMs likewise.
26240 (vt_emit_notes): Create and release the local cache.
26241 (vt_initialize, vt_finalize): Create and release the global
26242 cache, respectively.
26243 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26245 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26247 PR libmudflap/53359
26248 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26249 not found in the symtab.
26251 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26254 PR rtl-optimization/55547
26255 PR rtl-optimization/53827
26258 * alias.c (offset_overlap_p): New, factored out of...
26259 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26260 the conservative special case for symbolic constants. Don't
26261 adjust zero sizes on alignment.
26263 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26265 PR rtl-optimization/52573
26266 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26267 REG_UNUSED for the same register.
26269 2013-01-17 Richard Biener <rguenther@suse.de>
26270 Marek Polacek <polacek@redhat.com>
26272 PR rtl-optimization/55833
26273 * loop-unswitch.c (unswitch_loops): Move loop verification...
26274 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26275 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26276 Set it to true when we're removing a loop from hierarchy tree in
26277 an irreducible region.
26278 (fix_bb_placements): Adjust caller.
26279 (fix_loop_placements): Likewise.
26281 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26283 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26284 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26285 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26286 Remove ID. Adjust comments.
26287 * config/avr/avr-c.c (avr_builtin_name): Remove.
26288 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26289 * config/avr/avr.c (avr_tolower): New static function.
26290 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26291 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26292 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26295 2013-01-17 Jan Hubicka <jh@suse.cz>
26297 PR tree-optimization/55273
26298 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26300 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26303 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26304 store through atomic_store<mode>_1.
26305 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26307 2013-01-17 Martin Jambor <mjambor@suse.cz>
26309 PR tree-optimizations/55264
26310 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26311 for virtual methods.
26312 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26313 virtual methods before inlining is over.
26314 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26316 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26319 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26321 PR rtl-optimization/56005
26322 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26323 pending reads for prefetch.
26325 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26327 * config/aarch64/aarch64.md
26328 (*cstoresi_neg_uxtw): New pattern.
26329 (*cmovsi_insn_uxtw): New pattern.
26330 (*<optab>si3_uxtw): New pattern.
26331 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26332 (*<optab>si3_insn_uxtw): New pattern.
26333 (*bswapsi2_uxtw): New pattern.
26335 2013-01-16 Richard Biener <rguenther@suse.de>
26337 * tree-inline.c (tree_function_versioning): Remove set but
26338 never used variable.
26340 2013-01-16 Richard Biener <rguenther@suse.de>
26342 PR tree-optimization/55964
26343 * tree-flow.h (rename_variables_in_loop): Remove.
26344 (rename_variables_in_bb): Likewise.
26345 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26346 (copy_loop_before): Adjust and delete update-ssa status.
26347 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26348 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26349 (rename_variables_in_loop): Remove.
26350 (slpeel_update_phis_for_duplicate_loop): Likewise.
26351 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26352 use available cfg machinery instead of duplicating it.
26353 Update PHI nodes and perform poor-mans SSA update here.
26354 (slpeel_tree_peel_loop_to_edge): Adjust.
26356 2013-01-16 Richard Biener <rguenther@suse.de>
26358 PR tree-optimization/54767
26359 PR tree-optimization/53465
26360 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26361 (vrp_visit_phi_node): For PHI arguments coming via backedges
26362 drop all symbolical range information.
26363 (execute_vrp): Compute backedges.
26365 2013-01-16 Richard Biener <rguenther@suse.de>
26367 * doc/install.texi: Update CLooG and ISL requirements to
26370 2013-01-16 Christian Bruel <christian.bruel@st.com>
26373 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26374 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26375 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26377 2013-01-16 DJ Delorie <dj@redhat.com>
26379 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26380 (UNSPECV_SP_SWITCH_E): New.
26381 (sp_switch_1): Change to an unspec.
26382 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26385 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26387 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26388 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26389 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26390 (expand_mem_thread_fence): Ditto.
26391 (expand_mem_signal_fence): Ditto.
26392 (expand_atomic_load): Ditto.
26393 (expand_atomic_store): Ditto.
26395 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26397 PR rtl-optimization/55547
26398 PR rtl-optimization/53827
26401 * alias.c (memrefs_conflict_p): Set sizes to negative after
26404 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26407 * function.c (thread_prologue_and_epilogue_insns): Always
26408 add crtl->drap_reg to set_up_by_prologue.set, even if
26409 stack_realign_drap is false.
26411 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26413 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26414 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26415 *call): Fix indention.
26417 2013-01-15 Tom de Vries <tom@codesourcery.com>
26420 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26423 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26425 PR rtl-optimization/55153
26426 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26428 2013-01-15 Martin Jambor <mjambor@suse.cz>
26430 PR tree-optimization/55920
26431 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26432 accesses as grp_to_be_debug_replaced.
26434 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26436 PR tree-optimization/55920
26437 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26438 there is non-useless type conversion needed from debug rhs to lhs,
26439 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26441 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26442 Mikael Pettersson <mikpe@it.uu.se>
26445 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26447 (ASM_OUTPUT_CASE_LABEL): Remove.
26448 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26449 * final.c (shorten_branches): Update alignment of labels before
26450 jump tables if CASE_VECTOR_SHORTEN_MODE.
26452 2013-01-15 Richard Biener <rguenther@suse.de>
26455 * system.h: Do not include gmp.h for building host tools.
26457 2013-01-15 Richard Biener <rguenther@suse.de>
26459 PR middle-end/55882
26460 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26461 account for bitpos when computing alignment.
26463 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26465 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26466 (ix86_target_macros_internal): Likewise.
26468 * config/i386/i386.c (m_CORE2I7): Removed.
26469 (m_CORE_HASWELL): New macro.
26470 (m_CORE_ALL): Likewise.
26471 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26472 (initial_ix86_arch_features): Likewise.
26473 (processor_target_table): Initializations for Core avx2.
26474 (cpu_names): New names "core-avx2".
26475 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26476 PROCESSOR_CORE_HASWELL.
26477 (ix86_issue_rate): New case.
26478 (ia32_multipass_dfa_lookahead): Likewise.
26479 (ix86_sched_init_global): Likewise.
26481 * config/i386/i386.h (TARGET_HASWELL): New macro.
26482 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26483 (processor_type): New PROCESSOR_HASWELL.
26485 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26487 PR tree-optimization/55955
26488 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26489 *SHIFT_EXPR and *ROTATE_EXPR codes.
26491 PR tree-optimization/48766
26492 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26493 -ftrapv disable -fwrapv.
26495 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26498 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26499 etc. to 1 and not to __flash.
26500 Use LL suffix for __INT24_MAX__ with -mint8.
26501 Use ULL suffix for __UINT24_MAX__ with -mint8.
26503 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26505 * config/avr/avr-arch.h
26506 (struct base_arch_s): Use typedef avr_arch_t instead.
26507 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26508 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26509 * config/avr/avr.c: Same.
26510 * config/avr/avr-devices.c: Same.
26511 * config/avr/driver-avr.c: Same.
26512 * config/avr/gen-avr-mmcu-texi.c: Same.
26513 * config/avr/avr-mcus.def: Adjust comment.
26515 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26517 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26518 * config/aarch64/iterators.md (VALLDI): New.
26520 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26521 Andi Kleen <ak@linux.intel.com>
26524 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26525 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26528 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26530 * config/avr/avr-stdint.h: Remove trailing blanks.
26531 * config/avr/avr-log.h: Same.
26532 * config/avr/avr-arch.h: Same.
26533 * config/avr/avr-devices.c: Same.
26534 * config/avr/avr-dimode.md: Same.
26535 * config/avr/predicates.md: Same.
26536 * config/avr/avr-c.c: Same. And fix typo.
26538 * config/avr/avr-protos.h: Same. And:
26539 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26540 (init_cumulative_args): Rename to avr_init_cumulative_args.
26541 (expand_prologue): Rename to avr_expand_prologue.
26542 (expand_epilogue): Rename to avr_expand_epilogue.
26543 (adjust_insn_length): Rename to avr_adjust_insn_length.
26544 (notice_update_cc): Rename to avr_notice_update_cc.
26545 (final_prescan_insn): Rename to avr_final_prescan_insn.
26546 * config/avr/avr.c: Same.
26547 * config/avr/avr.h: Same.
26548 * config/avr/avr.md: Remove trailing blanks.
26549 (prologue): Use avr_expand_prologue.
26550 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26552 2013-01-14 Richard Biener <rguenther@suse.de>
26554 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26555 verify_location, collect_subblocks): New functions.
26556 (verify_gimple_in_cfg): Verify that locations only reference
26557 BLOCKs in the functions BLOCK tree.
26559 2013-01-14 Richard Biener <rguenther@suse.de>
26561 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26563 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26565 (insert_out_of_ssa_copy_on_edge): Likewise.
26566 (rewrite_close_phi_out_of_ssa): Likewise.
26567 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26569 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26570 propagated constants.
26571 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26574 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26576 * config/avr/avr-modes.def: Add GPL copyright notice.
26578 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26580 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26581 MEMMODEL_MASK to determine memory model.
26582 (atomic_store<mode>): Ditto from operands[2].
26583 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26585 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26588 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26589 (fold_gimple_assign): Don't call unshare_expr here.
26590 (fold_ctor_reference): Call unshare_expr.
26592 2013-01-13 Terry Guo <terry.guo@arm.com>
26594 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26595 * doc/fragments.texi: Document MULTILIB_REUSE.
26596 * gcc.c (multilib_reuse): New internal spec.
26597 (set_multilib_dir): Also search multilib from multilib_reuse.
26598 * genmultilib (tmpmultilib3): Refactor code.
26599 (tmpmultilib4): Ditto.
26600 (multilib_reuse): New multilib argument.
26602 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26604 * Makefile.in: Update copyright.
26606 2013-01-12 Tom de Vries <tom@codesourcery.com>
26608 PR middle-end/55890
26609 * calls.c (expand_call): Check if arg_nr is valid.
26611 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26613 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26614 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26615 documentation. Add missing '__' in front of
26616 __builtin_ia32_packssdw256.
26618 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26621 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26622 NO_REGS for larl operands.
26623 (s390_reload_larl_operand): Use s390_load_address instead of
26626 2013-01-11 Richard Biener <rguenther@suse.de>
26628 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26629 (verify_node_sharing): ... here.
26630 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26632 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26634 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26635 Substitute TREECHECKING.
26636 * configure: Regenerate.
26637 * Makefile.in (TREECHECKING): New.
26639 2013-01-11 Richard Guenther <rguenther@suse.de>
26641 PR tree-optimization/44061
26642 * tree-vrp.c (extract_range_basic): Compute zero as
26643 value-range for __builtin_constant_p of function parameters.
26645 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26647 Update copyright years.
26649 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26651 PR rtl-optimization/55672
26652 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26653 const to be eliminable.
26655 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26657 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26658 * configure: Regenerate.
26660 2013-01-10 Richard Biener <rguenther@suse.de>
26662 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26664 2013-01-10 Richard Biener <rguenther@suse.de>
26667 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26668 locations for virtual PHI arguments.
26669 (rewrite_update_phi_arguments): Likewise.
26671 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26673 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26676 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26678 PR tree-optimization/55921
26679 * tree-complex.c (expand_complex_asm): New function.
26680 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26682 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26685 * config/s390/s390.c (s390_symref_operand_p)
26686 (s390_loadrelative_operand_p): Merge the two functions.
26687 (s390_check_qrst_address, print_operand_address): Add parameters
26688 to s390_loadrelative_operand_p invokation.
26689 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26690 (s390_reload_larl_operand, s390_secondary_reload): Use
26691 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26692 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26694 2013-01-09 Mike Stump <mikestump@comcast.net>
26696 * dse.c (record_store): Remove unnecessary assert.
26698 2013-01-09 Jan Hubicka <jh@suse.cz>
26700 PR tree-optimization/55569
26701 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26702 * cfgloop.h (scale_loop_profile): Likewise.
26704 2013-01-09 Jan Hubicka <jh@suse.cz>
26707 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26709 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26711 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26713 PR middle-end/55114
26714 * expr.h (maybe_emit_group_store): Declare.
26715 * expr.c (maybe_emit_group_store): New function.
26716 * builtins.c (expand_builtin_int_roundingfn): Call it.
26717 (expand_builtin_int_roundingfn_2): Likewise.
26719 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26721 PR rtl-optimization/55829
26722 * lra-constraints.c (match_reload): Add code for absent output.
26723 (curr_insn_transform): Add code for reloads of matched inputs
26726 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26728 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26729 attribute of movddup insn to DF.
26730 (*vec_interleave_lowv2df): Ditto.
26731 (vec_dupv2df): Ditto.
26733 2013-01-09 Jan Hubicka <jh@suse.cz>
26735 PR tree-optimiation/55875
26736 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26737 EVERY_ITERATION parameter.
26738 (number_of_iterations_exit): Check if exit is executed every iteration.
26739 (idx_infer_loop_bounds): Similarly here.
26740 (n_of_executions_at_most): Simplify
26741 to only test for cases where statement is dominated by the
26742 particular bound; handle correctly the "postdominance" test.
26743 (scev_probably_wraps_p): Use max loop iterations info
26744 as a global bound first.
26746 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26747 Nick Clifton <nickc@redhat.com>
26749 * config/v850/v850.md (cbranchsf4): New pattern.
26750 (cstoresf4): New pattern.
26751 (cbranchdf4): New pattern.
26752 (cstoredf4): New pattern.
26753 (movsicc): Disallow floating point comparisons.
26754 (cmpsf_le_insn): Fix order of operators.
26755 (cmpsf_lt_insn): Likewise.
26756 (cmpsf_eq_insn): Likewise.
26757 (cmpdf_le_insn): Likewise.
26758 (cmpdf_lt_insn): Likewise.
26759 (cmpdf_eq_insn): Likewise.
26760 (cmpsf_ge_insn): Use LE comparison.
26761 (cmpdf_ge_insn): Likewise.
26762 (cmpsf_gt_insn): Use LT comparison.
26763 (cmpdf_gt_insn): Likewise.
26764 (cmpsf_ne_insn): Delete pattern.
26765 (cmpdf_ne_insn): Delete pattern.
26766 * config/v850/v850.c (v850_gen_float_compare): Use
26767 gen_cmpdf_eq_insn for NE comparison.
26768 (v850_float_z_comparison_operator)
26769 (v850_float_nz_comparison_operator): Move from here ...
26770 * config/v850/predicates.md: ... to here. Move GT and GE
26771 comparisons into v850_float_z_comparison_operator.
26772 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26774 (v850_float_nz_comparison_operator): Likewise.
26776 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26778 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26779 with calls to gen_insvsi/gen_insvdi.
26781 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26783 * config/i386/i386.c (initial_ix86_tune_features): Set up
26784 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26786 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26787 Jakub Jelinek <jakub@redhat.com>
26789 PR tree-optimization/48189
26790 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26791 If nitercst is 0, don't predict the exit edge.
26793 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26795 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26796 in asm_fprintf with reg_names.
26797 (aarch64_print_operand_address): Likewise.
26798 (aarch64_return_addr): Likewise.
26799 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26801 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26803 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26804 (INT_U6_BITS): Likewise.
26805 * config/pa/predicates.md (uint6_operand): New predicate.
26806 (shift5_operand, shift6_operand): Likewise.
26807 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26809 (lshrdi3): Use shift6_operand.
26810 (shrpsi4, shrpdi4): New insn patterns.
26811 (extzv): Delete expander.
26812 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26813 predicates in unamed zero extract patterns. Tighten common constraint.
26814 (extv): Delete expander.
26815 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26816 predicates in unamed sign extract patterns. Tighten common constraint.
26817 (insv): Delete expander.
26818 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26819 predicates in unamed insert patterns. Tighten common constraint.
26820 Change uint32_operand predicate to uint6_operand predicate in unamed
26821 DImode pattern to insert constant values of type 1...1xxxx.
26823 2013-01-04 Jan Hubicka <jh@suse.cz>
26825 PR tree-optimization/55823
26826 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26829 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26830 Uros Bizjak <ubizjak@gmail.com>
26832 PR rtl-optimization/55845
26833 * df-problems.c (can_move_insns_across): Stop scanning at
26834 volatile_insn_p source instruction or give up if
26835 across_from .. across_to range contains any volatile_insn_p
26838 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26840 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26841 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26843 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26844 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26846 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26849 * asan.c (asan_clear_shadow): New function.
26850 (asan_emit_stack_protection): Use it.
26852 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26854 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26855 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26856 with tab instead of space.
26858 2013-01-08 Nick Clifton <nickc@redhat.com>
26860 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26861 register bank 0 at the start of an interrupt handler.
26862 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26865 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26867 * config/aarch64/aarch64-simd.md
26868 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26869 (aarch64_simd_bsl): Likewise.
26870 (aarch64_vcond_internal<mode>): Likewise.
26871 (vcond<mode><mode>): Likewise.
26872 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26873 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26875 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26877 * config/aarch64/aarch64-builtins.c
26878 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26880 2013-01-08 Martin Jambor <mjambor@suse.cz>
26883 * tree-sra.c (analyze_access_subtree): Return true also after
26884 potentially creating a debug-only replacement.
26886 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26888 PR middle-end/55890
26889 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26891 PR tree-optimization/54120
26892 * tree-vrp.c (range_fits_type_p): Don't allow
26893 src_precision < precision from signed vr to unsigned_p
26894 if vr->min or vr->max is negative.
26895 (simplify_float_conversion_using_ranges): Test can_float_p
26896 against CODE_FOR_nothing.
26898 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26899 Richard Biener <rguenther@suse.de>
26901 PR middle-end/55851
26902 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26903 types instead of just INTEGER_TYPE types.
26905 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26907 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26910 2013-01-07 Steve Ellcey <sellcey@mips.com>
26913 * config/mips/mips.opt: Change mad to mmad to match documentation.
26915 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26918 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26919 .progmemx.data now.
26921 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26924 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26925 (avr_addrspace_t): Add .section_name field.
26926 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26928 (avr_addrspace): Same. Initialize .section_name. Remove last
26929 NULL entry. Put __memx into .progmemx.data.
26930 (progmem_section_prefix): Remove.
26931 (avr_asm_init_sections): No need to initialize progmem_section.
26932 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26933 section name prefix.
26934 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26935 retrieve the progmem section.
26936 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26937 boundary to run over avr_addrspace[].
26938 (avr_register_target_pragmas): Ditto.
26940 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26942 * varasm.c (output_constant_def_contents): For asan_protect_global
26943 protected strings, adjust DECL_ALIGN if needed, before testing for
26945 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26946 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26948 (output_object_block): For asan protected decls, emit asan padding
26949 after their contents.
26950 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26951 (asan_finish_file): Test it here instead.
26953 2013-01-07 Nick Clifton <nickc@redhat.com>
26954 Matthias Klose <doko@debian.org>
26955 Doug Kwan <dougkwan@google.com>
26956 H.J. Lu <hongjiu.lu@intel.com>
26959 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26961 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26963 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26965 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26967 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26969 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26972 * doc/install.texi (Cross-Compiler-Specific Options): Document
26975 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26977 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26978 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26979 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26980 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26981 vqmovun_high_s64): Fix source operand number and update copyright.
26983 2013-01-07 Richard Biener <rguenther@suse.de>
26985 PR middle-end/55890
26986 * gimple.h (gimple_call_builtin_p): New overload.
26987 * gimple.c (validate_call): New function.
26988 (gimple_call_builtin_p): Likewise.
26989 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26990 Use gimple_call_builtin_p.
26991 (find_func_clobbers): Likewise.
26992 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26993 (strlen_optimize_stmt): Likewise.
26995 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26997 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26998 (vld1q_dup_*): Likewise.
26999 (vld1_*): Likewise.
27000 (vld1q_*): Likewise.
27001 (vld1_lane_*): Likewise.
27002 (vld1q_lane_*): Likewise.
27004 2013-01-07 Richard Biener <rguenther@suse.de>
27006 * lto-streamer.h (LTO_minor_version): Bump to 2.
27008 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27010 * config/aarch64/aarch64-protos.h
27011 (aarch64_const_double_zero_rtx_p): Rename to...
27012 (aarch64_float_const_zero_rtx_p): ...this.
27013 (aarch64_float_const_representable_p): New.
27014 (aarch64_output_simd_mov_immediate): Likewise.
27015 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27016 move immediate case.
27017 * config/aarch64/aarch64.c
27018 (aarch64_const_double_zero_rtx_p): Rename to...
27019 (aarch64_float_const_zero_rtx_p): ...this.
27020 (aarch64_print_operand): Allow printing of new constants.
27021 (aarch64_valid_floating_const): New.
27022 (aarch64_legitimate_constant_p): Check for valid floating-point
27024 (aarch64_simd_valid_immediate): Likewise.
27025 (aarch64_vect_float_const_representable_p): New.
27026 (aarch64_float_const_representable_p): Likewise.
27027 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27028 (aarch64_output_simd_mov_immediate): New.
27029 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27030 (*movdf_aarch64): Likewise.
27031 * config/aarch64/constraints.md (Ufc): New.
27032 (Y): call aarch64_float_const_zero_rtx.
27033 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27035 2013-01-07 Richard Biener <rguenther@suse.de>
27037 PR tree-optimization/55888
27038 PR tree-optimization/55862
27039 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27040 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27041 not if it is contained therein.
27043 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27045 * config/avr/t-avr: Typo.
27047 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27050 * config/avr/t-avr: Don't automatically rebuild
27051 $(srcdir)/config/avr/t-multilib
27052 $(srcdir)/config/avr/avr-tables.opt
27053 $(srcdir)/doc/avr-mmcu.texi
27054 (avr-mcus): New phony target to build them on request.
27055 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27056 * avr/avr-mcus.def: Adjust comments.
27058 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27060 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27062 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27064 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27066 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27068 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27070 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27072 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27073 to generate profiling.
27074 * config/rs6000/aix64.h (LIB_SPEC): Same.
27076 2013-01-04 Andrew Pinski <apinski@cavium.com>
27078 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27080 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27082 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27084 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27086 (ix86_expand_move): Ditto.
27087 (ix86_zero_extend_to_Pmode): Ditto.
27088 (ix86_expand_call): Ditto.
27089 (ix86_expand_special_args_builtin): Ditto.
27090 (ix86_expand_builtin): Ditto.
27092 2013-01-04 Richard Biener <rguenther@suse.de>
27094 PR tree-optimization/55862
27095 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27096 translating them through PHI nodes.
27098 2013-01-04 Martin Jambor <mjambor@suse.cz>
27100 PR tree-optimization/55755
27101 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27102 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27104 2013-01-04 Marek Polacek <polacek@redhat.com>
27106 PR middle-end/55859
27107 * opts.c (default_options_optimization): Clarify error message.
27109 2013-01-04 Richard Biener <rguenther@suse.de>
27111 PR middle-end/55863
27112 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27115 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27118 * config/pa/pa.md (movsi): Revert previous change.
27119 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27122 2013-01-03 Richard Henderson <rth@redhat.com>
27124 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27125 after eliminating TLS symbols.
27127 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27130 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27131 * graphite-poly.c (debug_gmp_value): Likewise.
27133 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27136 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27137 selected code model, define __code_mode_small__, __code_model_medium__,
27138 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27139 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27140 xchg temporary register with %k. Declare temporary register as
27142 [__x86_64__]: For medium and large code models, preserve %rbx register.
27144 2013-01-03 Richard Biener <rguenther@suse.de>
27146 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27147 (dump_subscript): Adjust.
27148 (finalize_ddr_dependent): Do not dump redundant info.
27149 (analyze_siv_subscript): Adjust.
27150 (subscript_dependence_tester): Likewise.
27151 (compute_affine_dependence): Likewise.
27153 2013-01-03 Richard Biener <rguenther@suse.de>
27156 2013-01-03 Richard Biener <rguenther@suse.de>
27158 PR tree-optimization/55857
27159 * tree-vect-stmts.c (vectorizable_load): Do not setup
27160 re-alignment for invariant loads.
27162 2013-01-02 Richard Biener <rguenther@suse.de>
27164 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27165 invariant load do not generate a vector load from the scalar location.
27167 2013-01-03 Richard Biener <rguenther@suse.de>
27169 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27170 for not vectorizing.
27171 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27172 not build INDIRECT_REFs, call get_name once only.
27173 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27174 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27176 2013-01-03 Richard Biener <rguenther@suse.de>
27178 PR tree-optimization/55857
27179 * tree-vect-stmts.c (vectorizable_load): Do not setup
27180 re-alignment for invariant loads.
27182 2013-01-03 Richard Biener <rguenther@suse.de>
27185 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27186 prefer a built-in decl.
27188 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27190 * gcc.c (process_command): Update copyright notice dates.
27191 * gcov.c (print_version): Likewise.
27192 * gcov-dump.c (print_version): Likewise.
27194 PR rtl-optimization/55838
27195 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27196 iv0.step, iv1.step and step.
27198 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27199 Marc Glisse <marc.glisse@inria.fr>
27201 PR tree-optimization/55832
27202 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27203 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27204 integer_{one,zero}_node.
27206 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27209 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27210 * var-tracking.c (reverse_op): Don't add reverse ops to
27211 VALUEs that have already
27212 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27214 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27216 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27218 2013-01-02 Teresa Johnson <tejohnson@google.com>
27220 * dumpfile.c (dump_loc): Print filename with location.
27221 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27222 new location_t parameter to emit complete unroll message with
27223 new dump framework.
27224 (canonicalize_loop_induction_variables): Compute loops location
27225 and pass to try_unroll_loop_completely.
27226 * loop-unroll.c (report_unroll_peel): New function.
27227 (peel_loops_completely): Use new dump format with location
27228 for main dumpfile message, and invoke report_unroll_peel on success.
27229 (decide_unrolling_and_peeling): Ditto.
27230 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27231 by report_unroll_peel.
27232 (decide_peel_completely): Ditto.
27233 (decide_unroll_constant_iterations): Ditto.
27234 (decide_unroll_runtime_iterations): Ditto.
27235 (decide_peel_simple): Ditto.
27236 (decide_unroll_stupid): Ditto.
27237 * cfgloop.c (get_loop_location): New function.
27238 * cfgloop.h (get_loop_location): Declare.
27240 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27242 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27245 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27247 PR middle-end/55198
27248 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27249 BLKmode objects when EXPAND_MEMORY is specified.
27251 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27253 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27255 (fold_builtin_cpu): Do not share cpu model decls across statements.
27257 2013-01-02 Jason Merrill <jason@redhat.com>
27260 * tree.c (build_array_type_1): Revert earlier change.
27262 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27264 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27266 * config/aarch64/aarch64-tune.md: Re-generate.
27268 2013-01-02 Richard Biener <rguenther@suse.de>
27270 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27271 invariant load do not generate a vector load from the scalar location.
27273 2013-01-02 Richard Biener <rguenther@suse.de>
27276 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27277 * configure: Regenerate.
27279 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27281 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27282 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27283 (expand_builtin_int_roundingfn_2): Keep the original target around
27284 for the fallback case.
27286 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27288 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27289 to be clear for sign changes.
27291 2013-01-01 Jan Hubicka <jh@suse.cz>
27293 * ipa-inline-analysis.c: Fix formatting.
27295 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27297 PR tree-optimization/55831
27298 * tree-vect-loop.c (get_initial_def_for_induction): Use
27299 gsi_after_labels instead of gsi_start_bb.
27301 Copyright (C) 2013 Free Software Foundation, Inc.
27303 Copying and distribution of this file, with or without modification,
27304 are permitted in any medium without royalty provided the copyright
27305 notice and this notice are preserved.