1 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3 * tree-ssa-ter.h: Remove duplicate copy of file contents.
5 2013-10-21 Marek Polacek <polacek@redhat.com>
8 * fold-const.c (fold_range_test): Return 0 if the type is not
11 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
13 * system.h: Move hwint.h include further down.
14 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
16 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
18 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
20 Fix volatile issues in optimize_bit_field_compare.
21 * fold-const.c (optimize_bit_field_compare): Bail out if
22 lvolatilep or rvolatilep.
24 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
26 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
27 and get_inner_reference returning different pmode for non-volatile
28 bit-field members dependent on flag_strict_volatile_bitfields.
29 * stor-layout.c (layout_decl): Remove special handling of
30 flag_strict_volatile_bitfields.
31 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
32 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
34 2013-10-21 Paulo Matos <pmatos@broadcom.com>
36 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
39 2013-10-21 Jeff Law <law@redhat.com>
41 * tree-ssa-threadedge.c (thread_through_normal_block): New argument VISITED.
42 Remove VISISTED as a local variable. When we have a threadable jump, verify
43 the destination of the jump has not been visised.
44 (thread_across_edge): Allocate VISITED bitmap once at function scope and
45 use it throughout. Make sure to set appropriate bits in VISITED for E (start
48 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading through
49 a joiner if any edge on the path has a recorded jump thread.
51 2013-10-21 Ian Lance Taylor <iant@google.com>
53 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
54 don't imply that attributes can solve all problems.
55 (Directory Options): Fix typo.
57 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
59 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
60 extend and extend_add.
62 2013-10-21 Richard Biener <rguenther@suse.de>
64 PR tree-optimization/58794
65 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
66 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
68 2013-10-21 Richard Biener <rguenther@suse.de>
71 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
72 to (T) X for sign-changing conversions (or no conversion).
74 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
76 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
78 2013-10-20 Jan Hubicka <jh@suse.cz>
80 * config/i386/i386-tune.def: Add comment; organize into categories
82 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
84 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
85 argument. Update function comment.
86 (expand_set_or_movmem_via_rep): New function combining
87 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
88 (expand_movmem_via_rep_mov): Remove.
89 expand_setmem_via_rep_stos): Remove.
90 (expand_movmem_epilogue): Update calls correspondingly.
91 (expand_setmem_epilogue_via_loop): Likewise.
93 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
94 (expand_set_or_movmem_prologue): New function combining
95 expand_movmem_prologue and expand_setmem_prologue.
96 (expand_movmem_prologue): Remove.
97 (expand_setmem_prologue): Remove.
98 (expand_set_or_movmem_constant_prologue): New function combining
99 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
100 (expand_constant_movmem_prologue): Remove.
101 (expand_constant_setmem_prologue): Remove.
102 (promote_duplicated_reg): Allow vector-const0 value.
103 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
104 and ix86_expand_setmem.
105 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
106 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
108 2013-10-21 Diego Novillo <dnovillo@google.com>
110 * asan.c: Include tree.h
111 * bb-reorder.c: Likewise.
112 * cfgcleanup.c: Likewise.
113 * cfgloopmanip.c: Likewise.
114 * data-streamer-in.c: Likewise.
115 * data-streamer-out.c: Likewise.
116 * data-streamer.c: Likewise.
117 * dwarf2cfi.c: Likewise.
118 * graphite-blocking.c: Likewise.
119 * graphite-clast-to-gimple.c: Likewise.
120 * graphite-dependences.c: Likewise.
121 * graphite-interchange.c: Likewise.
122 * graphite-optimize-isl.c: Likewise.
123 * graphite-poly.c: Likewise.
124 * graphite-scop-detection.c: Likewise.
125 * graphite-sese-to-poly.c: Likewise.
126 * graphite.c: Likewise.
127 * ipa-devirt.c: Likewise.
128 * ipa-profile.c: Likewise.
131 * loop-init.c: Likewise.
132 * loop-unroll.c: Likewise.
133 * lower-subreg.c: Likewise.
134 * lto/lto-object.c: Likewise.
136 * reginfo.c: Likewise.
137 * tree-loop-distribution.c: Likewise.
138 * tree-parloops.c: Likewise.
139 * tree-ssa-strlen.c: Likewise.
140 * tree-streamer.c: Likewise.
141 * value-prof.c: Likewise.
142 * target-globals.c: Likewise.
143 * expr.h: Include tree-core.h instead of tree.h.
144 * gimple.h: Likewise.
145 * ipa-prop.h: Likewise.
146 * ipa-utils.h: Likewise.
147 * lto-streamer.h: Likewise.
148 * streamer-hooks.h: Likewise.
149 * ipa-reference.h: Include cgraph.h instead of tree.h.
150 * cgraph.h: Include basic-block.h instead of tree.h.
151 * tree-streamer.h: Do not include tree.h.
152 * genattrtab.c (write_header): Generate inclusion of
154 * genautomata.c (main): Likewise.
155 * genemit.c: Likewise.
156 * genopinit.c: Likewise.
157 * genoutput.c (output_prologue): Likewise.
158 * genpeep.c: Likewise.
160 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
162 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
164 (vec_unpacku_hi_v8hi): Likewise.
165 (vec_unpacku_lo_v16qi): Likewise.
166 (vec_unpacku_lo_v8hi): Likewise.
168 2013-10-20 Jan Hubicka <jh@suse.cz>
170 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
171 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
172 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
174 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
176 * config/mips/mips.h (ISA_HAS_WSBH): Define.
177 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
179 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
181 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
184 * system.h: Undef m_slot.
186 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
188 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
189 all elements for both endian flavors.
191 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
194 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
195 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
196 and SI_REG for 64bit SYSV ABI targets.
198 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
200 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
201 to multi_reg_return. Clarify that we are skipping USEs of multiple
202 return registers. Use bool type where appropriate.
204 2013-10-18 Jan Hubicka <jh@suse.cz>
206 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
208 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
209 (X86_TUNE_PUSH_MEMORY): Likewise.
210 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
211 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
212 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
214 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
215 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
217 (ix86_option_override_internal): Update to use tune features instead
220 2013-10-18 Cong Hou <congh@google.com>
222 PR tree-optimization/58508
223 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
224 statement that contains data refs with zero-step.
226 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
228 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
229 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
230 * gimple-low.c (gimple_check_call_arg,
231 gimple_check_call_matching_types): Move to cgraph.c.
232 * gimple-low.h: Remove prototype.
233 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
234 Relocate from gimple-low.c.
235 * cgraph.h: Add prototype. Don't include basic-block.h.
236 * gimplify.c: Add gimple-low to include list.
237 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
238 * tree-eh.c: Add gimple-low to include list.
239 * tree-nested.c: Likewise.
240 * cfgexpand.c: Add tree-ssa-address.h to include list.
242 * gimple-fold.c: Likewise.
243 * gimple-ssa-strength-reduction.c: Likewise.
244 * trans-mem.c: Likewise.
245 * tree-mudflap.c: Likewise.
246 * tree-ssa-loop-ivopts.c: Likewise.
247 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
248 (degenerate_phi_result): Move to tree-phinodes.c.
249 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
250 * tree-ssa-threadedge.c: Likewise.
251 * tree-vrp.c: Likewise.
252 * tree-phinodes.c (degenerate_phi_result): Relocate here.
253 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
254 * tree-phinodes.h (degenerate_phi_result): Add prototype.
255 * tree-ssa-copy.c: Include tree-ssa-dom.h.
256 * tree-ssa-forwprop.c: Likewise.
257 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
258 pass_data_cleanup_cfg_post_optimizing,
259 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
260 * tree-optimize.c: Delete File.
261 * graphite.c: Include tree-cfgcleanup.h.
262 * passes.c: Likewise.
263 * tree-cfg.c: Likewise.
264 * tree-profile.c: Likewise.
265 * tree-ssa-dse.c: Likewise.
266 * tree-ssa-loop-ivcanon.c: Likewise.
267 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
268 * tree-outof-ssa.c: Include sbitmap.h.
269 * tree-ssa-live.c: Likewise.
270 * tree-ssa-propagate.c: Likewise.
271 * tree-ssa-structalias.c: Likewise.
272 * tree-stdarg.c: Likewise.
273 * Makefile.in (OBJS): Delete tree-optimize.o.
274 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
275 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
276 * varasm.c: Include basic-block.h.
277 * cfgloop.h: Include function.h instead of basic-block.h
278 (bb_loop_depth): Move to cfgloop.c.
279 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
281 2013-10-18 Teresa Johnson <tejohnson@google.com>
283 * predict.c (probably_never_executed): Compare frequency-based
284 count to number of training runs.
285 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
287 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
289 * config/arm/arm.c (cortexa9_extra_costs): New table.
290 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
292 2013-10-18 Jeff Law <law@redhat.com>
294 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
296 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
297 (register_jump_thread): Add "registered_jump_thread" debug
299 * dbgcnt.def (registered_jump_thread): New debug counter.
301 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
303 * config/rs6000/rs6000.c: Include cgraph.h.
305 2013-10-18 Teresa Johnson <tejohnson@google.com>
307 * tree-ssa-tail-merge.c (replace_block_by): Update edge
308 weights during merging.
310 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
312 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
313 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
314 * tree-ssa-operands.h: Remove prototype.
315 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
316 * gimple.c (virtual_operand_p): Relocate from gimple.c.
317 * gimple.h: Add prototype.
318 * gimple-ssa.h: Include tree-ssa-operands.h.
319 * tree-dump.c: Add tree-cfg.h to include list.
320 * tree-ssa-alias.c: Add ipa-reference.h to include list.
321 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
322 * config/i386/i386.c: Don't include tree-flow.h.
323 * config/rs6000/rs6000.c: Likewise.
325 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
327 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
329 2013-10-18 Richard Biener <rguenther@suse.de>
331 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
332 or TYPE_UNSIGNED of integral types.
333 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
334 NULL_TREE for zero-precision integral types.
336 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
338 * config/aarch64/arm_neon.h
339 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
341 2013-10-17 Sriraman Tallam <tmsriram@google.com>
343 * opth-gen.awk: Define target_flags_explicit.
345 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
347 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
348 fields to the reg_addr array that describes the valid addressing
349 mode for any register, general purpose registers, floating point
350 registers, and Altivec registers.
351 (FIRST_RELOAD_REG_CLASS): Likewise.
352 (LAST_RELOAD_REG_CLASS): Likewise.
353 (struct reload_reg_map_type): Likewise.
354 (reload_reg_map_type): Likewise.
355 (RELOAD_REG_VALID): Likewise.
356 (RELOAD_REG_MULTIPLE): Likewise.
357 (RELOAD_REG_INDEXED): Likewise.
358 (RELOAD_REG_OFFSET): Likewise.
359 (RELOAD_REG_PRE_INCDEC): Likewise.
360 (RELOAD_REG_PRE_MODIFY): Likewise.
361 (reg_addr): Likewise.
362 (mode_supports_pre_incdec_p): New helper functions to say whether
363 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
364 (mode_supports_pre_modify_p): Likewise.
365 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
366 print the valid address mode bits for each mode.
367 (rs6000_debug_print_mode): Likewise.
368 (rs6000_debug_reg_global): Likewise.
369 (rs6000_setup_reg_addr_masks): New function to set up the address
370 mask bits for each type.
371 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
372 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
373 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
374 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
375 PRE_MODIFY are supported.
376 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
377 registers, instead of {src,dest}_av_p.
378 (rs6000_print_options_internal): Tweak the debug output slightly.
380 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
382 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
385 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
387 * tree-flow.h (struct omp_region): Move to omp-low.c.
388 Remove omp_ prototypes and variables.
389 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
390 (copy_var_decl): Relocate prototype from tree-flow.h.
391 * gimple.c (copy_var_decl): Relocate from omp-low.c.
392 * tree.h: Move prototype to omp-low.h.
393 * omp-low.h: New File. Relocate prototypes here.
394 * omp-low.c (struct omp_region): Make local here.
395 (root_omp_region): Make static.
396 (copy_var_decl) Move to gimple.c.
397 (new_omp_region): Make static.
398 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
399 * tree-cfg.c: Include omp-low.h.
400 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
401 * gimplify.c: Include omp-low.h.
402 * tree-parloops.c: Likewise.
404 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
406 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
407 operands are in memory, prefer to force non-matched operand 1 to
410 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
413 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
414 restrict TImode addresses to single indirect registers if both
415 -mquad-memory and -mvsx-timode are used.
416 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
417 we should emit load/store quad. Remove using %y for quad memory
420 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
421 constraints to allow load/store quad on machines where TImode is
422 not allowed in VSX registers. Use 'n' instead of 'F' constraint
423 for TImode to load integer constants.
425 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
427 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
429 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
431 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
432 handling of STACK_REG.
434 2013-10-17 Richard Biener <rguenther@suse.de>
436 PR tree-optimization/58143
437 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
439 (rewrite_to_defined_overflow): Likewise.
440 (move_computations_dom_walker::before_dom): Rewrite stmts
441 with undefined signed overflow that are not always executed
442 into unsigned arithmetic.
444 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
447 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
448 explicit isa flag to be an options variable, instead of using
449 global_options_set. Remove define from rs6000.h.
450 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
452 * config/rs6000/rs6000.c (rs6000_option_override_internal):
453 Initialize rs6000_isa_flags_explicit.
454 (rs6000_function_specific_save): Add gcc_options* parameter, so
455 that the powerpc builds after the 2013-10-15 changes.
456 (rs6000_function_specific_restore): Likewise.
458 2013-10-16 DJ Delorie <dj@redhat.com>
460 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
461 op is a REG before checking REGNO.
462 (rl78_alloc_physical_registers): Verify pattern is a SET before
465 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
467 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
469 (vec_unpacks_lo_v4sf): Likewise.
470 (vec_unpacks_float_hi_v4si): Likewise.
471 (vec_unpacks_float_lo_v4si): Likewise.
472 (vec_unpacku_float_hi_v4si): Likewise.
473 (vec_unpacku_float_lo_v4si): Likewise.
475 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
477 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
478 (vsx_concat_v2sf): Likewise.
480 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
482 * config/aarch64/aarch64.md
483 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
485 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
487 PR tree-optimization/58697
488 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
489 estimated_loop_iterations_int.
490 (max_stmt_executions_int): Call get_max_loop_iterations_int.
491 (get_max_loop_iterations_int): New. HWINT version of
492 get_max_loop_iterations.
493 * cfgloop.h: Add prototypes.
494 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
495 * loop-unroll.c (decide_peel_once_rolling): Call
496 get_estimated_loop_iterations_int.
497 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
498 * tree-ssa-loop-niter.h: Tweak prototypes.
500 2013-10-16 David Malcolm <dmalcolm@redhat.com>
502 * gengtype-parse.c (struct_field_seq): Ignore access-control
503 keywords ("public:" etc).
505 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
507 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
508 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
510 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
512 * config/arm/arm.opt (mlra): New option.
513 * config/arm/arm.c (arm_lra_p): New function.
514 (TARGET_LRA_P): Define.
516 2013-10-16 Paulo Matos <pmatos@broadcom.com>
518 * tree-core.h (tree_code_name): Remove.
519 * tree.h (get_tree_code_name): New prototype.
520 * tree.c (tree_code_name): Make static.
521 (get_tree_code_name): New function.
522 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
523 tree_class_check_failed, tree_range_check_failed,
524 tree_not_class_check_failed, omp_clause_check_failed,
525 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
526 wrapper get_tree_code_name instead of calling tree_code_name directly.
527 * tree-vrp.c (dump_asserts_for): Likewise.
528 * tree-dump.c (dequeue_and_dump): Likewise.
529 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
530 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
531 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
532 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
533 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
534 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
535 dump_gimple_omp_for): Likewise.
536 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
537 * tree-ssa-pre.c (print_pre_expr): Likewise.
538 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
539 * print-tree.c (print_node_brief, print_node): Likewise.
540 * gimple.c (gimple_check_failed): Likewise.
541 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
542 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
543 * config/mep/mep.c (mep_validate_vliw): Likewise.
544 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
545 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
547 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
549 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
552 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
554 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
555 (MULTILIB_MATCHES): Add multilib for -march=v8.
557 2013-10-15 Sriraman Tallam <tmsriram@google.com>
560 * optc-save-gen.awk: Add extra parameter to the save and restore
562 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
563 * tree.c (build_optimization_node): New parameter. Add extra parameter
564 to call to cl_optimization_save.
565 (build_target_option_node): New parameter. Add extra parameter
566 to call to cl_target_option_save.
567 * tree.h (build_optimization_node): New parameter.
568 (build_target_option_node): New parameter.
569 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
570 build_optimization_node and build_target_option_node.
571 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
572 (handle_pragma_push_options): Ditto.
573 * toplev.c (process_options): Ditto.
574 * opts.c (init_options_struct): Check for opts_set non-null.
575 * target.def (target_option.save): New parameter.
576 (target_option.restore): New parameter.
578 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
579 (ix86_pragma_target_parse): Ditto.
580 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
582 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
583 to build_optimization_node and build_target_option_node.
584 (rs6000_valid_attribute_p): Ditto.
585 (rs6000_pragma_target_parse): Ditto.
586 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
588 * config/i386/i386.h:
589 TARGET_64BIT_P: New Macro
590 TARGET_MMX_P: New Macro.
591 TARGET_3DNOW_P: New Macro.
592 TARGET_3DNOW_A_P: New Macro.
593 TARGET_SSE_P: New Macro.
594 TARGET_SSE2_P: New Macro.
595 TARGET_SSE3_P: New Macro.
596 TARGET_SSSE3_P: New Macro.
597 TARGET_SSE4_1_P: New Macro.
598 TARGET_SSE4_2_P: New Macro.
599 TARGET_AVX_P: New Macro.
600 TARGET_AVX2_P: New Macro.
601 TARGET_AVX512F_P: New Macro.
602 TARGET_AVX512PF_P: New Macro.
603 TARGET_AVX512ER_P: New Macro.
604 TARGET_AVX512CD_P: New Macro.
605 TARGET_FMA_P: New Macro.
606 TARGET_SSE4A_P: New Macro.
607 TARGET_FMA4_P: New Macro.
608 TARGET_XOP_P: New Macro.
609 TARGET_LWP_P: New Macro.
610 TARGET_ABM_P: New Macro.
611 TARGET_BMI_P: New Macro.
612 TARGET_BMI2_P: New Macro.
613 TARGET_LZCNT_P: New Macro.
614 TARGET_TBM_P: New Macro.
615 TARGET_POPCNT_P: New Macro.
616 TARGET_SAHF_P: New Macro.
617 TARGET_MOVBE_P: New Macro.
618 TARGET_CRC32_P: New Macro.
619 TARGET_AES_P: New Macro.
620 TARGET_PCLMUL_P: New Macro.
621 TARGET_CMPXCHG16B_P: New Macro.
622 TARGET_FSGSBASE_P: New Macro.
623 TARGET_RDRND_P: New Macro.
624 TARGET_F16C_P: New Macro.
625 TARGET_RTM_P: New Macro.
626 TARGET_HLE_P: New Macro.
627 TARGET_RDSEED_P: New Macro.
628 TARGET_PRFCHW_P: New Macro.
629 TARGET_ADX_P: New Macro.
630 TARGET_FXSR_P: New Macro.
631 TARGET_XSAVE_P: New Macro.
632 TARGET_XSAVEOPT_P: New Macro.
633 TARGET_LP64_P: New Macro.
634 TARGET_X32_P: New Macro.
635 TARGET_FPMATH_DEFAULT_P: New Macro.
636 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
637 * config/i386/i386.c (ix86_option_override_internal): New parameters.
639 Change ix86_tune_string to access opts->x_ix86_tune_string.
640 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
641 Change ix86_arch_string to access opts->x_ix86_arch_string.
642 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
643 Change ix86_pmode to access opts->x_ix86_pmode.
644 Change ix86_abi to access opts->x_ix86_abi.
645 Change ix86_cmodel to access opts->x_ix86_cmodel.
646 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
647 Change ix86_isa_flags_explicit to access
648 opts->x_ix86_isa_flags_explicit.
649 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
650 Change ix86_regparm to access opts->x_ix86_regparm.
651 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
652 Change ix86_preferred_stack_boundary_arg to access
653 opts->x_ix86_preferred_stack_boundary_arg.
654 Change ix86_force_align_arg_pointer to access
655 opts->x_ix86_force_align_arg_pointer.
656 Change ix86_incoming_stack_boundar_arg to access
657 opts->x_ix86_incoming_stack_boundar_arg.
658 Change ix86_fpmath to access opts->x_ix86_fpmath.
659 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
660 Change ix86_recip_name to access opts->x_ix86_recip_name.
661 Change ix86_stack_protector_guard to access
662 opts->x_ix86_stack_protector_guard.
663 Change ix86_tune_memcpy_strategy to access
664 opts->x_ix86_tune_memcpy_strategy.
665 Change ix86_tune_memset_strategy to access
666 opts->x_ix86_tune_memset_strategy.
667 Change global_options to access opts.
668 Change global_options_set to access opts_set.
669 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
670 Change TARGET_MMX to TARGET_MMX_P (opts->...).
671 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
672 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
673 Change TARGET_SSE to TARGET_SSE_P (opts->...).
674 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
675 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
676 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
677 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
678 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
679 Change TARGET_AVX to TARGET_AVX_P (opts->...).
680 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
681 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
682 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
683 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
684 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
685 Change TARGET_FMA to TARGET_FMA_P (opts->...).
686 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
687 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
688 Change TARGET_XOP to TARGET_XOP_P (opts->...).
689 Change TARGET_LWP to TARGET_LWP_P (opts->...).
690 Change TARGET_ABM to TARGET_ABM_P (opts->...).
691 Change TARGET_BMI to TARGET_BMI_P (opts->...).
692 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
693 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
694 Change TARGET_TBM to TARGET_TBM_P (opts->...).
695 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
696 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
697 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
698 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
699 Change TARGET_AES to TARGET_AES_P (opts->...).
700 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
701 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
702 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
703 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
704 Change TARGET_F16C to TARGET_F16C_P (opts->...).
705 Change TARGET_RTM to TARGET_RTM_P (opts->...).
706 Change TARGET_HLE to TARGET_HLE_P (opts->...).
707 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
708 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
709 Change TARGET_ADX to TARGET_ADX_P (opts->...).
710 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
711 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
712 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
713 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
714 Change TARGET_X32 to TARGET_X32_P (opts->...).
715 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
716 Change TARGET_FLOAT_RETURNS_IN_80387 to
717 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
718 (ix86_function_specific_save): New parameter. Use opts-> fields
719 to replace global fields.
720 (ix86_function_specific_restore): Ditto.
721 (ix86_valid_target_attribute_inner_p): New parameters.
723 Fix call to ix86_handle_option and set_option.
724 (ix86_valid_target_attribute_tree): New parameters.
725 Change global_options to access opts.
726 Change global_options_set to access opts_set.
727 Fix call to ix86_valid_target_attribute_inner_p.
728 Change ix86_tune_string to access opts->x_ix86_tune_string.
729 Change ix86_arch_string to access opts->x_ix86_arch_string.
730 Change ix86_fpmath to access opts->x_ix86_fpmath
731 Fix call to ix86_option_override_internal.
732 Fix call to ix86_add_new_builtins.
733 Fix calls to build_optimization_node and build_target_option_node.
734 (ix86_valid_target_attribute_p): Remove access to global_options.
735 Use new gcc_options structure func_options.
736 Fix call to ix86_valid_target_attribute_tree.
737 Fix call to build_optimization_node.
738 (get_builtin_code_for_version): Fix call to
739 ix86_valid_target_attribute_tree.
741 2013-10-15 David Malcolm <dmalcolm@redhat.com>
743 * Makefile.in (PICFLAG): New.
744 (enable_host_shared): New.
745 (INTERNAL_CFLAGS): Use PICFLAG.
746 (LIBIBERTY): Use pic build of libiberty.a if configured with
747 --enable-host-shared.
748 * configure.ac: Add --enable-host-shared, setting up new
750 * configure: Regenerate.
751 * doc/install.texi (--enable-shared): Add note contrasting it with ...
752 (--enable-host-shared): New option.
754 2013-10-15 Richard Biener <rguenther@suse.de>
756 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
757 for built-in functions.
759 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
761 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
762 (optimize_range_tests_1): New function,
763 extracted from optimize_range_tests.
764 (optimize_range_tests_xor): Similarly.
765 (optimize_range_tests_diff): New function.
766 (optimize_range_tests): Use optimize_range_tests_1.
768 2013-10-15 Cong Hou <congh@google.com>
770 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
771 requirement of the reduction pattern so that one operand of the
772 reduction operation can come from outside of the loop.
774 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
776 * config/arm/neon-schedgen.ml: Remove.
777 * config/arm/cortex-a9-neon.md: Remove comment regarding
780 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
782 * config/arm/types: Remove old neon types.
784 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
786 * config/arm/cortex-a7.md
787 (cortex_a7_neon_type): New.
788 (cortex_a7_neon_mul): Update for new types.
789 (cortex_a7_neon_mla): Likewise.
790 (cortex_a7_neon): Likewise.
792 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
794 * config/arm/cortex-a15-neon.md
795 (cortex_a15_neon_type): New,
797 (cortex_a15_neon_int_1): Remove.
798 (cortex_a15_neon_int_2): Likewise.
799 (cortex_a15_neon_int_3): Likewise.
800 (cortex_a15_neon_int_4): Likewise.
801 (cortex_a15_neon_int_5): Likewise.
802 (cortex_a15_neon_vqneg_vqabs): Likewise.
803 (cortex_a15_neon_vmov): Likewise.
804 (cortex_a15_neon_vaba): Likewise.
805 (cortex_a15_neon_vaba_qqq): Likewise.
806 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
807 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
808 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
810 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
811 (cortex_a15_neon_mla_qqq_8_16): Likewise.
812 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
813 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
814 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
815 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
816 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
817 (cortex_a15_neon_shift_1): Likewise.
818 (cortex_a15_neon_shift_2): Likewise.
819 (cortex_a15_neon_shift_3): Likewise.
820 (cortex_a15_neon_vshl_ddd): Likewise.
821 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
822 (cortex_a15_neon_vsra_vrsra): Likewise.
823 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
824 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
825 (cortex_a15_neon_bp_3cycle): Likewise.
826 (cortex_a15_neon_ldm_2): Likewise.
827 (cortex_a15_neon_stm_2): Likewise.
828 (cortex_a15_neon_mcr): Likewise.
829 (cortex_a15_neon_mrc): Likewise.
830 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
831 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
832 (cortex_a15_neon_fp_vmul_ddd): Likewise.
833 (cortex_a15_neon_fp_vmul_qqd): Likewise.
834 (cortex_a15_neon_fp_vmla_ddd): Likewise.
835 (cortex_a15_neon_fp_vmla_qqq): Likewise.
836 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
837 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
838 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
839 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
840 (cortex_a15_neon_bp_simple): Likewise.
841 (cortex_a15_neon_bp_2cycle): Likewise.
842 (cortex_a15_neon_bp_3cycle): Likewise.
843 (cortex_a15_neon_vld1_1_2_regs): Likewise.
844 (cortex_a15_neon_vld1_3_4_regs): Likewise.
845 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
846 (cortex_a15_neon_vld2_4_regs): Likewise.
847 (cortex_a15_neon_vld3_vld4): Likewise.
848 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
849 (cortex_a15_neon_vst1_3_4_regs): Likewise.
850 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
851 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
852 (cortex_a15_neon_vst3_vst4): Rename to...
853 (cortex_a15_neon_vst4): This, update for new attributes.
854 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
855 (cortex_a15_neon_vld3_vld4_lane): Likewise.
856 (cortex_a15_neon_vst1_vst2_lane): Likewise.
857 (cortex_a15_neon_vst3_vst4_lane): Likewise.
858 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
859 (cortex_a15_neon_ldm_2): Likewise.
860 (cortex_a15_neon_stm_2): Likewise.
861 (cortex_a15_neon_mcr): Likewise.
862 (cortex_a15_neon_mcr_2_mcrr): Likewise.
863 (cortex_a15_neon_mrc): Likewise.
864 (cortex_a15_neon_mrrc): Likewise.
866 (cortex_a15_neon_abd): New.
867 (cortex_a15_neon_abd_q): Likewise.
868 (cortex_a15_neon_aba): Likewise.
869 (cortex_a15_neon_aba_q): Likewise.
870 (cortex_a15_neon_acc): Likewise.
871 (cortex_a15_neon_acc_q): Likewise.
872 (cortex_a15_neon_arith_basic): Likewise.
873 (cortex_a15_neon_arith_complex): Likewise.
874 (cortex_a15_neon_multiply): Likewise.
875 (cortex_a15_neon_multiply_q): Likewise.
876 (cortex_a15_neon_mla): Likewise.
877 (cortex_a15_neon_mla_q): Likewise.
878 (cortex_a15_neon_sat_mla_long): Likewise.
879 (cortex_a15_neon_shift_acc): Likewise.
880 (cortex_a15_neon_shift_imm_basic): Likewise.
881 (cortex_a15_neon_shift_imm_complex): Likewise.
882 (cortex_a15_neon_shift_reg_basic): Likewise.
883 (cortex_a15_neon_shift_reg_basic_q): Likewise.
884 (cortex_a15_neon_shift_reg_complex): Likewise.
885 (cortex_a15_neon_shift_reg_complex_q): Likewise.
886 (cortex_a15_neon_fp_negabs): Likewise
887 (cortex_a15_neon_fp_arith): Likewise
888 (cortex_a15_neon_fp_arith_q): Likewise
889 (cortex_a15_neon_fp_cvt_int): Likewise
890 (cortex_a15_neon_fp_cvt_int_q): Likewise
891 (cortex_a15_neon_fp_cvt_16): Likewise
892 (cortex_a15_neon_fp_mul): Likewise
893 (cortex_a15_neon_fp_mul_q): Likewise
894 (cortex_a15_neon_fp_mla): Likewise
895 (cortex_a15_neon_fp_mla_q): Likewise
896 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
897 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
898 (cortex_a15_neon_bitops): Likewise.
899 (cortex_a15_neon_bitops_q): Likewise.
900 (cortex_a15_neon_from_gp): Likewise.
901 (cortex_a15_neon_from_gp_q): Likewise.
902 (cortex_a15_neon_tbl3_tbl4): Likewise.
903 (cortex_a15_neon_zip_q): Likewise.
904 (cortex_a15_neon_to_gp): Likewise.
905 (cortex_a15_neon_load_a): Likewise.
906 (cortex_a15_neon_load_b): Likewise.
907 (cortex_a15_neon_load_c): Likewise.
908 (cortex_a15_neon_load_d): Likewise.
909 (cortex_a15_neon_load_e): Likewise.
910 (cortex_a15_neon_load_f): Likewise.
911 (cortex_a15_neon_store_a): Likewise.
912 (cortex_a15_neon_store_b): Likewise.
913 (cortex_a15_neon_store_c): Likewise.
914 (cortex_a15_neon_store_d): Likewise.
915 (cortex_a15_neon_store_e): Likewise.
916 (cortex_a15_neon_store_f): Likewise.
917 (cortex_a15_neon_store_g): Likewise.
918 (cortex_a15_neon_store_h): Likewise.
919 (cortex_a15_vfp_to_from_gp): Likewise.
921 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
923 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
925 (cortex_a9_neon_vshl_ddd): Remove.
926 (cortex_a9_neon_vst3_vst4): Likewise.
927 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
929 (cortex_a9_neon_bit_ops_q): New.
931 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
932 (cortex_a9_neon_int_2): Likewise.
933 (cortex_a9_neon_int_3): Likewise.
934 (cortex_a9_neon_int_4): Likewise.
935 (cortex_a9_neon_int_5): Likewise.
936 (cortex_a9_neon_vqneg_vqabs): Likewise.
937 (cortex_a9_neon_vmov): Likewise.
938 (cortex_a9_neon_vaba): Likewise.
939 (cortex_a9_neon_vaba_qqq): Likewise.
940 (cortex_a9_neon_shift_1): Likewise.
941 (cortex_a9_neon_shift_2): Likewise.
942 (cortex_a9_neon_shift_3): Likewise.
943 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
944 (cortex_a9_neon_vsra_vrsra): Likewise.
945 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
946 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
947 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
949 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
950 (cortex_a9_neon_mla_qqq_8_16): Likewise.
951 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
953 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
954 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
955 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
956 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
957 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
958 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
959 (cortex_a9_neon_fp_vsum): Likewise.
960 (cortex_a9_neon_fp_vmul_ddd): Likewise.
961 (cortex_a9_neon_fp_vmul_qqd): Likewise.
962 (cortex_a9_neon_fp_vmla_ddd): Likewise.
963 (cortex_a9_neon_fp_vmla_qqq): Likewise.
964 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
965 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
966 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
967 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
968 (cortex_a9_neon_bp_simple): Likewise.
969 (cortex_a9_neon_bp_2cycle): Likewise.
970 (cortex_a9_neon_bp_3cycle): Likewise.
971 (cortex_a9_neon_ldr): Likewise.
972 (cortex_a9_neon_str): Likewise.
973 (cortex_a9_neon_vld1_1_2_regs): Likewise.
974 (cortex_a9_neon_vld1_3_4_regs): Likewise.
975 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
976 (cortex_a9_neon_vld2_4_regs): Likewise.
977 (cortex_a9_neon_vld3_vld4): Likewise.
978 (cortex_a9_neon_vld1_vld2_lane): Likewise.
979 (cortex_a9_neon_vld3_vld4_lane): Likewise.
980 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
981 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
982 (cortex_a9_neon_vst1_3_4_regs): Likewise.
983 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
984 (cortex_a9_neon_vst1_vst2_lane): Likewise.
985 (cortex_a9_neon_vst3_vst4_lane): Likewise.
986 (cortex_a9_neon_mcr): Likewise.
987 (cortex_a9_neon_mcr_2_mcrr): Likewise.
988 (cortex_a9_neon_mrc): Likewise.
989 (cortex_a9_neon_mrrc): Likewise.
991 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
993 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
995 (cortex_a8_neon_vshl_ddd): Remove.
996 (cortex_a8_neon_vst3_vst4): Likewise.
997 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
999 (cortex_a8_neon_bit_ops_q): New.
1001 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1002 (cortex_a8_neon_int_2): Likewise..
1003 (cortex_a8_neon_int_3): Likewise.
1004 (cortex_a8_neon_int_5): Likewise.
1005 (cortex_a8_neon_vqneg_vqabs): Likewise.
1006 (cortex_a8_neon_int_4): Likewise.
1007 (cortex_a8_neon_vaba): Likewise.
1008 (cortex_a8_neon_vaba_qqq): Likewise.
1009 (cortex_a8_neon_shift_1): Likewise.
1010 (cortex_a8_neon_shift_2): Likewise.
1011 (cortex_a8_neon_shift_3): Likewise.
1012 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1013 (cortex_a8_neon_vsra_vrsra): Likewise.
1014 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1015 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1016 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1018 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1019 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1020 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1022 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1023 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1024 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1025 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1026 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1027 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1028 (cortex_a8_neon_fp_vsum): Likewise.
1029 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1030 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1031 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1032 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1033 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1034 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1035 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1036 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1037 (cortex_a8_neon_bp_simple): Likewise.
1038 (cortex_a8_neon_bp_2cycle): Likewise.
1039 (cortex_a8_neon_bp_3cycle): Likewise.
1040 (cortex_a8_neon_ldr): Likewise.
1041 (cortex_a8_neon_str): Likewise.
1042 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1043 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1044 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1045 (cortex_a8_neon_vld2_4_regs): Likewise.
1046 (cortex_a8_neon_vld3_vld4): Likewise.
1047 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1048 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1049 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1050 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1051 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1052 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1053 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1054 (cortex_a8_neon_mcr): Likewise.
1055 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1056 (cortex_a8_neon_mrc): Likewise.
1057 (cortex_a8_neon_mrrc): Likewise.
1059 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1061 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1063 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1064 (aarch64_simd_dup<mode>): Add "type" attribute.
1065 (aarch64_dup_lane<mode>): Likewise.
1066 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1067 (*aarch64_simd_mov<mode>): Likewise.
1068 (aarch64_simd_mov_from_<mode>low): Likewise.
1069 (aarch64_simd_mov_from_<mode>high): Likewise.
1070 (orn<mode>3): Likewise.
1071 (bic<mode>3): Likewise.
1072 (add<mode>3): Likewise.
1073 (sub<mode>3): Likewise.
1074 (mul<mode>3): Likewise.
1075 (*aarch64_mul3_elt<mode>): Likewise.
1076 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1077 (*aarch64_mul3_elt_to_128df): Likewise.
1078 (*aarch64_mul3_elt_to_64v2df): Likewise.
1079 (neg<mode>2): Likewise.
1080 (abs<mode>2): Likewise.
1081 (abd<mode>_3): Likewise.
1082 (aba<mode>_3): Likewise.
1083 (fabd<mode>_3): Likewise.
1084 (*fabd_scalar<mode>3): Likewise.
1085 (and<mode>3): Likewise.
1086 (ior<mode>3): Likewise.
1087 (xor<mode>3): Likewise.
1088 (one_cmpl<mode>2): Likewise.
1089 (aarch64_simd_vec_set<mode>): Likewise.
1090 (aarch64_simd_lshr<mode>): Likewise.
1091 (aarch64_simd_ashr<mode>): Likewise.
1092 (aarch64_simd_imm_shl<mode>): Likewise.
1093 (aarch64_simd_reg_sshl<mode): Likewise.
1094 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1095 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1096 (aarch64_simd_vec_setv2di): Likewise.
1097 (aarch64_simd_vec_set<mode>): Likewise.
1098 (aarch64_mla<mode>): Likewise.
1099 (*aarch64_mla_elt<mode>): Likewise.
1100 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1101 (aarch64_mls<mode>): Likewise.
1102 (*aarch64_mls_elt<mode>): Likewise.
1103 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1104 (<su><maxmin><mode>3): Likewise.
1105 (move_lo_quad_<mode>): Likewise.
1106 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1107 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1108 (vec_pack_trunc_<mode>): Likewise.
1109 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1110 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1111 (*aarch64_<su>mlal_lo<mode>): Likewise.
1112 (*aarch64_<su>mlal_hi<mode>): Likewise.
1113 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1114 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1115 (*aarch64_<su>mlal<mode>): Likewise.
1116 (*aarch64_<su>mlsl<mode>): Likewise.
1117 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1118 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1119 (add<mode>3): Likewise.
1120 (sub<mode>3): Likewise.
1121 (mul<mode>3): Likewise.
1122 (div<mode>3): Likewise.
1123 (neg<mode>2): Likewise.
1124 (abs<mode>2): Likewise.
1125 (fma<mode>4): Likewise.
1126 (*aarch64_fma4_elt<mode>): Likewise.
1127 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1128 (*aarch64_fma4_elt_to_128df): Likewise.
1129 (*aarch64_fma4_elt_to_64v2df): Likewise.
1130 (fnma<mode>4): Likewise.
1131 (*aarch64_fnma4_elt<mode>): Likewise.
1132 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1133 (*aarch64_fnma4_elt_to_128df): Likewise.
1134 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1135 (<frint_pattern><mode>2): Likewise.
1136 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1137 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1138 (vec_unpacks_lo_v4sf): Likewise.
1139 (aarch64_float_extend_lo_v2df): Likewise.
1140 (vec_unpacks_hi_v4sf): Likewise.
1141 (aarch64_float_truncate_lo_v2sf): Likewise.
1142 (aarch64_float_truncate_hi_v4sf): Likewise.
1143 (aarch64_vmls<mode>): Likewise.
1144 (<su><maxmin><mode>3): Likewise.
1145 (<maxmin_uns><mode>3): Likewise.
1146 (reduc_<sur>plus_<mode>): Likewise.
1147 (reduc_<sur>plus_v2di): Likewise.
1148 (reduc_<sur>plus_v2si): Likewise.
1149 (reduc_<sur>plus_<mode>): Likewise.
1150 (aarch64_addpv4sf): Likewise.
1151 (clz<mode>2): Likewise.
1152 (reduc_<maxmin_uns>_<mode>): Likewise.
1153 (reduc_<maxmin_uns>_v2di): Likewise.
1154 (reduc_<maxmin_uns>_v2si): Likewise.
1155 (reduc_<maxmin_uns>_<mode>): Likewise.
1156 (reduc_<maxmin_uns>_v4sf): Likewise.
1157 (aarch64_simd_bsl<mode>_internal): Likewise.
1158 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1159 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1160 (aarch64_get_lane<mode>): Likewise.
1161 (*aarch64_combinez<mode>): Likewise.
1162 (aarch64_combine<mode>): Likewise.
1163 (aarch64_simd_combine<mode>): Likewise.
1164 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1165 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1166 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1167 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1168 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1169 (aarch64_<sur>h<addsub><mode>): Likewise.
1170 (aarch64_<sur><addsub>hn<mode>): Likewise.
1171 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1172 (aarch64_pmul<mode>): Likewise.
1173 (aarch64_<su_optab><optab><mode>): Likewise.
1174 (aarch64_<sur>qadd<mode>): Likewise.
1175 (aarch64_sqmovun<mode>): Likewise.
1176 (aarch64_<sur>qmovn<mode>): Likewise.
1177 (aarch64_s<optab><mode>): Likewise.
1178 (aarch64_sq<r>dmulh<mode>): Likewise.
1179 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1180 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1181 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1182 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1183 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1184 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1185 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1186 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1187 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1188 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1189 (aarch64_sqdmull<mode>): Likewise.
1190 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1191 (aarch64_sqdmull_n<mode>): Likewise.
1192 (aarch64_sqdmull2<mode>_internal): Likewise.
1193 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1194 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1195 (aarch64_<sur>shl<mode>): Likewise.
1196 (aarch64_<sur>q<r>shl<mode>
1197 (aarch64_<sur>shll_n<mode>): Likewise.
1198 (aarch64_<sur>shll2_n<mode>): Likewise.
1199 (aarch64_<sur>shr_n<mode>): Likewise.
1200 (aarch64_<sur>sra_n<mode>): Likewise.
1201 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1202 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1203 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1204 (aarch64_cm<optab><mode>): Likewise.
1205 (aarch64_cm<optab>di): Likewise.
1206 (aarch64_cm<optab><mode>): Likewise.
1207 (aarch64_cm<optab>di): Likewise.
1208 (aarch64_cmtst<mode>): Likewise.
1209 (aarch64_cmtstdi): Likewise.
1210 (aarch64_cm<optab><mode>): Likewise.
1211 (*aarch64_fac<optab><mode>): Likewise.
1212 (aarch64_addp<mode>): Likewise.
1213 (aarch64_addpdi): Likewise.
1214 (sqrt<mode>2): Likewise.
1215 (vec_load_lanesoi<mode>): Likewise.
1216 (vec_store_lanesoi<mode>): Likewise.
1217 (vec_load_lanesci<mode>): Likewise.
1218 (vec_store_lanesci<mode>): Likewise.
1219 (vec_load_lanesxi<mode>): Likewise.
1220 (vec_store_lanesxi<mode>): Likewise.
1221 (*aarch64_mov<mode>): Likewise.
1222 (aarch64_ld2<mode>_dreg): Likewise.
1223 (aarch64_ld2<mode>_dreg): Likewise.
1224 (aarch64_ld3<mode>_dreg): Likewise.
1225 (aarch64_ld3<mode>_dreg): Likewise.
1226 (aarch64_ld4<mode>_dreg): Likewise.
1227 (aarch64_ld4<mode>_dreg): Likewise.
1228 (aarch64_tbl1<mode>): Likewise.
1229 (aarch64_tbl2v16qi): Likewise.
1230 (aarch64_combinev16qi): Likewise.
1231 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1232 (aarch64_st2<mode>_dreg): Likewise.
1233 (aarch64_st2<mode>_dreg): Likewise.
1234 (aarch64_st3<mode>_dreg): Likewise.
1235 (aarch64_st3<mode>_dreg): Likewise.
1236 (aarch64_st4<mode>_dreg): Likewise.
1237 (aarch64_st4<mode>_dreg): Likewise.
1238 (*aarch64_simd_ld1r<mode>): Likewise.
1239 (aarch64_frecpe<mode>): Likewise.
1240 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1241 (aarch64_frecps<mode>): Likewise.
1243 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1245 * config/arm/iterators.md (V_elem_ch): New.
1247 (VQH_type): Likewise.
1248 * config/arm/arm.md (is_neon_type): New.
1249 (conds): Use is_neon_type.
1250 (anddi3_insn): Update type attribute.
1251 (xordi3_insn): Likewise.
1252 (one_cmpldi2): Likewise.
1253 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1254 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1255 (*movmisalign<mode>_neon_store): Likewise.
1256 (*movmisalign<mode>_neon_load): Likewise.
1257 (vec_set<mode>_internal): Likewise.
1258 (vec_set<mode>_internal): Likewise.
1259 (vec_setv2di_internal): Likewise.
1260 (vec_extract<mode>): Likewise.
1261 (vec_extract<mode>): Likewise.
1262 (vec_extractv2di): Likewise.
1263 (*add<mode>3_neon): Likewise.
1264 (adddi3_neon): Likewise.
1265 (*sub<mode>3_neon): Likewise.
1266 (subdi3_neon): Likewise.
1267 (fma<VCVTF:mode>4): Likewise.
1268 (fma<VCVTF:mode>4_intrinsic): Likewise.
1269 (*fmsub<VCVTF:mode>4): Likewise.
1270 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1271 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1272 (ior<mode>3): Likewise.
1273 (and<mode>3): Likewise.
1274 (orn<mode>3_neon): Likewise.
1275 (orndi3_neon): Likewise.
1276 (bic<mode>3_neon): Likewise.
1277 (bicdi3_neon): Likewise.
1278 (xor<mode>3): Likewise.
1279 (one_cmpl<mode>2): Likewise.
1280 (abs<mode>2): Likewise.
1281 (neg<mode>2): Likewise.
1282 (negdi2_neon): Likewise.
1283 (*umin<mode>3_neon): Likewise.
1284 (*umax<mode>3_neon): Likewise.
1285 (*smin<mode>3_neon): Likewise.
1286 (*smax<mode>3_neon): Likewise.
1287 (vashl<mode>3): Likewise.
1288 (vashr<mode>3_imm): Likewise.
1289 (vlshr<mode>3_imm): Likewise.
1290 (ashl<mode>3_signed): Likewise.
1291 (ashl<mode>3_unsigned): Likewise.
1292 (neon_load_count): Likewise.
1293 (ashldi3_neon_noclobber): Likewise.
1294 (ashldi3_neon): Likewise.
1295 (signed_shift_di3_neon): Likewise.
1296 (unsigned_shift_di3_neon): Likewise.
1297 (ashrdi3_neon_imm_noclobber): Likewise.
1298 (lshrdi3_neon_imm_noclobber): Likewise.
1299 (<shift>di3_neon): Likewise.
1300 (widen_ssum<mode>3): Likewise.
1301 (widen_usum<mode>3): Likewise.
1302 (quad_halves_<code>v4si): Likewise.
1303 (quad_halves_<code>v4sf): Likewise.
1304 (quad_halves_<code>v8hi): Likewise.
1305 (quad_halves_<code>v16qi): Likewise.
1306 (reduc_splus_v2di): Likewise.
1307 (neon_vpadd_internal<mode>): Likewise.
1308 (neon_vpsmin<mode>): Likewise.
1309 (neon_vpsmax<mode>): Likewise.
1310 (neon_vpumin<mode>): Likewise.
1311 (neon_vpumax<mode>): Likewise.
1312 (*ss_add<mode>_neon): Likewise.
1313 (*us_add<mode>_neon): Likewise.
1314 (*ss_sub<mode>_neon): Likewise.
1315 (*us_sub<mode>_neon): Likewise.
1316 (neon_vadd<mode>_unspec): Likewise.
1317 (neon_vaddl<mode>): Likewise.
1318 (neon_vaddw<mode>): Likewise.
1319 (neon_vhadd<mode>): Likewise.
1320 (neon_vqadd<mode>): Likewise.
1321 (neon_vaddhn<mode>): Likewise.
1322 (neon_vmul<mode>): Likewise.
1323 (neon_vfms<VCVTF:mode>): Likewise.
1324 (neon_vmlal<mode>): Likewise.
1325 (neon_vmls<mode>): Likewise.
1326 (neon_vmlsl<mode>): Likewise.
1327 (neon_vqdmulh<mode>): Likewise.
1328 (neon_vqdmlal<mode>): Likewise.
1329 (neon_vqdmlsl<mode>): Likewise.
1330 (neon_vmull<mode>): Likewise.
1331 (neon_vqdmull<mode>): Likewise.
1332 (neon_vsub<mode>_unspec): Likewise.
1333 (neon_vsubl<mode>): Likewise.
1334 (neon_vsubw<mode>): Likewise.
1335 (neon_vqsub<mode>): Likewise.
1336 (neon_vhsub<mode>): Likewise.
1337 (neon_vsubhn<mode>): Likewise.
1338 (neon_vceq<mode>): Likewise.
1339 (neon_vcge<mode>): Likewise.
1340 (neon_vcgeu<mode>): Likewise.
1341 (neon_vcgt<mode>): Likewise.
1342 (neon_vcgtu<mode>): Likewise.
1343 (neon_vcle<mode>): Likewise.
1344 (neon_vclt<mode>): Likewise.
1345 (neon_vcage<mode>): Likewise.
1346 (neon_vcagt<mode>): Likewise.
1347 (neon_vtst<mode>): Likewise.
1348 (neon_vabd<mode>): Likewise.
1349 (neon_vabdl<mode>): Likewise.
1350 (neon_vaba<mode>): Likewise.
1351 (neon_vabal<mode>): Likewise.
1352 (neon_vmax<mode>): Likewise.
1353 (neon_vmin<mode>): Likewise.
1354 (neon_vpaddl<mode>): Likewise.
1355 (neon_vpadal<mode>): Likewise.
1356 (neon_vpmax<mode>): Likewise.
1357 (neon_vpmin<mode>): Likewise.
1358 (neon_vrecps<mode>): Likewise.
1359 (neon_vrsqrts<mode>): Likewise.
1360 (neon_vqabs<mode>): Likewise.
1361 (neon_vqneg<mode>): Likewise.
1362 (neon_vcls<mode>): Likewise.
1363 (clz<mode>2): Likewise.
1364 (popcount<mode>2): Likewise.
1365 (neon_vrecpe<mode>): Likewise.
1366 (neon_vrsqrte<mode>): Likewise.
1367 (neon_vget_lane<mode>_sext_internal): Likewise.
1368 (neon_vget_lane<mode>_zext_internal): Likewise.
1369 (neon_vdup_n<mode>): Likewise.
1370 (neon_vdup_n<mode>): Likewise.
1371 (neon_vdup_nv2di): Likewise.
1372 (neon_vdup_lane<mode>_interal): Likewise.
1373 (*neon_vswp<mode>): Likewise.
1374 (neon_vcombine<mode>): Likewise.
1375 (float<mode><V_cvtto>2): Likewise.
1376 (floatuns<mode><V_cvtto>2): Likewise.
1377 (fix_trunc<mode><V_cvtto>2): Likewise.
1378 (fixuns_trunc<mode><V_cvtto>2
1379 (neon_vcvt<mode>): Likewise.
1380 (neon_vcvt<mode>): Likewise.
1381 (neon_vcvtv4sfv4hf): Likewise.
1382 (neon_vcvtv4hfv4sf): Likewise.
1383 (neon_vcvt_n<mode>): Likewise.
1384 (neon_vcvt_n<mode>): Likewise.
1385 (neon_vmovn<mode>): Likewise.
1386 (neon_vqmovn<mode>): Likewise.
1387 (neon_vqmovun<mode>): Likewise.
1388 (neon_vmovl<mode>): Likewise.
1389 (neon_vmul_lane<mode>): Likewise.
1390 (neon_vmul_lane<mode>): Likewise.
1391 (neon_vmull_lane<mode>): Likewise.
1392 (neon_vqdmull_lane<mode>): Likewise.
1393 (neon_vqdmulh_lane<mode>): Likewise.
1394 (neon_vqdmulh_lane<mode>): Likewise.
1395 (neon_vmla_lane<mode>): Likewise.
1396 (neon_vmla_lane<mode>): Likewise.
1397 (neon_vmlal_lane<mode>): Likewise.
1398 (neon_vqdmlal_lane<mode>): Likewise.
1399 (neon_vmls_lane<mode>): Likewise.
1400 (neon_vmls_lane<mode>): Likewise.
1401 (neon_vmlsl_lane<mode>): Likewise.
1402 (neon_vqdmlsl_lane<mode>): Likewise.
1403 (neon_vext<mode>): Likewise.
1404 (neon_vrev64<mode>): Likewise.
1405 (neon_vrev32<mode>): Likewise.
1406 (neon_vrev16<mode>): Likewise.
1407 (neon_vbsl<mode>_internal): Likewise.
1408 (neon_vshl<mode>): Likewise.
1409 (neon_vqshl<mode>): Likewise.
1410 (neon_vshr_n<mode>): Likewise.
1411 (neon_vshrn_n<mode>): Likewise.
1412 (neon_vqshrn_n<mode>): Likewise.
1413 (neon_vqshrun_n<mode>): Likewise.
1414 (neon_vshl_n<mode>): Likewise.
1415 (neon_vqshl_n<mode>): Likewise.
1416 (neon_vqshlu_n<mode>): Likewise.
1417 (neon_vshll_n<mode>): Likewise.
1418 (neon_vsra_n<mode>): Likewise.
1419 (neon_vsri_n<mode>): Likewise.
1420 (neon_vsli_n<mode>): Likewise.
1421 (neon_vtbl1v8qi): Likewise.
1422 (neon_vtbl2v8qi): Likewise.
1423 (neon_vtbl3v8qi): Likewise.
1424 (neon_vtbl4v8qi): Likewise.
1425 (neon_vtbl1v16qi): Likewise.
1426 (neon_vtbl2v16qi): Likewise.
1427 (neon_vcombinev16qi): Likewise.
1428 (neon_vtbx1v8qi): Likewise.
1429 (neon_vtbx2v8qi): Likewise.
1430 (neon_vtbx3v8qi): Likewise.
1431 (neon_vtbx4v8qi): Likewise.
1432 (*neon_vtrn<mode>_insn): Likewise.
1433 (*neon_vzip<mode>_insn): Likewise.
1434 (*neon_vuzp<mode>_insn): Likewise.
1435 (neon_vld1<mode>): Likewise.
1436 (neon_vld1_lane<mode>): Likewise.
1437 (neon_vld1_lane<mode>): Likewise.
1438 (neon_vld1_dup<mode>): Likewise.
1439 (neon_vld1_dup<mode>): Likewise.
1440 (neon_vld1_dupv2di): Likewise.
1441 (neon_vst1<mode>): Likewise.
1442 (neon_vst1_lane<mode>): Likewise.
1443 (neon_vst1_lane<mode>): Likewise.
1444 (neon_vld2<mode>): Likewise.
1445 (neon_vld2<mode>): Likewise.
1446 (neon_vld2_lane<mode>): Likewise.
1447 (neon_vld2_lane<mode>): Likewise.
1448 (neon_vld2_dup<mode>): Likewise.
1449 (neon_vst2<mode>): Likewise.
1450 (neon_vst2<mode>): Likewise.
1451 (neon_vst2_lane<mode>): Likewise.
1452 (neon_vst2_lane<mode>): Likewise.
1453 (neon_vld3<mode>): Likewise.
1454 (neon_vld3qa<mode>): Likewise.
1455 (neon_vld3qb<mode>): Likewise.
1456 (neon_vld3_lane<mode>): Likewise.
1457 (neon_vld3_lane<mode>): Likewise.
1458 (neon_vld3_dup<mode>): Likewise.
1459 (neon_vst3<mode>): Likewise.
1460 (neon_vst3qa<mode>): Likewise.
1461 (neon_vst3qb<mode>): Likewise.
1462 (neon_vst3_lane<mode>): Likewise.
1463 (neon_vst3_lane<mode>): Likewise.
1464 (neon_vld4<mode>): Likewise.
1465 (neon_vld4qa<mode>): Likewise.
1466 (neon_vld4qb<mode>): Likewise.
1467 (neon_vld4_lane<mode>): Likewise.
1468 (neon_vld4_lane<mode>): Likewise.
1469 (neon_vld4_dup<mode>): Likewise.
1470 (neon_vst4<mode>): Likewise.
1471 (neon_vst4qa<mode>): Likewise.
1472 (neon_vst4qb<mode>): Likewise.
1473 (neon_vst4_lane<mode>): Likewise.
1474 (neon_vst4_lane<mode>): Likewise.
1475 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1476 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1477 (neon_vec_<US>mult_lo_<mode>): Likewise.
1478 (neon_vec_<US>mult_hi_<mode>): Likewise.
1479 (neon_vec_<US>shiftl_<mode>): Likewise.
1480 (neon_unpack<US>_<mode>): Likewise.
1481 (neon_vec_<US>mult_<mode>): Likewise.
1482 (vec_pack_trunc_<mode>): Likewise.
1483 (neon_vec_pack_trunc_<mode>): Likewise.
1484 (neon_vabd<mode>_2): Likewise.
1485 (neon_vabd<mode>_3): Likewise.
1487 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1489 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1490 (load_pair): Update type attribute.
1491 (store_pair): Update type attribute.
1492 * config/aarch64/iterators.md (q): New.
1494 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1496 * config/arm/types.md: Add new types for Neon insns.
1498 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1499 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1500 Sergey Lega <sergey.s.lega@intel.com>
1501 Anna Tikhonova <anna.tikhonova@intel.com>
1502 Ilya Tocar <ilya.tocar@intel.com>
1503 Andrey Turetskiy <andrey.turetskiy@intel.com>
1504 Ilya Verbin <ilya.verbin@intel.com>
1505 Kirill Yukhin <kirill.yukhin@intel.com>
1506 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1508 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1509 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1510 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1512 (srcp14<mode>): Ditto.
1513 (rsqrt14<mode>): Ditto.
1514 (rsqrt14<mode>): Ditto.
1515 (avx512f_vmscalef<mode>): Ditto.
1516 (avx512f_scalef<mode>): Ditto.
1517 (avx512f_getexp<mode>): Ditto.
1518 (avx512f_sgetexp<mode>): Ditto.
1519 (avx512f_fixupimm<mode>): Ditto.
1520 (avx512f_sfixupimm<mode>): Ditto.
1521 (avx512f_rndscale<mode>): Ditto.
1522 (*avx512er_exp2<mode>): Ditto.
1523 (*avx512er_rcp28<mode>): Ditto.
1524 (avx512er_rsqrt28<mode>): Ditto.
1525 (avx512f_getmant<mode>): Ditto.
1526 (avx512f_getmant<mode>): Ditto.
1527 (avx512f_rndscale<mode>): Fix formatting.
1529 2013-10-15 Martin Jambor <mjambor@suse.cz>
1531 * ipa-utils.h (ipa_edge_within_scc): Declare.
1532 * ipa-cp.c (edge_within_scc): Moved...
1533 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1535 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1536 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1537 Sergey Lega <sergey.s.lega@intel.com>
1538 Anna Tikhonova <anna.tikhonova@intel.com>
1539 Ilya Tocar <ilya.tocar@intel.com>
1540 Andrey Turetskiy <andrey.turetskiy@intel.com>
1541 Ilya Verbin <ilya.verbin@intel.com>
1542 Kirill Yukhin <kirill.yukhin@intel.com>
1543 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1545 * config/i386/predicates.md (const_8_to_15_operand): New.
1546 (const_16_to_31_operand): Ditto.
1547 * config/i386/sse.md (V8FI): New.
1549 (reduc_splus_v8df): Ditto.
1550 (reduc_splus_v16sf): Ditto.
1551 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1552 (vec_extract_hi_<mode>): Ditto.
1553 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1554 (vec_set_lo_<mode>): Ditto.
1555 (vec_set_hi_<mode>): Ditto.
1556 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1557 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1558 (avx512f_pshufd_1): Ditto.
1559 (avx512f_broadcast<mode>): Ditto.
1560 (avx512f_broadcast<mode>): Ditto.
1561 (define_split): Split vec_extract_lo into move.
1562 (ssequartermode): Ditto.
1563 (ssedoublemode): Extened with wider modes.
1564 (vec_extract_lo_<mode>): Ditto.
1566 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1567 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1568 Sergey Lega <sergey.s.lega@intel.com>
1569 Anna Tikhonova <anna.tikhonova@intel.com>
1570 Ilya Tocar <ilya.tocar@intel.com>
1571 Andrey Turetskiy <andrey.turetskiy@intel.com>
1572 Ilya Verbin <ilya.verbin@intel.com>
1573 Kirill Yukhin <kirill.yukhin@intel.com>
1574 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1576 * config/i386/predicates.md (register_or_constm1_operand): New.
1577 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1578 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1579 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1580 UNSPEC_SCATTER_PREFETCH
1582 (avx512f_ucmp<mode>3): Ditto.
1583 (avx512f_vternlog<mode>): Ditto.
1584 (avx512f_align<mode>): Ditto.
1585 (<shift_insn><mode>3): Ditto.
1586 (avx512f_<rotate>v<mode>): Ditto.
1587 (avx512f_<rotate><mode>): Ditto.
1588 (avx512f_eq<mode>3): Ditto.
1589 (avx512f_eq<mode>3_1): Ditto.
1590 (avx512f_gt<mode>3): Ditto.
1591 (avx512f_testm<mode>3): Ditto.
1592 (avx512f_testnm<mode>3): Ditto.
1593 (avx512pf_gatherpf<mode>): Ditto.
1594 (*avx512pf_gatherpf<mode>_mask): Ditto.
1595 (*avx512pf_gatherpf<mode>): Ditto.
1596 (avx512pf_scatterpf<mode>): Ditto.
1597 (*avx512pf_scatterpf<mode>_mask): Ditto.
1598 (*avx512pf_scatterpf<mode>): Ditto.
1599 (avx512f_vec_dup_gpr<mode>): Ditto.
1600 (clz<mode>2): Ditto.
1601 (conflict<mode>): Ditto.
1602 (REDUC_SMINMAX_MODE): Extened with wider modes.
1603 (reduc_<code>_<mode>): Ditto.
1604 (vlshr<mode>3): Ditto.
1605 (vashl<mode>3): Ditto.
1607 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1608 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1609 Sergey Lega <sergey.s.lega@intel.com>
1610 Anna Tikhonova <anna.tikhonova@intel.com>
1611 Ilya Tocar <ilya.tocar@intel.com>
1612 Andrey Turetskiy <andrey.turetskiy@intel.com>
1613 Ilya Verbin <ilya.verbin@intel.com>
1614 Kirill Yukhin <kirill.yukhin@intel.com>
1615 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1617 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1620 (avx512fmaskmode): Ditto.
1621 (bcstscalarsuff): Ditto.
1622 (avx512f_blendm<mode>): Ditto.
1623 (cmp_imm_predicate): Ditto.
1624 (avx512f_cmp<mode>3): Ditto.
1625 (avx512f_vec_dup<mode>): Ditto.
1626 (avx512f_vec_dup_mem<mode>): Ditto.
1627 (avx512f_vpermi2var<mode>3): Ditto.
1628 (avx512f_vpermt2var<mode>3): Ditto.
1629 (vec_init<mode>): Ditto.
1630 (avx512f_gathersi<mode>): Ditto.
1631 (*avx512f_gathersi<mode>): Ditto.
1632 (*avx512f_gathersi<mode>_2): Ditto.
1633 (avx512f_gatherdi<mode>): Ditto.
1634 (*avx512f_gatherdi<mode>): Ditto.
1635 (*avx512f_gatherdi<mode>_2): Ditto.
1636 (avx512f_scattersi<mode>): Ditto.
1637 (*avx512f_scattersi<mode>): Ditto.
1638 (avx512f_scatterdi<mode>): Ditto.
1639 (*avx512f_scatterdi<mode>): Ditto.
1640 (sseintprefix): Extened with wider modes.
1641 (VEC_GATHER_IDXSI): Ditto.
1642 (VEC_GATHER_IDXDI): Ditto.
1643 (VEC_GATHER_SRCDI): Ditto.
1645 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1646 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1648 * config/arm/t-aprofile: New file.
1649 * config.gcc: Handle --with-multilib-list option.
1651 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1653 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1656 2013-10-15 Richard Biener <rguenther@suse.de>
1658 * tree-loop-distribution.c (build_empty_rdg): Inline into
1660 (rdg_flag_vertex): Inline into single user.
1661 (rdg_flag_vertex_and_dependent): Likewise.
1662 (build_rdg_partition_for_vertex): Remove processed bitmap.
1663 (rdg_build_partitions): Simplify.
1665 2013-10-15 Richard Biener <rguenther@suse.de>
1667 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1668 Restructure forwarding through conversions and copies to
1669 avoid performing copy-propagation the wrong way. Adjust
1670 recursion invocations.
1671 (forward_propagate_addr_expr): Add argument stating if we
1672 are recursing from a single-use.
1673 (ssa_forward_propagate_and_combine): Adjust.
1675 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1677 * dumpfile.h (gcc::dump_manager): New class, to hold state
1678 relating to dumpfile management.
1679 (get_dump_file_name): Remove in favor of method of dump_manager.
1680 (dump_initialized_p): Likewise.
1681 (dump_start): Likewise.
1682 (dump_finish): Likewise.
1683 (dump_switch_p): Likewise.
1684 (dump_register): Likewise.
1685 (get_dump_file_info): Likewise.
1686 * context.c (gcc::context::context): Construct the dump_manager
1688 * context.h (gcc::context::get_dumps): New.
1689 (gcc::context::m_dumps): New.
1690 * coverage.c (coverage_init): Port to dump_manager API.
1691 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1692 (extra_dump_files_in_use): Likewise.
1693 (extra_dump_files_alloced): Likewise.
1694 (gcc::dump_manager::dump_manager): New.
1695 (dump_register): Convert to...
1696 (gcc::dump_manager::dump_register): ...method, replacing
1697 function-static next_dump with m_next_dump field.
1698 (get_dump_file_info): Convert to...
1699 (gcc::dump_manager::get_dump_file_info): ...method.
1700 (get_dump_file_name): Convert to...
1701 (gcc::dump_manager::get_dump_file_name): ...method.
1702 (dump_start): Convert to...
1703 (gcc::dump_manager::dump_start): ...method.
1704 (dump_finish): Convert to...
1705 (gcc::dump_manager::dump_finish): ...method.
1706 (dump_begin): Replace body with...
1707 (gcc::dump_manager::dump_begin): ...new method.
1708 (dump_phase_enabled_p): Convert to...
1709 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1710 (dump_phase_enabled_p): Convert to...
1711 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1712 (dump_initialized_p): Convert to...
1713 (gcc::dump_manager::dump_initialized_p): ...method.
1714 (dump_flag_name): Replace body with...
1715 (gcc::dump_manager::dump_flag_name): ...new method.
1716 (dump_enable_all): Convert to...
1717 (gcc::dump_manager::dump_enable_all): ...new method.
1718 (opt_info_enable_passes): Convert to...
1719 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1720 (dump_switch_p_1): Convert to...
1721 (gcc::dump_manager::dump_switch_p_1): ...new method.
1722 (dump_switch_p): Convert to...
1723 (gcc::dump_manager::dump_switch_p): ...new method.
1724 (opt_info_switch_p): Port to dump_manager API.
1725 (enable_rtl_dump_file): Likewise.
1726 * opts-global.c (handle_common_deferred_options): Port to new
1728 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1729 (pass_manager::register_one_dump_file): Likewise.
1730 (pass_manager::register_pass): Likewise.
1731 (pass_init_dump_file): Likewise.
1732 (pass_fini_dump_file): Likewise.
1733 * statistics.c (statistics_early_init): Likewise.
1735 2013-10-14 Richard Biener <rguenther@suse.de>
1737 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1738 iterative_hash_canonical_type, gimple_canonical_type_hash,
1739 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1740 gimple_register_canonical_type, print_gimple_types_stats,
1741 free_gimple_type_tables): Move to lto/lto.c
1742 (gt-gimple.h): Do not include.
1743 * gimple.h (gimple_register_canonical_type,
1744 print_gimple_types_stats, free_gimple_type_tables): Remove.
1745 * Makefile.in (GTFILES): Remove gimple.c.
1747 2013-10-14 Travis Snoozy <quandary@remstate.com>
1750 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1751 scanning for msp430x targets.
1753 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1756 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1757 (registers_ok_for_ldd_peep): Move around.
1758 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1759 * config/sparc/sparc.md (widening peepholes): Use it.
1761 2013-10-14 Richard Biener <rguenther@suse.de>
1765 * gimple.c (iterative_hash_canonical_type): Make sure to
1766 record the hash into the correct hashtable slot.
1768 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1770 PR rtl-optimization/58662
1771 * combine.c (try_combine): Take into account death nodes on I2 when
1772 splitting a PARALLEL of two independent SETs. Fix dump message.
1774 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1777 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1778 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1779 * config/sh/t-sh (sh_treg_combine.o): New entry.
1780 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1781 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1782 (register_sh_passes): New function. Register sh_treg_combine pass.
1783 (sh_option_override): Invoke it.
1784 (sh_canonicalize_comparison): Handle op0_preserve_value.
1785 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1786 opportunities. Canonicalize branch condition.
1787 (nott): Allow only if pseudos can be created for non-SH2A.
1789 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1792 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1793 (ix86_expand_movmem): Replace copy_addr_to_reg with
1794 ix86_copy_addr_to_reg.
1795 (ix86_expand_setmem): Likewise.
1797 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1799 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1800 provided by ix86_fp_compare_mode instead of CCFPUmode.
1802 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1804 * config/aarch64/arm_neon.h
1805 (vtbx<1,3>_<psu>8): Fix register constriants.
1807 2013-10-11 Jeff Law <law@redhat.com>
1809 PR tree-optimization/58640
1810 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1811 threading paths that cross over two loop entry points.
1813 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1815 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1816 handle vector float as well.
1817 (*vsx_le_perm_load_v4si): Likewise.
1818 (*vsx_le_perm_store_v2di): Likewise.
1819 (*vsx_le_perm_store_v4si): Likewise.
1821 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1823 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1824 directly to circumvent subtract from splat{31} workaround.
1825 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1827 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1828 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1829 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1830 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1831 separate big and little endian logic.
1832 (*altivec_vperm_<mode>_internal): New define_insn.
1833 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1834 separate big and little endian logic.
1835 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1836 (vec_permv16qi): Add little endian logic.
1838 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1840 * doc/extend.texi (returns_nonnull): Remove arguments.
1842 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1843 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1844 Sergey Lega <sergey.s.lega@intel.com>
1845 Anna Tikhonova <anna.tikhonova@intel.com>
1846 Ilya Tocar <ilya.tocar@intel.com>
1847 Andrey Turetskiy <andrey.turetskiy@intel.com>
1848 Ilya Verbin <ilya.verbin@intel.com>
1849 Kirill Yukhin <kirill.yukhin@intel.com>
1850 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1852 * config/i386/sse.md (VI48F_256_512): New.
1853 (avx2_permvar<mode>): Change to ...
1854 (<avx2_avx512f>_permvar<mode>): This.
1856 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1857 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1858 Sergey Lega <sergey.s.lega@intel.com>
1859 Anna Tikhonova <anna.tikhonova@intel.com>
1860 Ilya Tocar <ilya.tocar@intel.com>
1861 Andrey Turetskiy <andrey.turetskiy@intel.com>
1862 Ilya Verbin <ilya.verbin@intel.com>
1863 Kirill Yukhin <kirill.yukhin@intel.com>
1864 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1866 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1867 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1868 * config/i386/sse.md (VI4_AVX): New.
1869 (sf2simodelower): Ditto.
1870 (sse2_cvtps2dq): Change to ...
1871 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1873 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1874 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1875 Sergey Lega <sergey.s.lega@intel.com>
1876 Anna Tikhonova <anna.tikhonova@intel.com>
1877 Ilya Tocar <ilya.tocar@intel.com>
1878 Andrey Turetskiy <andrey.turetskiy@intel.com>
1879 Ilya Verbin <ilya.verbin@intel.com>
1880 Kirill Yukhin <kirill.yukhin@intel.com>
1881 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1883 * config/i386/sse.md (V_512): New.
1885 (vcond<V_512:mode><VF_512:mode>): Ditto.
1886 (vcond<V_512:mode><VI_512:mode>): Ditto.
1887 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1889 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1890 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1891 Sergey Lega <sergey.s.lega@intel.com>
1892 Anna Tikhonova <anna.tikhonova@intel.com>
1893 Ilya Tocar <ilya.tocar@intel.com>
1894 Andrey Turetskiy <andrey.turetskiy@intel.com>
1895 Ilya Verbin <ilya.verbin@intel.com>
1896 Kirill Yukhin <kirill.yukhin@intel.com>
1897 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1899 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1900 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1902 (fma<mode>4): Removed condition.
1903 (fms<mode>4): Ditto.
1904 (fnma<mode>4): Ditto.
1905 (fnms<mode>4): Ditto.
1906 (fma4i_fmadd_<mode>): Ditto.
1907 (*fma_fmadd_<mode>): Ditto.
1908 (*fma_fmsub_<mode>): Ditto.
1909 (*fma_fnmadd_<mode>): Ditto.
1910 (*fma_fnmsub_<mode>): Ditto.
1911 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1912 (*fma_fmaddsub_<mode>): Ditto.
1913 (*fma_fmsubadd_<mode>): Ditto.
1914 (*fmai_fmadd_<mode>): Ditto.
1915 (*fmai_fmsub_<mode>): Ditto.
1916 (*fmai_fnmadd_<mode>): Ditto.
1917 (*fmai_fnmsub_<mode>): Ditto.
1919 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1920 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1921 Sergey Lega <sergey.s.lega@intel.com>
1922 Anna Tikhonova <anna.tikhonova@intel.com>
1923 Ilya Tocar <ilya.tocar@intel.com>
1924 Andrey Turetskiy <andrey.turetskiy@intel.com>
1925 Ilya Verbin <ilya.verbin@intel.com>
1926 Kirill Yukhin <kirill.yukhin@intel.com>
1927 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1929 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1930 (VI124_256): Changed to ...
1931 (VI124_256_48_512): This.
1932 (ssepackmode): Extended with wider modes.
1933 (<code><mode>3): Changed iterator.
1934 (*avx2_<code><mode>3): Ditto.
1935 (vec_pack_trunc_<mode>): Ditto.
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 (VI124_AVX2_48_AVX512F): New.
1948 (VI8F_256_512): Ditto.
1949 (abs<mode>2): Changed iterator.
1950 (avx2_perm<mode>): Changed to ...
1951 (<avx2_avx512f>_perm<mode>): This.
1952 (avx2_perm<mode>_1): Changed to ...
1953 (<avx2_avx512f>_perm<mode>_1): This.
1955 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1956 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1957 Sergey Lega <sergey.s.lega@intel.com>
1958 Anna Tikhonova <anna.tikhonova@intel.com>
1959 Ilya Tocar <ilya.tocar@intel.com>
1960 Andrey Turetskiy <andrey.turetskiy@intel.com>
1961 Ilya Verbin <ilya.verbin@intel.com>
1962 Kirill Yukhin <kirill.yukhin@intel.com>
1963 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1965 * config/i386/sse.md (VI48_AVX512F): New.
1966 (VI48_AVX2): Changed to ...
1967 (VI48_AVX2_48_AVX512F): This.
1968 (avx2_ashrv<mode>): Changed to ...
1969 (<avx2_avx512f>_ashrv<mode>): This.
1970 (avx2_<shift_insn>v<mode>): Changed to ...
1971 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1973 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1974 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1975 Sergey Lega <sergey.s.lega@intel.com>
1976 Anna Tikhonova <anna.tikhonova@intel.com>
1977 Ilya Tocar <ilya.tocar@intel.com>
1978 Andrey Turetskiy <andrey.turetskiy@intel.com>
1979 Ilya Verbin <ilya.verbin@intel.com>
1980 Kirill Yukhin <kirill.yukhin@intel.com>
1981 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1983 * config/i386/sse.md (VI4_AVX512F): New.
1984 (VI8_AVX2_AVX512F): Ditto.
1985 (mul<mode>3): Extended with wider modes.
1986 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1987 (mul<mode>3): Ditto.
1988 (vec_widen_<s>mult_odd_<mode>): Ditto.
1990 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1991 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1992 Sergey Lega <sergey.s.lega@intel.com>
1993 Anna Tikhonova <anna.tikhonova@intel.com>
1994 Ilya Tocar <ilya.tocar@intel.com>
1995 Andrey Turetskiy <andrey.turetskiy@intel.com>
1996 Ilya Verbin <ilya.verbin@intel.com>
1997 Kirill Yukhin <kirill.yukhin@intel.com>
1998 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2000 * config/i386/sse.md (VI2_AVX512F): New.
2001 (VI124_AVX512F): Ditto.
2002 (sseunpackmode): Extended with wider modes.
2003 (sseunpackfltmode): Ditto.
2004 (vec_unpacks_float_hi_<mode>): Ditto.
2005 (vec_unpacks_float_lo_<mode>): Ditto.
2006 (vec_unpacku_float_hi_<mode>): Ditto.
2007 (vec_unpacku_float_lo_<mode>): Ditto.
2008 (vec_unpacks_lo_<mode>): Ditto.
2009 (vec_unpacks_hi_<mode>): Ditto.
2010 (vec_unpacku_lo_<mode>): Ditto.
2011 (vec_unpacku_hi_<mode>): Ditto.
2013 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2014 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2015 Sergey Lega <sergey.s.lega@intel.com>
2016 Anna Tikhonova <anna.tikhonova@intel.com>
2017 Ilya Tocar <ilya.tocar@intel.com>
2018 Andrey Turetskiy <andrey.turetskiy@intel.com>
2019 Ilya Verbin <ilya.verbin@intel.com>
2020 Kirill Yukhin <kirill.yukhin@intel.com>
2021 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2023 * config/i386/i386.md (multdiv): New.
2024 (multdiv_mnemonic): Ditto.
2025 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2026 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2027 (<sse>_vmdiv<mode>3): Removed.
2029 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2030 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2031 Sergey Lega <sergey.s.lega@intel.com>
2032 Anna Tikhonova <anna.tikhonova@intel.com>
2033 Ilya Tocar <ilya.tocar@intel.com>
2034 Andrey Turetskiy <andrey.turetskiy@intel.com>
2035 Ilya Verbin <ilya.verbin@intel.com>
2036 Kirill Yukhin <kirill.yukhin@intel.com>
2037 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2039 * config/i386/sse.md (V): Extended with wider modes.
2041 (ssehalfvecmode): Ditto.
2043 (ssepackfltmode): Ditto.
2044 (avx_vec_concat<mode>): Ditto.
2045 (V_256_512): New iterator.
2046 (VF2_512_256): Ditto.
2047 (si2dfmode): New attribute.
2048 (si2dfmodelower): Ditto.
2050 (concat_tg_mode): Ditto.
2051 (floatv4siv4df2): Changed to ...
2052 (float<si2dfmodelower><mode>2): This.
2053 (avx_cvtps2pd256): Changed to ...
2054 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2055 (vec_pack_trunc_v4df): Changed to ...
2056 (vec_pack_trunc_<mode>): This.
2057 (avx_vpermil<mode>): Changed to ...
2058 (<sse2_avx_avx512f>_vpermil<mode>): This.
2059 (<fixsuffix>fix_truncv8dfv8si2): New.
2060 (vec_pack_sfix_trunc_v8df): Ditto.
2061 (avx512f_rndscale<mode>): Ditto.
2062 (avx512f_roundpd512): Ditto.
2063 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2065 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2066 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2067 Sergey Lega <sergey.s.lega@intel.com>
2068 Anna Tikhonova <anna.tikhonova@intel.com>
2069 Ilya Tocar <ilya.tocar@intel.com>
2070 Andrey Turetskiy <andrey.turetskiy@intel.com>
2071 Ilya Verbin <ilya.verbin@intel.com>
2072 Kirill Yukhin <kirill.yukhin@intel.com>
2073 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2075 * config/i386/i386.md (any_fix): New iterator.
2076 (fixsuffix): New attribute.
2077 * config/i386/sse.md (VF1): Extened with wider modes.
2081 (sseintvecmodelower): Ditto.
2082 (ssescalarmode): Ditto.
2083 (ssescalarnum): Ditto.
2085 (ssexmmmode): Ditto.
2086 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2087 (<sse>_rcp<mode>2): Change iterator.
2088 (rsqrt<mode>2): Ditto.
2089 (<sse>_rsqrt<mode>2): Ditto.
2090 (avx2_vec_dup<mode>): Ditto.
2091 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2092 (round<mode>2_sfix): Ditto.
2093 (avx2_pbroadcast<mode>): Ditto.
2094 (*andnot<mode>3): Handle XI mode.
2095 (*<code><mode>3): Ditto.
2096 (AVXTOSSEMODE): Removed.
2097 (avx_vpermil<mode>): Changed to ...
2098 (<sse2_avx_avx512f>_vpermil<mode>): This.
2100 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2101 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2102 Sergey Lega <sergey.s.lega@intel.com>
2103 Anna Tikhonova <anna.tikhonova@intel.com>
2104 Ilya Tocar <ilya.tocar@intel.com>
2105 Andrey Turetskiy <andrey.turetskiy@intel.com>
2106 Ilya Verbin <ilya.verbin@intel.com>
2107 Kirill Yukhin <kirill.yukhin@intel.com>
2108 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2110 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2111 (<sse>_comi): Ditto.
2112 (<sse>_ucomi): Ditto.
2113 (sse_cvtss2siq_2): Ditto.
2114 (sse2_cvtsd2si): Ditto.
2115 (sse2_cvtsd2siq): Ditto.
2116 (sse2_cvttsd2si): Ditto.
2117 (sse2_cvttsd2siq): Ditto.
2118 (<shift_insn><mode>3): Ditto.
2119 (sse2_cvtsi2sdq): Update constraint and prefix.
2120 (sse_cvtsi2ss): Update prefix.
2121 (sse_cvtsi2ssq): Ditto.
2123 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2125 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2126 ignore internal calls.
2128 2013-10-11 Richard Biener <rguenther@suse.de>
2130 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2131 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2132 TREE_OVERFLOW is set.
2134 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2136 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2138 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2139 * gimple.h: Likewise.
2141 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2143 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2146 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2148 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2149 * configure: Regenerate.
2151 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2153 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2154 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2155 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2156 similarly pp_right_paren.
2157 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2158 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2159 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2160 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2162 (omp_declare_simd_clauses_equal,
2163 omp_remove_redundant_declare_simd_attrs): New functions.
2164 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2165 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2166 * tree.h (OMP_LOOP_CHECK): Define.
2167 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2168 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2169 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2170 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2171 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2172 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2173 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2174 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2175 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2176 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2177 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2178 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2179 (omp_remove_redundant_declare_simd_attrs): New prototype.
2180 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2181 GIMPLE_OMP_TEAMS): New codes.
2182 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2183 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2185 (target_nesting_level): New variable.
2186 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2187 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2188 for collapse > 1 static schedule unless ordered.
2189 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2190 (determine_parallel_type): Adjust get_ws_args_for caller.
2191 (install_var_field): Handle mask & 4 for double indirection.
2192 (scan_sharing_clauses): Ignore shared clause on teams construct.
2193 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2194 (create_omp_child_function): If inside target or declare target
2195 constructs, set "omp declare target" attribute on the child function.
2196 (find_combined_for): New function.
2197 (scan_omp_parallel): Handle combined loops.
2198 (scan_omp_target, scan_omp_teams): New functions.
2199 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2200 restrictions and set ctx->cancellable for cancellable constructs.
2201 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2202 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2203 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2204 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2205 (omp_clause_aligned_alignment): New function.
2206 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2207 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2208 on teams constructs. Handle user defined reductions and new
2210 (lower_reduction_clauses): Don't set placeholder to address of ref
2211 if it has already the right type.
2212 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2213 (expand_parallel_call): Use the new non-_start suffixed builtins,
2214 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2215 and GOMP_parallel_end after the call.
2216 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2217 (expand_omp_for_init_counts): Handle combined loops.
2218 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2220 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2221 end of cancellable loops.
2222 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2223 Likewise. Handle collapse > 1 loops.
2224 (expand_omp_simd): Handle combined loops.
2225 (expand_omp_for): Add inner_stmt argument, adjust callers of
2226 expand_omp_for* functions, use expand_omp_for_static*chunk even
2227 for collapse > 1 unless ordered.
2228 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2229 of cancellable sections.
2230 (expand_omp_single): Remove need_barrier variable, just rely on
2231 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2232 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2233 (expand_omp_atomic_load, expand_omp_atomic_store,
2234 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2235 (expand_omp_target): New function.
2236 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2237 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2238 (build_omp_regions_1): Immediately close region for
2239 GF_OMP_TARGET_KIND_UPDATE.
2240 (maybe_add_implicit_barrier_cancel): New function.
2241 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2243 (lower_omp_single): Likewise. Add clobber after the barrier.
2244 (lower_omp_taskgroup): New function.
2245 (lower_omp_for): Handle combined loops. Adjust
2246 lower_rec_input_clauses caller. Handle cancellation.
2247 (lower_depend_clauses): New function.
2248 (lower_omp_taskreg): Lower depend clauses. Adjust
2249 lower_rec_input_clauses caller. Add clobber after the call. Handle
2251 (lower_omp_target, lower_omp_teams): New functions.
2252 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2253 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2254 and GOMP_cancellation_point calls.
2255 (lower_omp): Fold stmts inside of target region.
2256 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2257 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2258 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2259 (BT_FN_VOID_OMPFN_PTR_UINT,
2260 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2261 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2262 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2263 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2264 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2265 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2266 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2267 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2268 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2269 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2270 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2271 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2272 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2273 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2274 BUILT_IN_GOMP_PARALLEL_END.
2275 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2276 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2277 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2278 GF_OMP_FOR_KIND_DISTRIBUTE.
2279 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2280 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2281 (dump_gimple_omp_return): Print lhs if it has any.
2282 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2283 gimple_omp_atomic_seq_cst_p.
2284 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2285 and GIMPLE_OMP_TEAMS.
2286 * langhooks.c (lhd_omp_mappable_type): New function.
2287 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2288 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2290 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2291 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2292 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2293 (struct gimplify_omp_ctx): Add combined_loop field.
2294 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2295 on stmts inside of target region.
2296 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2297 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2298 ORT_TARGET and ORT_TARGET_DATA.
2299 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2301 (omp_notice_threadprivate_variable): Complain about threadprivate
2302 variables in target region.
2303 (omp_notice_variable): Complain about vars with non-mappable type
2304 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2305 (omp_check_private): Ignore ORT_TARGET* regions.
2306 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2307 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2308 (find_combined_omp_for): New function.
2309 (gimplify_omp_for): Handle gimplification of combined loops.
2310 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2312 (gimplify_omp_target_update): New function.
2313 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2314 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2315 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2316 (gimplify_body): If fndecl has "omp declare target" attribute, add
2317 implicit ORT_TARGET context around it.
2318 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2319 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2320 * tree-nested.c (convert_nonlocal_reference_stmt,
2321 convert_local_reference_stmt, convert_gimple_call): Handle
2322 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2323 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2324 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2325 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2326 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2327 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2328 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2329 BUILT_IN_GOMP_LOOP_END_CANCEL,
2330 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2331 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2332 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2333 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2334 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2335 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2336 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2337 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2338 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2339 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2340 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2341 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2342 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2343 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2344 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2345 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2346 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2347 gimple_build_omp_teams): New functions.
2348 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2349 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2350 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2351 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2352 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2353 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2354 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2355 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2356 GF_OMP_ATOMIC_SEQ_CST): New.
2357 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2358 gimple_build_omp_teams): New prototypes.
2359 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2360 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2361 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2362 GIMPLE_OMP_SINGLE as end of range.
2363 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2364 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2365 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2366 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2367 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2368 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2369 gimple_omp_target_kind, gimple_omp_target_set_kind,
2370 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2371 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2372 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2373 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2374 gimple_omp_teams_set_clauses): New inlines.
2375 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2376 and GIMPLE_OMP_TASKGROUP.
2377 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2378 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2379 enum omp_clause_proc_bind_kind): New.
2380 (union omp_clause_subcode): Add depend_kind, map_kind and
2381 proc_bind_kind fields.
2382 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2383 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2384 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2385 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2386 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2388 2013-10-10 Teresa Johnson <tejohnson@google.com>
2390 * predict.c (tree_estimate_probability): Add new parameter
2391 for estimate_bb_frequencies.
2392 (estimate_bb_frequencies): Add new parameter to force estimation.
2393 (rebuild_frequencies): When max frequency in function is small,
2394 recompute counts from frequencies.
2395 * predict.h (estimate_bb_frequencies): New parameter.
2397 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2399 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2400 optimizations are disabled.
2402 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2404 * coverage.c (coverage_finish): Fix leak of da_file_name.
2406 2013-10-10 Jan Hubicka <jh@suse.cz>
2408 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2409 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2412 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2415 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2416 if any labels are in FALLTHRU_BB, use a special label emitted
2417 immediately after the asm goto insn rather than label_rtx
2419 (expand_asm_stmt): Adjust caller.
2420 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2421 edge if the last insn in predecessor is a jump with single successor,
2422 but it isn't simplejump_p.
2424 2013-10-10 Richard Biener <rguenther@suse.de>
2426 PR tree-optimization/58656
2427 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2429 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2431 * gimplify.c: Include expr.h and tm_p.h for targets with special
2432 va-arg padding requirements.
2434 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2436 * tree-flow.h: Move some prototypes to gimple.h.
2437 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2438 * gimple.h: Relocate some prototypes from tree-flow.h
2439 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2441 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2442 (build_va_arg_indirect_ref): Relocate and make static.
2443 (std_gimplify_va_arg_expr): Relocate here.
2444 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2445 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2447 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2449 * doc/md.texi: Document the mnemonic attribute.
2451 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2454 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2455 (set_attr_alternative x ...) when searching for user defined
2458 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2460 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2462 2013-10-09 Easwaran Raman <eraman@google.com>
2464 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2465 * cfgexpand.c (defer_stack_allocation): ...use here
2466 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2468 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2470 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2471 (operand_equal_for_value_replacement): New function, extracted from
2472 value_replacement and enhanced to catch more cases.
2473 (value_replacement): Use operand_equal_for_value_replacement.
2475 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2477 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2478 get_max_loop_iterations.
2480 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2482 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2483 Place comment above function.
2485 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2487 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2488 are not related to tree-cfg.c.
2489 * tree-ssa-address.h: New file. Relocate prototypes.
2490 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2491 (addr_for_mem_ref): New. Combine call to get_address_description and
2492 return addr_for_mem_ref.
2493 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2494 * tree-ssa-live.h: Adjust prototypes.
2495 * passes.c: Include tree-ssa-live.h.
2496 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2497 * graphite.c (graphite_transform_loops): Make static.
2498 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2499 make_pass_graphite, pass_data_graphite_transforms,
2500 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2501 * ipa-pure-const.c (warn_function_noreturn): Make static.
2502 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2503 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2504 Relocate from tree-cfg.c
2505 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2507 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2508 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2509 Move to ipa-pure-const.c.
2510 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2511 Relocate from tree-optimize.c.
2512 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2513 make_pass_fixup_cfg): Move to tree-cfg.c.
2514 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2515 Relocate some prototypes.
2516 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2517 * tree-dump.c (dump_function_to_file): Remove prototype.
2518 Add tree-flow.h to the include file.
2519 * tree-dump.h: Remove prototype.
2520 * tree-parloops.h: New File. Add prototypes.
2521 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2522 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2523 from tree-ssa-loop.c.
2524 * tree-predcom.c (run_tree_predictive_commoning,
2525 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2526 Relocate here from tree-ssa-loop.c.
2527 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2528 ssa_name_values.release ().
2529 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2530 (ssa_name_values): Relocate from tree-flow.h.
2531 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2532 * tree-ssa-loop.c (run_tree_predictive_commoning,
2533 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2534 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2535 make_pass_graphite, pass_data_graphite_transforms,
2536 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2537 tree_parallelize_loops, pass_data_parallelize_loops,
2538 make_pass_parallelize_loops): Move to other files.
2539 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2541 * tree.h: Remove prototypes from tree-address.c.
2543 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2545 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2546 (struct int_tree_map): Move to tree-hasher.h
2547 (SCALE, LABEL, PERCENT): Move to gimple.h
2548 * tree-flow-inline.h: Delete. Move functions to other files.
2549 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2550 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2551 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2552 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2553 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2554 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2555 * tree-hasher.h: Don't include tree-flow.h.
2556 (struct int_tree_map): Relocate from tree-flow.h.
2557 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2558 tree-flow-inline.h and make static.
2559 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2561 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2562 tree-flow-inline.h and make static.
2563 * tree.h (is_global_var, may_be_aliased): Relocate from
2565 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2566 * value-prof.c: No longer include tree-flow-inline.h.
2567 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2569 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2571 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2572 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2573 (enum move_pos): Move to tree-ssa-loop-im.h
2574 * cfgloop.h: Move some prototypes.
2575 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2576 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2577 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2578 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2579 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2580 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2581 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2582 (enum move_pos): Relocate here.
2583 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2585 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2586 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2587 make_pass_lim): Relocate here from tree-ssa-loop.c.
2588 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2590 (loop_edge_to_cancel, unloop_loops): Make static.
2591 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2592 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2593 (tree_complete_unroll, gate_tree_complete_unroll,
2594 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2595 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2596 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2597 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2598 (stmt_invariant_in_loop_p): Remove unused function.
2599 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2600 * tree-ssa-loop-manip.h: New file. Add prototypes.
2601 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2602 (gcov_type_to_double_int): Move to cfgloop.h.
2603 (double_int_cmp, bound_index,
2604 estimate_numbers_of_iterations_loop): Make static.
2605 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2606 (max_loop_iterations): Factor out get_max_loop_iterations.
2607 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2609 * tree-ssa-loop-niter.h: New file. Add prototypes.
2610 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2611 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2612 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2613 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2614 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2615 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2616 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2617 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2618 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2619 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2620 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2621 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2622 pass_data_complete_unroll, make_pass_complete_unroll,
2623 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2624 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2625 tree-ssa-loop-ivcanon.c.
2626 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2627 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2628 tree-ssa-loop-prefetch.c.
2629 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2631 (get_lsm_tmp_name): Relocate and add suffix parameter.
2632 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2633 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2634 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2635 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2636 (get_estimated_loop_iterations): Factor out accessor from
2637 estimated_loop_iterations in tree-ssa-loop-niter.c.
2638 (get_max_loop_iterations): Factor out accessor from
2639 _max_loop_iterations in tree-ssa-niter.c.
2640 * loop-unroll.c (decide_unroll_constant_iterations,
2641 decide_unroll_runtime_iterations, decide_peel_simple,
2642 decide_unroll_stupid): Use new get_* accessors.
2644 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2646 PR tree-optimization/20318
2647 * doc/extend.texi (returns_nonnull): New function attribute.
2648 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2650 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2651 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2653 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2656 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2657 false if both components are bitfields.
2659 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2661 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2662 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2663 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2664 * config/aarch64/aarch64.h
2665 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2666 * config/aarch64/aarch64-simd-builtins.def
2667 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2669 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2671 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2673 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2675 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2677 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2679 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2680 (vneg_f64): New intrinsic.
2681 (vneg_s8): Asm replaced with C.
2682 (vneg_s16): Likewise.
2683 (vneg_s32): Likewise.
2684 (vneg_s64): New intrinsic.
2685 (vnegq_f32): Asm replaced with C.
2686 (vnegq_f64): Likewise.
2687 (vnegq_s8): Likewise.
2688 (vnegq_s16): Likewise.
2689 (vnegq_s32): Likewise.
2690 (vnegq_s64): Likewise.
2692 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2694 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2696 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2698 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2699 packed stack special handling.
2700 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2701 back to fixed stack slots for FPRs saved due to stdarg.
2703 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2705 * config/s390/s390.c (s390_frame_info): Restructure function.
2707 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2709 * config/s390/s390.c (struct s390_frame_layout): New field
2711 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2712 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2713 regs_ever_clobbered to char*.
2714 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2715 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2716 (s390_register_info): Enable FPR save slots. Move/Copy some
2717 functionality into ...
2718 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2719 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2721 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2722 not imply a stack frame.
2723 (s390_init_frame_layout): Remove variable clobbered_regs.
2724 (s390_update_register_info): Remove function.
2725 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2726 cannot be used for register renaming.
2727 (s390_hard_regno_scratch_ok): New function.
2728 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2729 (s390_initial_elimination_offset): Change offset calculation of
2730 the return address pointer.
2731 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2732 (restore_gprs): Set frame related flag.
2733 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2734 (s390_emit_prologue): Call s390_register_info instead of
2735 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2736 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2737 (s390_optimize_prologue): Call s390_optimize_register_info.
2738 Try to remove also FPR slot save/restore INSNs. Remove frame
2739 related flags from restore INSNs.
2741 2013-10-08 DJ Delorie <dj@redhat.com>
2743 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2746 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2747 avoid conflict with the MI use of %c.
2748 * config/rl78/rl78-real.md: change %c to %C throughout.
2749 * config/rl78/rl78-virt.md: Likewise.
2751 2013-10-08 Jan Hubicka <jh@suse.cz>
2753 * config/i386/i386.c (ix86_option_override_internal): Switch
2754 to SSE math for -ffast-math when target ISA supports SSE2.
2756 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2758 * tree-flow.h: Remove some prototypes.
2759 * tree.h: Remove some protypes, add a couple.
2760 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2761 using_eh_for_cleanups_p): Add interface routines for front ends.
2762 * tree-eh.h: New file. Add protoptyes.
2763 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2764 (add_stmt_to_eh_lp_fn): Make static.
2765 (lower_try_finally): Use new using_eh_for_cleanups_p.
2766 * emit-rtl.c: Include tree-eh.h.
2767 * gimple.h: Include tree-eh.h.
2769 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2771 PR tree-optimization/58480
2772 * tree-vrp.c (infer_nonnull_range): New function.
2773 (infer_value_range): Call infer_nonnull_range.
2775 2013-10-08 Dehao Chen <dehao@google.com>
2777 PR tree-optimization/58619
2778 * tree-inline.c (copy_phis_for_bb): Combine location data
2781 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2784 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2785 RTX_FRAME_RELATED_P on INSN.
2787 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2789 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2790 (altivec_expand_vec_perm_const): Call it.
2792 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2794 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2795 sequences for LE VSX loads and stores at expand time.
2796 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2798 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2799 (rs6000_gen_le_vsx_permute): New.
2800 (rs6000_gen_le_vsx_load): New.
2801 (rs6000_gen_le_vsx_store): New.
2802 (rs6000_gen_le_vsx_move): New.
2803 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2804 (*vsx_le_perm_load_v4si): New.
2805 (*vsx_le_perm_load_v8hi): New.
2806 (*vsx_le_perm_load_v16qi): New.
2807 (*vsx_le_perm_store_v2di): New.
2808 (*vsx_le_perm_store_v4si): New.
2809 (*vsx_le_perm_store_v8hi): New.
2810 (*vsx_le_perm_store_v16qi): New.
2811 (*vsx_xxpermdi2_le_<mode>): New.
2812 (*vsx_xxpermdi4_le_<mode>): New.
2813 (*vsx_xxpermdi8_le_V8HI): New.
2814 (*vsx_xxpermdi16_le_V16QI): New.
2815 (*vsx_lxvd2x2_le_<mode>): New.
2816 (*vsx_lxvd2x4_le_<mode>): New.
2817 (*vsx_lxvd2x8_le_V8HI): New.
2818 (*vsx_lxvd2x16_le_V16QI): New.
2819 (*vsx_stxvd2x2_le_<mode>): New.
2820 (*vsx_stxvd2x4_le_<mode>): New.
2821 (*vsx_stxvd2x8_le_V8HI): New.
2822 (*vsx_stxvd2x16_le_V16QI): New.
2824 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2826 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2828 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2830 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2831 loop to work also for 31bit ABI.
2832 Save the stack pointer for frame_size > 0.
2834 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2836 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2837 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2838 constraint letters from expanders.
2839 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2840 retry count to general_operand.
2841 ("tabort"): Give operand 0 a mode.
2842 ("tabort_1"): Add mode and constraint letter for operand 0.
2843 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2845 2013-10-04 Jeff Law <law@redhat.com>
2847 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2849 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2851 (thread_across_edge): Here. Call it.
2853 2013-10-04 Cary Coutant <ccoutant@google.com>
2855 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2856 discarding a location list expression (or a piece of one).
2858 2013-10-03 Jan Hubicka <jh@suse.cz>
2860 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2861 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2862 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2864 2013-10-03 Jan Hubicka <jh@suse.cz>
2866 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2867 accumulate-outgoing-args when producing unwind info.
2869 2013-10-03 Wei Mi <wmi@google.com>
2871 * lra-constraints.c (insert_move_for_subreg): New function
2872 extracted from simplify_operand_subreg.
2873 (simplify_operand_subreg): Add reload for paradoxical subreg.
2875 2013-10-03 Rong Xu <xur@google.com>
2877 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2878 the candidate of builtin_expect such that we should fix the
2879 size/time estimation.
2880 (estimate_function_body_sizes): Do the acutally size/time fix-up
2883 2013-10-03 Rong Xu <xur@google.com>
2885 * predict.c (tree_predict_by_opcode): Get the probability
2886 for builtin_expect from param builtin_expect_probability.
2887 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2888 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2889 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2891 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2894 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2895 * calls.c (alloca_call_p): Use get_callee_fndecl.
2896 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2897 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2899 (vrp_visit_stmt): Remove duplicated code.
2901 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2903 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2904 ceildf2, btruncdf2, instead of vsx_* name.
2906 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2907 iterators to only do V2DF and V4SF here. Move the DF code to
2908 rs6000.md where it is combined with SF mode. Replace <VSv> with
2909 just 'v' since only vector operations are handled with these insns
2910 after moving the DF support to rs6000.md.
2911 (vsx_sub<mode>3): Likewise.
2912 (vsx_mul<mode>3): Likewise.
2913 (vsx_div<mode>3): Likewise.
2914 (vsx_fre<mode>2): Likewise.
2915 (vsx_neg<mode>2): Likewise.
2916 (vsx_abs<mode>2): Likewise.
2917 (vsx_nabs<mode>2): Likewise.
2918 (vsx_smax<mode>3): Likewise.
2919 (vsx_smin<mode>3): Likewise.
2920 (vsx_sqrt<mode>2): Likewise.
2921 (vsx_rsqrte<mode>2): Likewise.
2922 (vsx_fms<mode>4): Likewise.
2923 (vsx_nfma<mode>4): Likewise.
2924 (vsx_copysign<mode>3): Likewise.
2925 (vsx_btrunc<mode>2): Likewise.
2926 (vsx_floor<mode>2): Likewise.
2927 (vsx_ceil<mode>2): Likewise.
2928 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2929 (vsx_sminsf3): Likewise.
2930 (vsx_fmadf4): Likewise.
2931 (vsx_fmsdf4): Likewise.
2932 (vsx_nfmadf4): Likewise.
2933 (vsx_nfmsdf4): Likewise.
2934 (vsx_cmpdf_internal1): Likewise.
2936 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2937 simpler to select whether a target has SPE or traditional floating
2938 point support in iterators.
2939 (TARGET_DF_SPE): Likewise.
2940 (TARGET_SF_FPR): Likewise.
2941 (TARGET_DF_FPR): Likewise.
2942 (TARGET_SF_INSN): Macros to say whether floating point support
2943 exists for a given operation for expanders.
2944 (TARGET_DF_INSN): Likewise.
2946 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2947 combining of SF/DF mode operations, using both traditional and VSX
2955 (abs<mode>2): Combine SF/DF modes using traditional floating point
2956 instructions. Add support for using the upper DF registers with
2957 VSX support, and SF registers with power8-vector support. Update
2958 expanders for operations supported by both the SPE and traditional
2959 floating point units.
2960 (abs<mode>2_fpr): Likewise.
2961 (nabs<mode>2): Likewise.
2962 (nabs<mode>2_fpr): Likewise.
2963 (neg<mode>2): Likewise.
2964 (neg<mode>2_fpr): Likewise.
2965 (add<mode>3): Likewise.
2966 (add<mode>3_fpr): Likewise.
2967 (sub<mode>3): Likewise.
2968 (sub<mode>3_fpr): Likewise.
2969 (mul<mode>3): Likewise.
2970 (mul<mode>3_fpr): Likewise.
2971 (div<mode>3): Likewise.
2972 (div<mode>3_fpr): Likewise.
2973 (sqrt<mode>3): Likewise.
2974 (sqrt<mode>3_fpr): Likewise.
2975 (fre<Fs>): Likewise.
2976 (rsqrt<mode>2): Likewise.
2977 (cmp<mode>_fpr): Likewise.
2978 (smax<mode>3): Likewise.
2979 (smin<mode>3): Likewise.
2980 (smax<mode>3_vsx): Likewise.
2981 (smin<mode>3_vsx): Likewise.
2982 (negsf2): Delete SF operations that are merged with DF.
2989 (fmasf4_fpr): Likewise.
2990 (fmssf4_fpr): Likewise.
2991 (nfmasf4_fpr): Likewise.
2992 (nfmssf4_fpr): Likewise.
2993 (sqrtsf2): Likewise.
2994 (rsqrtsf_internal1): Likewise.
2995 (smaxsf3): Likewise.
2996 (sminsf3): Likewise.
2997 (cmpsf_internal1): Likewise.
2998 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2999 (negdf2): Delete DF operations that are merged with SF.
3001 (nabsdf2): Likewise.
3007 (rsqrtdf_internal1): Likewise.
3008 (fmadf4_fpr): Likewise.
3009 (fmsdf4_fpr): Likewise.
3010 (nfmadf4_fpr): Likewise.
3011 (nfmsdf4_fpr): Likewise.
3012 (sqrtdf2): Likewise.
3013 (smaxdf3): Likewise.
3014 (smindf3): Likewise.
3015 (cmpdf_internal1): Likewise.
3016 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3017 (btrunc<mode>2): Delete separate expander, and combine with the
3018 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3019 (btrunc<mode>2_fpr): Likewise.
3020 (ceil<mode>2): Likewise.
3021 (ceil<mode>2_fpr): Likewise.
3022 (floor<mode>2): Likewise.
3023 (floor<mode>2_fpr): Likewise.
3024 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3025 Add support for using the upper registers with VSX and
3026 power8-vector. Move insns to be closer to the define_expands. On
3027 VSX systems, prefer the traditional form of FMA over the VSX
3028 version, since the traditional form allows the target not to
3029 overlap with the inputs.
3030 (fms<mode>4_fpr): Likewise.
3031 (nfma<mode>4_fpr): Likewise.
3032 (nfms<mode>4_fpr): Likewise.
3034 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3035 Richard Earnshaw <richard.earnshaw@arm.com>
3037 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3038 (struct mult_cost_table): Likewise.
3039 (struct mem_cost_table): Likewise.
3040 (struct fp_cost_table): Likewise.
3041 (struct vector_cost_table): Likewise.
3042 (cpu_cost_table): Likewise.
3043 * config/arm/arm.opt (mold-rts-costs): New option.
3044 (mnew-generic-costs): Likewise.
3045 * config/arm/arm.c (generic_extra_costs): New table.
3046 (cortexa15_extra_costs): Likewise.
3047 (arm_slowmul_tune): Use NULL as new costs.
3048 (arm_fastmul_tune): Likewise.
3049 (arm_strongarm_tune): Likewise.
3050 (arm_xscale_tune): Likewise.
3051 (arm_9e_tune): Likewise.
3052 (arm_v6t2_tune): Likewise.
3053 (arm_cortex_a5_tune): Likewise.
3054 (arm_cortex_a9_tune): Likewise.
3055 (arm_v6m_tune): Likewise.
3056 (arm_fa726te_tune): Likewise.
3057 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3058 (arm_cortex_tune): Use generict_extra_costs.
3059 (shifter_op_p): New function.
3060 (arm_unspec_cost): Likewise.
3061 (LIBCALL_COST): Define.
3062 (arm_new_rtx_costs): New function.
3063 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3064 table is available. Use old costs otherwise unless mnew-generic-costs
3066 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3067 (cpu_cost_table): Declare.
3069 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3072 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3073 (*subs_mul_imm_<mode>)
3074 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3075 (*sub_<shift>_<mode>)
3076 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3077 Remove k constraint.
3079 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3081 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3083 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3085 2013-10-02 Teresa Johnson <tejohnson@google.com>
3087 * predict.c (probably_never_executed): New function.
3088 (probably_never_executed_bb_p): Invoke probably_never_executed.
3089 (probably_never_executed_edge_p): Ditto.
3090 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3091 Treat profile insanities conservatively.
3093 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3095 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3097 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3099 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3100 use it. Use smaller increase for scratch. Don't increase reject
3101 for early clobber scratch.
3102 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3103 setting eliminated regs except setting fp from hfp.
3104 (lra_eliminate): Check lra_insn_recog_data on NULL.
3106 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3109 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3110 setting -mvsx-timode by default until the underlying problem is fixed.
3111 (RS6000_CPU, power7 defaults): Likewise.
3113 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3115 * config/x-linux (host-linux.o): Remove header dependencies.
3116 Use $(COMPILE) and $(POSTCOMPILE).
3117 * config/t-linux-android (linux-android.o): Ditto.
3119 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3121 * Makefile.in (expmed.o-warn): Remove.
3123 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3125 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3126 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3128 2013-10-02 Teresa Johnson <tejohnson@google.com>
3130 * dojump.c (do_jump_1): Divide probability between
3131 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3133 2013-10-02 Tom Tromey <tromey@redhat.com>
3135 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3137 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3139 * tree-flow.h: Remove some prototypes.
3140 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3141 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3142 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3143 mark_virtual_phi_result_for_renaming): Relocate here.
3144 * tree-into-ssa.h: Add prototypes.
3145 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3146 single_pred_before_succ_order.
3147 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3148 (nonfreeing_call_p) Move to gimple.c.
3149 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3151 * basic-block.h (single_pred_before_succ_order): Add prototype.
3152 * gimple.c (nonfreeing_call_p): Relocate here.
3153 * gimple.h: Add prototype.
3154 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3155 * tree-ssa-dom.h: New file. Relocate prototypes here.
3156 * tree-ssa.h: Include tree-ssa-dom.h.
3158 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3160 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3161 Use $(COMPILE) and $(POSTCOMPILE).
3163 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3165 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3167 * tree-flow.h: Remove some prototypes.
3168 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3169 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3170 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3171 propagate_tree_value*): Move from here to...
3172 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3173 propagate_tree_value*): Relocate here.
3174 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3175 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3176 * gimple-fold.c: Remove gimple-fold.h from include list.
3177 * tree-vrp.c: Remove gimple-fold.h from include list.
3178 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3179 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3180 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3181 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3182 * sese.c: Add tree-ssa-propagate.h to include list.
3184 2013-10-02 Richard Biener <rguenther@suse.de>
3186 * tree-loop-distribution.c: Include tree-vectorizer.h for
3188 (enum partition_kind): Remove PKIND_REDUCTION.
3189 (struct partition_s): Remove has_writes member, add reduction_p member.
3190 (partition_alloc): Adjust.
3191 (partition_builtin_p): Likewise.
3192 (partition_has_writes): Remove.
3193 (partition_reduction_p): New function.
3194 (partition_merge_into): Likewise.
3195 (generate_code_for_partition): Commonize builtin partition
3197 (rdg_cannot_recompute_vertex_p): Remove.
3198 (already_processed_vertex_p): Likewise.
3199 (rdg_flag_vertex): Do not set has_writes.
3200 (classify_partition): Adjust.
3201 (rdg_build_partitions): Do not set has_writes, treat all
3202 partitions as useful.
3203 (distribute_loop): Record number of library calls generated. Adjust.
3204 (tree_loop_distribution): Report number of loops and library
3205 calls generated as opt-info.
3207 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3209 * tree-flow.h: Include new .h files. Move prototypes.
3210 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3211 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3212 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3213 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3214 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3215 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3216 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3217 * tree.h (get_ref_base_and_extent): Move prototype out.
3218 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3220 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3221 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3222 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3223 * tree-scalar-evolution.c: Include tree.h.
3224 * sese.c: Include tree.h.
3225 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3226 * dwarf2out.c: Include tree-dfa.h.
3227 * tree-chrec.c: Include tree.h.
3228 * tree-data-ref.c: Include tree.h.
3230 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3232 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3235 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3237 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3238 Use $(COMPILE) and $(POSTCOMPILE).
3239 (sol2-cxx.o): Likewise.
3240 (sol2-stubs.o): Likewise.
3242 * config/x-solaris (host-solaris.o): Likewise.
3244 * config/sparc/t-sparc (sparc.o): Remove.
3245 (sparc-c.o): Remove header dependencies.
3246 Use $(COMPILE) and $(POSTCOMPILE).
3247 * config/sparc/x-sparc: Likewise.
3249 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3251 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3252 * config/arc/arc700.md: Likewise.
3253 * config/arc/arc-modes.def: Likewise.
3254 * config/arc/arc-simd.h: Likewise.
3255 * config/arc/t-arc-uClibc: Likewise.
3256 * config/arc/t-arc-newlib: Likewise.
3258 2013-10-02 Renlin Li <renlin.li@arm.com>
3260 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3262 (aarch64_expand_epilogue): Likewise.
3264 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3265 Yufeng Zhang <yufeng.zhang@arm.com>
3267 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3269 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3270 'base_in' represent a conversion and legal_cast_p_1 holds; set
3271 'base_in' with the returned value from get_unwidened.
3273 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3275 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3276 plus_constant is not used.
3278 2013-10-01 Wei Mi <wmi@google.com>
3280 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3281 * config/i386/i386.md: Add define_peephole2 to
3282 break partial reg stall for cvtss2sd/cvtsd2ss.
3284 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3286 * config/arc/arc.c (pass_arc_ifcvt::clone):
3287 Update for ctxt_ -> m_ctxt change.
3289 2013-10-01 Jeff Law <law@redhat.com>
3291 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3292 outgoing_edge and intermediate_edge fields. Instead store the path.
3293 (redirection_data::hash): Hash on the last edge's destination index.
3294 (redirection_data::equal): Check the entire thread path.
3295 (lookup_redirectio_data): Corresponding changes.
3296 (create_edge_and_update_destination_phis): Likewise.
3297 (thread_single_edge): Likewise.
3299 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3300 Diego Novillo <dnovillo@google.com>
3302 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3303 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3304 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3305 versions of these patterns.
3307 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3308 (__builtin_arc_aligned): Likewise.
3310 * config/arc/arc.md: Expand adc_0 comment stating the intended
3311 purpose and why it isn't ready.
3312 Replace commented out call_value_via_label_mixed with a
3313 plain comment about bl_s.
3315 * config/arc/arc.c (stdio.h): Don't include directly.
3316 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3317 Remove if (1) condition.
3318 (arc_encode_section_info): Fix comment.
3320 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3322 * config/arc/arc.c (arc_conditional_register_usage):
3323 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3324 Also set reg_alloc_order for DMA config regs.
3326 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3327 Jeremy Bennett <jeremy.bennett@embecosm.com>
3329 * doc/install.texi (--with-cpu): Mention ARC.
3330 (arc-*-elf32): New paragraph.
3331 (arc-linux-uclibc): Likewise.
3332 * doc/md.texi (Machine Constraints): Add ARC part.
3333 * doc/invoke.texi: (menu): Add ARC Options.
3334 (Machine Dependent Options) <ARC Options>: Add synopsis.
3335 (node ARC Options): Add.
3336 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3337 (ARC Built-in Functions): New section defining
3338 generic ARC built-in functions.
3339 (ARC SIMD Built-in Functions): New section defining SIMD specific
3341 (Declaring Attributes of Functions): Extended
3342 description of short_call and long_call attributes for ARC and
3343 added index entries.
3345 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3346 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3347 Joern Rennecke <joern.rennecke@embecosm.com>
3348 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3349 Brendan Kehoe <brendan@zen.org>
3350 Michael Eager <eager@eagercon.com>
3351 Simon Cook <simon.cook@embecosm.com>
3352 Jeremy Bennett <jeremy.bennett@embecosm.com>
3354 * config/arc, common/config/arc: New directories.
3356 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3357 Brendan Kehoe <brendan@zen.org>
3358 Simon Cook <simon.cook@embecosm.com>
3360 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3362 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3364 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3365 * tree-ssa-coalesce.h: New. Move prototype to here.
3366 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3367 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3368 (gimple_can_coalesce_p): Move to...
3369 * gimple.c (gimple_can_coalesce_p): Here.
3371 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3373 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3374 (dump_decl_set): Move to gimple.c.
3375 * gimple.h: Don't include tree-ssa-operands.h.
3376 (dump_decl_set): Add prototype.
3377 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3378 Move to gimple-ssa.h.
3379 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3380 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3381 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3382 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3383 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3385 (dump_decl_set): Relocate here.
3386 * gimple-ssa.h: New file.
3387 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3388 Relocate from gimple.h.
3389 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3390 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3391 swap_tree_operands and remove non-ssa path.
3392 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3393 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3395 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3396 vect_is_simple_reduction_1): Use swap_ssa_operands.
3397 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3398 (enum need_phi_state): Move to tree-into-ssa.c.
3399 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3400 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3401 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3402 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3403 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3404 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3405 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3406 num_imm_uses): Move to ssa-iterators.h.
3407 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3408 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3410 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3411 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3412 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3413 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3414 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3415 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3416 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3417 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3418 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3419 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3420 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3421 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3422 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3423 (set_phi_nodes): Move to tree-phinodes.h.
3424 * tree-ssa-operands.h (enum ssa_op_iter_type,
3425 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3426 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3427 (dump_decl_set): Remove prototype.
3428 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3429 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3430 (set_phi_nodes): Relocate from tree-flow-inline.h.
3431 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3433 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3434 include list. Temporarily add gimple.h to include list.
3435 * ssa-iterators.h: New file.
3436 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3437 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3438 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3439 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3440 Relocate from tree-ssa-operands.h.
3441 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3442 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3443 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3444 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3445 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3446 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3447 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3448 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3449 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3450 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3451 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3452 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3453 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3454 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3455 Relocate from tree-flow-inline.h.
3456 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3458 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3461 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3462 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3463 Insert '\t' to output template.
3464 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3465 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3466 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3467 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3469 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3471 * doc/install.texi (Host/target specific installation notes for GCC):
3472 Put @anchor before @heading.
3473 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3474 Use @email for email addresses.
3476 2013-10-01 Jeff Law <law@redhat.com>
3478 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3479 a vec. Only delete the path if we create one without successfully
3480 registering a jump thread.
3481 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3483 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3484 (paths): New vector of jump threading paths.
3485 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3486 (THREAD_PATH): New accessor macro for the entire thread path.
3487 (lookup_redirection_data): Get intermediate and final outgoing edge
3488 from the thread path.
3489 (create_edge_and_update_destination_phis): Copy the threading path.
3490 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3491 jump threading path.
3492 (ssa_redirect_edges): Get edges and block types from the jump threading
3493 path. Free the path vector.
3494 (thread_block): Get edges from the jump threading path. Look at the
3495 entire path to see if we thread to a loop exit. If we cancel a jump
3496 thread request, then free the path vector.
3497 (thread_single_edge): Get edges and block types from the jump threading
3498 path. Free the path vector.
3499 (thread_through_loop_header): Get edges and block types from the jump
3500 threading path. Free the path vector.
3501 (mark_threaded_blocks): Iterate over the vector of paths and store
3502 the path on the appropriate edge. Get edges and block types from the
3503 jump threading path.
3504 (mark_threaded_blocks): Get edges and block types from the jump
3505 threading path. Free the path vector.
3506 (thread_through_all_blocks): Use the vector of paths rather than
3507 a vector of 3-edge sets.
3508 (register_jump_thread): Accept pointer to a path vector rather
3509 than the path vector itself. Store the path vector for later use.
3512 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3513 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3516 * config/s390/s390.c (s390_split_branches): Modify check for table
3518 (s390_chunkify_start): Rearrange table jump insn check in order to
3519 deal with compare and branch insns correctly.
3521 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3525 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3526 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3527 define_insn_and_split.
3528 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3529 (shiftsi3_compare): New pattern.
3531 * config/arm/unspecs.md (UNSPEC_RRX): New.
3533 2013-10-01 Alan Modra <amodra@gmail.com>
3535 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3536 casing inout operands.
3538 2013-10-01 Richard Biener <rguenther@suse.de>
3540 PR tree-optimization/58553
3541 * tree-loop-distribution.c (struct partition_s): Add niter member.
3542 (classify_partition): Populate niter member for the partition
3543 and properly identify whether the relevant store happens before
3544 or after the loop exit.
3545 (generate_memset_builtin): Use niter member from the partition.
3546 (generate_memcpy_builtin): Likewise.
3548 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3550 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3551 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3553 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3555 * basic-block.h (edge_list): Prefix member names with "m_".
3556 * context.h (context): Likewise.
3557 * domwalk.h (dom_walker): Likewise.
3558 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3559 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3560 * hash-table.h (hash_table): Likewise.
3561 * machmode.h (bit_field_mode_iterator): Likewise.
3562 * pass_manager.h (pass_list): Likewise.
3563 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3564 * tree-pass.h (pass_data): Likewise.
3565 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3566 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3567 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3568 * asan.c (pass_data_asan): Update accordingly.
3569 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3570 (control_dependences::control_dependences): Likewise.
3571 (control_dependences::~control_dependences): Likewise.
3572 (control_dependences::~control_dependences): Likewise.
3573 (control_dependences::get_edges_dependent_on): Likewise.
3574 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3575 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3576 * context.c (gcc::context::context): Likewise.
3577 * cprop.c (pass_rtl_cprop::clone): Likewise.
3578 * domwalk.c (dom_walker::walk): Likewise.
3579 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3580 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3581 * mode-switching.c (pass_mode_switching::clone): Likewise.
3582 * passes.c (opt_pass::opt_pass): Likewise.
3583 (pass_manager::pass_manager): Likewise.
3584 * predict.c (pass_strip_predict_hints::clone): Likewise.
3585 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3586 (pass_split_all_insns::clone): Likewise.
3587 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3589 (bit_field_mode_iterator::next_mode): Likewise.
3590 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3591 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3592 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3593 * tree-complex.c (pass_lower_complex::clone): Likewise.
3594 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3595 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3596 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3597 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3598 (pass_fold_builtins::clone): Likewise.
3599 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3600 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3601 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3602 (pass_cd_dce::clone): Likewise.
3603 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3604 (pass_phi_only_cprop::clone): Likewise.
3605 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3606 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3607 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3608 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3609 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3610 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3611 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3612 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3613 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3614 * tree-vrp.c (pass_vrp::clone): Likewise.
3615 * tsan.c (pass_tsan::clone): Likewise.
3617 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3620 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3621 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3624 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3625 optimization, punt if sign_bit_p looked through any zero extension.
3627 2013-09-30 Teresa Johnson <tejohnson@google.com>
3629 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3630 Update redirected out edge count in joiner case.
3631 (ssa_redirect_edges): Common the joiner and non-joiner cases
3632 so that joiner case gets profile updates.
3634 2013-09-30 Richard Biener <rguenther@suse.de>
3636 PR tree-optimization/58554
3637 * tree-loop-distribution.c (classify_partition): Require
3638 unconditionally executed stores for memcpy and memset recognition.
3639 (tree_loop_distribution): Calculate dominance info.
3641 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3643 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3644 (NO_PROFILE_COUNTERS): Likewise.
3645 (PROFILE_HOOK): Likewise.
3646 (FUNCTION_PROFILER): Likewise.
3647 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3649 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3651 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3652 calls and defines in TARGET_MACHO conditional.
3653 (load_macho_picbase_di): Likewise.
3654 (reload_macho_picbase): Likewise.
3655 (reload_macho_picbase_si): Likewise.
3656 (reload_macho_picbase_di): Likewise.
3657 (nonlocal_goto_receiver): Likewise.
3659 2013-09-30 Nick Clifton <nickc@redhat.com>
3661 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3662 that use the MSP430X ISA.
3663 (msp430_option_override): Scan -mmcu command line option for any
3664 MCU name that supports the MSP430X ISA.
3665 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3666 -mmcu options which enable the MSP430X ISA.
3668 2013-09-30 Richard Biener <rguenther@suse.de>
3671 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3672 before looking for setjmp-like calls.
3674 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3677 * config/darwin-protos.h (machopic_get_function_picbase): New.
3678 * config/darwin.c (machopic_get_function_picbase): New.
3679 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3680 label for a new func. (load_macho_picbase_di): Likewise.
3681 (reload_macho_picbase): New expand.
3682 (reload_macho_picbase_si): New insn.
3683 (reload_macho_picbase_di): New insn.
3684 (nonlocal_goto_receiver): New define and split.
3685 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3686 (unspecv enum): Add UNSPECV_NLGR.
3688 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3690 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3691 that altivec registers are correctly sized on Darwin.
3693 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3695 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3696 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3697 * config/x-darwin (host-darwin.o): Likewise.
3698 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3699 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3700 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3702 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3704 * doc/invoke.texi: Fix usage of @tie{} command.
3706 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3708 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3710 2013-09-28 Jan Hubicka <jh@suse.cz>
3712 * config/i386/x86-tune.def: Add documentation for each of the options;
3715 2013-09-28 Jan Hubicka <jh@suse.cz>
3717 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3719 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3720 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3721 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3722 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3723 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3725 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3727 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3728 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3729 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3730 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3731 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3732 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3733 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3734 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3735 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3736 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3737 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3738 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3739 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3740 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3741 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3742 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3743 gimple.h, godump.c, graphite-clast-to-gimple.c,
3744 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3745 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3746 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3747 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3748 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3749 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3750 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3751 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3752 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3753 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3754 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3755 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3756 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3757 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3758 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3759 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3760 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3761 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3762 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3763 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3764 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3765 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3766 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3767 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3768 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3769 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3770 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3771 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3772 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3773 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3774 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3775 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3776 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3777 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3778 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3779 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3780 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3781 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3782 whitespace before "(".
3784 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3786 * expr.h (extract_bit_field): Remove packedp parameter.
3787 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3788 from forward declaration.
3789 (store_split_bit_field): Remove packedp arg from calls to
3790 extract_fixed_bit_field.
3791 (extract_bit_field_1): Remove packedp parameter and packedp
3792 argument from recursive calls and calls to extract_fixed_bit_field.
3793 (extract_bit_field): Remove packedp parameter and corresponding
3794 arg to extract_bit_field_1.
3795 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3797 (extract_split_bit_field): Remove packedp arg from call to
3798 extract_fixed_bit_field.
3799 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3800 (copy_blkmode_from_reg): Likewise.
3801 (copy_blkmode_to_reg): Likewise.
3802 (read_complex_part): Likewise.
3803 (store_field): Likewise.
3804 (expand_expr_real_1): Likewise.
3805 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3806 to extract_bit_field.
3807 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3808 call to extract_bit_field.
3809 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3810 call to extract_bit_field.
3811 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3812 and special packedp behavior from -fstrict-volatile-bitfields
3815 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3817 * lra-eliminations.c (init_elim_table): Guard value_p.
3819 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3821 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3822 DFmode, DImode, and SFmode in the upper VSX registers based on the
3823 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3824 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3825 of the VSX register handling.
3827 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3829 (zero_extendsidi2_lfiwzx): Add support for loading into the
3830 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3831 only do VSX memory options on Altivec registers.
3832 (extendsidi2_lfiwax): Likewise.
3833 (extendsfdf2_fpr): Likewise.
3834 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3835 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3836 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3837 (movdi_internal64): Likewise.
3839 2013-09-27 Xinliang David Li <davidxl@google.com>
3841 * opts.c (finish_options): Adjust parameters
3842 according to vect cost model.
3843 (common_handle_option): Set dynamic vect cost
3845 targhooks.c (default_add_stmt_cost): Compute stmt cost
3847 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3848 Use helper function.
3849 * tree-vectorizer.h (unlimited_cost_model): New function.
3850 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3851 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3853 (vect_enhance_data_refs_alignment): Ditto.
3854 * flag-types.h: New enum.
3855 * common/config/i386/i386-common.c (ix86_option_init_struct):
3856 No need to initialize vect_cost_model flag.
3857 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3860 2013-09-27 Diego Novillo <dnovillo@google.com>
3862 * gimple.h (enum ssa_mode): Remove.
3864 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3866 * cfgloop.h (number_of_loops): Fix typo in check for null.
3868 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3871 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3872 are children of outermost saved_cfun's loop, and set it up to
3873 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3874 if loop != loop0 and SESE region contains bbs that belong to loop0.
3876 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3878 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3879 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3880 (set_address_segment, set_address_base, set_address_index)
3881 (set_address_disp): Accept the argument unconditionally.
3882 (baseness): Remove must_be_base_p and must_be_index_p checks.
3883 (decompose_normal_address): Classify as much as possible in the
3886 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3888 * cse.c (count_reg_usage): Handle INT_LIST.
3889 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3890 * reginfo.c (reg_scan_mark_refs): Likewise.
3891 * reload1.c (eliminate_regs_1): Likewise.
3893 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3896 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3899 2013-09-27 Richard Biener <rguenther@suse.de>
3901 PR tree-optimization/58459
3902 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3903 restriction not propagating into loops.
3905 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3907 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3908 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3909 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3911 2013-09-26 Richard Biener <rguenther@suse.de>
3913 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3915 2013-09-26 Richard Biener <rguenther@suse.de>
3917 * alias.h (component_uses_parent_alias_set): Rename to ...
3918 (component_uses_parent_alias_set_from): ... this.
3919 * alias.c (component_uses_parent_alias_set): Rename to ...
3920 (component_uses_parent_alias_set_from): ... this and return
3922 (reference_alias_ptr_type_1): Use the result from
3923 component_uses_parent_alias_set_from instead of stripping
3924 components one at a time.
3925 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3927 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3929 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3930 Move prototypes to...
3931 * tree-ssa-ter.h: New File. Move prototypes here.
3932 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3933 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3934 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3935 from is_replaceable_p.
3936 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3937 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3938 (process_replaceable): Use ter_is_replaceable_p.
3939 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3940 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3941 newly refactored ssa_is_replaceable_p.
3942 * cfgexpand.c: Include tree-outof-ssa.h.
3943 * ssaexpand.h: Delete.
3945 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3947 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3948 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3950 (create_gimple_tmp): Delete.
3951 (get_expr_type, build_assign, build_type_cast): Move to...
3952 * gimple-builder.c: New File.
3953 (get_expr_type): Relocate from gimple.c.
3954 (build_assign, build_type_cast): Change to only create ssanames.
3955 * gimple.h: Move prototypes to...
3956 * gimple-builder.h: New File. Here.
3957 * tree-ssa.h: And here.
3958 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3959 count_uses_and_derefs): Relocate from gimple.c.
3960 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3961 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3962 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3963 gimple_replace_ssa_lhs.
3964 * asan.c: Include gimple-builder.h.
3965 * Makefile.in: Add gimple-builder.o.
3967 2013-09-26 Richard Biener <rguenther@suse.de>
3969 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3970 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3971 (loe_visit_block): Use gcc_checking_assert.
3972 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3973 gimple_assign_ssa_name_copy_p.
3974 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3977 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3979 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3980 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3982 2013-09-26 Richard Biener <rguenther@suse.de>
3984 PR tree-optimization/58539
3985 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3986 the fact that debug statements are not taking part in loop-closed
3989 2013-09-26 Nick Clifton <nickc@redhat.com>
3991 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3992 time warning message.
3993 (msp430_print_operand_raw): Delete unused letter parameter.
3994 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3995 (msp430_print_operand_address): New function.
3996 (msp430_print_operand): Move address printing code from here to
3998 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4000 (zero_extendpsisi2): Likewise.
4001 (extendpsisi2): New pattern.
4002 (andneghi3): New pattern.
4004 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4006 * config/aarch64/aarch64.opt (mlra): New option.
4007 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4008 (TARGET_LRA_P): Define.
4010 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4012 * expr.c (expand_assignment): Remove obsolete comment.
4014 2013-09-25 Jeff Law <law@redhat.com>
4016 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4017 tree-ssa-threadupdate.h.
4018 (register_jump_thread): Similarly.
4019 * tree-ssa-threadupdate.h: New header file.
4020 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4021 * tree-vrp.c: Likewise.
4022 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4023 (thread_around_empty_blocks): Change type of path vector argument to
4024 an edge,type pair from just an edge. Initialize both elements when
4025 appending to a jump threading path. Tweak references to elements
4027 (thread_across_edge): Similarly. Release memory for the elements
4029 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4030 (dump_jump_thread_path): New function broken out from
4031 register_jump_thread.
4032 (register_jump_thread): Use dump_jump_thread_path. Change type of
4033 path vector entries. Search the path for NULL edges and dump
4034 the path if one is found. Tweak the conversion of path to 3-edge
4035 form to use the block copy type information embedded in the path.
4037 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4039 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4041 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4042 Vladimir Makarov <vmakarov@redhat.com>
4044 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4045 from the least significant bit.
4046 (strip_address_mutations): Add bitfield operations handling.
4047 (must_be_index_p): Add shifting and rotate operations handling.
4048 (set_address_base): Use must_be_base_p predicate.
4049 (set_address_index): Use must_be_index_p predicate.
4051 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4052 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4053 Sergey Lega <sergey.s.lega@intel.com>
4054 Anna Tikhonova <anna.tikhonova@intel.com>
4055 Ilya Tocar <ilya.tocar@intel.com>
4056 Andrey Turetskiy <andrey.turetskiy@intel.com>
4057 Ilya Verbin <ilya.verbin@intel.com>
4058 Kirill Yukhin <kirill.yukhin@intel.com>
4059 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4061 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4063 (ix86_expand_vector_move_misalign): Support new unaligned load and
4064 stores and use new names.
4065 (CODE_FOR_sse2_storedqu): Rename to ...
4066 (CODE_FOR_sse2_storedquv16qi): ... this.
4067 (CODE_FOR_sse2_loaddqu): Rename to ...
4068 (CODE_FOR_sse2_loaddquv16qi): ... this.
4069 (CODE_FOR_avx_loaddqu256): Rename to ...
4070 (CODE_FOR_avx_loaddquv32qi): ... this.
4071 (CODE_FOR_avx_storedqu256): Rename to ...
4072 (CODE_FOR_avx_storedquv32qi): ... this.
4073 * config/i386/i386.md (fpint_logic): New.
4074 * config/i386/sse.md (VMOVE): Extend for AVX512.
4078 (VI_UNALIGNED_LOADSTORE): Ditto.
4079 (sse2_avx_avx512f): Ditto.
4080 (sse2_avx2): Extend for AVX512.
4081 (sse4_1_avx2): Ditto.
4082 (avx2_avx512f): New.
4083 (sse): Extend for AVX512.
4086 (avxsizesuffix): Ditto.
4087 (sseintvecmode): Ditto.
4089 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4090 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4091 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4092 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4093 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4094 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4095 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4096 (STORENT_MODE): Extend for AVX512.
4097 (*absneg<mode>2): Replace constraint "x" with "v".
4098 (*mul<mode>3): Ditto.
4099 (*ieee_smin<mode>3): Ditto.
4100 (*ieee_smax<mode>3): Ditto.
4101 (avx_cmp<mode>3): Replace VF with VF_128_256.
4102 (*<sse>_maskcmp<mode>3_comm): Ditto.
4103 (<sse>_maskcmp<mode>3): Ditto.
4104 (<sse>_andnot<mode>3): Extend for AVX512.
4105 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4106 (<code><mode>3, fpint_logic): New.
4107 (*<code><mode>3): Extend for AVX512.
4108 (avx512flogicsuff): New.
4109 (avx512f_<logic><mode>): Ditto.
4110 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4112 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4113 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4114 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4115 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4116 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4117 (xop_vpermil2<mode>3): Ditto.
4118 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4119 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4120 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4121 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4123 2013-09-25 Tom Tromey <tromey@redhat.com>
4125 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4126 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4127 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4128 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4129 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4130 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4131 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4132 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4133 (GRAPHITE_HTAB_H): Remove.
4135 2013-09-25 Tom Tromey <tromey@redhat.com>
4137 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4139 2013-09-25 Tom Tromey <tromey@redhat.com>
4141 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4143 2013-09-25 Tom Tromey <tromey@redhat.com>
4145 * config/i386/t-i386 (i386.o): Remove.
4146 (i386-c.o): Use COMPILE and POSTCOMPILE.
4148 2013-09-25 Tom Tromey <tromey@redhat.com>
4150 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4152 2013-09-25 Tom Tromey <tromey@redhat.com>
4154 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4155 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4156 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4157 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4158 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4159 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4160 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4161 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4162 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4163 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4164 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4165 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4166 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4167 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4168 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4169 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4170 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4171 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4172 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4173 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4174 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4175 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4176 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4177 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4178 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4179 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4180 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4181 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4182 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4183 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4184 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4185 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4186 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4187 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4188 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4189 (graphite-clast-to-gimple.o, graphite-dependences.o)
4190 (graphite-interchange.o, graphite-poly.o)
4191 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4192 (graphite-optimize-isl.o, tree-vect-loop.o)
4193 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4194 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4195 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4196 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4197 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4198 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4199 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4200 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4201 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4202 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4203 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4204 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4205 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4206 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4207 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4208 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4209 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4210 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4211 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4212 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4213 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4214 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4215 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4216 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4217 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4218 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4219 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4220 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4221 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4222 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4223 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4224 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4225 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4226 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4227 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4228 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4229 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4230 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4231 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4232 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4233 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4234 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4235 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4236 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4237 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4238 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4239 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4240 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4241 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4242 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4243 (gcov-dump.o): Remove.
4244 (default-c.o): Use COMPILE and POSTCOMPILE.
4245 (CFLAGS-gcc.o): New variable.
4246 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4248 2013-09-25 Tom Tromey <tromey@redhat.com>
4250 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4251 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4252 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4253 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4254 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4255 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4256 (c-family/array-notation-common.o, c-family/stub-objc.o)
4257 (c-family/c-ubsan.o): Remove.
4259 2013-09-25 Tom Tromey <tromey@redhat.com>
4261 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4263 2013-09-25 Tom Tromey <tromey@redhat.com>
4265 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4266 to add -DENABLE_SHARED_LIBGCC.
4267 (gcc.o): Don't use subshell.
4269 2013-09-25 Tom Tromey <tromey@redhat.com>
4271 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4272 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4273 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4274 * configure, config.in: Rebuild.
4276 2013-09-25 Tom Tromey <tromey@redhat.com>
4278 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4279 (COMPILE, POSTCOMPILE): New variables.
4280 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4281 (DEPFILES): New variable.
4282 Include ".Po" files.
4283 * configure.ac: Add checks for dependency checking.
4284 * configure, aclocal.m4: Regenerate.
4286 2013-09-25 Tom Tromey <tromey@redhat.com>
4288 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4289 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4291 2013-09-25 Tom Tromey <tromey@redhat.com>
4293 * Makefile.in (generated_files): Add options.h,
4294 target-hooks-def.h, insn-opinit.h,
4295 common/common-target-hooks-def.h, pass-instances.def,
4296 c-family/c-target-hooks-def.h.
4298 2013-09-25 Jeff Law <law@redhat.com>
4300 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4301 than foo[foo.length () - 1] to access last member in a vec.
4302 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4304 2013-09-25 Richard Biener <rguenther@suse.de>
4307 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4309 2013-09-25 Jan Hubicka <jh@suse.cz>
4311 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4314 2013-09-25 Marek Polacek <polacek@redhat.com>
4317 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4318 when determining the type name.
4320 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4322 * config/sh/sh.md: Fix formatting.
4324 2013-09-24 Xinliang David Li <davidxl@google.com>
4326 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4327 max peel iterations parameter.
4328 * param.def: New parameter.
4329 * doc/invoke.texi: Document New parameter.
4331 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4333 * gimple-pretty-print.c: Various whitespace tweaks.
4334 * tree-core.h: Likewise.
4335 * tree-pretty-print.c: Likewise.
4336 * tree-ssa-alias.c: Likewise.
4337 * tree-ssa-copy.c: Likewise.
4338 * tree-ssanames.c: Likewise.
4339 * tree-ssanames.h: Likewise.
4340 * tree-vrp.c: Likewise.
4342 2013-09-24 Alan Modra <amodra@gmail.com>
4346 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4347 for output operands that disallow regs. Don't use EXPAND_WRITE on
4350 2013-09-24 Richard Biener <rguenther@suse.de>
4353 * tree.c (reference_alias_ptr_type): Move ...
4354 * alias.c (reference_alias_ptr_type): ... here and implement
4355 in terms of the new reference_alias_ptr_type_1.
4356 (ref_all_alias_ptr_type_p): New helper.
4357 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4358 use ref_all_alias_ptr_type_p.
4359 (get_deref_alias_set): Add flag_strict_aliasing check here.
4360 (reference_alias_ptr_type_1): New function, split out from ...
4361 (get_alias_set): ... here.
4362 (alias_ptr_types_compatible_p): New function.
4363 * alias.h (reference_alias_ptr_type): Declare.
4364 (alias_ptr_types_compatible_p): Likewise.
4365 * tree.h (reference_alias_ptr_type): Remove.
4366 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4367 to compare MEM_REF alias types.
4369 2013-09-24 Richard Biener <rguenther@suse.de>
4371 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4373 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4375 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4376 reload helper function arrays into a single array reg_addr.
4377 (reload_fpr_gpr): Likewise.
4378 (reload_gpr_vsx): Likewise.
4379 (reload_vsx_gpr): Likewise.
4380 (struct rs6000_reg_addr): Likewise.
4381 (reg_addr): Likewise.
4382 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4383 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4384 (rs6000_init_hard_regno_mode_ok): Likewise.
4385 (rs6000_secondary_reload_direct_move): Likewise.
4386 (rs6000_secondary_reload): Likewise.
4388 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4389 constraints: wu, ww, and wy. Repurpose wv constraint added during
4390 power8 changes. Put wg constraint in alphabetical order.
4392 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4393 for future work to add ISA 2.07 VSX single precision support.
4394 (-mvsx-scalar-double): Change default from -1 to 1, update
4395 documentation comment.
4396 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4397 (-mupper-regs-df): New debug switch to control whether DF values
4398 can go in the traditional Altivec registers.
4399 (-mupper-regs-sf): New debug switch to control whether SF values
4400 can go in the traditional Altivec registers.
4402 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4404 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4405 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4406 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4407 (rs6000_debug_legitimate_address_p): Print if we are running
4408 before, during, or after reload.
4409 (rs6000_secondary_reload): Add a comment.
4410 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4412 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4413 constraints. Update documentation string.
4414 (wd constraint): Likewise.
4415 (wf constraint): Likewise.
4416 (wg constraint): Likewise.
4417 (wn constraint): Likewise.
4418 (ws constraint): Likewise.
4419 (wt constraint): Likewise.
4420 (wx constraint): Likewise.
4421 (wz constraint): Likewise.
4422 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4424 (ww constraint): Likewise.
4425 (wy constraint): Likewise.
4426 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4427 in the previous submissions.
4428 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4430 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4432 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4433 (REG_BR_PROB): Say that the probability is stored in an int_list.
4434 * reg-notes.def: Update commentary to mention INT_LIST.
4435 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4436 (INT_LIST): New rtx.
4437 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4438 * rtlanal.c (int_reg_note_p): New function.
4439 (alloc_reg_note): Assert that the note does not have an int argument.
4440 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4441 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4442 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4443 rather than an INSN_LIST. Handle INT_LIST.
4444 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4445 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4446 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4447 Manipulate them as ints rather than rtxes.
4448 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4449 * regmove.c (copy_src_to_dest): Likewise.
4450 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4452 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4453 into the cases that need it.
4454 * config/arm/arm.c (arm_unwind_emit): Likewise.
4456 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4457 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4458 * loop-doloop.c (add_test, doloop_modify): Likewise.
4459 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4460 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4461 * predict.c (combine_predictions_for_insn): Likewise.
4462 * print-rtl.c (print_rtx): Handle INT_LIST.
4463 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4464 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4465 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4466 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4467 (ix86_split_fp_branch, predict_jump): Likewise.
4468 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4469 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4470 * config/spu/spu.c (ea_load_store_inline): Likewise.
4472 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4473 value of a REG_BR_PROB note.
4474 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4475 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4476 * emit-rtl.c (try_split): Likewise.
4477 * predict.c (br_prob_note_reliable_p): Likewise.
4478 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4479 * reorg.c (mostly_true_jump): Likewise.
4480 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4481 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4482 * config/i386/i386.c (ix86_print_operand): Likewise.
4483 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4484 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4485 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4486 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4487 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4488 * config/sparc/sparc.c (output_cbranch): Likewise.
4489 * config/spu/spu.c (get_branch_target): Likewise.
4490 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4491 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4493 2013-09-23 Jan Hubicka <jh@suse.cz>
4495 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4497 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4498 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4500 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4501 (gimple_fold_call): Dump inconsistent devirtualizations; add
4502 sanity check for type based devirtualizations.
4503 * ipa-prop.c: Include ipa-utils.h
4504 (ipa_intraprocedural_devirtualization): Add sanity check.
4505 (try_make_edge_direct_virtual_call): Likewise.
4507 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4509 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4510 assignment statements.
4512 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4514 * gimple-pretty-print.c (dump_ssaname_info): New function.
4515 (dump_gimple_phi): Call it.
4516 (pp_gimple_stmt_1): Likewise.
4517 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4518 (range_info_def): Declare.
4519 * tree-pretty-print.c (pp_double_int): New function.
4520 (dump_generic_node): Call it.
4521 * tree-pretty-print.h (pp_double_int): Declare.
4522 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4523 * tree-ssanames.h (range_info_def): New structure.
4524 (value_range_type): Move definition here.
4525 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4527 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4529 (set_range_info): New function.
4530 (get_range_info): Likewise.
4531 (duplicate_ssa_name_range_info): Likewise.
4532 (duplicate_ssa_name_fn): Check pointer type and call
4533 duplicate_ssa_name_range_info.
4534 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4535 * tree-vrp.c (value_range_type): Remove definition, now in
4537 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4538 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4539 (SSA_NAME_RANGE_INFO): New macro.
4541 2013-09-23 Richard Biener <rguenther@suse.de>
4543 PR tree-optimization/58464
4544 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4545 (phi_trans_add): Change to add conditionally on being not present.
4546 (phi_translate_1): Remove recursion detection here.
4547 (phi_translate): Pre-seed the cache with NULL to catch
4548 recursion here in a more generic way.
4549 (bitmap_find_leader): Adjust comment.
4550 (get_representative_for): Dump value-numbers.
4551 (create_expression_by_pieces): Likewise.
4552 (insert_into_preds_of_block): Likewise.
4554 2013-09-23 Christian Bruel <christian.bruel@st.com>
4557 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4558 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4560 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4563 2013-09-20 Renlin Li <renlin.li@arm.com>
4565 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4567 (aarch64_expand_epilogue): Likewise.
4568 (aarch64_legitimize_reload_address): Likewise.
4570 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4572 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4573 NULL_TREE before pushing them onto the vector. Likewise for labels.
4575 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4577 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4578 comparison operators when -fno-trapping-math is in effect.
4579 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4580 comparison operators in TFmode and assert that unsupported operators
4582 (ia64_print_operand): Likewise.
4584 2013-09-21 Jan Hubicka <jh@suse.cz>
4586 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4587 (sse_typeless_stores): Enable for core
4588 (sse_load0_by_pxor): Likewise.
4589 (four_jump_limit): Disable for core.
4590 (pad_returns): Likewise.
4591 (avoid_vector_decode): Likewise.
4592 (fuse_cmp_and_branch): Enable for cores.
4593 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4595 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4598 * config/pa/pa.c (pa_option_override): Disable auto increment and
4599 decrement instructions until reload is completed.
4601 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4602 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4603 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4605 2013-09-20 DJ Delorie <dj@redhat.com>
4606 Nick Clifton <nickc@redhat.com>
4608 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4609 (need_to_save): Save bank 0 on interrupts.
4610 (characterize_address): Strip far address wrappers.
4611 (rl78_as_legitimate_address): Likewise.
4612 (transcode_memory_rtx): Likewise.
4613 (rl78_peep_movhi_p): Disable this peephole after devirt.
4614 (rl78_propogate_register_origins): Forget all origins when a
4616 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4617 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4619 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4620 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4621 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4623 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4625 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4627 * config/pa/pa.md: In "scc" insn patterns, change output template to
4628 handle const0_rtx in reg_or_0_operand operands.
4630 2013-09-20 Martin Husemann <martin@NetBSD.org>
4633 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4634 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4636 2013-09-20 Richard Biener <rguenther@suse.de>
4639 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4640 remembering SSA name version and block index.
4641 (new_scev_info_str): Adjust.
4642 (hash_scev_info): Likewise. Also hash the block index.
4643 (eq_scev_info): Adjust.
4644 (find_var_scev_info): Likewise.
4645 (struct instantiate_cache_entry): Remove.
4646 (struct instantiate_cache_type): Use a htab to map name, block
4648 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4649 (get_instantiated_value_entry): Likewise.
4650 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4651 (instantiate_scev_name): Adjust.
4653 2013-09-20 Jeff Law <law@redhat.com>
4655 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4657 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4659 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4660 Call aarch64_simd_expand_args to update op[argc].
4662 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4664 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4667 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4669 * gengtype.c (file_rules): Added rule for *.cc files.
4670 (get_output_file_with_visibility): Give fatal message when no
4673 2013-09-20 Renlin Li <renlin.li@arm.com>
4675 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4676 (aarch64_expand_epilogue): Likewise.
4677 (aarch64_legitimize_reload_address): Likewise.
4679 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4682 * expr.c (expand_assignment): Remove misalignp code path.
4684 2013-09-20 Marek Polacek <polacek@redhat.com>
4687 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4688 TYPE_PRECISION. Add asserts.
4690 2013-09-20 Richard Biener <rguenther@suse.de>
4692 PR tree-optimization/58453
4693 * tree-loop-distribution.c (distribute_loop): Apply the cost
4694 model for -ftree-loop-distribute-patterns, too.
4696 2013-09-20 Richard Biener <rguenther@suse.de>
4699 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4700 make type comparison less strict.
4702 2013-09-20 Alan Modra <amodra@gmail.com>
4704 * configure: Regenerate.
4705 * aclocal.m4: Regenerate.
4707 2013-09-20 Marek Polacek <polacek@redhat.com>
4710 * doc/extend.texi: Document that attribute used is meant to be used
4711 on variables with static storage duration.
4713 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4715 PR tree-optimization/58472
4716 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4717 simd_lane_access set inv_p = false.
4718 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4719 the simduid magic VAR_DECL.
4721 2013-09-19 Jan Hubicka <jh@suse.cz>
4723 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4725 2013-09-17 Jeff Law <law@redhat.com>
4727 * tree-ssa-dom.c (record_temporary_equivalences): New function
4728 split out of dom_opt_dom_walker::after_dom_children.
4729 (dom_opt_dom_walker::thread_across_edge): Move common code
4730 in here from dom_opt_dom_walker::after_dom_children.
4731 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4733 2013-09-19 Jan Hubicka <jh@suse.cz>
4735 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4736 (TARGET_GENERIC): Use PROCESOR_GENERIC
4737 (enum processor_type): Unify generic32 and 64.
4738 * i386.md (cpu): Likewise.
4739 * x86-tune.def (use_leave): Enable for generic32.
4740 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4741 * athlon.md: Change generic64 to generic in all occurences.
4742 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4743 (ix86_target_macros_internal): Likewise.
4744 * driver-i386.c (host_detect_local_cpu): Likewise.
4745 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4747 (generic_memcpy, generic_memset, generic_cost): This one.
4748 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4749 (m_GENERIC32, m_GENERIC64): Remove.
4750 (m_GENERIC): Turn into one flag.
4751 (processor_target): Unify generic tunnings.
4752 (ix86_option_override_internal): Replace generic32/64 by generic.
4753 (ix86_issue_rate): Likewise.
4754 (ix86_adjust_cost): Likewise.
4756 2013-09-19 Jan Hubicka <jh@suse.cz>
4758 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4759 of speculative flag.
4761 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4763 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4764 GOMP_sections_start, even if !exit_reachable.
4766 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4768 * lra-constraints.c (need_for_all_save_p): Use macro
4769 HARD_REGNO_CALL_PART_CLOBBERED.
4770 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4771 set up pseudo conflict hard regs.
4773 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4776 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4779 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4781 PR rtl-optimization/58438
4782 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4783 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4784 from previous subpasses.
4786 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4788 * arm.c (arm_get_frame_offsets): Validate architecture supports
4789 LDRD/STRD before accepting the tuning preference.
4790 (arm_expand_prologue): Likewise.
4791 (arm_expand_epilogue): Likewise.
4793 2013-09-18 Richard Biener <rguenther@suse.de>
4795 PR tree-optimization/58417
4796 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4797 have chrecs with symbols defined in the loop as operands.
4798 (chrec_fold_multiply): Likewise.
4799 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4800 parameters before folding binary operations.
4801 (struct instantiate_cache_entry_hasher): Remove.
4802 (struct instantiate_cache_type): Use a pointer-map.
4803 (instantiate_cache_type::instantiate_cache_type): New function.
4804 (instantiate_cache_type::get): Likewise.
4805 (instantiate_cache_type::set): Likewise.
4806 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4807 (get_instantiated_value_entry): Likewise.
4808 (global_cache): New global.
4809 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4810 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4811 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4812 (instantiate_scev_name): Adjust.
4813 (instantiate_scev): Construct global instead of local cache.
4814 (resolve_mixers): Likewise.
4816 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4817 Paolo Carlini <paolo.carlini@oracle.com>
4820 * doc/implement-cxx.texi: Fix references to the C++ standards.
4822 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4824 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4825 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4826 simd_lane_access drs, update also DR_ALIGNED_TO.
4828 2013-09-18 Marek Polacek <polacek@redhat.com>
4831 * doc/extend.texi: Document no_sanitize_undefined attribute.
4832 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4833 no_sanitize_undefined attribute.
4835 2013-09-18 Nick Clifton <nickc@redhat.com>
4837 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4838 (ASM_DECLARE_FUNCTION_NAME): Define.
4840 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4842 * compare-elim.c (find_comparison_dom_walker): New class
4843 (find_comparisons_in_bb): Rename to
4844 find_comparison_dom_walker::before_dom_children
4845 (find_comparisons): Adjust
4846 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4848 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4849 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4850 (init_walk_dominator_tree): Remove declaration.
4851 (fini_walk_dominator_tree): Remove declaration.
4852 * fwprop.c (single_def_use_dom_walker): New class
4853 (single_def_use_enter_block): Convert to
4854 single_def_use_dom_walker::before_dom_children.
4855 (single_def_use_leave_block): Convert to
4856 single_def_use_dom_walker::after_dom_children.
4857 (build_single_def_use_links): Adjust.
4858 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4860 (find_candidates_in_block): Convert to
4861 find_candidates_dom_walker::before_dom_children.
4862 (execute_strength_reduction): Adjust.
4863 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4864 (sese_dom_walker): New class.
4865 (sese_dom_walker::sese_dom_walker): New constructor.
4866 (sese_dom_walker::~sese_dom_walker): New destructor.
4867 (build_sese_conditions_before): Convert to
4868 sese_dom_walker::before_dom_children.
4869 (build_sese_conditions_after): Convert to
4870 sese_dom_walker::after_dom_children.
4871 (build_poly_scop): Adjust
4872 * tree-into-ssa.c (rewrite_dom_walker): New class
4873 (rewrite_enter_block): Convert to
4874 rewrite_dom_walker::before_dom_children.
4875 (rewrite_leave_block): Convert to
4876 rewrite_dom_walker::after_dom_children.
4877 (rewrite_update_dom_walker): New class.
4878 (rewrite_update_enter_block): Convert to
4879 rewrite_update_dom_walker::before_dom_children.
4880 (rewrite_update_leave_block): Convert to
4881 rewrite_update_dom_walker::after_dom_children.
4882 (rewrite_blocks, rewrite_into_ssa): Adjust.
4883 (mark_def_dom_walker): New class.
4884 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4885 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4886 (mark_def_sites_blocks): Convert to
4887 mark_def_dom_walker::before_dom_children.
4888 (mark_def_site_blocks): Remove.
4889 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4890 (tree_ssa_dominator_optimize): Adjust.
4891 (dom_thread_across_edge): Convert to method
4892 dom_opt_dom_walker::thread_across_edge.
4893 (dom_opt_enter_block): Convert to member function
4894 dom_opt_dom_walker::before_dom_children.
4895 (dom_opt_leave_block): Convert to member function
4896 dom_opt_dom_walker::after_dom_children.
4897 * tree-ssa-dse.c (dse_dom_walker): New class.
4898 (dse_enter_block): Convert to member function
4899 dse_dom_walker::before_dom_children.
4900 (tree_ssa_dse): Adjust.
4901 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4902 (determine_invariantness_stmt): Convert to method
4903 invariantness_dom_walker::before_dom_children.
4904 (determine_invariantness): Remove
4905 (move_computations_dom_walker): New class.
4906 (move_computations_stmt): Convert to method
4907 move_computations_dom_walker::before_dom_children.
4908 (move_computations, tree_ssa_lim): Adjust.
4909 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4910 (nt_init_block): Make method
4911 notrappping_dom_walker::before_dom_children.
4912 (nt_fini_block): Make
4913 method nontrapping_dom_walker::after_dom_children.
4914 (get_non_trapping): Adjust.
4915 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4916 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4917 (eliminate_leave_block): Make method.
4918 eliminate_dom_walker::after_dom_children.
4920 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4921 (strlen_enter_block): Make method
4922 strlen_dom_walker::before_dom_children.
4923 (strlen_leave_block): Make
4924 method strlen_dom_walker::after_dom_children.
4925 (tree_ssa_strlen): Adjust.
4926 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4927 (tree_ssa_uncprop): Adjust.
4928 (uncprop_leave_block): Make method
4929 uncprop_dom_walker::after_dom_children.
4930 (uncprop_leave_block): Make method
4931 uncprop_dom_walker::before_dom_children.
4933 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4935 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4936 branch in same basic block. Check both src and dest of the move insn.
4938 2013-09-17 Nick Clifton <nickc@redhat.com>
4940 * config/rl78/rl78-real.md (bf): New pattern.
4942 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4943 (rl78_print_operand): Do not put a # before a %B.
4944 * config/rl78/rl78.opt: Tweak doc strings.
4946 2013-09-17 DJ Delorie <dj@redhat.com>
4948 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4949 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4950 (rl78_option_override): Likewise, if -mallregs.
4951 (is_virtual_register): Likewise.
4952 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4953 (REGNO_OK_FOR_BASE_P): Likewise.
4954 * config/rl78/rl78.opt (-mallregs): New.
4956 2013-09-17 Nick Clifton <nickc@redhat.com>
4958 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4959 For interrupt functions: save all call clobbered registers if the
4960 interrupt handler is not a leaf function.
4961 (rl78_expand_prologue): Always recompute the frame information.
4962 For interrupt functions: only select bank 0 if one of the bank 0
4963 registers is going to be psuhed.
4965 2013-09-17 DJ Delorie <dj@redhat.com>
4967 * config/rl78/constraints.md: For each W* constraint, rename to C*
4968 and create a W* constraint that checks for an optional ES: prefix
4970 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4971 (es_addr): New. Used to wrap far addresses.
4972 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4973 (rl78_es_base): New.
4974 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4975 wrapped far addresses.
4976 (rl78_print_operand_1): Unwrap far addresses before processing.
4977 (rl78_lo16): Wrap far addresses in unspecs.
4978 (rl78_es_addr): New.
4979 (rl78_es_base): New.
4980 (insn_ok_now): Check for not-yet-wrapped far addresses.
4981 (transcode_memory_rtx): Properly re-wrap far addresses.
4983 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4985 * config/sparc/t-rtems: Add leon3 multilibs.
4987 2013-09-17 Cong Hou <congh@google.com>
4989 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4990 when checking the dot production pattern. The type of rhs operand
4991 of multiply is now checked correctly.
4993 2013-09-17 Jeff Law <law@redhat.com>
4995 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4996 edge implied equivalences into successor phis.
4997 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4998 here from tree-ssa-threadedge.c.
4999 (mark_threaded_blocks): When threading through a joiner, if both
5000 successors of the joiner's clone reach the same block, verify the
5001 PHI arguments are equal. If not, cancel the jump threading request.
5002 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5003 tree-ssa-threadupdate.c
5004 (thread_across_edge): Don't check PHI argument equality when
5005 threading through joiner block here.
5007 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5009 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5010 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5011 (warn_uninit, warn_uninitialized_vars,
5012 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5013 Move to tree-ssa-uninit.c.
5014 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5015 (has_undefined_value_p): New. Pass dependant parts of
5016 ssa_undefined_value_p.
5017 (uninit_undefined_value_p): Use has_undefined_value_p.
5018 (warn_uninit, warn_uninitialized_vars,
5019 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5020 Move from tree-ssa.c.
5021 * tree-ssa.h: Adjust prototypes.
5023 2013-09-17 Jan Hubicka <jh@suse.cz>
5026 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5027 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5028 FUNCTION_NOT_OPTIMIZED.
5029 * ipa-inline-analysis.c (compute_inline_parameters): Function
5030 not optimized is not inlinable unless it is alwaysinline.
5031 (inline_analyze_function): Force calls in not optimized
5032 function not inlinable.
5034 2013-09-17 Jan Hubicka <jh@suse.cz>
5037 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5039 * ipa.c (function_and_variable_visibility): Likewise.
5040 * ipa-profile.c (ipa_profile): Likewise.
5042 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5045 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5046 attribute, and don't inline the resolver in this case.
5048 2013-09-17 Teresa Johnson <tejohnson@google.com>
5050 * coverage.c (get_coverage_counts): Add missing newline.
5052 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5054 PR tree-optimization/58088
5055 * fold-const.c (mask_with_trailing_zeros): New function.
5056 (fold_binary_loc): Make sure we don't recurse infinitely
5057 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5058 Use mask_with_trailing_zeros where appropriate.
5060 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5062 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5063 of 'prev' var to get better distance estimation.
5065 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5067 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5068 * tree-inline.c (is_parameter_of): New predicate.
5069 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5070 a parameter has been remapped.
5071 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5072 (optimize_inline_calls): Initialize transform_parameter.
5073 (copy_gimple_seq_and_replace_locals): Likewise.
5074 (tree_function_versioning): Likewise.
5075 (maybe_inline_call_in_expr): Likewise.
5077 2013-09-17 Nick Clifton <nickc@redhat.com>
5079 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5080 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5082 (is_attr_func): New function.
5083 (msp430_is_interrupt_func): New function.
5084 (is_naked_func): New function.
5085 (is_reentrant_func): New function.
5086 (is_critical_func): New function.
5087 (msp430_start_function): Add annotations for function attributes.
5088 (msp430_attr): New function.
5089 (msp430_attribute_table): New.
5090 (msp430_function_section): New function.
5091 (TARGET_ASM_FUNCTION_SECTION): Define.
5092 (msp430_builtin): New enum.
5093 (msp430_init_builtins): New function.
5094 (msp430_builtin_devl): New function.
5095 (msp430_expand_builtin): New function.
5096 (TARGET_INIT_BUILTINS): Define.
5097 (TARGET_EXPAND_BUILTINS): Define.
5098 (TARGET_BUILTIN_DECL): Define.
5099 (msp430_expand_prologue): Add support for naked, interrupt,
5100 critical and reentrant functions.
5101 (msp430_expand_epilogue): Likewise.
5102 (msp430_print_operand): Handle 'O' character.
5103 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5105 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5106 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5107 (pushm): Use a 'n' rather than an 'i' constraint.
5108 (msp_return): Add generation of the interrupt return instruction.
5109 (disable_interrupts): New pattern.
5110 (enable_interrupts): New pattern.
5111 (push_intr_state): New pattern.
5112 (pop_intr_state): New pattern.
5113 (bic_SR): New pattern.
5114 (bis_SR): New pattern.
5115 * doc/extend.texi: Document MSP430 function attributes and builtin
5118 2013-09-17 Richard Biener <rguenther@suse.de>
5120 PR tree-optimization/58432
5121 * tree-loop-distribution.c (tree_loop_distribution): Also
5122 scan PHIs for outside loop uses and seed a partition from them.
5124 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5126 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5127 (restructure_reference): Call backtrace_base_for_ref.
5129 2013-09-17 Alan Modra <amodra@gmail.com>
5132 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5135 2013-09-16 DJ Delorie <dj@redhat.com>
5137 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5138 vregs location for RL78/G10.
5139 (rl78_expand_prologue): Avoid SEL on G10.
5140 (rl78_expand_epilogue): Likewise.
5141 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5142 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5143 __RL78_G10__ when appropriate.
5144 (ASM_SPEC): Pass -mg10 along to the assembler.
5145 * config/rl78/rl78.md (sel_rb): Disable for G10.
5146 * config/rl78/rl78.opt: Add -mg10 option.
5147 * config/rl78/t-rl78: Add -mg10 multilib.
5149 2013-09-16 Xinliang David Li <davidxl@google.com>
5151 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5152 * omp-low.c (omp_max_vf): Ditto.
5153 (expand_omp_simd): Ditto.
5154 * tree-vectorizer.c (vectorize_loops): Ditto.
5155 (gate_vect_slp): Ditto.
5156 (gate_increase_alignment): Ditto.
5157 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5158 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5159 (gate_tree_vectorize): Name change.
5160 (tree_vectorize): Ditto.
5161 (pass_vectorize::gate): Call new function.
5162 (pass_vectorize::execute): Ditto.
5163 * opts.c: O3 default setting change.
5164 (finish_options): Check new flag.
5165 * doc/invoke.texi: Document new flags.
5166 * common.opt: New flags.
5168 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5170 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5171 (Condition Code): Don't reference it.
5172 * doc/tm.texi: Regenerate.
5174 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5177 * lra-constraints.c (undo_optional_reloads): Consider all optional
5178 reload even if it did not get a hard reg.
5180 2013-09-16 Teresa Johnson <tejohnson@google.com>
5182 * dumpfile.c (dump_loc): Remove newline emission.
5183 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5184 emission to dump_printf_loc calls where missing.
5185 (vect_mark_for_runtime_alias_test): Ditto.
5186 (vect_analyze_data_ref_dependence): Ditto.
5187 (vect_analyze_data_ref_dependences): Ditto.
5188 (vect_slp_analyze_data_ref_dependence): Ditto.
5189 (vect_slp_analyze_data_ref_dependences): Ditto.
5190 (vect_compute_data_ref_alignment): Ditto.
5191 (vect_update_misalignment_for_peel): Ditto.
5192 (vect_verify_datarefs_alignment): Ditto.
5193 (vector_alignment_reachable_p): Ditto.
5194 (vect_get_data_access_cost): Ditto.
5195 (vect_enhance_data_refs_alignment): Ditto.
5196 (vect_find_same_alignment_drs): Ditto.
5197 (vect_analyze_data_refs_alignment): Ditto.
5198 (vect_analyze_group_access): Ditto.
5199 (vect_analyze_data_ref_access): Ditto.
5200 (vect_analyze_data_ref_accesses): Ditto.
5201 (vect_prune_runtime_alias_test_list): Ditto.
5202 (vect_analyze_data_refs): Ditto.
5203 (vect_create_addr_base_for_vector_ref): Ditto.
5204 (vect_create_data_ref_ptr): Ditto.
5205 (vect_grouped_store_supported): Ditto.
5206 (vect_grouped_load_supported): Ditto.
5207 * value-prof.c (check_counter): Ditto.
5208 (check_ic_target): Ditto.
5209 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5210 (vect_recog_widen_mult_pattern): Ditto.
5211 (vect_recog_widen_sum_pattern): Ditto.
5212 (vect_recog_over_widening_pattern): Ditto.
5213 (vect_recog_widen_shift_pattern): Ditto.
5214 (vect_recog_rotate_pattern): Ditto.
5215 (vect_recog_vector_vector_shift_pattern): Ditto.
5216 (vect_recog_divmod_pattern): Ditto.
5217 (vect_recog_mixed_size_cond_pattern): Ditto.
5218 (vect_recog_bool_pattern): Ditto.
5219 (vect_pattern_recog_1): Ditto.
5220 (vect_pattern_recog): Ditto.
5221 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5222 (vect_is_simple_iv_evolution): Ditto.
5223 (vect_analyze_scalar_cycles_1): Ditto.
5224 (vect_get_loop_niters): Ditto.
5225 (vect_analyze_loop_1): Ditto.
5226 (vect_analyze_loop_form): Ditto.
5227 (vect_analyze_loop_operations): Ditto.
5228 (vect_analyze_loop_2): Ditto.
5229 (vect_analyze_loop): Ditto.
5230 (report_vect_op): Ditto.
5231 (vect_is_slp_reduction): Ditto.
5232 (vect_is_simple_reduction_1): Ditto.
5233 (vect_get_known_peeling_cost): Ditto.
5234 (vect_estimate_min_profitable_iters): Ditto.
5235 (vect_model_reduction_cost): Ditto.
5236 (vect_model_induction_cost): Ditto.
5237 (get_initial_def_for_induction): Ditto.
5238 (vect_create_epilog_for_reduction): Ditto.
5239 (vectorizable_reduction): Ditto.
5240 (vectorizable_induction): Ditto.
5241 (vectorizable_live_operation): Ditto.
5242 (vect_loop_kill_debug_uses): Ditto.
5243 (vect_transform_loop): Ditto.
5244 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5245 (vect_stmt_relevant_p): Ditto.
5246 (process_use): Ditto.
5247 (vect_mark_stmts_to_be_vectorized): Ditto.
5248 (vect_model_simple_cost): Ditto.
5249 (vect_model_promotion_demotion_cost): Ditto.
5250 (vect_model_store_cost): Ditto.
5251 (vect_get_store_cost): Ditto.
5252 (vect_model_load_cost): Ditto.
5253 (vect_get_load_cost): Ditto.
5254 (vect_init_vector_1): Ditto.
5255 (vect_get_vec_def_for_operand): Ditto.
5256 (vect_finish_stmt_generation): Ditto.
5257 (vectorizable_call): Ditto.
5258 (vectorizable_conversion): Ditto.
5259 (vectorizable_assignment): Ditto.
5260 (vectorizable_shift): Ditto.
5261 (vectorizable_operation): Ditto.
5262 (vectorizable_store): Ditto.
5263 (vectorizable_load): Ditto.
5264 (vectorizable_condition): Ditto.
5265 (vect_analyze_stmt): Ditto.
5266 (vect_transform_stmt): Ditto.
5267 (vect_is_simple_use): Ditto.
5268 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5269 (vect_can_advance_ivs_p): Ditto.
5270 (vect_update_ivs_after_vectorizer): Ditto.
5271 (vect_do_peeling_for_loop_bound): Ditto.
5272 (vect_gen_niters_for_prolog_loop): Ditto.
5273 (vect_update_inits_of_drs): Ditto.
5274 (vect_create_cond_for_alias_checks): Ditto.
5275 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5276 (vect_build_slp_tree_1): Ditto.
5277 (vect_supported_load_permutation_p): Ditto.
5278 (vect_analyze_slp_instance): Ditto.
5279 (vect_analyze_slp): Ditto.
5280 (vect_make_slp_decision): Ditto.
5281 (vect_detect_hybrid_slp): Ditto.
5282 (vect_bb_vectorization_profitable_p): Ditto.
5283 (vect_slp_analyze_bb_1): Ditto.
5284 (vect_update_slp_costs_according_to_vf): Ditto.
5285 (vect_get_mask_element): Ditto.
5286 (vect_transform_slp_perm_load): Ditto.
5287 (vect_schedule_slp_instance): Ditto.
5288 (vect_schedule_slp): Ditto.
5289 (vect_slp_transform_bb): Ditto.
5290 * profile.c (read_profile_edge_counts): Ditto.
5291 (compute_branch_probabilities): Ditto.
5292 * coverage.c (get_coverage_counts): Ditto.
5294 2013-09-16 Diego Novillo <dnovillo@google.com>
5296 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5298 2013-09-16 Jan Hubicka <jh@suse.cz>
5300 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5301 abstract functions; for static functions check the presence of body.
5303 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5305 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5306 * config/aarch64/aarch64-simd.md
5307 (aarch64_mla_elt<mode>): New.
5308 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5309 (aarch64_mls_elt<mode>): Likewise.
5310 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5311 (aarch64_fma4_elt<mode>): Likewise.
5312 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5313 (aarch64_fma4_elt_to_128v2df): Likewise.
5314 (aarch64_fma4_elt_to_64df): Likewise.
5315 (fnma<mode>4): Likewise.
5316 (aarch64_fnma4_elt<mode>): Likewise.
5317 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5318 (aarch64_fnma4_elt_to_128v2df): Likewise.
5319 (aarch64_fnma4_elt_to_64df): Likewise.
5320 * config/aarch64/iterators.md (VDQSF): New.
5321 * config/aarch64/arm_neon.h
5322 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5323 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5325 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5327 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5328 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5329 (aarch64_mul3_elt_to_128df): Likewise.
5330 (aarch64_mul3_elt_to_64v2df): Likewise.
5331 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5333 (VMUL_CHANGE_NLANES) Likewise.
5336 * config/aarch64/arm_neon.h
5337 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5339 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5341 * config/aarch64/arm_neon.h
5342 (vcvtx_high_f32_f64): Fix parameters.
5344 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5345 Uros Bizjak <ubizjak@gmail.com>
5347 * config/alpha.c: Include tree-ssanames.h.
5349 2013-09-16 Richard Biener <rguenther@suse.de>
5351 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5352 (dot_rdg_1): Handle control_dd.
5353 (create_edge_for_control_dependence): New function.
5354 (create_rdg_edges): Add control dependences if asked for.
5355 (build_rdg): Likewise.
5356 (generate_loops_for_partition): If there are not necessary
5357 control stmts remove all their dependencies.
5358 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5359 (distribute_loop): Pass on control dependences.
5360 (tree_loop_distribution): Compute control dependences and remove
5361 restriction on number of loop nodes.
5363 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5365 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5368 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5370 * cse.c (try_const_anchors): Punt on CC modes.
5372 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5374 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5376 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5379 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5382 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5385 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5387 (darwin_objc1_section): Likewise.
5388 (darwin_file_end): Emit Image Info section when required.
5390 2013-09-14 Jan Hubicka <jh@suse.cz>
5392 * tree-into-ssa.c (gate_into_ssa): New.
5393 (pass_data_build_ssa): Use it.
5394 * cgraph.h (expand_thunk): Update prototype.
5395 * cgraphunit.c (analyze_function): Expand thunks early.
5396 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5397 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5398 set lowered flag; do not add new function.
5399 (assemble_thunks_and_aliases): Update.
5400 * tree-ssa.c (gate_init_datastructures): New gate.
5401 (pass_data_init_datastructures): Use it.
5403 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5406 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5407 xmm register set described in the psABI.
5409 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5411 * dwarf2out.c (should_emit_struct_debug): Add check
5412 for type_decl variable is not NULL.
5414 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5416 * config.gcc: Use new winnt-c.c target hooks
5417 * config/t-winnt: New file
5418 * config/winnt-c.c: New file
5419 * doc/tm.texi.in: Document new hook
5420 * doc/tm.texi: Regenerated
5422 2013-09-13 Jan Hubicka <jh@suse.cz>
5425 * ipa-inline.c (check_callers): New function.
5426 (check_caller_edge): Remove.
5427 (want_inline_function_to_all_callers_p): Also permit alises that are
5429 (inline_to_all_callers): Terminate the walk when devirtualization
5430 introduce new calls.
5432 2013-09-13 Jan Hubicka <jh@suse.cz>
5434 * ipa-inline-analysis.c (struct growth_data): Add node.
5435 (do_estimate_growth_1): Fix detection of recursion.
5437 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5439 PR tree-optimization/58392
5440 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5441 to avoid shadowing of outer loop variable. If
5442 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5443 replace_by_duplicate_decl simduid of loops that have it set and
5444 set dest_cfun->has_simduid_loops and/or
5445 dest_cfun->has_force_vect_loops.
5446 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5447 instead of maybe_lookup_decl.
5448 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5449 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5450 call. Copy over force_vect and copy and remap simduid. Set
5451 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5452 (copy_cfg_body): Remove blocks_to_copy argument. Use
5453 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5454 caller. Don't set cfun->has_simduid_loops and/or
5455 cfun->has_force_vect_loops here.
5456 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5458 (expand_call_inline, tree_function_versioning): Adjust copy_body
5461 2013-09-13 Martin Jambor <mjambor@suse.cz>
5464 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5465 the assert if the edge was a speculative one.
5467 2013-09-13 Richard Biener <rguenther@suse.de>
5469 * tree-data-ref.h (known_dependences_p): Move here ...
5470 * tree-loop-distribution.c (known_dependences_p): ... from here.
5471 (dump_rdg_component, debug_rdg_component): Remove.
5473 (generate_loops_for_partition): Use gimple_uid instead of
5474 relying on matching stmt visit order.
5475 (rdg_build_partitions): Take starting stmt vector.
5476 (ldist_gen): Merge into ...
5477 (distribute_loop): ... this function. Do not compute starting
5479 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5481 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5483 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5484 Set type attribute correctly. Set predicable_short_it attribute.
5485 (cmpsi_shiftsi): Remove %? from output template.
5487 2013-09-13 Richard Biener <rguenther@suse.de>
5489 * tree-loop-distribution.c (struct rdg_component,
5490 rdg_defs_used_in_other_loops_p, free_rdg_components,
5491 rdg_build_components): Remove.
5492 (stmts_from_loop): Do not record virtual PHIs.
5493 (generate_loops_for_partition): Skip virtual PHIs.
5494 (build_rdg_partition_for_component): Rename to ...
5495 (build_rdg_partition_for_vertex): ... this and adjust.
5496 (rdg_build_partitions): Take a vector of starting vertices
5497 instead of components. Remove unnecessary leftover handling.
5498 (ldist_gen): Do not build components or record other stores.
5499 (distribute_loop): Do not distribute loops containing stmts
5502 2013-09-13 Christian Bruel <christian.bruel@st.com>
5505 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5507 2013-09-13 Kai Tietz <ktietz@redhat.com>
5509 * config.gcc: Separate cases for mingw and cygwin targets,
5510 and add 64-bit cygwin target case.
5512 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5513 dll-export inline-functions.
5514 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5517 2013-09-13 Jeff Law <law@redhat.com>
5521 2013-09-06 Jeff Law <law@redhat.com>
5523 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5524 edge implied equivalences into successor phis.
5526 2013-09-12 DJ Delorie <dj@redhat.com>
5528 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5531 2013-09-12 Brooks Moses <bmoses@google.com>
5534 * Makefile.in: Do not install driver binaries in $(target)/bin.
5536 2013-09-12 DJ Delorie <dj@redhat.com>
5538 * config/rl78/rl78.opt (mrelax): New.
5539 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5540 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5542 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5544 (rl78_expand_epilogue): Likewise.
5546 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5549 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5550 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5551 pointer elimination with using remove_reg_equal_offset_note.
5553 2013-09-12 DJ Delorie <dj@redhat.com>
5555 * config/msp430/: New port.
5556 * config.gcc (msp430): Added.
5557 * doc/invoke.texi: Document MSP430 options.
5558 * doc/install.texi: Document msp430-elf
5559 * doc/md.texi: Document msp430-elf
5560 * doc/contrib.texi: Document msp430-elf
5562 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5565 2013-09-12 Martin Jambor <mjambor@suse.cz>
5568 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5569 reference descriptor is NULL.
5570 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5573 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5575 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5576 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5577 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5578 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5579 make_temp_ssa_name): move to tree-ssanames.h
5580 * tree-ssa-alias.h: Move prototype.
5581 * tree-ssa.h: Include tree-ssanames.h.
5582 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5583 * tree-ssanames.h: New. Move items from tree-flow*.h
5584 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5586 2013-09-12 Richard Biener <rguenther@suse.de>
5588 PR tree-optimization/58404
5589 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5590 propagate non-invariant addresses into dereferences wrapped
5591 in component references.
5593 2013-09-12 Richard Biener <rguenther@suse.de>
5595 PR tree-optimization/58402
5596 * passes.def: Move pass_late_warn_uninitialized later.
5598 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5600 * tree-ssa.h: New. Move content from tree-flow.h and
5602 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5603 Move prototypes belonging to tree-ssa.c.
5604 * tree-flow-inline.h (redirect_edge_var_map_def,
5605 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5607 * gimple.h: Adjust prototypes.
5608 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5610 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5611 * tree.h: Move prototype to tree-ssa.h.
5612 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5613 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5614 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5615 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5616 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5617 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5618 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5619 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5620 graphite-blocking.c, graphite-clast-to-gimple.c,
5621 graphite-dependences.c, graphite-interchange.c,
5622 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5623 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5624 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5625 ipa-reference.c, ipa-split.c, ipa-utils.c,
5626 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5627 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5628 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5629 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5630 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5631 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5632 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5633 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5634 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5635 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5636 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5637 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5638 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5639 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5640 value-prof.c, var-tracking.c,
5641 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5643 2013-09-12 Richard Biener <rguenther@suse.de>
5645 PR tree-optimization/58396
5646 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5647 (build_rdg): Take a loop-nest parameter, fix memleaks.
5648 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5650 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5652 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5655 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5657 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5658 The FMA instruction names should have a 'v' prefix.
5660 2013-09-12 Richard Biener <rguenther@suse.de>
5662 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5663 (dot_rdg): Use popen instead of system in optional code.
5664 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5665 (already_processed_vertex_p): Adjust.
5666 (has_anti_or_output_dependence, predecessor_has_mem_write,
5667 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5668 rdg_flag_uses): Remove.
5669 (rdg_flag_vertex): Simplify.
5670 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5672 (build_rdg_partition_for_component): Process the first vertex
5673 of a component only.
5674 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5676 2013-09-12 Alan Modra <amodra@gmail.com>
5678 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5680 2013-09-11 DJ Delorie <dj@redhat.com>
5681 Nick Clifton <nickc@redhat.com>
5683 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5684 (rl78_stack_based_mem): New.
5685 * config/rl78/constraints.md (Iv08): New.
5694 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5696 (movsi): Change from expand to insn-and-split.
5697 (ashrsi3): Clobber AX.
5701 * config/rl78/rl78.md (CC_REG): Fix.
5702 (addsi3): Allow memory and immediate operands.
5703 (addsi3_internal): Split into...
5704 (addsi3_internal_virt): ...new, and ...
5705 (addsi3_internal_real): ...new.
5707 (subsi3_internal_virt): New.
5708 (subsi3_internal_real): New.
5709 (mulsi3): Add memory operand.
5710 (mulsi3_rl78): Likewise.
5711 (mulsi3_g13): Likewise.
5712 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5713 (cbranchqi4_real): Add more constraint options.
5714 (cbranchhi4_real): Expand pattern.
5715 (cbranchhi4_real_signed): New.
5716 (cbranchhi4_real_inverted): New.
5717 (cbranchsi4_real_lt): New.
5718 (cbranchsi4_real_ge): New.
5719 (cbranchsi4_real_signed): New.
5720 (cbranchsi4_real): New.
5722 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5724 (lshrsi3_virt): Likewise.
5725 (ashlsi3_virt): Likewise.
5726 (cbranchqi4_virt_signed): New.
5727 (cbranchhi4_virt_signed): New.
5728 (cbranchsi4_virt): New.
5729 * config/rl78/rl78.c: Whitespace fixes throughout.
5730 (move_elim_pass): New.
5731 (pass_data_rl78_move_elim): New.
5732 (pass_rl78_move_elim): New.
5733 (make_pass_rl78_move_elim): New.
5734 (rl78_devirt_info): Run devirt earlier.
5735 (rl78_move_elim_info): New.
5736 (rl78_asm_file_start): Register it.
5737 (rl78_split_movsi): New.
5738 (rl78_as_legitimate_address): Allow virtual base registers when
5740 (rl78_addr_space_convert): Remove spurious debug stuff.
5741 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5742 (rl78_print_operand): More cases for not printing '#'.
5743 (rl78_expand_compare): Remove most of the logic.
5744 (content_memory): New.
5745 (clear_content_memory): New.
5746 (get_content_index): New.
5747 (get_content_name): New.
5748 (display_content_memory): New.
5749 (update_content): New.
5750 (record_content): New.
5751 (already_contains): New.
5752 (insn_ok_now): Re-recog insns with virtual registers.
5753 (add_postponed_content_update): New.
5754 (process_postponed_content_update): New.
5755 (gen_and_emit_move): New.
5756 (transcode_memory_rtx): Record new location content.
5757 Use gen_and_emit_move.
5758 (force_into_acc): New.
5759 (move_to_acc): Use gen_and_emit_move.
5760 (move_from_acc): Likewise.
5761 (move_acc_to_reg): Likewise.
5762 (move_to_x): Likewise.
5763 (move_to_hl): Likewise.
5764 (move_to_de): Likewise.
5765 (rl78_alloc_physical_registers_op1): Record location content.
5766 (has_constraint): New.
5767 (rl78_alloc_physical_registers_op2): Record location content.
5769 (rl78_alloc_physical_registers_ro1): Likewise.
5770 (rl78_alloc_physical_registers_cmp): Likewise.
5771 (rl78_alloc_physical_registers_umul): Likewise.
5772 (rl78_alloc_address_registers_macax): New.
5773 (rl78_alloc_physical_registers): Initialize and set location
5774 content memory as needed.
5775 (rl78_reorg): Make sure split2 is called.
5776 (rl78_rtx_costs): New.
5778 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5780 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5781 for (not (neg ...)) and (neg (not ...)) cases.
5783 2013-09-11 Richard Biener <rguenther@suse.de>
5786 * passes.def: Split critical edges before late uninit warning passes.
5787 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5789 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5791 PR tree-optimization/58385
5792 * fold-const.c (build_range_check): If both low and high are NULL,
5793 use omit_one_operand_loc to preserve exp side-effects.
5795 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5797 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5799 2013-09-11 Richard Biener <rguenther@suse.de>
5801 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5802 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5803 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5804 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5805 stmts_from_loop, known_dependences_p, build_empty_rdg,
5806 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5807 * tree-loop-distribution.c: ... here.
5808 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5809 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5810 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5811 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5812 * tree-loop-distribution.c: ... here.
5813 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5814 (struct partition_s): Add loops member.
5815 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5816 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5817 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5819 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5820 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5821 Sergey Lega <sergey.s.lega@intel.com>
5822 Anna Tikhonova <anna.tikhonova@intel.com>
5823 Ilya Tocar <ilya.tocar@intel.com>
5824 Andrey Turetskiy <andrey.turetskiy@intel.com>
5825 Ilya Verbin <ilya.verbin@intel.com>
5826 Kirill Yukhin <kirill.yukhin@intel.com>
5827 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5829 * config/i386/constraints.md (k): New.
5831 * config/i386/i386.c (const regclass_map): Add new mask registers.
5832 (dbx_register_map): Ditto.
5833 (dbx64_register_map): Ditto.
5834 (svr4_dbx_register_map): Ditto.
5835 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5837 (ix86_preferred_reload_class): Disable constants for mask registers.
5838 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5839 (ix86_hard_regno_mode_ok): Support new mask registers.
5840 (x86_order_regs_for_local_alloc): Ditto.
5841 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5842 (FIXED_REGISTERS): Add new mask registers.
5843 (CALL_USED_REGISTERS): Ditto.
5844 (REG_ALLOC_ORDER): Ditto.
5845 (VALID_MASK_REG_MODE): New.
5846 (FIRST_MASK_REG): Ditto.
5847 (LAST_MASK_REG): Ditto.
5848 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5849 (MAYBE_MASK_CLASS_P): New.
5850 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5851 (REG_CLASS_CONTENTS): Ditto.
5852 (MASK_REGNO_P): New.
5853 (ANY_MASK_REG_P): Ditto.
5854 (HI_REGISTER_NAMES): Add new mask registers.
5855 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5856 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5858 (attribute "type"): Add mskmov, msklog.
5859 (attribute "length_immediate"): Support them.
5860 (attribute "memory"): Ditto.
5861 (attribute "prefix_0f"): Ditto.
5862 (*movhi_internal): Support new mask registers.
5863 (*movqi_internal): Ditto.
5864 (define_split): Split out clobber pattern is a logic
5865 insn on mask registers.
5866 (*k<logic><mode>): New.
5867 (*andhi_1): Extend to support mask regs.
5868 (*andqi_1): Extend to support mask regs.
5870 (define_split): Split and-not to and and not if operands
5872 (*<code><mode>_1): Separate HI mode to new pattern...
5873 (*<code>hi_1): This.
5874 (*<code>qi_1): Extend to support mask regs.
5876 (kortestzhi): Ditto.
5877 (kortestchi): Ditto.
5879 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5880 (*one_cmplhi2_1): ...Here, now with mask registers support.
5881 (*one_cmplqi2_1): Support new mask registers.
5882 (HI/QImode arithmetics splitter): Don't split if mask registers
5884 (HI/QImode not splitter): Ditto.
5885 * config/i386/predicated.md (mask_reg_operand): New.
5886 (general_reg_operand): Ditto.
5888 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5890 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5891 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5893 2013-09-10 Jeff Law <law@redhat.com>
5895 PR tree-optimization/58380
5896 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5897 of threading through a buried loop header.
5899 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5900 return value for single successor case.
5902 2013-09-10 Jan Hubicka <jh@suse.cz>
5904 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5906 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5909 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5910 support conditional execution.
5911 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5913 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5915 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5917 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5918 Check destination too to check move insn.
5919 (undo_optional_reloads): Add check that the original peudo did not
5920 changed its allocation and the optional reload was inherited on last
5921 inheritance pass. Break loop after deciding to keep optional reload.
5922 (lra_undo_inheritance): Add check that inherited pseudo still in
5925 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5927 * config/aarch64/aarch64.md (generic_sched): New.
5928 * config/aarch64/aarch64-generic.md (load): Make conditional
5929 on generic_sched attribute.
5930 (nonload): Likewise.
5932 2013-09-10 Jan Hubicka <jh@suse.cz>
5934 * lto-cgraph.c: Include ipa-utils.h.
5935 (compute_ltrans_boundary): Also add possible targets into the boundary.
5937 2013-09-10 Jan Hubicka <jh@suse.cz>
5939 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5940 VAR_DECL of vtable rather than full expression.
5942 2013-09-10 Jan Hubicka <jh@suse.cz>
5943 Paolo Carlini <paolo.carlini@oracle.com>
5945 * cgraphunit.c (analyze_functions): Save input_location, set it
5946 to UNKNOWN_LOCATION and restore it at the end.
5948 2013-09-10 Martin Jambor <mjambor@suse.cz>
5950 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5951 represented by a thunk.
5953 2013-09-10 Jeff Law <law@redhat.com>
5955 PR tree-optimization/58343
5956 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5957 jump threading requests through loop headers buried in the middle
5958 of a jump threading path.
5960 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5961 in return value/type.
5963 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5965 PR rtl-optimization/58365
5966 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5967 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5970 2013-09-10 Richard Biener <rguenther@suse.de>
5972 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5973 * tree-data-ref.c (create_rdg_vertices): Collect all data
5974 references, signal failure to the caller, use data-ref API.
5975 (build_rdg): Compute data references only once. Maintain lifetime
5976 of data references and data dependences from within RDG.
5977 (free_rdg): Free dependence relations.
5978 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5979 inventing extra dependences.
5980 (distribute_loop): Update for RDG API changes.
5982 2013-09-10 Kai Tietz <ktietz@redhat.com>
5984 * doc/invoke.texi (fms-extensions): Document changed
5985 behavior for ms-abi targets.
5986 * config/i386/i386.c (ix86_option_override_internal):
5987 Set default value of option -fms-extension for ms-abi targets.
5989 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5991 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5993 2013-09-10 Alan Modra <amodra@gmail.com>
5996 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5998 2013-09-10 Alan Modra <amodra@gmail.com>
6000 * config/rs6000/predicates.md (add_cint_operand): New.
6001 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6002 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6003 using add_cint_operand.
6004 (largetoc_high_plus_aix): Likewise.
6006 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6008 PR tree-optimization/58364
6009 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6010 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6011 the current range can't be an unconditional true or false.
6013 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6015 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6017 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6019 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6021 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6024 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6026 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6028 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6029 NO_REGS for immediate that can't be moved directly into FP_REGS.
6031 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6033 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6034 comparison with negated operand.
6035 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6038 2013-09-09 Richard Biener <rguenther@suse.de>
6041 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6042 of a subloop record all its block as affecting loop-closed SSA form.
6044 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6046 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6047 of an rtx/bitpos pair.
6048 (store_fixed_bit_field): Update accordingly.
6050 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6052 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6054 * builtins.c (expand_errno_check): Likewise.
6055 * dwarf2cfi.c (init_return_column_size): Likewise.
6056 * except.c (sjlj_mark_call_sites): Likewise.
6057 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6058 * lra-constraints.c (emit_inc): Likewise.
6059 * ree.c (combine_set_extension): Likewise.
6060 * regmove.c (fixup_match_2): Likewise.
6061 * reload1.c (inc_for_reload): Likewise.
6063 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6065 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6066 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6067 Use gen_int_mode with the mode of the associated simplify_* call.
6068 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6069 * expmed.c (expand_shift_1): Likewise.
6070 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6071 * loop-iv.c (iv_number_of_iterations): Likewise.
6072 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6073 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6075 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6077 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6078 of the associated expand_* call.
6079 (asan_emit_stack_protection): Likewise.
6080 * builtins.c (round_trampoline_addr): Likewise.
6081 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6082 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6083 (emit_store_flag): Likewise.
6084 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6085 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6087 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6088 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6089 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6090 * modulo-sched.c (generate_prolog_epilog): Likewise.
6091 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6092 (expand_ctz, expand_ffs, expand_unop): Likewise.
6094 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6096 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6097 of the associated gen_rtx_* call.
6098 * caller-save.c (init_caller_save): Likewise.
6099 * combine.c (find_split_point, make_extraction): Likewise.
6100 (make_compound_operation): Likewise.
6101 * dwarf2out.c (mem_loc_descriptor): Likewise.
6102 * explow.c (plus_constant, probe_stack_range): Likewise.
6103 * expmed.c (expand_mult_const): Likewise.
6104 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6105 * reload1.c (init_reload): Likewise.
6106 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6107 * var-tracking.c (adjust_mems): Likewise.
6108 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6109 rather than gen_rtx_fmt_ee.
6111 2013-09-09 Jan Hubicka <jh@suse.cz>
6114 * value-prof.c (gimple_ic): Copy also abnormal edges.
6116 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6118 * asan.c (asan_shadow_cst): Use gen_int_mode.
6120 2013-09-08 Jan Hubicka <jh@suse.cz>
6122 * ipa-profile.c: Add toplevel comment.
6123 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6124 (contains_hot_call_p): New function.
6125 (ipa_propagate_frequency): Set frequencies based on counts when
6127 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6128 profile; do not tamper with profile after inlining if it is read.
6130 2013-09-08 Jan Hubicka <jh@suse.cz>
6132 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6135 2013-09-08 Jan Hubicka <jh@suse.cz>
6137 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6140 2013-09-08 Jeff Law <law@redhat.com>
6143 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6146 2013-09-08 Andi Kleen <ak@linux.intel.com>
6148 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6150 2013-09-08 Jan Hubicka <jh@suse.cz>
6152 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6154 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6156 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6157 for non-debug insns.
6158 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6159 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6160 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6163 2013-09-08 Jan Hubicka <jh@suse.cz>
6165 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6166 targets and devirtualize to BUILT_IN_UNREACHABLE.
6167 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6168 * ipa.c (walk_polymorphic_call_targets): New function.
6169 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6170 functions; use the new timevar.
6171 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6172 was removed from the program.
6173 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6174 not consider it in the walk when its vtable is dead.
6175 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6177 (devirt_variable_node_removal_hook): New function.
6178 (possible_polymorphic_call_targets): Also register
6179 devirt_variable_node_removal_hook.
6180 (ipa_devirt): Do not do non-speculative devirtualization.
6181 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6183 2013-09-08 Jan Hubicka <jh@suse.cz>
6185 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6186 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6187 varpool_remove_variable_insertion_hook): Declare.
6188 * varpool.c (varpool_node_hook_list): New structure.
6189 (first_varpool_node_removal_hook,
6190 first_varpool_variable_insertion_hook): New variables.
6191 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6192 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6193 varpool_remove_variable_insertion_hook,
6194 varpool_call_variable_insertion_hooks): New functions.
6195 (varpool_remove_node): Use it.
6197 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6200 * diagnostic.c (diagnostic_build_prefix): When s.file is
6201 "<built-in>" don't output line and column numbers.
6203 2013-09-06 Jan Hubicka <jh@suse.cz>
6205 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6206 * lto-streamer-out.c: Stream thunks, too.
6207 * lto-streamer-in.c (input_function): Pop cfun here
6208 (lto_read_body): Instead of here.
6210 2013-09-06 Caroline Tice <cmtice@google.com>
6212 * doc/install.texi: Add documentation for the --enable-vtable-verify
6213 and the --disable-libvtv configure options.
6215 2013-09-06 Jeff Law <law@redhat.com>
6217 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6218 edge implied equivalences into successor phis.
6220 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6222 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6224 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6226 * resource.c (mark_target_live_regs): Compute resources taking
6227 into account if a call is predicated or not.
6229 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6231 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6232 the compiler in the function names.
6234 2013-09-06 Jan Hubicka <jh@suse.cz>
6237 * ipa-inline.c (has_caller_p): New function.
6238 (want_inline_function_to_all_callers_p): Use it.
6239 (sum_callers, inline_to_all_callers): Break out from ...
6240 (ipa_inline): ... here.
6242 2013-09-06 Jan Hubicka <jh@suse.cz>
6244 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6245 only when AVX is enabled.
6247 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6249 * config/aarch64/aarch64.md
6250 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6251 is fpsimd_<load/store>2.
6252 (load_pair<mode>): Likewise.
6253 (store_pair<mode>): Likewise.
6255 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6257 * config/arm/types.md (type): Add "mrs" type.
6258 * config/aarch64/aarch64.md
6259 (aarch64_load_tp_hard): Make type "mrs".
6261 (load_tp_hard): Make type "mrs".
6262 * config/arm/cortex-a15.md: Update with new attributes.
6263 * config/arm/cortex-a5.md: Update with new attributes.
6264 * config/arm/cortex-a53.md: Update with new attributes.
6265 * config/arm/cortex-a7.md: Update with new attributes.
6266 * config/arm/cortex-a8.md: Update with new attributes.
6267 * config/arm/cortex-a9.md: Update with new attributes.
6268 * config/arm/cortex-m4.md: Update with new attributes.
6269 * config/arm/cortex-r4.md: Update with new attributes.
6270 * config/arm/fa526.md: Update with new attributes.
6271 * config/arm/fa606te.md: Update with new attributes.
6272 * config/arm/fa626te.md: Update with new attributes.
6273 * config/arm/fa726te.md: Update with new attributes.
6275 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6277 * config/aarch64/aarch64.md
6278 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6279 (*movtf_aarch64): Likewise.
6281 (thumb1_movdi_insn): Use "multiple" for type where more than one
6282 instruction is used for a move.
6283 (*arm32_movhf): Likewise.
6284 (*thumb_movdf_insn): Likewise.
6286 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6288 * config/arm/types.md (type): Rename fcpys to fmov.
6290 (*arm_movsi_vfp): Rename type fcpys as fmov.
6291 (*thumb2_movsi_vfp): Likewise
6292 (*movhf_vfp_neon): Likewise
6293 (*movhf_vfp): Likewise
6294 (*movsf_vfp): Likewise
6295 (*thumb2_movsf_vfp): Likewise
6296 (*movsfcc_vfp): Likewise
6297 (*thumb2_movsfcc_vfp): Likewise
6298 * config/aarch64/aarch64-simd.md
6299 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6300 * config/aarch64/aarch64.md
6301 (*movsi_aarch64): Replace type mov_reg with fmovs.
6302 (*movdi_aarch64): Likewise
6303 (*movsf_aarch64): Likewise
6304 (*movdf_aarch64): Likewise
6306 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6307 * config/arm/iwmmxt.md
6308 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6309 * config/arm/arm1020e.md: Update with new attributes.
6310 * config/arm/cortex-a15-neon.md: Update with new attributes.
6311 * config/arm/cortex-a5.md: Update with new attributes.
6312 * config/arm/cortex-a53.md: Update with new attributes.
6313 * config/arm/cortex-a7.md: Update with new attributes.
6314 * config/arm/cortex-a8-neon.md: Update with new attributes.
6315 * config/arm/cortex-a9.md: Update with new attributes.
6316 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6317 * config/arm/cortex-r4f.md: Update with new attributes.
6318 * config/arm/marvell-pj4.md: Update with new attributes.
6319 * config/arm/vfp11.md: Update with new attributes.
6321 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6323 * config/aarch64/aarch64.md
6324 (*madd<mode>): Fix type attribute.
6325 (*maddsi_uxtw): Likewise.
6326 (*msub<mode>): Likewise.
6327 (*msubsi_uxtw): Likewise.
6328 (<su_optab>maddsidi4): Likewise.
6329 (<su_optab>msubsidi4): Likewise.
6331 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6333 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6334 * config/arm/arm.md (core_cycles): Remove fdiv.
6335 * config/arm/vfp.md:
6336 (*sqrtsf2_vfp): Update for attribute changes.
6337 (*sqrtdf2_vfp): Likewise.
6338 * config/aarch64/aarch64.md:
6339 (sqrt<mode>2): Update for attribute changes.
6340 * config/arm/arm1020e.md: Update with new attributes.
6341 * config/arm/cortex-a15-neon.md: Update with new attributes.
6342 * config/arm/cortex-a5.md: Update with new attributes.
6343 * config/arm/cortex-a53.md: Update with new attributes.
6344 * config/arm/cortex-a7.md: Update with new attributes.
6345 * config/arm/cortex-a8-neon.md: Update with new attributes.
6346 * config/arm/cortex-a9.md: Update with new attributes.
6347 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6348 * config/arm/cortex-r4f.md: Update with new attributes.
6349 * config/arm/marvell-pj4.md: Update with new attributes.
6350 * config/arm/vfp11.md: Update with new attributes.
6352 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6354 * config/arm/types.md
6355 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6356 * config/aarch64/aarch64.md
6357 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6359 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6360 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6361 (float<GPI:mode><GPF:mode>2): Likewise.
6363 (*truncsisf2_vfp): Update with new attributes.
6364 (*truncsidf2_vfp): Likewise.
6365 (fixuns_truncsfsi2): Likewise.
6366 (fixuns_truncdfsi2): Likewise.
6367 (*floatsisf2_vfp): Likewise.
6368 (*floatsidf2_vfp): Likewise.
6369 (floatunssisf2): Likewise.
6370 (floatunssidf2): Likewise.
6371 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6372 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
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/arm_neon.h
6388 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6389 (vqtbx<1,2,3,4><q>_s8): Likewise.
6391 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6393 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6394 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6395 (add<mode>3): Add type attribute.
6396 (add<mode>3): Likewise.
6397 (usadd<mode>3): Likewise.
6398 (ssadd<mode>3): Likewise.
6399 (sub<mode>3): Likewise.
6400 (sub<mode>3): Likewise.
6401 (ussub<mode>3): Likewise.
6402 (sssub<mode>3): Likewise.
6403 (ssmulsa3): Likewise.
6404 (usmulusa3): Likewise.
6405 (arm_usatsihi): Likewise.
6407 (*movdi_vfp): Add types for all instructions.
6408 (*movdi_vfp_cortexa8): Likewise.
6409 (*movhf_vfp_neon): Likewise.
6410 (*movhf_vfp): Likewise.
6411 (*movdf_vfp): Likewise.
6412 (*thumb2_movdf_vfp): Likewise.
6413 (*thumb2_movdfcc_vfp): Likewise.
6414 * config/arm/arm.md: Add type attribute to all insn patterns.
6415 (*thumb1_adddi3): Add type attribute.
6416 (*arm_adddi3): Likewise.
6417 (*adddi_sesidi_di): Likewise.
6418 (*adddi_zesidi_di): Likewise.
6419 (*thumb1_addsi3): Likewise.
6420 (addsi3_compare0): Likewise.
6421 (*addsi3_compare0_scratch): Likewise.
6422 (*compare_negsi_si): Likewise.
6423 (cmpsi2_addneg): Likewise.
6424 (*addsi3_carryin_<optab>): Likewise.
6425 (*addsi3_carryin_alt2_<optab>): Likewise.
6426 (*addsi3_carryin_clobercc_<optab>): Likewise.
6427 (*subsi3_carryin): Likewise.
6428 (*subsi3_carryin_const): Likewise.
6429 (*subsi3_carryin_compare): Likewise.
6430 (*subsi3_carryin_compare_const): Likewise.
6431 (*arm_subdi3): Likewise.
6432 (*thumb_subdi3): Likewise.
6433 (*subdi_di_zesidi): Likewise.
6434 (*subdi_di_sesidi): Likewise.
6435 (*subdi_zesidi_di): Likewise.
6436 (*subdi_sesidi_di): Likewise.
6437 (*subdi_zesidi_ze): Likewise.
6438 (thumb1_subsi3_insn): Likewise.
6439 (*arm_subsi3_insn): Likewise.
6440 (*anddi3_insn): Likewise.
6441 (*anddi_zesidi_di): Likewise.
6442 (*anddi_sesdi_di): Likewise.
6443 (*ne_zeroextracts): Likewise.
6444 (*ne_zeroextracts): Likewise.
6445 (*ite_ne_zeroextr): Likewise.
6446 (*ite_ne_zeroextr): Likewise.
6447 (*anddi_notdi_di): Likewise.
6448 (*anddi_notzesidi): Likewise.
6449 (*anddi_notsesidi): Likewise.
6450 (andsi_notsi_si): Likewise.
6451 (thumb1_bicsi3): Likewise.
6452 (*iordi3_insn): Likewise.
6453 (*iordi_zesidi_di): Likewise.
6454 (*iordi_sesidi_di): Likewise.
6455 (*thumb1_iorsi3_insn): Likewise.
6456 (*xordi3_insn): Likewise.
6457 (*xordi_zesidi_di): Likewise.
6458 (*xordi_sesidi_di): Likewise.
6459 (*arm_xorsi3): Likewise.
6460 (*andsi_iorsi3_no): Likewise.
6461 (*smax_0): Likewise.
6462 (*smax_m1): Likewise.
6463 (*arm_smax_insn): Likewise.
6464 (*smin_0): Likewise.
6465 (*arm_smin_insn): Likewise.
6466 (*arm_umaxsi3): Likewise.
6467 (*arm_uminsi3): Likewise.
6468 (*minmax_arithsi): Likewise.
6469 (*minmax_arithsi_): Likewise.
6470 (*satsi_<SAT:code>): Likewise.
6471 (arm_ashldi3_1bit): Likewise.
6472 (arm_ashrdi3_1bit): Likewise.
6473 (arm_lshrdi3_1bit): Likewise.
6474 (*arm_negdi2): Likewise.
6475 (*thumb1_negdi2): Likewise.
6476 (*arm_negsi2): Likewise.
6477 (*thumb1_negsi2): Likewise.
6478 (*negdi_extendsid): Likewise.
6479 (*negdi_zero_extend): Likewise.
6480 (*arm_abssi2): Likewise.
6481 (*thumb1_abssi2): Likewise.
6482 (*arm_neg_abssi2): Likewise.
6483 (*thumb1_neg_abss): Likewise.
6484 (one_cmpldi2): Likewise.
6485 (extend<mode>di2): Likewise.
6486 (*compareqi_eq0): Likewise.
6487 (*arm_extendhisi2addsi): Likewise.
6488 (*arm_movdi): Likewise.
6489 (*thumb1_movdi_insn): Likewise.
6490 (*arm_movt): Likewise.
6491 (*thumb1_movsi_insn): Likewise.
6492 (pic_add_dot_plus_four): Likewise.
6493 (pic_add_dot_plus_eight): Likewise.
6494 (tls_load_dot_plus_eight): Likewise.
6495 (*thumb1_movhi_insn): Likewise.
6496 (*thumb1_movsf_insn): Likewise.
6497 (*movdf_soft_insn): Likewise.
6498 (*thumb_movdf_insn): Likewise.
6499 (cbranchsi4_insn): Likewise.
6500 (cbranchsi4_scratch): Likewise.
6501 (*negated_cbranchsi4): Likewise.
6502 (*tbit_cbranch): Likewise.
6503 (*tlobits_cbranch): Likewise.
6504 (*tstsi3_cbranch): Likewise.
6505 (*cbranchne_decr1): Likewise.
6506 (*addsi3_cbranch): Likewise.
6507 (*addsi3_cbranch_scratch): Likewise.
6508 (*arm_cmpdi_insn): Likewise.
6509 (*arm_cmpdi_unsig): Likewise.
6510 (*arm_cmpdi_zero): Likewise.
6511 (*thumb_cmpdi_zero): Likewise.
6512 (*deleted_compare): Likewise.
6513 (*mov_scc): Likewise.
6514 (*mov_negscc): Likewise.
6515 (*mov_notscc): Likewise.
6516 (*cstoresi_eq0_thumb1_insn): Likewise.
6517 (cstoresi_nltu_thumb1): Likewise.
6518 (cstoresi_ltu_thu): Likewise.
6519 (thumb1_addsi3_addgeu): Likewise.
6520 (*arm_jump): Likewise.
6521 (*thumb_jump): Likewise.
6522 (*check_arch2): Likewise.
6523 (arm_casesi_internal): Likewise.
6524 (thumb1_casesi_dispatch): Likewise.
6525 (*arm_indirect_jump): Likewise.
6526 (*thumb1_indirect_jump): Likewise.
6528 (*and_scc): Likewise.
6529 (*ior_scc): Likewise.
6530 (*compare_scc): Likewise.
6531 (*cond_move): Likewise.
6532 (*cond_arith): Likewise.
6533 (*cond_sub): Likewise.
6534 (*cmp_ite0): Likewise.
6535 (*cmp_ite1): Likewise.
6536 (*cmp_and): Likewise.
6537 (*cmp_ior): Likewise.
6538 (*ior_scc_scc): Likewise.
6539 (*ior_scc_scc_cmp): Likewise.
6540 (*and_scc_scc): Likewise.
6541 (*and_scc_scc_cmp): Likewise.
6542 (*and_scc_scc_nod): Likewise.
6543 (*negscc): Likewise.
6544 (movcond_addsi): Likewise.
6545 (movcond): Likewise.
6546 (*ifcompare_plus_move): Likewise.
6547 (*if_plus_move): Likewise.
6548 (*ifcompare_move_plus): Likewise.
6549 (*if_move_plus): Likewise.
6550 (*ifcompare_arith_arith): Likewise.
6551 (*if_arith_arith): Likewise.
6552 (*ifcompare_arith_move): Likewise.
6553 (*if_arith_move): Likewise.
6554 (*ifcompare_move_arith): Likewise.
6555 (*if_move_arith): Likewise.
6556 (*ifcompare_move_not): Likewise.
6557 (*if_move_not): Likewise.
6558 (*ifcompare_not_move): Likewise.
6559 (*if_not_move): Likewise.
6560 (*ifcompare_shift_move): Likewise.
6561 (*if_shift_move): Likewise.
6562 (*ifcompare_move_shift): Likewise.
6563 (*if_move_shift): Likewise.
6564 (*ifcompare_shift_shift): Likewise.
6565 (*ifcompare_not_arith): Likewise.
6566 (*ifcompare_arith_not): Likewise.
6567 (*if_arith_not): Likewise.
6568 (*ifcompare_neg_move): Likewise.
6569 (*if_neg_move): Likewise.
6570 (*ifcompare_move_neg): Likewise.
6571 (*if_move_neg): Likewise.
6572 (prologue_thumb1_interwork): Likewise.
6573 (*cond_move_not): Likewise.
6574 (*sign_extract_onebit): Likewise.
6575 (*not_signextract_onebit): Likewise.
6576 (stack_tie): Likewise.
6577 (align_4): Likewise.
6578 (align_8): Likewise.
6579 (consttable_end): Likewise.
6580 (consttable_1): Likewise.
6581 (consttable_2): Likewise.
6582 (consttable_4): Likewise.
6583 (consttable_8): Likewise.
6584 (consttable_16): Likewise.
6585 (*thumb1_tablejump): Likewise.
6586 (prefetch): Likewise.
6587 (force_register_use): Likewise.
6588 (thumb_eh_return): Likewise.
6589 (load_tp_hard): Likewise.
6590 (load_tp_soft): Likewise.
6591 (tlscall): Likewise.
6592 (*arm_movtas_ze): Likewise.
6593 (*arm_rev): Likewise.
6594 (*arm_revsh): Likewise.
6595 (*arm_rev16): Likewise.
6596 * config/arm/thumb2.md
6597 (*thumb2_smaxsi3): Likewise.
6598 (*thumb2_sminsi3): Likewise.
6599 (*thumb32_umaxsi3): Likewise.
6600 (*thumb2_uminsi3): Likewise.
6601 (*thumb2_negdi2): Likewise.
6602 (*thumb2_abssi2): Likewise.
6603 (*thumb2_neg_abss): Likewise.
6604 (*thumb2_movsi_insn): Likewise.
6605 (tls_load_dot_plus_four): Likewise.
6606 (*thumb2_movhi_insn): Likewise.
6607 (*thumb2_mov_scc): Likewise.
6608 (*thumb2_mov_negs): Likewise.
6609 (*thumb2_mov_negs): Likewise.
6610 (*thumb2_mov_nots): Likewise.
6611 (*thumb2_mov_nots): Likewise.
6612 (*thumb2_movsicc_): Likewise.
6613 (*thumb2_movsfcc_soft_insn): Likewise.
6614 (*thumb2_indirect_jump): Likewise.
6615 (*thumb2_and_scc): Likewise.
6616 (*thumb2_ior_scc): Likewise.
6617 (*thumb2_ior_scc_strict_it): Likewise.
6618 (*thumb2_cond_move): Likewise.
6619 (*thumb2_cond_arith): Likewise.
6620 (*thumb2_cond_ari): Likewise.
6621 (*thumb2_cond_sub): Likewise.
6622 (*thumb2_negscc): Likewise.
6623 (*thumb2_movcond): Likewise.
6624 (thumb2_casesi_internal): Likewise.
6625 (thumb2_casesi_internal_pic): Likewise.
6626 (*thumb2_alusi3_short): Likewise.
6627 (*thumb2_mov<mode>_shortim): Likewise.
6628 (*thumb2_addsi_short): Likewise.
6629 (*thumb2_subsi_short): Likewise.
6630 (thumb2_addsi3_compare0): Likewise.
6631 (*thumb2_cbz): Likewise.
6632 (*thumb2_cbnz): Likewise.
6633 (*thumb2_one_cmplsi2_short): Likewise.
6634 (*thumb2_negsi2_short): Likewise.
6635 (*orsi_notsi_si): Likewise.
6636 * config/arm/arm1020e.md: Update with new attributes.
6637 * config/arm/arm1026ejs.md: Update with new attributes.
6638 * config/arm/arm1136jfs.md: Update with new attributes.
6639 * config/arm/arm926ejs.md: Update with new attributes.
6640 * config/arm/cortex-a15.md: Update with new attributes.
6641 * config/arm/cortex-a5.md: Update with new attributes.
6642 * config/arm/cortex-a53.md: Update with new attributes.
6643 * config/arm/cortex-a7.md: Update with new attributes.
6644 * config/arm/cortex-a8.md: Update with new attributes.
6645 * config/arm/cortex-a9.md: Update with new attributes.
6646 * config/arm/cortex-m4.md: Update with new attributes.
6647 * config/arm/cortex-r4.md: Update with new attributes.
6648 * config/arm/fa526.md: Update with new attributes.
6649 * config/arm/fa606te.md: Update with new attributes.
6650 * config/arm/fa626te.md: Update with new attributes.
6651 * config/arm/fa726te.md: Update with new attributes.
6653 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6655 * config/aarch64/aarch64-simd.md
6656 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6657 <vwx> iterator to ensure correct register choice.
6658 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6659 (aarch64_sqdmull_n<mode>): Likewise.
6660 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6661 * config/aarch64/arm_neon.h
6662 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6663 (vml<as><q>_n_<su>16): Likewise.
6664 (vml<as>l_high_lane<q>_<su>16): Likewise.
6665 (vml<as>l_high_n_<su>16): Likewise.
6666 (vml<as>l_lane<q>_<su>16): Likewise.
6667 (vml<as>l_n_<su>16): Likewise.
6668 (vmul<q>_lane<q>_<su>16): Likewise.
6669 (vmul<q>_n_<su>16): Likewise.
6670 (vmull_lane<q>_<su>16): Likewise.
6671 (vmull_n_<su>16): Likewise.
6672 (vmull_high_lane<q>_<su>16): Likewise.
6673 (vmull_high_n_<su>16): Likewise.
6674 (vqrdmulh<q>_n_s16): Likewise.
6676 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6678 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6679 have the correct lane parameter.
6681 2013-09-06 Richard Biener <rguenther@suse.de>
6683 * cfganal.c (control_dependences::~control_dependences):
6684 Properly free all of the vector.
6686 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6689 * config/i386/i386.c (ix86_conditional_register_usage):
6690 Proper initialize extended SSE registers.
6692 2013-09-06 Jan Hubicka <jh@suse.cz>
6694 PR tree-optimization/58311
6695 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6697 2013-09-06 Jan Hubicka <jh@suse.cz>
6699 * Makefile.in (tree-sra.o): Update dependencies.
6700 * tree-sra.c: Include ipa-utils.h
6701 (scan_function): Use recursive_call_p.
6702 (has_caller_p): New function.
6703 (cgraph_for_node_and_aliases): Count also callers of aliases.
6705 2013-09-06 Jan Hubicka <jh@suse.cz>
6708 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6709 * tree-tailcall.c: Include ipa-utils.h.
6710 (find_tail_calls): Use it.
6711 * ipa-pure-const.c (check_call): Likewise.
6712 * ipa-utils.c (recursive_call_p): New function.
6713 * ipa-utils.h (recursive_call_p): Dclare.
6714 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6715 (symtab_semantically_equivalent_p): New function.
6716 * Makefile.in (tree-tailcall.o): Update dependencies.
6718 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6720 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6723 2013-09-06 Richard Biener <rguenther@suse.de>
6725 * lto-streamer.h (lto_global_var_decls): Remove.
6726 * Makefile.in (OBJS): Remove lto-symtab.o.
6727 (lto-symtab.o): Remove.
6728 (GTFILES): Remove lto-symtab.c
6729 * lto-symtab.c: Move to lto/
6731 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6733 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6734 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6735 (UNSPEC_FPINT_RINT): New constant definitions.
6736 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6737 definition with 2 attributes.
6738 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6739 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6742 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6744 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6745 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6746 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6747 attribute to "z196_alone".
6748 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6751 2013-09-06 Richard Biener <rguenther@suse.de>
6753 * basic-block.h (class control_dependences): New.
6754 * tree-ssa-dce.c (control_dependence_map): Remove.
6756 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6757 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6758 find_pdom, find_control_dependence, find_all_control_dependences):
6760 (mark_control_dependent_edges_necessary,
6761 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6762 tree_dce_done, perform_tree_ssa_dce): Adjust.
6763 * cfganal.c (set_control_dependence_map_bit,
6764 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6765 find_all_control_dependences): Move from tree-ssa-dce.c and
6766 implement as methods of control_dependences class.
6767 (control_dependences::control_dependences): New.
6768 (control_dependences::~control_dependences): Likewise.
6769 (control_dependences::get_edges_dependent_on): Likewise.
6770 (control_dependences::get_edge): Likewise.
6772 2013-09-04 Jan Hubicka <jh@suse.cz>
6774 * tree.c (types_same_for_odr): Drop overactive check.
6775 * ipa-devirt.c (hash_type_name): Likewise.
6777 2013-09-04 Jan Hubicka <jh@suse.cz>
6779 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6780 (analyze_functions): ... here.
6782 2013-09-04 Jan Hubicka <jh@suse.cz>
6785 * cgraphunit.c (analyze_functions): Clear AUX fields
6786 after processing; initialize assembler name has.
6788 2013-09-05 Jeff Law <law@redhat.com>
6790 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6791 from thread_around_empty_block. Record threading path into PATH.
6792 Recurse if threading through the initial block is successful.
6793 (thread_across_edge): Corresponding changes to slightly simplify.
6795 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6797 * config/aarch64/aarch64.md
6798 (type): Remove frecpe, frecps, frecpx.
6799 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6800 fix to be a TARGET_SIMD instruction.
6801 (aarch64_frecps): Remove.
6802 * config/aarch64/aarch64-simd.md
6803 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6804 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6806 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6807 Sofiane Naci <sofiane.naci@arm.com>
6809 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6810 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6811 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6812 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6813 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6814 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6815 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6816 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6817 into "clz, "rbit". Rename "shift" to "shift_imm".
6818 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6819 changes. Update for attribute changes all occurrences of arlo_* and
6821 * config/arm/arm-fixed.md: Update for attribute changes
6822 all occurrences of arlo_* types.
6823 * config/arm/thumb2.md: Update for attribute changes all occurrences
6825 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6826 (cortexa7_older_only): Likewise.
6827 (cortexa7_younger): Likewise.
6828 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6829 (1020alu_shift_op): Likewise.
6830 (1020alu_shift_reg_op): Likewise.
6831 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6832 (alu_shift_op): Likewise.
6833 (alu_shift_reg_op): Likewise.
6834 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6835 (11_alu_shift_op): Likewise.
6836 (11_alu_shift_reg_op): Likewise.
6837 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6838 (9_alu_shift_reg_op): Likewise.
6839 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6841 (cortex_a15_alu_shift): Likewise.
6842 (cortex_a15_alu_shift_reg): Likewise.
6843 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6845 (cortex_a5_alu_shift): Likewise.
6846 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6848 (cortex_a53_alu_shift): Likewise.
6849 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6851 (cortex_a7_alu_reg): Likewise.
6852 (cortex_a7_alu_shift): Likewise.
6853 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6855 (cortex_a8_alu_shift): Likewise.
6856 (cortex_a8_alu_shift_reg): Likewise.
6857 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6858 (cortex_a9_dp_shift): Likewise.
6859 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6861 * config/arm/cortex-r4.md
6862 (cortex_r4_alu): Update for attribute changes.
6863 (cortex_r4_mov): Likewise.
6864 (cortex_r4_alu_shift_reg): Likewise.
6865 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6866 (526_alu_shift_op): Likewise.
6867 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6868 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6869 (626te_alu_shift_op): Likewise.
6870 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6871 (726te_alu_shift_op): Likewise.
6872 (726te_alu_shift_reg_op): Likewise.
6873 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6874 (mp626_alu_shift_op): Likewise.
6875 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6876 (pj4_alu_conds): Likewise.
6877 (pj4_shift): Likewise.
6878 (pj4_shift_conds): Likewise.
6879 (pj4_alu_shift): Likewise.
6880 (pj4_alu_shift_conds): Likewise.
6881 * config/aarch64/aarch64.md: Update for attribute change
6882 all occurrences of arlo_* and shift* types.
6884 2013-09-05 Mike Stump <mikestump@comcast.net>
6886 * tree.h: Move documentation for tree_function_decl to tree-core.h
6887 with the declaration.
6889 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6892 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6893 looking for widest mode.
6895 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6897 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6899 2013-09-05 Richard Biener <rguenther@suse.de>
6901 PR tree-optimization/58137
6902 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6903 Do not create vectors of pointers.
6904 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6905 types for the components of the vector initializer.
6906 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6907 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6909 2013-09-05 Martin Jambor <mjambor@suse.cz>
6911 * ipa-prop.c (remove_described_reference): Accept missing references,
6912 return false if that hppens, otherwise return true.
6913 (cgraph_node_for_jfunc): New function.
6914 (try_decrement_rdesc_refcount): Likewise.
6915 (try_make_edge_direct_simple_call): Use them.
6916 (ipa_edge_removal_hook): Remove references from rdescs.
6917 (ipa_edge_duplication_hook): Clone rdescs and their references
6918 when the new edge has the same caller as the old one.
6919 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6920 reference before removing any edges.
6922 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6924 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6926 * thumb2.md (thumb2_storewb_parisi): New pattern.
6928 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6930 * config/aarch64/aarch64-option-extensions.def: Add
6931 AARCH64_OPT_EXTENSION of 'crc'.
6932 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6933 (AARCH64_ISA_CRC): Ditto.
6934 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6935 description of the CRC extension.
6937 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6939 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6940 * config/rs6000/linux.h: Ditto.
6941 * alpha/linux.h: Ditto.
6942 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6943 no_c99_libc_has_function.
6944 * config/c6x/uclinux-elf.h: Ditto.
6945 * config/lm32/uclinux-elf.h: Ditto.
6946 * config/m68k/uclinux.h: Ditto.
6947 * config/moxie/uclinux.h: Ditto.
6948 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6949 (crisv32-*-linux*, cris-*-linux*): Ditto.
6950 * config/bfin/bfin.c: Include "tm_p.h".
6952 2013-09-05 Richard Biener <rguenther@suse.de>
6954 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6955 check for a definition without a basic-block.
6957 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6958 Sofiane Naci <sofiane.naci@arm.com>
6960 * config/aarch64/aarch64.md
6961 (*movti_aarch64): Rename r_2_f and f_2_r.
6962 (*movsf_aarch64): Likewise.
6963 (*movdf_aarch64): Likewise.
6964 (*movtf_aarch64): Likewise.
6965 (aarch64_movdi_<mode>low): Likewise.
6966 (aarch64_movdi_<mode>high): Likewise.
6967 (aarch64_mov<mode>high_di): Likewise.
6968 (aarch64_mov<mode>low_di): Likewise.
6969 (aarch64_movtilow_tilow): Likewise.
6970 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6971 values to config/arm/types.md
6972 (attribute "conds"): Update for attribute change.
6973 (anddi3_insn): Likewise.
6974 (iordi3_insn): Likewise.
6975 (xordi3_insn): Likewise.
6976 (one_cmpldi2): Likewise.
6977 * config/arm/types.md (type): Add Neon types.
6978 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6979 use "type" attribute.
6980 (movmisalign<mode>_neon_store): Likewise.
6981 (movmisalign<mode>_neon_load): Likewise.
6982 (vec_set<mode>_internal): Likewise.
6983 (vec_setv2di_internal): Likewise.
6984 (vec_extract<mode>): Likewise.
6985 (vec_extractv2di): Likewise.
6986 (add<mode>3_neon): Likewise.
6987 (adddi3_neon): Likewise.
6988 (sub<mode>3_neon): Likewise.
6989 (subdi3_neon): Likewise.
6990 (mul<mode>3_neon): Likewise.
6991 (mul<mode>3add<mode>_neon): Likewise.
6992 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6993 (fma<VCVTF:mode>4)): Likewise.
6994 (fma<VCVTF:mode>4_intrinsic): Likewise.
6995 (fmsub<VCVTF:mode>4)): Likewise.
6996 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6997 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6998 (ior<mode>3): Likewise.
6999 (and<mode>3): Likewise.
7000 (anddi3_neon): Likewise.
7001 (orn<mode>3_neon): Likewise.
7002 (orndi3_neon): Likewise.
7003 (bic<mode>3_neon): Likewise.
7004 (bicdi3_neon): Likewise.
7005 (xor<mode>3): Likewise.
7006 (one_cmpl<mode>2): Likewise.
7007 (abs<mode>2): Likewise.
7008 (neg<mode>2): Likewise.
7009 (umin<mode>3_neon): Likewise.
7010 (umax<mode>3_neon): Likewise.
7011 (smin<mode>3_neon): Likewise.
7012 (smax<mode>3_neon): Likewise.
7013 (vashl<mode>3): Likewise.
7014 (vashr<mode>3_imm): Likewise.
7015 (vlshr<mode>3_imm): Likewise.
7016 (ashl<mode>3_signed): Likewise.
7017 (ashl<mode>3_unsigned): Likewise.
7018 (neon_load_count): Likewise.
7019 (ashldi3_neon_noclobber): Likewise.
7020 (signed_shift_di3_neon): Likewise.
7021 (unsigned_shift_di3_neon): Likewise.
7022 (ashrdi3_neon_imm_noclobber): Likewise.
7023 (lshrdi3_neon_imm_noclobber): Likewise.
7024 (widen_ssum<mode>3): Likewise.
7025 (widen_usum<mode>3): Likewise.
7026 (quad_halves_<code>v4si): Likewise.
7027 (quad_halves_<code>v4sf): Likewise.
7028 (quad_halves_<code>v8hi): Likewise.
7029 (quad_halves_<code>v16qi): Likewise.
7030 (reduc_splus_v2di): Likewise.
7031 (neon_vpadd_internal<mode>): Likewise.
7032 (neon_vpsmin<mode>): Likewise.
7033 (neon_vpsmax<mode>): Likewise.
7034 (neon_vpumin<mode>): Likewise.
7035 (neon_vpumax<mode>): Likewise.
7036 (ss_add<mode>_neon): Likewise.
7037 (us_add<mode>_neon): Likewise.
7038 (ss_sub<mode>_neon): Likewise.
7039 (us_sub<mode>_neon): Likewise.
7040 (neon_vadd<mode>_unspec): Likewise.
7041 (neon_vaddl<mode>): Likewise.
7042 (neon_vaddw<mode>): Likewise.
7043 (neon_vhadd<mode>): Likewise.
7044 (neon_vqadd<mode>): Likewise.
7045 (neon_vaddhn<mode>): Likewise.
7046 (neon_vmul<mode>): Likewise.
7047 (neon_vmla<mode>): Likewise.
7048 (neon_vmlal<mode>): Likewise.
7049 (neon_vmls<mode>): Likewise.
7050 (neon_vmlsl<mode>): Likewise.
7051 (neon_vqdmulh<mode>): Likewise.
7052 (neon_vqdmlal<mode>): Likewise.
7053 (neon_vqdmlsl<mode>): Likewise.
7054 (neon_vmull<mode>): Likewise.
7055 (neon_vqdmull<mode>): Likewise.
7056 (neon_vsub<mode>_unspec): Likewise.
7057 (neon_vsubl<mode>): Likewise.
7058 (neon_vsubw<mode>): Likewise.
7059 (neon_vqsub<mode>): Likewise.
7060 (neon_vhsub<mode>): Likewise.
7061 (neon_vsubhn<mode>): Likewise.
7062 (neon_vceq<mode>): Likewise.
7063 (neon_vcge<mode>): Likewise.
7064 (neon_vcgeu<mode>): Likewise.
7065 (neon_vcgt<mode>): Likewise.
7066 (neon_vcgtu<mode>): Likewise.
7067 (neon_vcle<mode>): Likewise.
7068 (neon_vclt<mode>): Likewise.
7069 (neon_vcage<mode>): Likewise.
7070 (neon_vcagt<mode>): Likewise.
7071 (neon_vtst<mode>): Likewise.
7072 (neon_vabd<mode>): Likewise.
7073 (neon_vabdl<mode>): Likewise.
7074 (neon_vaba<mode>): Likewise.
7075 (neon_vabal<mode>): Likewise.
7076 (neon_vmax<mode>): Likewise.
7077 (neon_vmin<mode>): Likewise.
7078 (neon_vpaddl<mode>): Likewise.
7079 (neon_vpadal<mode>): Likewise.
7080 (neon_vpmax<mode>): Likewise.
7081 (neon_vpmin<mode>): Likewise.
7082 (neon_vrecps<mode>): Likewise.
7083 (neon_vrsqrts<mode>): Likewise.
7084 (neon_vqabs<mode>): Likewise.
7085 (neon_vqneg<mode>): Likewise.
7086 (neon_vcls<mode>): Likewise.
7087 (clz<mode>2): Likewise.
7088 (popcount<mode>2): Likewise.
7089 (neon_vrecpe): Likewise.
7090 (neon_vrsqrte): Likewise.
7091 (neon_vget_lane<mode>_sext_internal): Likewise.
7092 (neon_vget_lane<mode>_zext_internal): Likewise.
7093 (neon_vdup_n<mode>): Likewise.
7094 (neon_vdup_nv2di): Likewise.
7095 (neon_vdpu_lane<mode>_internal): Likewise.
7096 (neon_vswp<mode>): Likewise.
7097 (float<mode><V_cvtto>2): Likewise.
7098 (floatuns<mode><V_cvtto>2): Likewise.
7099 (fix_trunc<mode><V_cvtto>)2): Likewise
7100 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7101 (neon_vcvt<mode>): Likewise.
7102 (neon_vcvtv4sfv4hf): Likewise.
7103 (neon_vcvtv4hfv4sf): Likewise.
7104 (neon_vcvt_n<mode>): Likewise.
7105 (neon_vmovn<mode>): Likewise.
7106 (neon_vqmovn<mode>): Likewise.
7107 (neon_vqmovun<mode>): Likewise.
7108 (neon_vmovl<mode>): Likewise.
7109 (neon_vmul_lane<mode>): Likewise.
7110 (neon_vmull_lane<mode>): Likewise.
7111 (neon_vqdmull_lane<mode>): Likewise.
7112 (neon_vqdmulh_lane<mode>): Likewise.
7113 (neon_vmla_lane<mode>): Likewise.
7114 (neon_vmlal_lane<mode>): Likewise.
7115 (neon_vqdmlal_lane<mode>): Likewise.
7116 (neon_vmls_lane<mode>): Likewise.
7117 (neon_vmlsl_lane<mode>): Likewise.
7118 (neon_vqdmlsl_lane<mode>): Likewise.
7119 (neon_vext<mode>): Likewise.
7120 (neon_vrev64<mode>): Likewise.
7121 (neon_vrev32<mode>): Likewise.
7122 (neon_vrev16<mode>): Likewise.
7123 (neon_vbsl<mode>_internal): Likewise.
7124 (neon_vshl<mode>): Likewise.
7125 (neon_vqshl<mode>): Likewise.
7126 (neon_vshr_n<mode>): Likewise.
7127 (neon_vshrn_n<mode>): Likewise.
7128 (neon_vqshrn_n<mode>): Likewise.
7129 (neon_vqshrun_n<mode>): Likewise.
7130 (neon_vshl_n<mode>): Likewise.
7131 (neon_vqshl_n<mode>): Likewise.
7132 (neon_vqshlu_n<mode>): Likewise.
7133 (neon_vshll_n<mode>): Likewise.
7134 (neon_vsra_n<mode>): Likewise.
7135 (neon_vsri_n<mode>): Likewise.
7136 (neon_vsli_n<mode>): Likewise.
7137 (neon_vtbl1v8qi): Likewise.
7138 (neon_vtbl2v8qi): Likewise.
7139 (neon_vtbl3v8qi): Likewise.
7140 (neon_vtbl4v8qi): Likewise.
7141 (neon_vtbx1v8qi): Likewise.
7142 (neon_vtbx2v8qi): Likewise.
7143 (neon_vtbx3v8qi): Likewise.
7144 (neon_vtbx4v8qi): Likewise.
7145 (neon_vtrn<mode>_internal): Likewise.
7146 (neon_vzip<mode>_internal): Likewise.
7147 (neon_vuzp<mode>_internal): Likewise.
7148 (neon_vld1<mode>): Likewise.
7149 (neon_vld1_lane<mode>): Likewise.
7150 (neon_vld1_dup<mode>): Likewise.
7151 (neon_vld1_dupv2di): Likewise.
7152 (neon_vst1<mode>): Likewise.
7153 (neon_vst1_lane<mode>): Likewise.
7154 (neon_vld2<mode>): Likewise.
7155 (neon_vld2_lane<mode>): Likewise.
7156 (neon_vld2_dup<mode>): Likewise.
7157 (neon_vst2<mode>): Likewise.
7158 (neon_vst2_lane<mode>): Likewise.
7159 (neon_vld3<mode>): Likewise.
7160 (neon_vld3qa<mode>): Likewise.
7161 (neon_vld3qb<mode>): Likewise.
7162 (neon_vld3_lane<mode>): Likewise.
7163 (neon_vld3_dup<mode>): Likewise.
7164 (neon_vst3<mode>): Likewise.
7165 (neon_vst3qa<mode>): Likewise.
7166 (neon_vst3qb<mode>): Likewise.
7167 (neon_vst3_lane<mode>): Likewise.
7168 (neon_vld4<mode>): Likewise.
7169 (neon_vld4qa<mode>): Likewise.
7170 (neon_vld4qb<mode>): Likewise.
7171 (neon_vld4_lane<mode>): Likewise.
7172 (neon_vld4_dup<mode>): Likewise.
7173 (neon_vst4<mode>): Likewise.
7174 (neon_vst4qa<mode>): Likewise.
7175 (neon_vst4qb<mode>): Likewise.
7176 (neon_vst4_lane<mode>): Likewise.
7177 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7178 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7179 (neon_vec_<US>mult_lo_<mode>): Likewise.
7180 (neon_vec_<US>mult_hi_<mode>): Likewise.
7181 (neon_vec_<US>shiftl_<mode>): Likewise.
7182 (neon_unpack<US>_<mode>): Likewise.
7183 (neon_vec_<US>mult_<mode>): Likewise.
7184 (vec_pack_trunc_<mode>): Likewise.
7185 (neon_vec_pack_trunk_<mode>): Likewise.
7186 (neon_vabd<mode>_2): Likewise.
7187 (neon_vabd<mode>_3): Likewise.
7188 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7189 (thumb2_movsi_vfp): Likewise.
7190 (movdi_vfp): Likewise.
7191 (movdi_vfp_cortexa8): Likewise.
7192 (movhf_vfp_neon): Likewise.
7193 (movhf_vfp): Likewiwse.
7194 (movsf_vfp): Likewiwse.
7195 (thumb2_movsf_vfp): Likewiwse.
7196 (movdf_vfp): Likewise.
7197 (thumb2_movdf_vfp): Likewise.
7198 (movsfcc_vfp): Likewise.
7199 (thumb2_movsfcc_vfp): Likewise.
7200 (movdfcc_vfp): Likewise.
7201 (thumb2_movdfcc_vfp): Likewise.
7202 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7203 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7204 (v10_v2c): Likewise.
7205 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7207 (cortex_a15_neon_int_2): Likewise.
7208 (cortex_a15_neon_int_3): Likewise.
7209 (cortex_a15_neon_int_4): Likewise.
7210 (cortex_a15_neon_int_5): Likewise.
7211 (cortex_a15_neon_vqneg_vqabs): Likewise.
7212 (cortex_a15_neon_vmov): Likewise.
7213 (cortex_a15_neon_vaba): Likewise.
7214 (cortex_a15_neon_vaba_qqq): Likewise.
7215 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7216 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7217 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7218 scalar_64_32_long_scalar): Likewise.
7219 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7220 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7221 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7222 lotype_qdd_64_32_long): Likewise.
7223 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7224 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7225 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7226 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7227 (cortex_a15_neon_shift_1): Likewise.
7228 (cortex_a15_neon_shift_2): Likewise.
7229 (cortex_a15_neon_shift_3): Likewise.
7230 (cortex_a15_neon_vshl_ddd): Likewise.
7231 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7232 (cortex_a15_neon_vsra_vrsra): Likewise.
7233 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7234 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7235 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7236 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7237 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7238 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7239 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7240 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7241 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7242 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7243 (cortex_a15_neon_bp_simple): Likewise.
7244 (cortex_a15_neon_bp_2cycle): Likewise.
7245 (cortex_a15_neon_bp_3cycle): Likewise.
7246 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7247 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7248 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7249 (cortex_a15_neon_vld2_4_regs): Likewise.
7250 (cortex_a15_neon_vld3_vld4): Likewise.
7251 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7252 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7253 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7254 (cortex_a15_neon_vst3_vst4): Likewise.
7255 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7256 (cortex_a15_neon_vld3_vld4_lane" 10
7257 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7258 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7259 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7260 (cortex_a15_neon_ldm_2): Likewise.0
7261 (cortex_a15_neon_stm_2): Likewise.
7262 (cortex_a15_neon_mcr): Likewise.
7263 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7264 (cortex_a15_neon_mrc): Likewise.
7265 (cortex_a15_neon_mrrc): Likewise.
7266 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7268 (cortex_a15_alu_shift): Likewise.
7269 (cortex_a15_alu_shift_reg): Likewise.
7270 (cortex_a15_mult32): Likewise.
7271 (cortex_a15_mult64): Likewise.
7272 (cortex_a15_block): Likewise.
7273 (cortex_a15_branch): Likewise.
7274 (cortex_a15_load1): Likewise.
7275 (cortex_a15_load3): Likewise.
7276 (cortex_a15_store1): Likewise.
7277 (cortex_a15_store3): Likewise.
7278 (cortex_a15_call): Likewise.
7279 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7280 (cortex_a5_f2r): Likewise.
7281 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7283 (cortex_a53_f2r): Likewise.
7284 * config/arm/cortex-a7.md
7285 (cortex_a7_branch): Update for attribute change.
7286 (cortex_a7_call): Likewise.
7287 (cortex_a7_alu_imm): Likewise.
7288 (cortex_a7_alu_reg): Likewise.
7289 (cortex_a7_alu_shift): Likewise.
7290 (cortex_a7_mul): Likewise.
7291 (cortex_a7_load1): Likewise.
7292 (cortex_a7_store1): Likewise.
7293 (cortex_a7_load2): Likewise.
7294 (cortex_a7_store2): Likewise.
7295 (cortex_a7_load3): Likewise.
7296 (cortex_a7_store3): Likewise.
7297 (cortex_a7_load4): Likewise.
7298 (cortex_a7_store4): Likewise.
7299 (cortex_a7_fpalu): Likewise.
7300 (cortex_a7_fconst): Likewise.
7301 (cortex_a7_fpmuls): Likewise.
7302 (cortex_a7_neon_mul): Likewise.
7303 (cortex_a7_fpmacs): Likewise.
7304 (cortex_a7_neon_mla: Likewise.
7305 (cortex_a7_fpmuld: Likewise.
7306 (cortex_a7_fpmacd: Likewise.
7307 (cortex_a7_fpfmad: Likewise.
7308 (cortex_a7_fdivs: Likewise.
7309 (cortex_a7_fdivd: Likewise.
7310 (cortex_a7_r2f: Likewise.
7311 (cortex_a7_f2r: Likewise.
7312 (cortex_a7_f_flags: Likewise.
7313 (cortex_a7_f_loads: Likewise.
7314 (cortex_a7_f_loadd: Likewise.
7315 (cortex_a7_f_stores: Likewise.
7316 (cortex_a7_f_stored: Likewise.
7317 (cortex_a7_neon): Likewise.
7318 * config/arm/cortex-a8-neon.md
7319 (cortex_a8_neon_mrc): Update for attribute change.
7320 (cortex_a8_neon_mrrc): Likewise.
7321 (cortex_a8_neon_int_1): Likewise.
7322 (cortex_a8_neon_int_2): Likewise.
7323 (cortex_a8_neon_int_3): Likewise.
7324 (cortex_a8_neon_int_4): Likewise.
7325 (cortex_a8_neon_int_5): Likewise.
7326 (cortex_a8_neon_vqneg_vqabs): Likewise.
7327 (cortex_a8_neon_vmov): Likewise.
7328 (cortex_a8_neon_vaba): Likewise.
7329 (cortex_a8_neon_vaba_qqq): Likewise.
7330 (cortex_a8_neon_vsma): Likewise.
7331 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7332 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7333 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7334 long_scalar): Likewise.
7335 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7336 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7337 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7338 long_scalar_qdd_64_32_long): Likewise.
7339 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7340 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7341 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7342 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7343 (cortex_a8_neon_shift_1): Likewise.
7344 (cortex_a8_neon_shift_2): Likewise.
7345 (cortex_a8_neon_shift_3): Likewise.
7346 (cortex_a8_neon_vshl_ddd): Likewise.
7347 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7348 (cortex_a8_neon_vsra_vrsra): Likewise.
7349 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7350 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7351 (cortex_a8_neon_fp_vsum): Likewise.
7352 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7353 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7354 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7355 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7356 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7357 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7358 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7359 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7360 (cortex_a8_neon_bp_simple): Likewise.
7361 (cortex_a8_neon_bp_2cycle): Likewise.
7362 (cortex_a8_neon_bp_3cycle): Likewise.
7363 (cortex_a8_neon_ldr): Likewise.
7364 (cortex_a8_neon_str): Likewise.
7365 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7366 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7367 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7368 (cortex_a8_neon_vld2_4_regs): Likewise.
7369 (cortex_a8_neon_vld3_vld4): Likewise.
7370 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7371 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7372 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7373 (cortex_a8_neon_vst3_vst4): Likewise.
7374 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7375 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7376 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7377 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7378 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7379 (cortex_a8_neon_mcr): Likewise.
7380 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7381 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7382 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7384 (ca9_neon_mrrc): Likewise.
7385 (cortex_a9_neon_int_1): Likewise.
7386 (cortex_a9_neon_int_2): Likewise.
7387 (cortex_a9_neon_int_3): Likewise.
7388 (cortex_a9_neon_int_4): Likewise.
7389 (cortex_a9_neon_int_5): Likewise.
7390 (cortex_a9_neon_vqneg_vqabs): Likewise.
7391 (cortex_a9_neon_vmov): Likewise.
7392 (cortex_a9_neon_vaba): Likewise.
7393 (cortex_a9_neon_vaba_qqq): Likewise.
7394 (cortex_a9_neon_vsma): Likewise.
7395 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7396 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7397 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7398 long_scalar): Likewise.
7399 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7400 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7401 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7402 long_scalar_qdd_64_32_long): Likewise.
7403 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7404 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7405 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7406 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7407 (cortex_a9_neon_shift_1): Likewise.
7408 (cortex_a9_neon_shift_2): Likewise.
7409 (cortex_a9_neon_shift_3): Likewise.
7410 (cortex_a9_neon_vshl_ddd): Likewise.
7411 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7412 (cortex_a9_neon_vsra_vrsra): Likewise.
7413 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7414 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7415 (cortex_a9_neon_fp_vsum): Likewise.
7416 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7417 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7418 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7419 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7420 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7421 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7422 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7423 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7424 (cortex_a9_neon_bp_simple): Likewise.
7425 (cortex_a9_neon_bp_2cycle): Likewise.
7426 (cortex_a9_neon_bp_3cycle): Likewise.
7427 (cortex_a9_neon_ldr): Likewise.
7428 (cortex_a9_neon_str): Likewise.
7429 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7430 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7431 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7432 (cortex_a9_neon_vld2_4_regs): Likewise.
7433 (cortex_a9_neon_vld3_vld4): Likewise.
7434 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7435 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7436 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7437 (cortex_a9_neon_vst3_vst4): Likewise.
7438 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7439 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7440 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7441 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7442 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7443 (cortex_a9_neon_mcr): Likewise.
7444 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7445 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7446 (cortex_a9_fps): Likewise.
7447 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7449 (cortex_m4_fmuls): Likewise.
7450 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7452 (cortex_r4_mrc): Likewise.
7453 * config/arm/iterators.md: Update comment referring to neon_type.
7454 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7455 (iwmmxt_movsi_insn): Likewise.
7456 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7458 (pj4_core_to_vfp): Likewise.
7459 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7461 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7462 (vfp_fstore): Likewise.
7463 * doc/md.texi: Change references to neon_type to refer to type.
7465 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7467 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7469 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7471 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7474 2013-09-04 Easwaran Raman <eraman@google.com>
7477 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7478 (build_and_add_sum): Use it.
7479 (appears_later_in_bb): Simplify code.
7481 2013-09-04 Teresa Johnson <tejohnson@google.com>
7483 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7485 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7487 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7489 2013-09-04 Jan Hubicka <jh@suse.cz>
7491 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7492 * ipa-devirt.c: Include diganostic.h
7493 (odr_type_d): Add types and types_set.
7494 (hash_type_name): Work for types with vtables during LTO.
7495 (odr_hasher::remove): Fix comment; destroy types_set.
7496 (add_type_duplicate): New function,
7497 (get_odr_type): Use it.
7498 (dump_type_inheritance_graph): Dump type duplicates.
7499 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7501 * tree.c (types_same_for_odr): Give exact answers on types with
7504 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7506 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7507 explaining their differences.
7509 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7511 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7513 2013-09-03 Jeff Law <law@redhat.com>
7515 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7516 when not threading through a joiner block. Pass joiner/no joiner
7517 state to register_jump_thread.
7518 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7519 state from argument rather than implying on path length.
7520 Dump the entire jump thread path into debugging dump.
7521 * tree-flow.h (register_jump_thread): Update prototype.
7523 2013-08-29 Xinliang David Li <davidxl@google.com>
7525 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7526 Remove a trivial gcc_assert.
7528 2013-08-29 Xinliang David Li <davidxl@google.com>
7530 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7531 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7532 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7533 Delay base decl alignment adjustment.
7534 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7535 * tree-vectorizer.h: New data structure.
7536 (set_dr_misalignment): New function.
7537 (dr_misalignment): Ditto.
7538 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7539 (vectorizable_load): Ditto.
7540 (ensure_base_align): New function.
7541 (vectorize_loops): Add dbg_cnt support.
7542 (execute_vect_slp): Ditto.
7543 * dbgcnt.def: New debug counter.
7544 * Makefile: New dependency.
7546 2013-09-03 Meador Inge <meadori@codesourcery.com>
7550 2013-08-30 Meador Inge <meadori@codesourcery.com>
7552 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7554 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7556 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7557 function descriptor.
7559 2013-09-03 Richard Biener <rguenther@suse.de>
7561 * tree-affine.c (add_elt_to_tree): Fix association issue,
7562 avoid useless converts and make sure to always return a
7563 properly typed result.
7565 2013-09-03 Richard Biener <rguenther@suse.de>
7568 * fold-const.c (negate_expr_p): Fix division case.
7569 (negate_expr): Likewise.
7571 2013-09-03 Richard Biener <rguenther@suse.de>
7574 * tree-streamer-out.c: Include tm.h.
7575 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7577 2013-09-03 Jan Hubicka <jh@suse.cz>
7579 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7581 2013-09-03 Alan Modra <amodra@gmail.com>
7583 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7584 multilibs to big-endian target and vice versa.
7585 * config/rs6000/t-linux64: Use := assignment on all vars.
7586 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7587 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7588 * config/rs6000/t-linux64le: New file.
7589 * config/rs6000/t-linux64bele: New file.
7590 * config/rs6000/t-linux64lebe: New file.
7592 2013-09-02 Jan Hubicka <jh@suse.cz>
7594 * ipa-inline-transform.c (inline_transform): Do not
7595 optimize_inline_calls when not optimizing.
7597 2013-09-02 Jan Hubicka <jh@suse.cz>
7599 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7600 duplicated nodes for assembler names.
7601 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7602 hard registers from assembler name hash.
7604 2013-09-02 Jan Hubicka <jh@suse.cz>
7606 * ipa-split.c (execute_split_functions): Split externally visible
7607 functions called once.
7609 2013-09-02 Martin Jambor <mjambor@suse.cz>
7612 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7613 linked list. When finding the correct duplicate, also consider also
7614 the caller in additon to its inlined_to node.
7616 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7618 * config/aarch64/aarch64-simd-builtins.def
7619 (dup_lane_scalar): Remove.
7620 * config/aarch64/aarch64-simd.md
7621 (aarch64_simd_dup): Add 'w->w' alternative.
7622 (aarch64_dup_lane<mode>): Allow for VALL.
7623 (aarch64_dup_lane_scalar<mode>): Remove.
7624 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7625 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7626 (aarch64_get_lane_unsigned<mode>): Likewise.
7627 (aarch64_get_lane<mode>): Likewise.
7628 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7629 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7630 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7631 (VCON): Change container of V2SF.
7632 (vswap_width_name): Likewise.
7633 * config/aarch64/arm_neon.h
7634 (__aarch64_vdup_lane_any): New.
7635 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7636 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7637 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7639 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7642 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7643 the source or the destination is a single hard register.
7645 2013-09-02 Richard Biener <rguenther@suse.de>
7648 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7651 2013-09-02 Richard Biener <rguenther@suse.de>
7653 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7654 arithmetic to sizetype.
7656 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7658 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7659 Find auto-increment use both before and after candidate.
7661 2013-09-02 Marek Polacek <polacek@redhat.com>
7663 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7665 2013-09-01 Jan Hubicka <jh@suse.cz>
7667 * Makefile.in: Add ipa-profile.o
7668 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7669 * cgraph.c (struct cgraph_propagate_frequency_data,
7670 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7671 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7672 * cgraph.h (cgraph_propagate_frequency): Remove.
7673 * ipa-inline-analysis.c: Include ipa-utils.h;
7674 drop duplicated cfgloop.h.
7675 (inline_update_callee_summaries): Update.
7676 * ipa-profile.c: New file.
7677 * ipa-utils.h (ipa_propagate_frequency): Declare.
7678 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7679 data-streamer.h, value-prof.h.
7680 (symtab_remove_unreachable_nodes): Update profile.
7681 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7682 account_time_size, cmp_counts, dump_histogram,
7683 ipa_profile_generate_summary, ipa_profile_write_summary,
7684 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7685 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7686 Move to ipa-profile.c.
7688 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7690 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7692 2013-09-01 Jan Hubicka <jh@suse.cz>
7694 * common.opt (fdevirtualize-speculatively): New function.
7695 * invoke.texi (fdevirtualize-speculatively): Document.
7696 * ipa-devirt.c: Include ipa-inline.h
7697 (likely_target_p): New function.
7698 (ipa_devirt): New function.
7699 (gate_ipa_devirt): New function.
7700 (pass_data_ipa_devirt): New static var.
7701 (pass_ipa_devirt): Likewise.
7702 (make_pass_ipa_devirt): New function.
7703 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7704 (common_handle_option): Disable devirtualization when
7705 value range profiling is available.
7706 * passes.def (pass_ipa_devirt): Add.
7707 * timever.def (TV_IPA_DEVIRT): New timevar.
7708 * tree-pass.h (make_pass_ipa_devirt):
7710 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7712 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7713 include sanitize(undefined).
7715 2013-08-31 Diego Novillo <dnovillo@google.com>
7717 * Makefile.in (TREE_CORE_H): Define.
7719 (GTFILES): Add tree-core.h.
7720 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7722 * tree-core.h: New file.
7723 Move all data structures, enum, typedefs, global
7724 declarations and constants from ...
7727 2013-08-31 Jan Hubicka <jh@suse.cz>
7729 * bulitins.c (expand_builtin): Do not early exit for gcov
7730 instrumented functions.
7732 2013-08-31 Marek Polacek <polacek@redhat.com>
7734 * ubsan.c: Include tm_p.h.
7736 2013-08-31 Jan Hubicka <jh@suse.cz>
7738 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7741 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7742 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7743 * tree-cfg.c (verify_gimple_label): ... here.
7744 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7745 (ipa_merge_profiles): New function.
7746 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7747 (lto_input_function_body): Likewise.
7748 * ipa-utils.h (ipa_merge_profiles): Declare.
7749 * lto-streamer.h (lto_input_function_body): Update prototype.
7750 (emit_label_in_global_context_p): Remove.
7751 * lto-symtab.c: Include ipa-utils.h
7752 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7754 2013-08-31 Jan Hubicka <jh@suse.cz>
7756 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7758 2013-08-31 Jan Hubicka <jh@suse.cz>
7760 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7762 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7764 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7765 "cmp" RTX before signed_comparison_operator check to account
7768 2013-08-30 Jan Hubicka <jh@suse.cz>
7770 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7771 (detect_type_change_1): Rename to ...
7772 (detect_type_change): ... this one; early return on non-polymorphic
7774 (detect_type_change_ssa): Add comp_type parameter; update
7775 use of detect_type_change.
7776 (compute_complex_assign_jump_func): Add param_type parameter;
7777 update use of detect_type_change_ssa.
7778 (compute_complex_ancestor_jump_func): Likewise.
7779 (ipa_get_callee_param_type): New function.
7780 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7781 update calls to the jump function computation functions.
7783 2013-08-30 Teresa Johnson <tejohnson@google.com>
7784 Steven Bosscher <steven@gcc.gnu.org>
7786 * cfgrtl.c (fixup_new_cold_bb): New routine.
7787 (commit_edge_insertions): Invoke fixup_partitions.
7788 (find_partition_fixes): New routine.
7789 (fixup_partitions): Ditto.
7790 (verify_hot_cold_block_grouping): Update comments.
7791 (rtl_verify_edges): Invoke find_partition_fixes.
7792 (rtl_verify_bb_pointers): Update comments.
7793 (rtl_verify_bb_layout): Ditto.
7794 * basic-block.h (probably_never_executed_edge_p): Declare.
7795 (fixup_partitions): Ditto.
7796 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7797 * bb-reorder.c (sanitize_hot_paths): New function.
7798 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7800 * predict.c (probably_never_executed_edge_p): New routine.
7801 * cfg.c (check_bb_profile): Add partition insanity warnings.
7803 2013-08-30 Meador Inge <meadori@codesourcery.com>
7805 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7807 2013-08-30 Marek Polacek <polacek@redhat.com>
7809 * Makefile.in (ubsan.o): Add.
7810 (c-family/c-ubsan.o): Add.
7811 (builtins.o): Add ubsan.h dependency.
7812 * ubsan.h: New file.
7813 * ubsan.c: New file.
7814 * common.opt: Add -fsanitize=undefined option.
7815 (flag_sanitize): Add variable.
7816 (fsanitize=): Add option. Add Driver.
7817 (fsanitize=thread): Remove option.
7818 (fsanitize=address): Likewise.
7819 (static-libubsan): New option.
7820 * doc/invoke.texi: Document the new flag and -static-libubsan.
7821 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7822 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7823 * builtin-attrs.def (ATTR_COLD): Define.
7824 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7825 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7826 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7827 * flag-types.h (sanitize_code): New enum.
7828 * opts.c (common_handle_option): Parse command line arguments
7829 of -fsanitize=. Add -fsanitize=unreachable option.
7830 * varasm.c (get_variable_section): Adjust.
7831 (assemble_noswitch_variable): Likewise.
7832 (assemble_variable): Likewise.
7833 (output_constant_def_contents): Likewise.
7834 (categorize_decl_for_section): Likewise.
7835 (place_block_symbol): Likewise.
7836 (output_object_block): Likewise.
7837 * builtins.def: Likewise.
7838 * toplev.c (compile_file): Likewise.
7839 (process_options): Likewise.
7840 * cppbuiltin.c: Likewise.
7841 * tsan.c (tsan_pass): Likewise.
7842 (tsan_gate): Likewise.
7843 (tsan_gate_O0): Likewise.
7844 * cfgexpand.c (partition_stack_vars): Likewise.
7845 (expand_stack_vars): Likewise.
7846 (defer_stack_allocation): Likewise.
7847 (expand_used_vars): Likewise.
7848 * cfgcleanup.c (old_insns_match_p): Likewise.
7849 * asan.c (asan_finish_file): Likewise.
7850 (asan_instrument): Likewise.
7851 (gate_asan): Likewise.
7852 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7853 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7854 (asan_global_struct): Use pointer_sized_int_node instead
7855 calling build_nonstandard_integer_type.
7856 (initialize_sanitizer_builtins): Likewise.
7857 (asan_finish_file): Likewise.
7858 * gcc.c: Document %{%:function(args):X}.
7859 (static_spec_functions): Add sanitize.
7860 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7861 store funcval != NULL there.
7862 (do_spec_1): Adjust handle_spec_function caller.
7863 (handle_braces): Allow %:function(args) as condition.
7864 (sanitize_spec_function): New function.
7865 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7866 (LIBUBSAN_SPEC): Likewise.
7867 (LIBUBSAN_EARLY_SPEC): Likewise.
7868 (SANITIZER_SPEC): Handle libubsan.
7869 (SANITIZER_EARLY_SPEC): Likewise.
7870 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7871 instead of fsanitize=address.
7872 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7873 instead of fsanitize=address*.
7874 * builtins.c: Include ubsan.h.
7875 (fold_builtin_0): Instrument __builtin_unreachable.
7876 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7877 instead of flag_asan.
7878 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7879 (pointer_sized_int_node): Define.
7880 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7882 2013-08-30 Mike Stump <mikestump@comcast.net>
7884 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7887 2013-08-29 Jan Hubicka <jh@suse.cz>
7889 * cgraph.c (cgraph_function_body_availability): Handle weakref
7891 * passes.def: Remove pass_fixup_cfg.
7892 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7893 track when we need to remove functions.
7894 (gate_ipa_inline): Execute inlining always; add comment why.
7895 (pass_data_ipa_inline): Remove TODO_remove_functions.
7896 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7897 do not produce summaries.
7898 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7899 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7900 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7901 constant pool and vtable.
7903 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7905 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7906 New arm_neon.h's internal macros to specify 64-bit constants.
7907 Avoid using stdint.h's macros.
7909 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7911 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7912 before applying REG_P.
7914 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7916 PR tree-optimization/58277
7917 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7918 after seeing too many stmts with vdef in between dombb and current
7919 bb, invalidate everything.
7921 2013-08-30 Richard Biener <rguenther@suse.de>
7923 * fold-const.c (fold_single_bit_test): Fix overflow test.
7925 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7927 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7928 and which can live in a register, always retrieve the value on entry.
7929 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7930 passed by invisible reference specially.
7931 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7932 (vt_add_function_parameter): Correctly deal with a parameter passed by
7933 invisible reference.
7935 2013-08-30 Jan Hubicka <jh@suse.cz>
7937 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7939 2013-08-30 Richard Biener <rguenther@suse.de>
7941 PR tree-optimization/58228
7942 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7943 allow invariant loads in nested loop vectorization.
7945 2013-08-30 Richard Biener <rguenther@suse.de>
7947 PR tree-optimization/58223
7948 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7949 (has_anti_or_output_dependence): ... this and adjust to also
7950 look for output dependences.
7951 (mark_nodes_having_upstream_mem_writes): Adjust.
7952 (rdg_flag_uses): Likewise.
7954 2013-08-30 Richard Biener <rguenther@suse.de>
7956 PR tree-optimization/58010
7957 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7958 assert that we have a loop-closed PHI.
7960 2013-08-29 Jan Hubicka <jh@suse.cz>
7962 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7963 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7964 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7965 (lto_free_function_in_decl_state_for_node): New function.
7967 2013-08-29 Xinliang David Li <davidxl@google.com>
7969 * loop-unroll.c (report_unroll_peel): Minor message change.
7970 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7971 Emit alignment peeling message with default -fopt-info.
7972 (vect_loop_versioning): Emit loop version info message.
7973 * tree-vectorizer.c (vectorize_loops): Minor message change.
7974 (execute_vect_slp): Ditto.
7976 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7978 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7979 of the clone from the DECL_NAME of the original function.
7981 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7983 * passes.c (register_pass): Add overload.
7984 * tree-pass.h (register_pass): Forward declare it. Add comment.
7986 2013-08-29 Jan Hubicka <jh@suse.cz>
7988 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7989 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7990 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7991 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7992 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7993 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7994 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7995 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7997 2013-08-29 Teresa Johnson <tejohnson@google.com>
7999 * dumpfile.c (dump_loc): Output column number.
8000 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8001 * doc/invoke.texi: Document optall -fopt-info flag.
8002 * profile.c (read_profile_edge_counts): Use new dump framework.
8003 (compute_branch_probabilities): Ditto.
8004 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8005 when pass not in any opt group.
8006 * pass_manager.h (pass_manager::get_pass_profile): New method.
8007 * value-prof.c (check_counter): Use new dump framework.
8008 (check_ic_target): Ditto.
8009 * coverage.c (get_coverage_counts): Ditto.
8010 (coverage_init): Setup new dump framework.
8012 2013-08-29 Richard Biener <rguenther@suse.de>
8014 PR tree-optimization/58246
8015 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8016 handle the dominance check inside a basic-block.
8018 2013-08-29 Richard Biener <rguenther@suse.de>
8021 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8022 of default defs that appear in abnormal PHI nodes.
8024 2013-08-29 Richard Biener <rguenther@suse.de>
8026 PR tree-optimization/57685
8027 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8028 single-use operands to avoid exponential complexity.
8030 2013-08-28 Dehao Chen <dehao@google.com>
8032 * ipa-inline.c (edge_badness): Fix integer underflow.
8034 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8036 * gtm-builtins.def (_ITM_free): Declare leaf.
8038 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8041 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8042 (*tls_local_dynamic_base_64_largepic): Likewise.
8043 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8044 Remove predicate from call operand.
8045 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8046 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8048 2013-08-28 Jeff Law <law@redhat.com>
8050 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8051 checks for the number of predecessors and successors allowed.
8052 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8053 which require copying a joiner block if there is a request which
8054 is a subpath that requires no joiner block copying.
8056 2013-08-28 Jan Hubicka <jh@suse.cz>
8058 * lto-streamer-out.c (DFS_write_tree_body): Drop
8059 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8060 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8061 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8062 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8063 Do not read DECL_ERROR_ISSUED.
8064 (unpack_ts_decl_with_vis_value_fields): Do not read
8066 (lto_input_ts_binfo_tree_pointers): Do not read
8067 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8068 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8069 write DECL_ERROR_ISSUED..
8070 (pack_ts_decl_with_vis_value_fields): Do not write
8072 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8073 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8074 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8075 * tree.h (tree_decl_common): Update comment.
8076 (DECL_ERROR_ISSUED): Remove.
8078 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8081 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8083 2013-08-28 Jan Hubicka <jh@suse.cz>
8085 * builtins.def (free): Declare leaf.
8087 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8090 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8091 * configure: Regenerate.
8093 2013-08-27 Martin Jambor <mjambor@suse.cz>
8095 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8096 (ipa_ancestor_jf_data): Likewise.
8097 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8098 (ipa_get_jf_pass_through_type_preserved): New function.
8099 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8100 (ipa_get_jf_ancestor_type_preserved): New function.
8101 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8102 (ipa_get_jf_ancestor_result): Likewise.
8103 (propagate_vals_accross_pass_through): Use
8104 ipa_get_jf_pass_through_result to do all the value mappings.
8105 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8106 type_preserved flag.
8107 (ipa_set_jf_cst_copy): New function.
8108 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8109 (ipa_set_jf_arith_pass_through): Likewise.
8110 (ipa_set_ancestor_jf): Likewise.
8111 (compute_complex_assign_jump_func): Set type_preserved instead of
8113 (ipa_compute_jump_functions_for_edge): Likewise.
8114 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8115 (update_jump_functions_after_inlining): Update type_preserved.
8116 Explicitely create jump functions when combining one with pass_through.
8117 (ipa_write_jump_function): Stream the type_preserved flags.
8118 (ipa_read_jump_function): Likewise.
8120 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8121 Aldy Hernandez <aldyh@redhat.com>
8123 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8124 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8125 * function.h (struct function): Add has_force_vect_loops and
8127 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8128 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8130 * gimple.def: Update CLAUSES comments.
8131 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8132 (gimple_build_omp_for): Add argument to prototype.
8133 (gimple_omp_for_kind): New.
8134 (gimple_omp_for_set_kind): New.
8135 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8136 GOVD_DATA_SHARE_CLASS.
8137 (enum omp_region_type): Add ORT_SIMD.
8138 (gimple_add_tmp_var): Handle ORT_SIMD.
8139 (gimplify_var_or_parm_decl): Same.
8140 (is_gimple_stmt): Same.
8141 (omp_firstprivatize_variable): Same.
8142 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8143 (omp_notice_variable): Handle ORT_SIMD.
8144 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8145 (omp_check_private): Handle ORT_SIMD.
8146 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8148 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8149 Handle OMP_CLAUSE_LASTPRIVATE.
8150 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8152 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8153 (gimplify_expr): Handle OMP_SIMD.
8154 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8155 (expand_GOMP_SIMD_VF): New.
8156 (expand_GOMP_SIMD_LAST_LANE): New.
8157 * internal-fn.def (GOMP_SIMD_LANE): New.
8158 (GOMP_SIMD_VF): New.
8159 (GOMP_SIMD_LAST_LANE): New.
8160 * omp-low.c: Include target.h.
8161 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8163 (check_omp_nesting_restrictions): Same.
8165 (lower_rec_simd_input_clauses): New.
8166 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8168 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8169 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8170 (expand_omp_build_assign): New.
8171 (expand_omp_for_init_counts): New.
8172 (expand_omp_for_init_vars): New.
8173 (extract_omp_for_update_vars): New.
8174 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8175 and rewrite accordingly.
8176 (expand_omp_simd): New.
8177 (expand_omp_for): Use expand_omp_simd.
8178 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8179 (lower_omp_for): Do not lower the body.
8180 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8182 * tree-flow.h (find_omp_clause): Remove prototype.
8183 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8184 forcing vectorization of the loop, or if flag_tree_vectorize.
8185 (gate_tree_if_conversion): Similarly.
8186 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8187 gimple_build_omp_for.
8188 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8189 * tree-parloops (create_parallel_loop): Pass kind argument to
8190 gimple_build_omp_for.
8191 * tree-pretty-print.c (dump_omp_clause): Add cases for
8192 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8193 OMP_CLAUSE__SIMDUID_.
8194 (dump_generic_node): Handle OMP_SIMD.
8195 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8196 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8197 unroll OMP_SIMD loops here.
8198 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8199 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8201 (vect_analyze_data_refs): Handle simd loops.
8202 * tree-vect-loop.c (vectorizable_live_operation): Handle
8204 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8205 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8206 (vectorizable_load): Same.
8207 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8208 (struct simduid_to_vf): New.
8209 (simduid_to_vf::hash): New.
8210 (simduid_to-vf::equal): New.
8211 (struct simd_array_to_simduid): New.
8212 (simd_array_to_simduid::hash): New.
8213 (simd_array_to_simduid::equal): New.
8214 (adjust_simduid_builtins): New.
8215 (struct note_simd_array_uses_struct): New.
8216 (note_simd_array_uses_cb): New.
8217 (note_simd_array_uses): New.
8218 (vectorize_loops): Handle simd hints and adjust simd builtins
8220 * tree-vectorizer.h (struct _stmt_vec_info): Add
8221 simd_lane_access_p field.
8222 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8223 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8224 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8225 (omp_clause_code_name): Same.
8226 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8227 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8228 * tree.def (OMP_SIMD): New entry.
8229 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8230 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8231 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8232 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8233 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8234 (OMP_CLAUSE_LINEAR_STEP): New.
8235 (OMP_CLAUSE_SAFELEN_EXPR): New.
8236 (OMP_CLAUSE__SIMDUID__DECL): New.
8237 (find_omp_clause): New prototype.
8239 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8241 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8242 Haswell processor detection.
8244 2013-08-27 Christian Widmer <shadow@umbrox.de>
8247 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8248 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8249 AVX2 capable processors.
8251 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8253 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8254 of vget_low_* with implementations in terms of other intrinsics.
8256 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8259 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8260 values to -1, 0 and 1.
8262 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8264 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8265 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8266 (<optab><mode>3_insn): Remove.
8267 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8268 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8269 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8270 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8271 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8272 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8273 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8274 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8275 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8276 (ror<mode>3_insn): Likewise.
8277 * config/aarch64/predicates.md (aarch64_simd_register): New.
8279 2013-08-27 Richard Biener <rguenther@suse.de>
8281 PR tree-optimization/57521
8282 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8283 one edge is non-critical.
8284 (find_phi_replacement_condition): Make sure to use a non-critical
8285 edge. Cleanup and remove old bug workarounds.
8286 (bb_postdominates_preds): Remove.
8287 (if_convertible_loop_p_1): Do not compute post-dominators.
8288 (combine_blocks): Do not free post-dominators.
8289 (main_tree_if_conversion): Likewise.
8290 (pass_data_if_conversion): Add TODO_verify_ssa.
8292 2013-08-27 DJ Delorie <dj@redhat.com>
8294 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8296 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8298 * function.c (assign_parm_find_data_types): Set passed_mode and
8299 nominal_mode to the TYPE_MODE of nominal_type for the built
8300 pointer type in case of the struct-pass-by-reference.
8302 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8304 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8305 (UINT16_TYPE): Change default to "unsigned int".
8307 * config/avr/avr.opt (mfract-convert-truncate): New option.
8308 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8309 is set, round negative fractional integers according to n1169
8310 when converting to integer types.
8312 2013-08-26 Jan Hubicka <jh@suse.cz>
8314 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8315 methods can not be called indirectly when their address is not taken.
8317 2013-08-26 Jan Hubicka <jh@suse.cz>
8319 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8322 2013-08-26 Jan Hubicka <jh@suse.cz>
8324 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8327 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8329 * reload.c (find_valid_class): Allow classes that do not include
8330 FIRST_PSEUDO_REGISTER - 1.
8332 2013-08-26 Jan Hubicka <jh@suse.cz>
8334 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8335 fix edge count/frequency when speculation failed; fix type check
8336 for the direct call.
8338 2013-08-26 Jan Hubicka <jh@suse.cz>
8340 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8342 2013-08-26 Jan Hubicka <jh@suse.cz>
8344 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8345 to be changed by edge redirection.
8347 2013-08-26 Jan Hubicka <jh@suse.cz>
8349 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8350 formating; add sanity check.
8351 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8352 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8353 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8354 (resolve_noninline_speculation): Update callee keys, too.
8356 2013-08-26 Jan Hubicka <jh@suse.cz>
8358 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8359 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8361 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8363 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8366 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8368 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8369 (x86_64_elf_select_section): Ditto.
8370 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8372 (ix86_pass_by_reference): Ditto.
8373 (output_set_got): Ditto.
8374 (ix86_unary_operator_ok): Ditto.
8375 (ix86_expand_builtin): Ditto.
8377 2013-08-23 Jan Hubicka <jh@suse.cz>
8379 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8381 2013-08-23 Jan Hubicka <jh@suse.cz>
8383 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8384 (tree_decl_with_vis): Add FINAL field.
8386 2013-08-23 Jeff Law <law@redhat.com>
8388 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8389 the debugging dump when the expression is fully redundant.
8391 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8393 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8394 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8395 * pretty-print.c (pp_formatted_text_data): Likewise.
8396 (pp_write_text_to_stream): Likewise.
8397 (pp_write_text_as_dot_label_to_stream): Likewise.
8398 (pp_append_r): Likewise.
8399 (pp_format): Likewise.
8400 (pp_flush): Likewise.
8401 (pp_clear_output_area): Likewise.
8402 (pp_append_text): Likewise.
8403 (pp_formatted_text): Likewise.
8404 (pp_remaining_character_count_for_line): Likewise.
8405 (pp_newline): Likewise.
8406 (pp_character): Likewise.
8407 (output_buffer::~output_buffer): Define.
8408 (pretty_printer::~pretty_printer): Destruct output buffer.
8409 * pretty-print.h (output_buffer::~output_buffer): Declare.
8410 (pretty_printer::~pretty_printer): Declare virtual.
8412 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8415 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8416 HOST_WIDE_INT_M1U): New macros.
8417 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8418 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8419 unsigned -1 for lshift.
8420 * cse.c (cse_insn): Likewise.
8421 * double-int.c (rshift_double, lshift_double): Likewise.
8422 * builtins.c (fold_builtin_bitop): Likewise.
8423 * combine.c (force_to_mode): Likewise.
8424 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8425 * simplify-rtx.c (simplify_const_unary_operation,
8426 simplify_const_binary_operation): Likewise.
8427 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8428 check_va_list_escapes): Likewise.
8429 * rtlanal.c (nonzero_bits1): Likewise.
8430 * expmed.c (expand_smod_pow2): Likewise.
8431 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8433 2013-08-23 Jan Hubicka <jh@suse.cz>
8435 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8436 as having address taken.
8438 2013-08-23 Jan Hubicka <jh@suse.cz>
8440 * ipa-utils.h (method_class_type): Declare.
8441 * ipa-devirt.c (method_class_type): Export.
8443 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8444 do not walk base classes of anonymous types.
8446 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8448 PR rtl-optimization/58220
8450 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8451 handle SEQUENCE insns properly.
8453 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8455 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8457 (pp_newline_and_indent): Likewise.
8458 (pp_separate_with): Likewise.
8459 * pretty-print.c (pp_newline_and_flush): Define.
8460 (pp_newline_and_indent): Likewise.
8461 (pp_separate_with): Likewise.
8463 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8466 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8467 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8469 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8471 * config/i386/predicates.md (ext_sse_reg_operand): New.
8472 * config/i386/i386.md (*movti_internal): Use
8473 predicate to determine if EVEX is needed.
8474 (*movsi_internal): Ditto.
8475 (*movdf_internal): Ditto.
8476 (*movsf_internal): Ditto.
8477 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8479 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8481 PR tree-optimization/58209
8482 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8483 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8484 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8485 emit POINTER_PLUS_EXPR.
8486 (create_tailcall_accumulator): For pointer result type accumulate in
8489 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8491 * configure.ac: Add backslashes missing from the last change.
8492 * configure: Regenerate.
8494 2013-08-22 Jan Hubicka <jh@susue.cz>
8496 * ipa.c (function_and_variable_visibility): First remember function
8497 was global and then make it local.
8499 2013-08-22 Julian Brown <julian@codesourcery.com>
8501 * configure.ac: Add aarch64 to list of arches which use "nop" in
8503 * configure: Regenerate.
8505 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8507 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8508 gnu_libc_has_function.
8509 * config/s390/tpf.h: Likewise.
8511 2013-08-22 Jan Hubicka <jh@susue.cz>
8513 * timevar.c (validate_phases): Add cast.
8515 2013-08-22 Jan Hubicka <jh@susue.cz>
8517 * timevar.c (validate_phases): Use size_t for memory.
8518 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8520 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8522 * pretty-print.h (output_buffer::output_buffer): Declare.
8523 (pretty_printer::pretty_printer): Likewise.
8524 (pp_construct): Remove.
8525 * pretty-print.c (output_buffer::output_buffer): Define.
8526 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8527 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8528 (print_gimple_expr): Likewise.
8529 (print_gimple_seq): Likewise.
8530 (gimple_dump_bb): Likewise.
8531 * sched-vis.c (dump_value_slim): Likewise.
8532 (dump_insn_slim): Likewise.
8533 (dump_rtl_slim): Likewise.
8534 (str_pattern_slim): Likewise.
8535 * tree-mudflap.c (mf_varname_tree): Likewise.
8536 * graph.c (print_graph_cfg): Likewise.
8537 (start_graph_dump): Likewise.
8538 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8540 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8541 pretty printer initialization.
8542 * coretypes.h (diagnostic_context): Remove superflous type alias
8544 (pretty_printer): Likewise. Declare directly as a class.
8545 (pretty_print_info): Remove declaration as class.
8546 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8547 and pp_clear_output_area.
8548 (asan_add_global): Likewise.
8550 2013-08-22 Jan Hubicka <jh@suse.cz>
8552 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8553 * ipa-utils.h (update_type_inheritance_graph): Declare.
8554 (possible_polymorphic_call_target_p): Declare.
8555 (possible_polymorphic_call_target_p): New.
8556 * ipa-devirt.c: Update toplevel comments.
8557 (cached_polymorphic_call_targets): Move up.
8558 (odr_type_d): Move ID down.
8559 (polymorphic_type_binfo_p): Update comment.
8560 (odr_hasher::remove): Likewise;
8561 (get_odr_type): Set anonymous_namespace.
8562 (dump_odr_type): Dump it.
8563 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8564 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8565 (record_binfo): Add comment.
8566 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8568 (devirt_node_removal_hook): Do not iCE when cache is freed.
8569 (possible_polymorphic_call_target_p): New predicate.
8570 (update_type_inheritance_graph): New function.
8572 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8573 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8574 Sergey Lega <sergey.s.lega@intel.com>
8575 Anna Tikhonova <anna.tikhonova@intel.com>
8576 Ilya Tocar <ilya.tocar@intel.com>
8577 Andrey Turetskiy <andrey.turetskiy@intel.com>
8578 Ilya Verbin <ilya.verbin@intel.com>
8579 Kirill Yukhin <kirill.yukhin@intel.com>
8580 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8582 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8583 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8584 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8585 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8586 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8587 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8588 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8589 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8590 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8591 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8592 OPT_mavx512pf, OPT_mavx512er cases.
8593 * config/i386/constraints.md (v): New constraint.
8594 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8595 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8596 (bit_AVX512CD): New.
8597 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8598 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8599 * config/i386/i386-c.c (ix86_target_macros_internal):
8600 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8602 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8603 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8604 * config/i386/i386.c (regclass_map, dbx_register_map)
8605 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8606 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8607 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8609 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8610 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8611 -mavx512pf options. Fix formatting.
8612 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8613 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8614 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8615 -mavx512cd, -mavx512pf options.
8616 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8617 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8618 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8620 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8621 (ix86_set_reg_reg_cost): Ditto.
8622 (x86_order_regs_for_local_alloc): Ditto.
8623 (MAX_VECT_LEN): Extend to 64-byte.
8624 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8625 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8626 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8627 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8628 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8629 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8630 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8631 (SSE_REG_MODE_P): Support new modes.
8632 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8633 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8634 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8635 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8636 (REG_CLASS_CONTENTS): Add new registers.
8637 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8638 (EXT_REX_SSE_REGNO_P): New.
8639 (HI_REGISTER_NAMES): Add new registers.
8640 * config/i386/i386.md: Define constants for new registers.
8641 (mode): Add new 512-bit modes.
8642 (prefix): Support evex prefix.
8643 (isa): Support avx512f, noavx512f, fma_avx512f.
8644 (ssemodesuffix): Add new 512-bit modes.
8646 (*movxi_internal_avx512f): Ditto.
8647 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8649 (*movsi_internal): Likewise.
8650 (*movdf_internal): Likewise.
8651 (*movsf_internal): Likewise.
8652 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8653 (<code><mode>3): Likewise.
8654 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8656 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8657 with the new constraint "v".
8658 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8660 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8661 with the new constraint "v".
8662 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8663 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8664 (*<plusminus_insn><mode>3): Likewise.
8665 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8666 (*mul<mode>3): Likewise.
8667 (<sse>_vmmul<mode>3): Likewise.
8668 (<sse>_div<mode>3): Likewise.
8669 (<sse>_vmdiv<mode>3): Likewise.
8670 (<sse>_sqrt<mode>2): Likewise.
8671 (<sse>_vmsqrt<mode>2): Likewise.
8672 (*<code><mode>3_finite): Likewise.
8673 (*<code><mode>3) <smaxmin>: Likewise.
8674 (<sse>_vm<code><mode>3): Likewise.
8675 (*<code><mode>3) <any_logic>: Likewise.
8676 (*fma_fmadd_<mode>): Likewise.
8677 (*fma_fmsub_<mode>): Likewise.
8678 (*fma_fnmadd_<mode>): Likewise.
8679 (*fma_fnmsub_<mode>): Likewise.
8680 (*fma_fmaddsub_<mode>): Likewise.
8681 (*fma_fmsubadd_<mode>): Likewise.
8682 (*fmai_fmadd_<mode>): Likewise.
8683 (*fmai_fmsub_<mode>): Likewise.
8684 (*fmai_fnmadd_<mode>): Likewise.
8685 (*fmai_fnmsub_<mode>): Likewise.
8686 (sse_cvtsi2ss): Likewise.
8687 (sse_cvtsi2ssq): Likewise.
8688 (sse_cvtss2si): Likewise.
8689 (sse_cvtss2si_2): Likewise.
8690 (sse_cvtss2siq): Likewise.
8691 (sse_cvtss2siq_2): Likewise.
8692 (sse_cvttss2si): Likewise.
8693 (sse_cvtss2siq_2): Likewise.
8694 (float<sseintvecmodelower><mode>2): Likewise.
8695 (sse2_cvtsd2si_2): Likewise.
8696 (sse2_cvtsd2siq_2): Likewise.
8697 (*<plusminus_insn><mode>3): Likewise.
8698 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8699 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8700 (ashr<mode>3): Likewise.
8701 (<shift_insn><mode>3): Likewise.
8702 (avx2_<code><mode>3): Likewise.
8703 (*avx2_<code><mode>3): Likewise.
8704 (*andnot<mode>3): Likewise.
8705 (*<code><mode>3) <any_logic>: Likewise.
8706 (abs<mode>2): Likewise.
8707 (avx2_permvar<mode>): Likewise.
8708 (avx2_perm<mode>_1): Likewise.
8709 (*avx_vpermilp<mode>): Likewise.
8710 (avx_vpermilvar<mode>3): Likewise.
8711 (avx2_ashrv<mode>): Likewise.
8712 (avx2_<shift_insn>v<mode>): Likewise.
8713 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8715 * doc/rtl.texi: Document XImode.
8717 2013-08-21 Jeff Law <law@redhat.com>
8719 * tree-flow.h (register_jump_thread): Pass vector of edges
8720 instead of each important edge.
8721 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8722 thread path into a vector and pass that to register_jump_thread.
8723 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8724 passed in edge vector to the current 3-edge form.
8727 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8729 * dce.c (fini_dce): Call df_analyze again just in case
8730 delete_unmarked_insns removed anything.
8732 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8734 * reload.h (struct reg_equivs): Rename to ..
8735 (struct reg_equivs_s): .. this.
8737 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8739 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8741 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8743 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8745 2013-08-21 Jeff Law <law@redhat.com>
8747 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8748 simplify assignments too. If the RHS collapses to a singleton
8749 range, then return the value for the range.
8751 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8753 * config/i386/sse.md (V16): Rename to...
8755 (mov<mode>): Update iterator name.
8756 (*mov<mode>_internal): Ditto.
8757 (push<mode>1): Ditto.
8758 (movmisalign<mode>): Ditto.
8760 2013-08-20 Jan Hubicka <jh@suse.cz>
8763 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8764 entry for direct edges.
8765 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8767 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8769 Revert my last two changes, r201865 and r201864:
8772 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8774 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8775 instances can own GC refs.
8777 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8778 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8779 (gcc::context::gt_pch_nx): Likewise.
8780 (gcc::context::gt_pch_nx): Likewise.
8781 * ggc.h (gt_ggc_mx <T>): New.
8782 (gt_pch_nx_with_op <T>): New.
8783 (gt_pch_nx <T>): New.
8784 * passes.c (opt_pass::gt_ggc_mx): New.
8785 (opt_pass::gt_pch_nx): New.
8786 (opt_pass::gt_pch_nx_with_op): New.
8787 (pass_manager::gt_ggc_mx): New.
8788 (pass_manager::gt_pch_nx): New.
8789 (pass_manager::gt_pch_nx_with_op): New.
8790 (pass_manager::operator new): Use
8791 ggc_internal_cleared_alloc_stat rather than xcalloc.
8792 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8793 (pass_manager::gt_ggc_mx): New.
8794 (pass_manager::gt_pch_nx): New.
8795 (pass_manager::gt_pch_nx_with_op): New.
8796 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8797 (opt_pass::operator new): New.
8798 (opt_pass::gt_ggc_mx): New.
8799 (opt_pass::gt_pch_nx): New.
8800 (opt_pass::gt_pch_nx_with_op): New.
8803 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8805 * Makefile.in (GTFILES): Add context.h.
8806 * context.c (gcc::context::operator new): New.
8807 (gcc::context::gt_ggc_mx): New.
8808 (gcc::context::gt_pch_nx): New.
8809 (gcc::context::gt_pch_nx): New.
8810 * context.h (gcc::context): Add GTY((user)) marking.
8811 (gcc::context::operator new): New.
8812 (gcc::context::gt_ggc_mx): New.
8813 (gcc::context::gt_pch_nx): New.
8814 (gcc::context::gt_pch_nx): New.
8815 (g): Add GTY marking.
8816 (gt_ggc_mx (gcc::context *)): New.
8817 (gt_pch_nx (gcc::context *)): New.
8818 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8819 void *cookie)): New.
8820 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8822 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8824 * dce.c (fini_dce): Call df_analyze again just in case
8825 delete_unmarked_insns removed anything.
8827 2013-08-20 Teresa Johnson <tejohnson@google.com>
8829 PR rtl-optimizations/57451
8830 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8831 from crossing split section boundaries.
8833 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8835 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8836 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8837 with MULTLIB_DEFAULTS.
8839 2013-08-20 Nick Clifton <nickc@redhat.com>
8841 * target.def (narrow_volatile_bitfield): Note that the default
8842 value is false, not !TARGET_STRICT_ALIGN.
8843 * doc/tm.texi: Regenerate.
8845 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8847 Fix LIB_SPEC for systems without libpthread.
8849 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8850 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8852 * config/i386/linux-common.h: Likewise.
8853 * config/mips/linux-common.h: Likewise.
8855 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8857 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8860 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8862 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8863 instances can own GC refs.
8865 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8866 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8867 (gcc::context::gt_pch_nx): Likewise.
8868 (gcc::context::gt_pch_nx): Likewise.
8869 * ggc.h (gt_ggc_mx <T>): New.
8870 (gt_pch_nx_with_op <T>): New.
8871 (gt_pch_nx <T>): New.
8872 * passes.c (opt_pass::gt_ggc_mx): New.
8873 (opt_pass::gt_pch_nx): New.
8874 (opt_pass::gt_pch_nx_with_op): New.
8875 (pass_manager::gt_ggc_mx): New.
8876 (pass_manager::gt_pch_nx): New.
8877 (pass_manager::gt_pch_nx_with_op): New.
8878 (pass_manager::operator new): Use
8879 ggc_internal_cleared_alloc_stat rather than xcalloc.
8880 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8881 (pass_manager::gt_ggc_mx): New.
8882 (pass_manager::gt_pch_nx): New.
8883 (pass_manager::gt_pch_nx_with_op): New.
8884 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8885 (opt_pass::operator new): New.
8886 (opt_pass::gt_ggc_mx): New.
8887 (opt_pass::gt_pch_nx): New.
8888 (opt_pass::gt_pch_nx_with_op): New.
8890 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8892 * Makefile.in (GTFILES): Add context.h.
8893 * context.c (gcc::context::operator new): New.
8894 (gcc::context::gt_ggc_mx): New.
8895 (gcc::context::gt_pch_nx): New.
8896 (gcc::context::gt_pch_nx): New.
8897 * context.h (gcc::context): Add GTY((user)) marking.
8898 (gcc::context::operator new): New.
8899 (gcc::context::gt_ggc_mx): New.
8900 (gcc::context::gt_pch_nx): New.
8901 (gcc::context::gt_pch_nx): New.
8902 (g): Add GTY marking.
8903 (gt_ggc_mx (gcc::context *)): New.
8904 (gt_pch_nx (gcc::context *)): New.
8905 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8906 void *cookie)): New.
8907 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8909 2013-08-20 Alan Modra <amodra@gmail.com>
8912 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8913 (rs6000_emit_epilogue): Likewise.
8915 2013-08-19 Dehao Chen <dehao@google.com>
8917 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8919 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8920 Jakub Jelinek <jakub@redhat.com>
8922 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8923 (BUILT_IN_FABSD64): Likewise.
8924 (BUILT_IN_FABSD128): Likewise.
8925 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8926 (fold_builtin_1): Likewise.
8927 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8928 destination and source operands.
8929 (*abstd2_fpr): Likewise.
8930 (*nabstd2_fpr): Likewise.
8932 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8934 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8937 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8939 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8942 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8944 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8945 * builtins.c (default_libc_has_function): New.
8946 (gnu_libc_has_function): Ditto.
8947 (no_c99_libc_has_function): Ditto.
8948 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8949 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8950 (fold_builtin_sincos): Likewise.
8951 (fold_builtin_cexp): Likewise.
8952 * builtins.def (DEF_C94_BUILTIN): Likewise.
8953 (DEF_C99_BUILTIN): Likewise.
8954 (DEF_C99_C90RES_BUILTIN): Likewise.
8955 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8956 definitions to using this define.
8957 * config/darwin-protos.h (darwin_libc_has_function): New.
8958 * config/darwin.c (darwin_libc_has_function): Ditto.
8959 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8960 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8961 * config/darwin.h: Ditto.
8962 * config/elfos.h: Ditto.
8963 * config/freebsd.h: Ditto.
8964 * config/i386/cygming.h: Ditto.
8965 * config/i386/djgpp.h: Ditto.
8966 * config/i386/i386-interix.h: Ditto.
8967 * config/microblaze/microblaze.h: Ditto.
8968 * config/mmix/mmix.h: Ditto.
8969 * config/gnu-user.h: Ditto.
8970 * config/ia64/hpux.h: Ditto.
8971 * config/pa/pa-hpux.h: Ditto.
8972 * config/pdp11/pdp11.h: Ditto.
8973 * config/picochip/picochip.h: Ditto.
8974 * config/linux.h: Ditto.
8975 * config/netbsd.h: Ditto.
8976 * config/openbsd.h: Ditto.
8977 * config/rs6000/aix43.h: Ditto.
8978 * config/rs6000/aix51.h: Ditto.
8979 * config/rs6000/aix52.h: Ditto.
8980 * config/rs6000/aix53.h: Ditto.
8981 * config/rs6000/aix61.h: Ditto.
8982 * config/rs6000/darwin.h: Ditto.
8983 * config/rs6000/linux.h: Ditto.
8984 * config/rs6000/linux64.h: Ditto.
8985 * config/s390/tpf.h: Ditto.
8986 * config/sol2-10.h: Ditto.
8987 * config/sol2.h: Ditto.
8988 * config/vms/vms.h: Ditto.
8989 * config/vxworks.h: Ditto.
8990 * config/linux-android.c (linux_android_libc_has_function):
8991 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8992 * config/linux-protos.h (linux_android_libc_has_function):
8994 * config/i386/i386.c (ix86_libc_has_function): New.
8995 * config/i386/i386-protos.h
8996 (ix86_libc_has_function): New declaration.
8997 * config/i386/i386.md
8998 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8999 ("isinf<mode>2): Likewise.
9000 * convert.c (convert_to_integer): Using new target hook
9001 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9002 TARGET_C99_FUNCTIONS.
9003 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9004 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9005 * coretypes.h (function_class): New enum for different
9006 classes of functions.
9007 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9008 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9009 (TARGET_HAS_SINCOS): Likewise.
9010 (TARGET_LIBC_HAS_FUNCTION): New.
9011 * doc/tm.texi: Regenerated.
9012 * targhooks.h (default_libc_has_function): New declaration.
9013 (no_c99_libc_has_function): Ditto.
9014 (gnu_libc_has_function): Ditto.
9015 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9016 and TARGET_HAS_SINCOS.
9018 2013-08-18 Jan Hubicka <jh@suse.cz>
9020 * Makeifle-in (ipa-devirt.o): New.
9021 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9022 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9023 (analyze_functions): Look into possible targets of polymorphic call.
9024 * dumpfile.c (dump_files): Add type-inheritance dump.
9025 * dumpfile.h (TDI_inheritance): New.
9026 * ipa-devirt.c: New file.
9027 * ipa-utils.h (odr_type_d): Forward declare.
9028 (odr_type): New type.
9029 (build_type_inheritance_graph): Declare.
9030 (possible_polymorphic_call_targets): Declare and introduce inline
9031 variant when only edge is pased.
9032 (dump_possible_polymorphic_call_targets): Likewise.
9033 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9034 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9035 (types_same_for_odr): ... here.
9036 * tree.h (type_in_anonymous_namespace_p): Declare.
9038 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9040 PR tree-optimization/58006
9041 * tree-parloops.c (take_address_of): Don't ICE if get_name
9043 (eliminate_local_variables_stmt): Remove clobber stmts.
9045 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9047 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9048 error message is issued for an alias to undefined symbol.
9050 2013-08-18 Jan Hubicka <jh@suse.cz>
9052 * cgraph.c (cgraph_create_indirect_edge): Discover
9053 polymorphic calls and record basic info into indirect_info.
9054 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9055 devirtualization, ignore objc function calls.
9056 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9057 call with no parm index info.
9058 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9059 * tree.c (virtual_method_call_p): New function.
9060 * tree.h (virtual_method_call_p): Declare.
9062 2013-08-16 Jan Hubicka <jh@suse.cz>
9065 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9067 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9069 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9072 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9074 * gengtype.c (type_for_name): Add special-case support for
9075 locating types within the "gcc::" namespace.
9076 (open_base_files): Emit a "using namespace gcc" directive.
9078 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9081 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9082 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9084 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9085 array instead of each individual operand as a separate argument.
9086 (emit_fusion_gpr_load): Likewise.
9087 (expand_fusion_gpr_load): Add new function declaration.
9089 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9090 signature to have the operands passed as an array, instead of as
9091 separate arguments. Allow ZERO_EXTEND to be in the memory
9092 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9093 depend on the register live/dead flags when peepholes are run.
9094 (expand_fusion_gpr_load): New function to be called from the
9095 peephole2 pass, to change the register that addis sets to be the
9097 (emit_fusion_gpr_load): Change the calling signature to have the
9098 operands passed as an array, instead of as separate arguments.
9099 Allow ZERO_EXTEND to be in the memory address, and also
9100 SIGN_EXTEND if -mpower8-fusion-sign.
9102 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9104 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9105 adjust the register addis loads up in the peephole2 pass. Do not
9106 depend on the register live/dead state when the peephole pass is done.
9108 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9110 * gengtype.c (create_user_defined_type): Ensure that the kind
9111 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9112 declaration is seen before the GTY((user)) marking.
9114 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9117 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9119 2013-08-16 Jan Hubicka <jh@suse.cz>
9121 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9122 arugment expected_type.
9123 (gimple_fold_call): Use it.
9124 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9125 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9126 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9127 (try_make_edge_direct_virtual_call): Likewise.
9128 * tree.c (obj_type_ref_class): New.
9129 * tree.h (obj_type_ref_class): Use it.
9131 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9133 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9134 (rtl_slim_pp): Likewise.
9135 (init_rtl_slim_pretty_print): Likewise.
9136 (dump_value_slim): Don't call it. Use local pretty printer.
9137 (dump_insn_slim): Likewise.
9138 (dump_rtl_slim): Likewise.
9139 (str_pattern_slim): Likewise.
9140 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9143 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9145 PR tree-optimization/58164
9146 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9147 walk gimple_goto_dest of GIMPLE_GOTO.
9149 PR tree-optimization/58165
9150 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9151 bi_call must be the last stmt in a bb, don't split_block, instead
9152 use fallthru edge from it and give up if there is none.
9153 Release conds vector when returning early.
9155 2013-08-14 Xinliang David Li <davidxl@google.com>
9157 * config/i386/i386.c (ix86_option_override_internal):
9158 Remove unused variable and field.
9160 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9163 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9164 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9165 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9166 and Linux, correct BLKmode alignment when 128-bit alignment is
9167 required and compatibility flag is not set.
9168 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9169 for zero-size arguments when compatibility flag is not set.
9171 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9173 PR tree-optimization/58145
9174 * tree-sra.c (build_ref_for_offset): If prev_base has
9175 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9177 2013-08-14 Xinliang David Li <davidxl@google.com>
9179 * config/i386/i386.c (ix86_option_override_internal):
9180 Fix uninitialized variable error.
9182 2013-08-14 Xinliang David Li <davidxl@google.com>
9184 * config/i386/i386.opt: Define two new options.
9185 * config/i386/x86-tune.def: Add arch selector field in macros.
9186 * config/i386/i386.h: Adjust macro definition.
9187 * config/i386/i386.c (ix86_option_override_internal):
9189 (parse_mtune_ctrl_str): New function.
9190 (set_ix86_tune_features): New function.
9191 (ix86_function_specific_restore): Call the new helper function.
9193 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9195 PR rtl-optimization/57662
9196 * sel-sched.c (code_motion_process_successors): When the current insn
9197 is removed after the recursive traversal, break from the loop.
9198 Add comments and debug printouts.
9200 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9201 Alexandre Oliva <aoliva@redhat.com>
9204 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9205 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9206 there also UNSPEC_PLTOFF.
9208 2013-08-14 Marek Polacek <polacek@redhat.com>
9210 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9211 NUM_CONDITIONS bit positions.
9213 2013-08-13 Cary Coutant <ccoutant@google.com>
9215 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9216 (attr_checksum): Hash vector contents instead of pointer.
9217 (attr_checksum_ordered): Likewise.
9219 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9221 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9222 when Pmode != word_mode. Add length_address attribute.
9223 (sse3_monitor_<mode>): Merge from sse3_monitor and
9224 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9225 Pmode != word_mode. Update insn length attribute.
9226 * config/i386/i386.c (ix86_option_override_internal): Update
9227 ix86_gen_monitor selection for merged sse3_monitor insn.
9229 2013-08-13 Julian Brown <julian@codesourcery.com>
9231 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9232 perform invalid legitimization on greater-than-word-size modes for
9235 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9237 * ira.c (setup_class_translate_array): Use aclass instead of cl
9238 for classes not fully covered by allocno classes.
9240 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9242 PR tree-optimization/57661
9243 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9244 * tree-inline.c (tree_function_versioning): Initialize it.
9245 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9246 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9247 that is not being copied.
9250 * asan.c (instrument_strlen_call): Fix typo in comment.
9251 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9253 2013-08-13 Steve Ellcey <sellcey@mips.com>
9255 * config/mips/mips.md (prefetch): Use lw instead of ld on
9256 loongson in 32bit mode.
9258 2013-08-13 Nick Clifton <nickc@redhat.com>
9260 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9262 2013-08-13 Jan Hubicka <jh@suse.cz>
9264 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9265 introduced edge; fix typo in sanity check.
9266 (cgraph_resolve_speculation): Export; improve diagnostic.
9267 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9268 speculation at type mismatch.
9269 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9270 (cgraph_resolve_speculation): Declare.
9271 (symtab_can_be_discarded): New function.
9272 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9273 * ipa-inline-transform.c (speculation_removed): New global var.
9274 (clone_inlined_nodes): See if speculation can be removed.
9275 (inline_call): If speculations was removed, we growths may not match.
9276 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9277 (speculation_useful_p): New function.
9278 (resolve_noninline_speculation): New function.
9279 (inline_small_functions): Resolve useless speculations.
9280 * ipa-inline.h (speculation_useful_p): Declare
9281 * ipa.c (can_replace_by_local_alias): Simplify.
9282 (ipa_profile): Produce speculative calls in non-lto, too;
9283 add simple cost model; produce local aliases.
9285 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9287 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9290 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9292 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9294 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9296 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9297 iterator instead of X87MODEF.
9299 2013-08-12 Perez Read <netfirewall@gmail.com>
9302 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9303 operand 0 for intel asm alternative.
9304 (*movabs<mode>_2): Ditto for operand 1.
9306 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9308 * config/aarch64/arm_none.h
9309 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9311 2013-08-12 Nick Clifton <nickc@redhat.com>
9313 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9315 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9317 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9318 expand for QI/HImode operand to produce more effictive code for
9319 unsigned char(short) --> float(double) conversion.
9321 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9323 * doc/invoke.texi: Mention that -ftls-model does not force the final
9326 2013-08-12 Marek Polacek <polacek@redhat.com>
9327 Marc Glisse <marc.glisse@inria.fr>
9329 PR tree-optimization/57980
9330 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9331 when creating -1 constant.
9333 2013-08-10 Jan Hubicka <jh@suse.cz>
9335 Workaround binutils PR14342.
9336 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9337 (gimple_init_edge_profiler): Likewise.
9338 (gimple_gen_ic_func_profiler): Likewise.
9340 2013-08-09 Jan Hubicka <jh@suse.cz>
9342 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9344 2013-08-09 Xinliang David Li <davidxl@google.com>
9346 * config/i386/stringop.def: New file.
9347 * config/i386/stringop.opt: New file.
9348 * config/i386/i386-opts.h: Include stringopt.def.
9349 * config/i386/i386.opt: Include stringopt.opt.
9350 * config/i386/i386.c (ix86_option_override_internal):
9351 Override default size based stringop inline strategies with options.
9352 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9355 2013-08-09 Jan Hubicka <jh@suse.cz>
9357 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9359 2013-08-09 Jan Hubicka <jh@suse.cz>
9361 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9363 (dump_cgraph_node): Dump profile-id.
9364 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9365 and common_target_probability.
9366 * lto-cgraph.c (lto_output_edge): Stream common targets.
9367 (lto_output_node): Stream profile ids.
9368 (input_node): Stream profile ids.
9369 (input_edge): Stream common targets.
9370 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9371 * ipa.c: Include value-prof.h
9372 (ipa_profile_generate_summary): Turn indirect call statement histograms
9373 into common targets.
9374 (ipa_profile): Turn common targets into speculative edges.
9376 2013-08-09 Jan Hubicka <jh@suse.cz>
9378 * cgraph.h (cgraph_node): Add profile_id.
9379 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9380 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9381 (del_node_map): Update.
9382 (find_func_by_funcdef_no): Replace by ...
9383 (find_func_by_profile_id): ... this one.
9384 (gimple_ic_transform): Do not remove useful histograms when
9385 speculation is not done; dump info when indirect call removal
9387 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9388 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9389 (__gcov_indirect_call_profiler_v2): .. this one.
9390 * profile.h (init_node_map): Update.
9391 * coverage.c (coverage_compute_profile_id): New function.
9392 * coverage.h (coverage_compute_profile_id): Declare.
9393 * tree-profile.c (init_ic_make_global_vars): Make
9394 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9395 (gimple_init_edge_profiler): Update prototype of
9396 __gcov_indirect_call_profiler.
9397 (gimple_gen_ic_func_profiler): Simplify.
9398 (tree_profiling): Use init_node_map
9400 2013-08-09 Jan Hubicka <jh@suse.cz>
9402 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9403 non-speculative refs.
9404 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9405 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9406 (cgraph_set_call_stmt): Likewise.
9407 (cgraph_create_edge_1): Fix release checking compilatoin;
9409 (cgraph_free_edge): Free indirect info.
9410 (cgraph_turn_edge_to_speculative): New function.
9411 (cgraph_speculative_call_info): New function.
9412 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9413 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9414 (dump_cgraph_node): Dump speculation.
9415 (verify_edge_count_and_frequency): Accept speculative edges.
9416 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9417 (verify_cgraph_node): Handle speculation.
9418 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9419 (cgraph_set_call_stmt): Update prototype.
9420 (cgraph_make_edge_direct): Update prototype.
9421 (cgraph_speculative_call_info): Declare.
9422 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9423 to change; update call of ipa_find_references.
9424 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9425 lto_stmt_uid and speculative flags.
9426 (ipa_dump_references): Dump speculation.
9427 (ipa_clone_references): Clone speculative flag.
9428 (ipa_clone_referring): Likewise.
9429 (ipa_clone_ref): New function.
9430 (ipa_find_reference): Look into lto_stmt_uids
9431 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9432 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9433 (ipa_find_reference): Update declaration.
9434 (ipa_clone_ref): Declare.
9435 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9436 stream speculative flag.
9437 (lto_output_ref): Stream statements uids and speculation.
9438 (input_ref): Likewise.
9439 (input_edge): Stream speuclation.
9440 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9441 (cgraph_set_call_stmt_including_clones): Handle speculation.
9442 * ipa-inline.c (heap_edge_removal_hook): New function.
9443 (inline_small_functions): Register it.
9444 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9445 also initialize refs.
9446 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9448 (try_make_edge_direct_simple_call): Likewise.
9449 (try_make_edge_direct_simple_call): Likewise.
9450 (update_indirect_edges_after_inlining): Likewise.
9451 (remove_described_reference): Look proper lto_stmt_uid.
9452 (propagate_controlled_uses): Likewise.
9453 (propagate_controlled_uses): Liekwise.
9454 * tree-inline.c (copy_bb): Copy speculative edges.
9455 (redirect_all_calls): New function.
9456 (copy_cfg_body): Do redirection after loop info is updated.
9457 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9459 2013-08-09 Jan Hubicka <jh@suse.cz>
9461 * lto-streamer-out.c (output_function): Renumber PHIs.
9462 * lto-streamer-in.c (input_function): Likewise.
9464 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9466 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9467 (get_lane_unsigned): Likewise.
9468 (dup_lane_scalar): Likewise.
9469 (get_lane): enable for VALL.
9470 * config/aarch64/aarch64-simd.md
9471 (aarch64_dup_lane_scalar<mode>): Remove.
9472 (aarch64_get_lane_signed<mode>): Likewise.
9473 (aarch64_get_lane_unsigned<mode>): Likewise.
9474 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9475 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9476 (aarch64_get_lane<mode>): Enable for all vector modes.
9477 (aarch64_get_lanedi): Remove misleading constraints.
9478 * config/aarch64/arm_neon.h
9479 (__aarch64_vget_lane_any): Define.
9480 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9481 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9482 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9483 * config/aarch64/iterators.md (VDQQH): New.
9487 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9489 * configure.ac: Add GAS check for LEON instructions on SPARC.
9490 * configure: Regenerate.
9491 * config.in: Likewise.
9492 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9494 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9495 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9497 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9498 (AS_LEON_FLAG): New macro.
9499 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9500 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9501 Deal with LEON and LEON3 for the memory model.
9502 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9503 (atomic_compare_and_swap<mode>_1): Likewise.
9504 (*atomic_compare_and_swap<mode>_1): Likewise.
9506 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9508 * config/arm/neon.md (vcond): Fix floating-point vector
9509 comparisons against 0.
9511 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9513 * lra-constraints.c (emit_spill_move): Remove assert.
9514 (process_alt_operands): Add more debugging
9515 output. Increase reject for spilling into memory. Decrease
9516 reject for reloading scratch.
9517 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9519 2013-08-08 Steve Ellcey <sellcey@mips.com>
9521 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9522 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9523 micromips incompatible. Add nan2008.
9524 (MULTILIB_DIRNAMES): Add nan2008.
9525 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9526 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9527 and micromips incompatible. Add nan2008.
9528 (MULTILIB_DIRNAMES): Add nan2008.
9529 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9531 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9533 PR rtl-optimization/58079
9534 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9535 simplify_comparison has widened a comparison with an integer.
9537 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9539 * config/arm/neon.md (movmisalign<mode>): Disable when we
9540 don't allow unaligned accesses.
9541 (*movmisalign<mode>_neon_store): Likewise.
9542 (*movmisalign<mode>_neon_load): Likewise.
9543 (*movmisalign<mode>_neon_store): Likewise.
9544 (*movmisalign<mode>_neon_load): Likewise.
9546 2013-08-08 Jan Hubicka <jh@suse.cz>
9548 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9549 (make_pass_rebuild_cgraph_edges): Also clear references.
9550 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9551 * ipa-inline-transform.c (inline_transform): Remove all references
9553 * cgraphunit.c (expand_function): Remove all references after
9555 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9556 (ipa_find_reference): Rewrite to iterator.
9557 (remove_stmt_references): Likewise.
9558 (ipa_clear_stmts_in_references): New function.
9559 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9560 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9562 * ipa-split.c (split_function): Remove references in split function.
9564 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9567 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9568 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9570 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9573 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9574 suggested mode for the assignment isn't compatible with the
9577 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9580 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9582 2013-08-07 Xinliang David Li <davidxl@google.com>
9584 * config/i386/i386.opt: New option -mtune-ctrl=.
9585 * config/i386/x86-tune.def: New file.
9586 * config/i386/i386.h: include x86-tune.def.
9587 * config/i386/i386.c (ix86_option_override_internal):
9588 Parsing -mtune-ctrl= option and set tune features.
9590 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9593 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9594 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9597 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9599 * rtl.h (update_alignments): Declare.
9600 * final.c (grow_label_align): New function extracted from...
9601 (shorten_branches): ...here. Call it.
9602 (update_alignments): New function.
9603 * reorg.c (sibling_labels): New variable.
9604 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9605 the new label along with it onto the sibling_labels vector.
9606 (fill_simple_delay_slots): Adjust call to get_label_before.
9607 (fill_slots_from_thread): Likewise.
9608 (relax_delay_slots): Likewise.
9609 (make_return_insns): Likewise.
9610 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9612 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9614 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9615 document its semantics.
9616 (diagnostic_report_diagnostic): Adjust accordingly.
9618 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9620 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9621 (sparc_option_override): ...and port to new C++ pass API.
9622 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9624 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9626 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9628 2013-08-06 Caroline Tice <cmtice@google.com>
9630 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9631 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9632 * tree-pass.h: Add pass_vtable_verify.
9633 * varasm.c (assemble_variable): Add code to properly set the comdat
9634 section and name for the .vtable_map_vars section.
9635 (assemble_vtyv_preinit_initializer): New function.
9636 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9638 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9639 * vtable-verify.c: New file.
9640 * vtable-verify.h: New file.
9641 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9642 initialiation levels.
9643 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9644 * passes.def: Insert pass_vtable_verify.
9645 * aclocal.m4: Reorder includes.
9646 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9647 -fvtv-counts options.
9648 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9649 as appropriate, if -fvtable-verify=... is used.
9650 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9651 -fvtable-verify=... is used.
9652 * Makefile.in (OBJS): Add vtable-verify.o to list.
9653 (vtable-verify.o): Add new build rule.
9654 (GTFILES): Add vtable-verify.c to list.
9655 * common.opt (fvtable-verify=): New flag.
9656 (vtv_priority): Values for fvtable-verify= flag.
9657 (fvtv-counts): New flag.
9658 (fvtv-debug): New flag.
9659 * tree.h (save_vtable_map_decl): New extern function decl.
9661 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9663 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9664 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9665 (pass_data_rl78_devirt): ...new pass_data instance and...
9666 (make_pass_rl78_devirt): ...new function.
9667 (rl78_asm_file_start): Port pass registration to new C++ API.
9669 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9671 * coretypes.h (rtl_opt_pass): Add.
9672 (gcc::context): Add.
9673 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9674 (epiphany_init): Port to new C++ pass API.
9675 (epiphany_optimize_mode_switching): Likewise.
9676 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9677 (pass_manager::get_pass_mode_switching): New.
9678 (pass_manager::get_pass_peephole2): New.
9679 * mode-switching.c (pass_mode_switching): Add clone method.
9680 * recog.c (pass_peephole2): Add clone method.
9681 (pass_split_all_insns): Add clone method.
9683 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9685 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9686 (mips_option_override): ...here, porting to new C++ API for passes.
9688 2013-08-06 Jan Hubicka <jh@suse.cz>
9690 * cgraph.c (cgraph_get_body): New function based on lto.c
9692 * cgraph.h (cgraph_get_body): Declare.
9693 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9695 * cgraphunit.c (expand_function): Get body prior expanding.
9696 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9697 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9699 * passes.c (do_per_function_toporder): Get body.
9700 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9701 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9703 2013-08-06 Martin Jambor <mjambor@suse.cz>
9706 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9707 is not re-finalized. Rename second parameter to no_collect.
9709 2013-08-06 Martin Jambor <mjambor@suse.cz>
9712 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9713 MEM_REF has proper alignment information.
9715 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9718 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9720 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9721 rtx (*) (rtx, ...) with insn_gen_fn.
9722 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9723 insn_gen_fn::stored_funcptr. Add initializer braces.
9725 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9727 Rewrite how instances of passes are cloned to remove assumptions
9728 about their sizes (thus allowing pass subclasses to have
9729 additional data fields, albeit non-GC-managed ones at this point).
9731 * passes.c (make_pass_instance): Now that passes have clone
9732 methods, rewrite this function to eliminate XNEW and memcpy
9733 calls that used hardcoded sizes. Since this function no longer
9734 creates pass instances, rename it to...
9735 (add_pass_instance): ...this. Document the old way that passes were
9736 numbered and flagged, and rework this function to continue using it.
9737 (next_pass_1): Add an initial_pass argument for use by
9739 (position_pass): When adding multiple instances of a pass, use
9740 the pass's clone method, rather than relying on the XNEW/memcpy
9741 within the former make_pass_instance (now add_pass_instance).
9742 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9743 the initial instance of the current pass within the pass manager.
9745 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9747 This is the automated part of the conversion of passes from C
9748 structs to C++ classes.
9750 Patch autogenerated by refactor_passes.py from
9751 https://github.com/davidmalcolm/gcc-refactoring-scripts
9752 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9754 * asan.c (pass_asan): Convert from a global struct to a subclass of
9755 gimple_opt_pass along with...
9756 (pass_data_asan): ...new pass_data instance and...
9757 (make_pass_asan): ...new function.
9758 (pass_asan_O0): Convert from a global struct to a subclass of
9759 gimple_opt_pass along with...
9760 (pass_data_asan_O0): ...new pass_data instance and...
9761 (make_pass_asan_O0): ...new function.
9762 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9763 subclass of rtl_opt_pass along with...
9764 (pass_data_inc_dec): ...new pass_data instance and...
9765 (make_pass_inc_dec): ...new function.
9766 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9767 a subclass of rtl_opt_pass along with...
9768 (pass_data_reorder_blocks): ...new pass_data instance and...
9769 (make_pass_reorder_blocks): ...new function.
9770 (pass_duplicate_computed_gotos): Convert from a global struct to a
9771 subclass of rtl_opt_pass along with...
9772 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9773 (make_pass_duplicate_computed_gotos): ...new function.
9774 (pass_partition_blocks): Convert from a global struct to a subclass of
9775 rtl_opt_pass along with...
9776 (pass_data_partition_blocks): ...new pass_data instance and...
9777 (make_pass_partition_blocks): ...new function.
9778 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9779 struct to a subclass of rtl_opt_pass along with...
9780 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9782 (make_pass_branch_target_load_optimize1): ...new function.
9783 (pass_branch_target_load_optimize2): Convert from a global struct to a
9784 subclass of rtl_opt_pass along with...
9785 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9787 (make_pass_branch_target_load_optimize2): ...new function.
9788 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9789 of rtl_opt_pass along with...
9790 (pass_data_jump): ...new pass_data instance and...
9791 (make_pass_jump): ...new function.
9792 (pass_jump2): Convert from a global struct to a subclass of
9793 rtl_opt_pass along with...
9794 (pass_data_jump2): ...new pass_data instance and...
9795 (make_pass_jump2): ...new function.
9796 * cfgexpand.c (pass_expand): Convert from a global struct to a
9797 subclass of rtl_opt_pass along with...
9798 (pass_data_expand): ...new pass_data instance and...
9799 (make_pass_expand): ...new function.
9800 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9801 of rtl_opt_pass along with...
9802 (pass_data_free_cfg): ...new pass_data instance and...
9803 (make_pass_free_cfg): ...new function.
9804 (pass_into_cfg_layout_mode): Convert from a global struct to a
9805 subclass of rtl_opt_pass along with...
9806 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9807 (make_pass_into_cfg_layout_mode): ...new function.
9808 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9809 subclass of rtl_opt_pass along with...
9810 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9811 (make_pass_outof_cfg_layout_mode): ...new function.
9812 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9813 struct to a subclass of gimple_opt_pass along with...
9814 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9815 (make_pass_build_cgraph_edges): ...new function.
9816 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9817 subclass of gimple_opt_pass along with...
9818 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9819 (make_pass_rebuild_cgraph_edges): ...new function.
9820 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9821 subclass of gimple_opt_pass along with...
9822 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9824 (make_pass_remove_cgraph_callee_edges): ...new function.
9825 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9826 struct to a subclass of rtl_opt_pass along with...
9827 (pass_data_stack_adjustments): ...new pass_data instance and...
9828 (make_pass_stack_adjustments): ...new function.
9829 * combine.c (pass_combine): Convert from a global struct to a subclass
9830 of rtl_opt_pass along with...
9831 (pass_data_combine): ...new pass_data instance and...
9832 (make_pass_combine): ...new function.
9833 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9834 global struct to a subclass of rtl_opt_pass along with...
9835 (pass_data_compare_elim_after_reload): ...new pass_data instance
9837 (make_pass_compare_elim_after_reload): ...new function.
9838 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9839 of rtl_opt_pass along with...
9840 (pass_data_rtl_cprop): ...new pass_data instance and...
9841 (make_pass_rtl_cprop): ...new function.
9842 * cse.c (pass_cse): Convert from a global struct to a subclass of
9843 rtl_opt_pass along with...
9844 (pass_data_cse): ...new pass_data instance and...
9845 (make_pass_cse): ...new function.
9846 (pass_cse2): Convert from a global struct to a subclass of
9847 rtl_opt_pass along with...
9848 (pass_data_cse2): ...new pass_data instance and...
9849 (make_pass_cse2): ...new function.
9850 (pass_cse_after_global_opts): Convert from a global struct to a
9851 subclass of rtl_opt_pass along with...
9852 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9853 (make_pass_cse_after_global_opts): ...new function.
9854 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9855 of rtl_opt_pass along with...
9856 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9857 (make_pass_ud_rtl_dce): ...new function.
9858 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9859 rtl_opt_pass along with...
9860 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9861 (make_pass_fast_rtl_dce): ...new function.
9862 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9863 a subclass of rtl_opt_pass along with...
9864 (pass_data_df_initialize_opt): ...new pass_data instance and...
9865 (make_pass_df_initialize_opt): ...new function.
9866 (pass_df_initialize_no_opt): Convert from a global struct to a
9867 subclass of rtl_opt_pass along with...
9868 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9869 (make_pass_df_initialize_no_opt): ...new function.
9870 (pass_df_finish): Convert from a global struct to a subclass of
9871 rtl_opt_pass along with...
9872 (pass_data_df_finish): ...new pass_data instance and...
9873 (make_pass_df_finish): ...new function.
9874 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9875 rtl_opt_pass along with...
9876 (pass_data_rtl_dse1): ...new pass_data instance and...
9877 (make_pass_rtl_dse1): ...new function.
9878 (pass_rtl_dse2): Convert from a global struct to a subclass of
9879 rtl_opt_pass along with...
9880 (pass_data_rtl_dse2): ...new pass_data instance and...
9881 (make_pass_rtl_dse2): ...new function.
9882 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9883 subclass of rtl_opt_pass along with...
9884 (pass_data_dwarf2_frame): ...new pass_data instance and...
9885 (make_pass_dwarf2_frame): ...new function.
9886 * except.c (pass_set_nothrow_function_flags): Convert from a global
9887 struct to a subclass of rtl_opt_pass along with...
9888 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9890 (make_pass_set_nothrow_function_flags): ...new function.
9891 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9892 subclass of rtl_opt_pass along with...
9893 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9895 (make_pass_convert_to_eh_region_ranges): ...new function.
9896 * final.c (pass_compute_alignments): Convert from a global struct to a
9897 subclass of rtl_opt_pass along with...
9898 (pass_data_compute_alignments): ...new pass_data instance and...
9899 (make_pass_compute_alignments): ...new function.
9900 (pass_final): Convert from a global struct to a subclass of
9901 rtl_opt_pass along with...
9902 (pass_data_final): ...new pass_data instance and...
9903 (make_pass_final): ...new function.
9904 (pass_shorten_branches): Convert from a global struct to a subclass of
9905 rtl_opt_pass along with...
9906 (pass_data_shorten_branches): ...new pass_data instance and...
9907 (make_pass_shorten_branches): ...new function.
9908 (pass_clean_state): Convert from a global struct to a subclass of
9909 rtl_opt_pass along with...
9910 (pass_data_clean_state): ...new pass_data instance and...
9911 (make_pass_clean_state): ...new function.
9912 * function.c (pass_instantiate_virtual_regs): Convert from a global
9913 struct to a subclass of rtl_opt_pass along with...
9914 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9915 (make_pass_instantiate_virtual_regs): ...new function.
9916 (pass_leaf_regs): Convert from a global struct to a subclass of
9917 rtl_opt_pass along with...
9918 (pass_data_leaf_regs): ...new pass_data instance and...
9919 (make_pass_leaf_regs): ...new function.
9920 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9921 subclass of rtl_opt_pass along with...
9922 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9924 (make_pass_thread_prologue_and_epilogue): ...new function.
9925 (pass_match_asm_constraints): Convert from a global struct to a
9926 subclass of rtl_opt_pass along with...
9927 (pass_data_match_asm_constraints): ...new pass_data instance and...
9928 (make_pass_match_asm_constraints): ...new function.
9929 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9930 subclass of rtl_opt_pass along with...
9931 (pass_data_rtl_fwprop): ...new pass_data instance and...
9932 (make_pass_rtl_fwprop): ...new function.
9933 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9934 rtl_opt_pass along with...
9935 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9936 (make_pass_rtl_fwprop_addr): ...new function.
9937 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9938 rtl_opt_pass along with...
9939 (pass_data_rtl_pre): ...new pass_data instance and...
9940 (make_pass_rtl_pre): ...new function.
9941 (pass_rtl_hoist): Convert from a global struct to a subclass of
9942 rtl_opt_pass along with...
9943 (pass_data_rtl_hoist): ...new pass_data instance and...
9944 (make_pass_rtl_hoist): ...new function.
9945 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9946 subclass of gimple_opt_pass along with...
9947 (pass_data_lower_cf): ...new pass_data instance and...
9948 (make_pass_lower_cf): ...new function.
9949 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9950 from a global struct to a subclass of gimple_opt_pass along with...
9951 (pass_data_strength_reduction): ...new pass_data instance and...
9952 (make_pass_strength_reduction): ...new function.
9953 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9954 of rtl_opt_pass along with...
9955 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9956 (make_pass_rtl_ifcvt): ...new function.
9957 (pass_if_after_combine): Convert from a global struct to a subclass of
9958 rtl_opt_pass along with...
9959 (pass_data_if_after_combine): ...new pass_data instance and...
9960 (make_pass_if_after_combine): ...new function.
9961 (pass_if_after_reload): Convert from a global struct to a subclass of
9962 rtl_opt_pass along with...
9963 (pass_data_if_after_reload): ...new pass_data instance and...
9964 (make_pass_if_after_reload): ...new function.
9965 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9966 a subclass of rtl_opt_pass along with...
9967 (pass_data_initialize_regs): ...new pass_data instance and...
9968 (make_pass_initialize_regs): ...new function.
9969 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9970 of ipa_opt_pass_d along with...
9971 (pass_data_ipa_cp): ...new pass_data instance and...
9972 (make_pass_ipa_cp): ...new function.
9973 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9974 global struct to a subclass of gimple_opt_pass along with...
9975 (pass_data_inline_parameters): ...new pass_data instance and...
9976 (make_pass_inline_parameters): ...new function.
9977 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9978 subclass of gimple_opt_pass along with...
9979 (pass_data_early_inline): ...new pass_data instance and...
9980 (make_pass_early_inline): ...new function.
9981 (pass_ipa_inline): Convert from a global struct to a subclass of
9982 ipa_opt_pass_d along with...
9983 (pass_data_ipa_inline): ...new pass_data instance and...
9984 (make_pass_ipa_inline): ...new function.
9985 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9986 struct to a subclass of gimple_opt_pass along with...
9987 (pass_data_local_pure_const): ...new pass_data instance and...
9988 (make_pass_local_pure_const): ...new function.
9989 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9990 ipa_opt_pass_d along with...
9991 (pass_data_ipa_pure_const): ...new pass_data instance and...
9992 (make_pass_ipa_pure_const): ...new function.
9993 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9994 to a subclass of ipa_opt_pass_d along with...
9995 (pass_data_ipa_reference): ...new pass_data instance and...
9996 (make_pass_ipa_reference): ...new function.
9997 * ipa-split.c (pass_split_functions): Convert from a global struct to
9998 a subclass of gimple_opt_pass along with...
9999 (pass_data_split_functions): ...new pass_data instance and...
10000 (make_pass_split_functions): ...new function.
10001 (pass_feedback_split_functions): Convert from a global struct to a
10002 subclass of gimple_opt_pass along with...
10003 (pass_data_feedback_split_functions): ...new pass_data instance and...
10004 (make_pass_feedback_split_functions): ...new function.
10005 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10006 global struct to a subclass of simple_ipa_opt_pass along with...
10007 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10009 (make_pass_ipa_function_and_variable_visibility): ...new function.
10010 (pass_ipa_free_inline_summary): Convert from a global struct to a
10011 subclass of simple_ipa_opt_pass along with...
10012 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10013 (make_pass_ipa_free_inline_summary): ...new function.
10014 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10015 subclass of ipa_opt_pass_d along with...
10016 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10018 (make_pass_ipa_whole_program_visibility): ...new function.
10019 (pass_ipa_profile): Convert from a global struct to a subclass of
10020 ipa_opt_pass_d along with...
10021 (pass_data_ipa_profile): ...new pass_data instance and...
10022 (make_pass_ipa_profile): ...new function.
10023 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10024 ipa_opt_pass_d along with...
10025 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10026 (make_pass_ipa_cdtor_merge): ...new function.
10027 * ira.c (pass_ira): Convert from a global struct to a subclass of
10028 rtl_opt_pass along with...
10029 (pass_data_ira): ...new pass_data instance and...
10030 (make_pass_ira): ...new function.
10031 (pass_reload): Convert from a global struct to a subclass of
10032 rtl_opt_pass along with...
10033 (pass_data_reload): ...new pass_data instance and...
10034 (make_pass_reload): ...new function.
10035 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10036 subclass of rtl_opt_pass along with...
10037 (pass_data_cleanup_barriers): ...new pass_data instance and...
10038 (make_pass_cleanup_barriers): ...new function.
10039 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10040 of rtl_opt_pass along with...
10041 (pass_data_loop2): ...new pass_data instance and...
10042 (make_pass_loop2): ...new function.
10043 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10044 rtl_opt_pass along with...
10045 (pass_data_rtl_loop_init): ...new pass_data instance and...
10046 (make_pass_rtl_loop_init): ...new function.
10047 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10048 rtl_opt_pass along with...
10049 (pass_data_rtl_loop_done): ...new pass_data instance and...
10050 (make_pass_rtl_loop_done): ...new function.
10051 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10052 subclass of rtl_opt_pass along with...
10053 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10054 (make_pass_rtl_move_loop_invariants): ...new function.
10055 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10056 rtl_opt_pass along with...
10057 (pass_data_rtl_unswitch): ...new pass_data instance and...
10058 (make_pass_rtl_unswitch): ...new function.
10059 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10060 subclass of rtl_opt_pass along with...
10061 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10063 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10064 (pass_rtl_doloop): Convert from a global struct to a subclass of
10065 rtl_opt_pass along with...
10066 (pass_data_rtl_doloop): ...new pass_data instance and...
10067 (make_pass_rtl_doloop): ...new function.
10068 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10069 a subclass of rtl_opt_pass along with...
10070 (pass_data_lower_subreg): ...new pass_data instance and...
10071 (make_pass_lower_subreg): ...new function.
10072 (pass_lower_subreg2): Convert from a global struct to a subclass of
10073 rtl_opt_pass along with...
10074 (pass_data_lower_subreg2): ...new pass_data instance and...
10075 (make_pass_lower_subreg2): ...new function.
10076 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10077 struct to a subclass of ipa_opt_pass_d along with...
10078 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10079 (make_pass_ipa_lto_gimple_out): ...new function.
10080 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10081 of ipa_opt_pass_d along with...
10082 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10083 (make_pass_ipa_lto_finish_out): ...new function.
10084 * mode-switching.c (pass_mode_switching): Convert from a global struct
10085 to a subclass of rtl_opt_pass along with...
10086 (pass_data_mode_switching): ...new pass_data instance and...
10087 (make_pass_mode_switching): ...new function.
10088 * modulo-sched.c (pass_sms): Convert from a global struct to a
10089 subclass of rtl_opt_pass along with...
10090 (pass_data_sms): ...new pass_data instance and...
10091 (make_pass_sms): ...new function.
10092 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10093 subclass of gimple_opt_pass along with...
10094 (pass_data_expand_omp): ...new pass_data instance and...
10095 (make_pass_expand_omp): ...new function.
10096 (pass_lower_omp): Convert from a global struct to a subclass of
10097 gimple_opt_pass along with...
10098 (pass_data_lower_omp): ...new pass_data instance and...
10099 (make_pass_lower_omp): ...new function.
10100 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10101 of gimple_opt_pass along with...
10102 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10103 (make_pass_diagnose_omp_blocks): ...new function.
10104 * passes.c (pass_early_local_passes): Convert from a global struct to
10105 a subclass of simple_ipa_opt_pass along with...
10106 (pass_data_early_local_passes): ...new pass_data instance and...
10107 (make_pass_early_local_passes): ...new function.
10108 (pass_all_early_optimizations): Convert from a global struct to a
10109 subclass of gimple_opt_pass along with...
10110 (pass_data_all_early_optimizations): ...new pass_data instance and...
10111 (make_pass_all_early_optimizations): ...new function.
10112 (pass_all_optimizations): Convert from a global struct to a subclass
10113 of gimple_opt_pass along with...
10114 (pass_data_all_optimizations): ...new pass_data instance and...
10115 (make_pass_all_optimizations): ...new function.
10116 (pass_all_optimizations_g): Convert from a global struct to a subclass
10117 of gimple_opt_pass along with...
10118 (pass_data_all_optimizations_g): ...new pass_data instance and...
10119 (make_pass_all_optimizations_g): ...new function.
10120 (pass_rest_of_compilation): Convert from a global struct to a subclass
10121 of rtl_opt_pass along with...
10122 (pass_data_rest_of_compilation): ...new pass_data instance and...
10123 (make_pass_rest_of_compilation): ...new function.
10124 (pass_postreload): Convert from a global struct to a subclass of
10125 rtl_opt_pass along with...
10126 (pass_data_postreload): ...new pass_data instance and...
10127 (make_pass_postreload): ...new function.
10128 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10129 subclass of rtl_opt_pass along with...
10130 (pass_data_gcse2): ...new pass_data instance and...
10131 (make_pass_gcse2): ...new function.
10132 * postreload.c (pass_postreload_cse): Convert from a global struct to
10133 a subclass of rtl_opt_pass along with...
10134 (pass_data_postreload_cse): ...new pass_data instance and...
10135 (make_pass_postreload_cse): ...new function.
10136 * predict.c (pass_profile): Convert from a global struct to a subclass
10137 of gimple_opt_pass along with...
10138 (pass_data_profile): ...new pass_data instance and...
10139 (make_pass_profile): ...new function.
10140 (pass_strip_predict_hints): Convert from a global struct to a subclass
10141 of gimple_opt_pass along with...
10142 (pass_data_strip_predict_hints): ...new pass_data instance and...
10143 (make_pass_strip_predict_hints): ...new function.
10144 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10145 of rtl_opt_pass along with...
10146 (pass_data_peephole2): ...new pass_data instance and...
10147 (make_pass_peephole2): ...new function.
10148 (pass_split_all_insns): Convert from a global struct to a subclass of
10149 rtl_opt_pass along with...
10150 (pass_data_split_all_insns): ...new pass_data instance and...
10151 (make_pass_split_all_insns): ...new function.
10152 (pass_split_after_reload): Convert from a global struct to a subclass
10153 of rtl_opt_pass along with...
10154 (pass_data_split_after_reload): ...new pass_data instance and...
10155 (make_pass_split_after_reload): ...new function.
10156 (pass_split_before_regstack): Convert from a global struct to a
10157 subclass of rtl_opt_pass along with...
10158 (pass_data_split_before_regstack): ...new pass_data instance and...
10159 (make_pass_split_before_regstack): ...new function.
10160 (pass_split_before_sched2): Convert from a global struct to a subclass
10161 of rtl_opt_pass along with...
10162 (pass_data_split_before_sched2): ...new pass_data instance and...
10163 (make_pass_split_before_sched2): ...new function.
10164 (pass_split_for_shorten_branches): Convert from a global struct to a
10165 subclass of rtl_opt_pass along with...
10166 (pass_data_split_for_shorten_branches): ...new pass_data instance
10168 (make_pass_split_for_shorten_branches): ...new function.
10169 * ree.c (pass_ree): Convert from a global struct to a subclass of
10170 rtl_opt_pass along with...
10171 (pass_data_ree): ...new pass_data instance and...
10172 (make_pass_ree): ...new function.
10173 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10174 subclass of rtl_opt_pass along with...
10175 (pass_data_stack_regs): ...new pass_data instance and...
10176 (make_pass_stack_regs): ...new function.
10177 (pass_stack_regs_run): Convert from a global struct to a subclass of
10178 rtl_opt_pass along with...
10179 (pass_data_stack_regs_run): ...new pass_data instance and...
10180 (make_pass_stack_regs_run): ...new function.
10181 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10182 subclass of rtl_opt_pass along with...
10183 (pass_data_cprop_hardreg): ...new pass_data instance and...
10184 (make_pass_cprop_hardreg): ...new function.
10185 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10186 subclass of rtl_opt_pass along with...
10187 (pass_data_reginfo_init): ...new pass_data instance and...
10188 (make_pass_reginfo_init): ...new function.
10189 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10190 of rtl_opt_pass along with...
10191 (pass_data_regmove): ...new pass_data instance and...
10192 (make_pass_regmove): ...new function.
10193 * regrename.c (pass_regrename): Convert from a global struct to a
10194 subclass of rtl_opt_pass along with...
10195 (pass_data_regrename): ...new pass_data instance and...
10196 (make_pass_regrename): ...new function.
10197 * reorg.c (pass_delay_slots): Convert from a global struct to a
10198 subclass of rtl_opt_pass along with...
10199 (pass_data_delay_slots): ...new pass_data instance and...
10200 (make_pass_delay_slots): ...new function.
10201 (pass_machine_reorg): Convert from a global struct to a subclass of
10202 rtl_opt_pass along with...
10203 (pass_data_machine_reorg): ...new pass_data instance and...
10204 (make_pass_machine_reorg): ...new function.
10205 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10206 of rtl_opt_pass along with...
10207 (pass_data_sched): ...new pass_data instance and...
10208 (make_pass_sched): ...new function.
10209 (pass_sched2): Convert from a global struct to a subclass of
10210 rtl_opt_pass along with...
10211 (pass_data_sched2): ...new pass_data instance and...
10212 (make_pass_sched2): ...new function.
10213 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10214 to a subclass of rtl_opt_pass along with...
10215 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10216 (make_pass_stack_ptr_mod): ...new function.
10217 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10218 to a subclass of rtl_opt_pass along with...
10219 (pass_data_rtl_store_motion): ...new pass_data instance and...
10220 (make_pass_rtl_store_motion): ...new function.
10221 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10222 of gimple_opt_pass along with...
10223 (pass_data_tracer): ...new pass_data instance and...
10224 (make_pass_tracer): ...new function.
10225 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10226 to a subclass of gimple_opt_pass along with...
10227 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10228 (make_pass_diagnose_tm_blocks): ...new function.
10229 (pass_lower_tm): Convert from a global struct to a subclass of
10230 gimple_opt_pass along with...
10231 (pass_data_lower_tm): ...new pass_data instance and...
10232 (make_pass_lower_tm): ...new function.
10233 (pass_tm_init): Convert from a global struct to a subclass of
10234 gimple_opt_pass along with...
10235 (pass_data_tm_init): ...new pass_data instance and...
10236 (make_pass_tm_init): ...new function.
10237 (pass_tm_mark): Convert from a global struct to a subclass of
10238 gimple_opt_pass along with...
10239 (pass_data_tm_mark): ...new pass_data instance and...
10240 (make_pass_tm_mark): ...new function.
10241 (pass_tm_edges): Convert from a global struct to a subclass of
10242 gimple_opt_pass along with...
10243 (pass_data_tm_edges): ...new pass_data instance and...
10244 (make_pass_tm_edges): ...new function.
10245 (pass_tm_memopt): Convert from a global struct to a subclass of
10246 gimple_opt_pass along with...
10247 (pass_data_tm_memopt): ...new pass_data instance and...
10248 (make_pass_tm_memopt): ...new function.
10249 (pass_ipa_tm): Convert from a global struct to a subclass of
10250 simple_ipa_opt_pass along with...
10251 (pass_data_ipa_tm): ...new pass_data instance and...
10252 (make_pass_ipa_tm): ...new function.
10253 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10254 subclass of gimple_opt_pass along with...
10255 (pass_data_call_cdce): ...new pass_data instance and...
10256 (make_pass_call_cdce): ...new function.
10257 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10258 subclass of gimple_opt_pass along with...
10259 (pass_data_build_cfg): ...new pass_data instance and...
10260 (make_pass_build_cfg): ...new function.
10261 (pass_split_crit_edges): Convert from a global struct to a subclass of
10262 gimple_opt_pass along with...
10263 (pass_data_split_crit_edges): ...new pass_data instance and...
10264 (make_pass_split_crit_edges): ...new function.
10265 (pass_warn_function_return): Convert from a global struct to a
10266 subclass of gimple_opt_pass along with...
10267 (pass_data_warn_function_return): ...new pass_data instance and...
10268 (make_pass_warn_function_return): ...new function.
10269 (pass_warn_function_noreturn): Convert from a global struct to a
10270 subclass of gimple_opt_pass along with...
10271 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10272 (make_pass_warn_function_noreturn): ...new function.
10273 (pass_warn_unused_result): Convert from a global struct to a subclass
10274 of gimple_opt_pass along with...
10275 (pass_data_warn_unused_result): ...new pass_data instance and...
10276 (make_pass_warn_unused_result): ...new function.
10277 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10278 a subclass of gimple_opt_pass along with...
10279 (pass_data_merge_phi): ...new pass_data instance and...
10280 (make_pass_merge_phi): ...new function.
10281 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10282 a subclass of gimple_opt_pass along with...
10283 (pass_data_lower_complex): ...new pass_data instance and...
10284 (make_pass_lower_complex): ...new function.
10285 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10286 gimple_opt_pass along with...
10287 (pass_data_lower_complex_O0): ...new pass_data instance and...
10288 (make_pass_lower_complex_O0): ...new function.
10289 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10290 subclass of gimple_opt_pass along with...
10291 (pass_data_lower_eh): ...new pass_data instance and...
10292 (make_pass_lower_eh): ...new function.
10293 (pass_refactor_eh): Convert from a global struct to a subclass of
10294 gimple_opt_pass along with...
10295 (pass_data_refactor_eh): ...new pass_data instance and...
10296 (make_pass_refactor_eh): ...new function.
10297 (pass_lower_resx): Convert from a global struct to a subclass of
10298 gimple_opt_pass along with...
10299 (pass_data_lower_resx): ...new pass_data instance and...
10300 (make_pass_lower_resx): ...new function.
10301 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10302 of gimple_opt_pass along with...
10303 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10304 (make_pass_lower_eh_dispatch): ...new function.
10305 (pass_cleanup_eh): Convert from a global struct to a subclass of
10306 gimple_opt_pass along with...
10307 (pass_data_cleanup_eh): ...new pass_data instance and...
10308 (make_pass_cleanup_eh): ...new function.
10309 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10310 to a subclass of simple_ipa_opt_pass along with...
10311 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10312 (make_pass_ipa_lower_emutls): ...new function.
10313 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10314 a subclass of gimple_opt_pass along with...
10315 (pass_data_if_conversion): ...new pass_data instance and...
10316 (make_pass_if_conversion): ...new function.
10317 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10318 subclass of gimple_opt_pass along with...
10319 (pass_data_build_ssa): ...new pass_data instance and...
10320 (make_pass_build_ssa): ...new function.
10321 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10322 global struct to a subclass of gimple_opt_pass along with...
10323 (pass_data_loop_distribution): ...new pass_data instance and...
10324 (make_pass_loop_distribution): ...new function.
10325 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10326 subclass of gimple_opt_pass along with...
10327 (pass_data_mudflap_1): ...new pass_data instance and...
10328 (make_pass_mudflap_1): ...new function.
10329 (pass_mudflap_2): Convert from a global struct to a subclass of
10330 gimple_opt_pass along with...
10331 (pass_data_mudflap_2): ...new pass_data instance and...
10332 (make_pass_mudflap_2): ...new function.
10333 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10334 subclass of gimple_opt_pass along with...
10335 (pass_data_mudflap_1): ...new pass_data instance and...
10336 (make_pass_mudflap_1): ...new function.
10337 (pass_mudflap_2): Convert from a global struct to a subclass of
10338 gimple_opt_pass along with...
10339 (pass_data_mudflap_2): ...new pass_data instance and...
10340 (make_pass_mudflap_2): ...new function.
10341 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10342 gimple_opt_pass along with...
10343 (pass_data_nrv): ...new pass_data instance and...
10344 (make_pass_nrv): ...new function.
10345 (pass_return_slot): Convert from a global struct to a subclass of
10346 gimple_opt_pass along with...
10347 (pass_data_return_slot): ...new pass_data instance and...
10348 (make_pass_return_slot): ...new function.
10349 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10350 to a subclass of gimple_opt_pass along with...
10351 (pass_data_object_sizes): ...new pass_data instance and...
10352 (make_pass_object_sizes): ...new function.
10353 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10354 global struct to a subclass of gimple_opt_pass along with...
10355 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10357 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10358 (pass_fixup_cfg): Convert from a global struct to a subclass of
10359 gimple_opt_pass along with...
10360 (pass_data_fixup_cfg): ...new pass_data instance and...
10361 (make_pass_fixup_cfg): ...new function.
10362 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10363 (make_pass_mudflap_1): ...new function.
10364 (pass_mudflap_2): Replace declaration with that of...
10365 (make_pass_mudflap_2): ...new function.
10366 (pass_asan): Replace declaration with that of...
10367 (make_pass_asan): ...new function.
10368 (pass_asan_O0): Replace declaration with that of...
10369 (make_pass_asan_O0): ...new function.
10370 (pass_tsan): Replace declaration with that of...
10371 (make_pass_tsan): ...new function.
10372 (pass_tsan_O0): Replace declaration with that of...
10373 (make_pass_tsan_O0): ...new function.
10374 (pass_lower_cf): Replace declaration with that of...
10375 (make_pass_lower_cf): ...new function.
10376 (pass_refactor_eh): Replace declaration with that of...
10377 (make_pass_refactor_eh): ...new function.
10378 (pass_lower_eh): Replace declaration with that of...
10379 (make_pass_lower_eh): ...new function.
10380 (pass_lower_eh_dispatch): Replace declaration with that of...
10381 (make_pass_lower_eh_dispatch): ...new function.
10382 (pass_lower_resx): Replace declaration with that of...
10383 (make_pass_lower_resx): ...new function.
10384 (pass_build_cfg): Replace declaration with that of...
10385 (make_pass_build_cfg): ...new function.
10386 (pass_early_tree_profile): Replace declaration with that of...
10387 (make_pass_early_tree_profile): ...new function.
10388 (pass_cleanup_eh): Replace declaration with that of...
10389 (make_pass_cleanup_eh): ...new function.
10390 (pass_sra): Replace declaration with that of...
10391 (make_pass_sra): ...new function.
10392 (pass_sra_early): Replace declaration with that of...
10393 (make_pass_sra_early): ...new function.
10394 (pass_early_ipa_sra): Replace declaration with that of...
10395 (make_pass_early_ipa_sra): ...new function.
10396 (pass_tail_recursion): Replace declaration with that of...
10397 (make_pass_tail_recursion): ...new function.
10398 (pass_tail_calls): Replace declaration with that of...
10399 (make_pass_tail_calls): ...new function.
10400 (pass_tree_loop): Replace declaration with that of...
10401 (make_pass_tree_loop): ...new function.
10402 (pass_tree_loop_init): Replace declaration with that of...
10403 (make_pass_tree_loop_init): ...new function.
10404 (pass_lim): Replace declaration with that of...
10405 (make_pass_lim): ...new function.
10406 (pass_tree_unswitch): Replace declaration with that of...
10407 (make_pass_tree_unswitch): ...new function.
10408 (pass_predcom): Replace declaration with that of...
10409 (make_pass_predcom): ...new function.
10410 (pass_iv_canon): Replace declaration with that of...
10411 (make_pass_iv_canon): ...new function.
10412 (pass_scev_cprop): Replace declaration with that of...
10413 (make_pass_scev_cprop): ...new function.
10414 (pass_empty_loop): Replace declaration with that of...
10415 (make_pass_empty_loop): ...new function.
10416 (pass_record_bounds): Replace declaration with that of...
10417 (make_pass_record_bounds): ...new function.
10418 (pass_graphite): Replace declaration with that of...
10419 (make_pass_graphite): ...new function.
10420 (pass_graphite_transforms): Replace declaration with that of...
10421 (make_pass_graphite_transforms): ...new function.
10422 (pass_if_conversion): Replace declaration with that of...
10423 (make_pass_if_conversion): ...new function.
10424 (pass_loop_distribution): Replace declaration with that of...
10425 (make_pass_loop_distribution): ...new function.
10426 (pass_vectorize): Replace declaration with that of...
10427 (make_pass_vectorize): ...new function.
10428 (pass_slp_vectorize): Replace declaration with that of...
10429 (make_pass_slp_vectorize): ...new function.
10430 (pass_complete_unroll): Replace declaration with that of...
10431 (make_pass_complete_unroll): ...new function.
10432 (pass_complete_unrolli): Replace declaration with that of...
10433 (make_pass_complete_unrolli): ...new function.
10434 (pass_parallelize_loops): Replace declaration with that of...
10435 (make_pass_parallelize_loops): ...new function.
10436 (pass_loop_prefetch): Replace declaration with that of...
10437 (make_pass_loop_prefetch): ...new function.
10438 (pass_iv_optimize): Replace declaration with that of...
10439 (make_pass_iv_optimize): ...new function.
10440 (pass_tree_loop_done): Replace declaration with that of...
10441 (make_pass_tree_loop_done): ...new function.
10442 (pass_ch): Replace declaration with that of...
10443 (make_pass_ch): ...new function.
10444 (pass_ccp): Replace declaration with that of...
10445 (make_pass_ccp): ...new function.
10446 (pass_phi_only_cprop): Replace declaration with that of...
10447 (make_pass_phi_only_cprop): ...new function.
10448 (pass_build_ssa): Replace declaration with that of...
10449 (make_pass_build_ssa): ...new function.
10450 (pass_build_alias): Replace declaration with that of...
10451 (make_pass_build_alias): ...new function.
10452 (pass_build_ealias): Replace declaration with that of...
10453 (make_pass_build_ealias): ...new function.
10454 (pass_dominator): Replace declaration with that of...
10455 (make_pass_dominator): ...new function.
10456 (pass_dce): Replace declaration with that of...
10457 (make_pass_dce): ...new function.
10458 (pass_dce_loop): Replace declaration with that of...
10459 (make_pass_dce_loop): ...new function.
10460 (pass_cd_dce): Replace declaration with that of...
10461 (make_pass_cd_dce): ...new function.
10462 (pass_call_cdce): Replace declaration with that of...
10463 (make_pass_call_cdce): ...new function.
10464 (pass_merge_phi): Replace declaration with that of...
10465 (make_pass_merge_phi): ...new function.
10466 (pass_split_crit_edges): Replace declaration with that of...
10467 (make_pass_split_crit_edges): ...new function.
10468 (pass_pre): Replace declaration with that of...
10469 (make_pass_pre): ...new function.
10470 (pass_profile): Replace declaration with that of...
10471 (make_pass_profile): ...new function.
10472 (pass_strip_predict_hints): Replace declaration with that of...
10473 (make_pass_strip_predict_hints): ...new function.
10474 (pass_lower_complex_O0): Replace declaration with that of...
10475 (make_pass_lower_complex_O0): ...new function.
10476 (pass_lower_complex): Replace declaration with that of...
10477 (make_pass_lower_complex): ...new function.
10478 (pass_lower_vector): Replace declaration with that of...
10479 (make_pass_lower_vector): ...new function.
10480 (pass_lower_vector_ssa): Replace declaration with that of...
10481 (make_pass_lower_vector_ssa): ...new function.
10482 (pass_lower_omp): Replace declaration with that of...
10483 (make_pass_lower_omp): ...new function.
10484 (pass_diagnose_omp_blocks): Replace declaration with that of...
10485 (make_pass_diagnose_omp_blocks): ...new function.
10486 (pass_expand_omp): Replace declaration with that of...
10487 (make_pass_expand_omp): ...new function.
10488 (pass_expand_omp_ssa): Replace declaration with that of...
10489 (make_pass_expand_omp_ssa): ...new function.
10490 (pass_object_sizes): Replace declaration with that of...
10491 (make_pass_object_sizes): ...new function.
10492 (pass_strlen): Replace declaration with that of...
10493 (make_pass_strlen): ...new function.
10494 (pass_fold_builtins): Replace declaration with that of...
10495 (make_pass_fold_builtins): ...new function.
10496 (pass_stdarg): Replace declaration with that of...
10497 (make_pass_stdarg): ...new function.
10498 (pass_early_warn_uninitialized): Replace declaration with that of...
10499 (make_pass_early_warn_uninitialized): ...new function.
10500 (pass_late_warn_uninitialized): Replace declaration with that of...
10501 (make_pass_late_warn_uninitialized): ...new function.
10502 (pass_cse_reciprocals): Replace declaration with that of...
10503 (make_pass_cse_reciprocals): ...new function.
10504 (pass_cse_sincos): Replace declaration with that of...
10505 (make_pass_cse_sincos): ...new function.
10506 (pass_optimize_bswap): Replace declaration with that of...
10507 (make_pass_optimize_bswap): ...new function.
10508 (pass_optimize_widening_mul): Replace declaration with that of...
10509 (make_pass_optimize_widening_mul): ...new function.
10510 (pass_warn_function_return): Replace declaration with that of...
10511 (make_pass_warn_function_return): ...new function.
10512 (pass_warn_function_noreturn): Replace declaration with that of...
10513 (make_pass_warn_function_noreturn): ...new function.
10514 (pass_cselim): Replace declaration with that of...
10515 (make_pass_cselim): ...new function.
10516 (pass_phiopt): Replace declaration with that of...
10517 (make_pass_phiopt): ...new function.
10518 (pass_forwprop): Replace declaration with that of...
10519 (make_pass_forwprop): ...new function.
10520 (pass_phiprop): Replace declaration with that of...
10521 (make_pass_phiprop): ...new function.
10522 (pass_tree_ifcombine): Replace declaration with that of...
10523 (make_pass_tree_ifcombine): ...new function.
10524 (pass_dse): Replace declaration with that of...
10525 (make_pass_dse): ...new function.
10526 (pass_nrv): Replace declaration with that of...
10527 (make_pass_nrv): ...new function.
10528 (pass_rename_ssa_copies): Replace declaration with that of...
10529 (make_pass_rename_ssa_copies): ...new function.
10530 (pass_sink_code): Replace declaration with that of...
10531 (make_pass_sink_code): ...new function.
10532 (pass_fre): Replace declaration with that of...
10533 (make_pass_fre): ...new function.
10534 (pass_check_data_deps): Replace declaration with that of...
10535 (make_pass_check_data_deps): ...new function.
10536 (pass_copy_prop): Replace declaration with that of...
10537 (make_pass_copy_prop): ...new function.
10538 (pass_vrp): Replace declaration with that of...
10539 (make_pass_vrp): ...new function.
10540 (pass_uncprop): Replace declaration with that of...
10541 (make_pass_uncprop): ...new function.
10542 (pass_return_slot): Replace declaration with that of...
10543 (make_pass_return_slot): ...new function.
10544 (pass_reassoc): Replace declaration with that of...
10545 (make_pass_reassoc): ...new function.
10546 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10547 (make_pass_rebuild_cgraph_edges): ...new function.
10548 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10549 (make_pass_remove_cgraph_callee_edges): ...new function.
10550 (pass_build_cgraph_edges): Replace declaration with that of...
10551 (make_pass_build_cgraph_edges): ...new function.
10552 (pass_local_pure_const): Replace declaration with that of...
10553 (make_pass_local_pure_const): ...new function.
10554 (pass_tracer): Replace declaration with that of...
10555 (make_pass_tracer): ...new function.
10556 (pass_warn_unused_result): Replace declaration with that of...
10557 (make_pass_warn_unused_result): ...new function.
10558 (pass_diagnose_tm_blocks): Replace declaration with that of...
10559 (make_pass_diagnose_tm_blocks): ...new function.
10560 (pass_lower_tm): Replace declaration with that of...
10561 (make_pass_lower_tm): ...new function.
10562 (pass_tm_init): Replace declaration with that of...
10563 (make_pass_tm_init): ...new function.
10564 (pass_tm_mark): Replace declaration with that of...
10565 (make_pass_tm_mark): ...new function.
10566 (pass_tm_memopt): Replace declaration with that of...
10567 (make_pass_tm_memopt): ...new function.
10568 (pass_tm_edges): Replace declaration with that of...
10569 (make_pass_tm_edges): ...new function.
10570 (pass_split_functions): Replace declaration with that of...
10571 (make_pass_split_functions): ...new function.
10572 (pass_feedback_split_functions): Replace declaration with that of...
10573 (make_pass_feedback_split_functions): ...new function.
10574 (pass_strength_reduction): Replace declaration with that of...
10575 (make_pass_strength_reduction): ...new function.
10576 (pass_ipa_lower_emutls): Replace declaration with that of...
10577 (make_pass_ipa_lower_emutls): ...new function.
10578 (pass_ipa_function_and_variable_visibility): Replace declaration with
10580 (make_pass_ipa_function_and_variable_visibility): ...new function.
10581 (pass_ipa_tree_profile): Replace declaration with that of...
10582 (make_pass_ipa_tree_profile): ...new function.
10583 (pass_early_local_passes): Replace declaration with that of...
10584 (make_pass_early_local_passes): ...new function.
10585 (pass_ipa_whole_program_visibility): Replace declaration with that
10587 (make_pass_ipa_whole_program_visibility): ...new function.
10588 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10589 (make_pass_ipa_lto_gimple_out): ...new function.
10590 (pass_ipa_increase_alignment): Replace declaration with that of...
10591 (make_pass_ipa_increase_alignment): ...new function.
10592 (pass_ipa_inline): Replace declaration with that of...
10593 (make_pass_ipa_inline): ...new function.
10594 (pass_ipa_free_lang_data): Replace declaration with that of...
10595 (make_pass_ipa_free_lang_data): ...new function.
10596 (pass_ipa_free_inline_summary): Replace declaration with that of...
10597 (make_pass_ipa_free_inline_summary): ...new function.
10598 (pass_ipa_cp): Replace declaration with that of...
10599 (make_pass_ipa_cp): ...new function.
10600 (pass_ipa_reference): Replace declaration with that of...
10601 (make_pass_ipa_reference): ...new function.
10602 (pass_ipa_pure_const): Replace declaration with that of...
10603 (make_pass_ipa_pure_const): ...new function.
10604 (pass_ipa_pta): Replace declaration with that of...
10605 (make_pass_ipa_pta): ...new function.
10606 (pass_ipa_lto_finish_out): Replace declaration with that of...
10607 (make_pass_ipa_lto_finish_out): ...new function.
10608 (pass_ipa_tm): Replace declaration with that of...
10609 (make_pass_ipa_tm): ...new function.
10610 (pass_ipa_profile): Replace declaration with that of...
10611 (make_pass_ipa_profile): ...new function.
10612 (pass_ipa_cdtor_merge): Replace declaration with that of...
10613 (make_pass_ipa_cdtor_merge): ...new function.
10614 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10616 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10617 (pass_init_datastructures): Replace declaration with that of...
10618 (make_pass_init_datastructures): ...new function.
10619 (pass_fixup_cfg): Replace declaration with that of...
10620 (make_pass_fixup_cfg): ...new function.
10621 (pass_expand): Replace declaration with that of...
10622 (make_pass_expand): ...new function.
10623 (pass_instantiate_virtual_regs): Replace declaration with that of...
10624 (make_pass_instantiate_virtual_regs): ...new function.
10625 (pass_rtl_fwprop): Replace declaration with that of...
10626 (make_pass_rtl_fwprop): ...new function.
10627 (pass_rtl_fwprop_addr): Replace declaration with that of...
10628 (make_pass_rtl_fwprop_addr): ...new function.
10629 (pass_jump): Replace declaration with that of...
10630 (make_pass_jump): ...new function.
10631 (pass_jump2): Replace declaration with that of...
10632 (make_pass_jump2): ...new function.
10633 (pass_lower_subreg): Replace declaration with that of...
10634 (make_pass_lower_subreg): ...new function.
10635 (pass_cse): Replace declaration with that of...
10636 (make_pass_cse): ...new function.
10637 (pass_fast_rtl_dce): Replace declaration with that of...
10638 (make_pass_fast_rtl_dce): ...new function.
10639 (pass_ud_rtl_dce): Replace declaration with that of...
10640 (make_pass_ud_rtl_dce): ...new function.
10641 (pass_rtl_dce): Replace declaration with that of...
10642 (make_pass_rtl_dce): ...new function.
10643 (pass_rtl_dse1): Replace declaration with that of...
10644 (make_pass_rtl_dse1): ...new function.
10645 (pass_rtl_dse2): Replace declaration with that of...
10646 (make_pass_rtl_dse2): ...new function.
10647 (pass_rtl_dse3): Replace declaration with that of...
10648 (make_pass_rtl_dse3): ...new function.
10649 (pass_rtl_cprop): Replace declaration with that of...
10650 (make_pass_rtl_cprop): ...new function.
10651 (pass_rtl_pre): Replace declaration with that of...
10652 (make_pass_rtl_pre): ...new function.
10653 (pass_rtl_hoist): Replace declaration with that of...
10654 (make_pass_rtl_hoist): ...new function.
10655 (pass_rtl_store_motion): Replace declaration with that of...
10656 (make_pass_rtl_store_motion): ...new function.
10657 (pass_cse_after_global_opts): Replace declaration with that of...
10658 (make_pass_cse_after_global_opts): ...new function.
10659 (pass_rtl_ifcvt): Replace declaration with that of...
10660 (make_pass_rtl_ifcvt): ...new function.
10661 (pass_into_cfg_layout_mode): Replace declaration with that of...
10662 (make_pass_into_cfg_layout_mode): ...new function.
10663 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10664 (make_pass_outof_cfg_layout_mode): ...new function.
10665 (pass_loop2): Replace declaration with that of...
10666 (make_pass_loop2): ...new function.
10667 (pass_rtl_loop_init): Replace declaration with that of...
10668 (make_pass_rtl_loop_init): ...new function.
10669 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10670 (make_pass_rtl_move_loop_invariants): ...new function.
10671 (pass_rtl_unswitch): Replace declaration with that of...
10672 (make_pass_rtl_unswitch): ...new function.
10673 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10674 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10675 (pass_rtl_doloop): Replace declaration with that of...
10676 (make_pass_rtl_doloop): ...new function.
10677 (pass_rtl_loop_done): Replace declaration with that of...
10678 (make_pass_rtl_loop_done): ...new function.
10679 (pass_web): Replace declaration with that of...
10680 (make_pass_web): ...new function.
10681 (pass_cse2): Replace declaration with that of...
10682 (make_pass_cse2): ...new function.
10683 (pass_df_initialize_opt): Replace declaration with that of...
10684 (make_pass_df_initialize_opt): ...new function.
10685 (pass_df_initialize_no_opt): Replace declaration with that of...
10686 (make_pass_df_initialize_no_opt): ...new function.
10687 (pass_reginfo_init): Replace declaration with that of...
10688 (make_pass_reginfo_init): ...new function.
10689 (pass_inc_dec): Replace declaration with that of...
10690 (make_pass_inc_dec): ...new function.
10691 (pass_stack_ptr_mod): Replace declaration with that of...
10692 (make_pass_stack_ptr_mod): ...new function.
10693 (pass_initialize_regs): Replace declaration with that of...
10694 (make_pass_initialize_regs): ...new function.
10695 (pass_combine): Replace declaration with that of...
10696 (make_pass_combine): ...new function.
10697 (pass_if_after_combine): Replace declaration with that of...
10698 (make_pass_if_after_combine): ...new function.
10699 (pass_ree): Replace declaration with that of...
10700 (make_pass_ree): ...new function.
10701 (pass_partition_blocks): Replace declaration with that of...
10702 (make_pass_partition_blocks): ...new function.
10703 (pass_match_asm_constraints): Replace declaration with that of...
10704 (make_pass_match_asm_constraints): ...new function.
10705 (pass_regmove): Replace declaration with that of...
10706 (make_pass_regmove): ...new function.
10707 (pass_split_all_insns): Replace declaration with that of...
10708 (make_pass_split_all_insns): ...new function.
10709 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10710 (make_pass_fast_rtl_byte_dce): ...new function.
10711 (pass_lower_subreg2): Replace declaration with that of...
10712 (make_pass_lower_subreg2): ...new function.
10713 (pass_mode_switching): Replace declaration with that of...
10714 (make_pass_mode_switching): ...new function.
10715 (pass_sms): Replace declaration with that of...
10716 (make_pass_sms): ...new function.
10717 (pass_sched): Replace declaration with that of...
10718 (make_pass_sched): ...new function.
10719 (pass_ira): Replace declaration with that of...
10720 (make_pass_ira): ...new function.
10721 (pass_reload): Replace declaration with that of...
10722 (make_pass_reload): ...new function.
10723 (pass_clean_state): Replace declaration with that of...
10724 (make_pass_clean_state): ...new function.
10725 (pass_branch_prob): Replace declaration with that of...
10726 (make_pass_branch_prob): ...new function.
10727 (pass_value_profile_transformations): Replace declaration with that
10729 (make_pass_value_profile_transformations): ...new function.
10730 (pass_postreload_cse): Replace declaration with that of...
10731 (make_pass_postreload_cse): ...new function.
10732 (pass_gcse2): Replace declaration with that of...
10733 (make_pass_gcse2): ...new function.
10734 (pass_split_after_reload): Replace declaration with that of...
10735 (make_pass_split_after_reload): ...new function.
10736 (pass_branch_target_load_optimize1): Replace declaration with that
10738 (make_pass_branch_target_load_optimize1): ...new function.
10739 (pass_thread_prologue_and_epilogue): Replace declaration with that
10741 (make_pass_thread_prologue_and_epilogue): ...new function.
10742 (pass_stack_adjustments): Replace declaration with that of...
10743 (make_pass_stack_adjustments): ...new function.
10744 (pass_peephole2): Replace declaration with that of...
10745 (make_pass_peephole2): ...new function.
10746 (pass_if_after_reload): Replace declaration with that of...
10747 (make_pass_if_after_reload): ...new function.
10748 (pass_regrename): Replace declaration with that of...
10749 (make_pass_regrename): ...new function.
10750 (pass_cprop_hardreg): Replace declaration with that of...
10751 (make_pass_cprop_hardreg): ...new function.
10752 (pass_reorder_blocks): Replace declaration with that of...
10753 (make_pass_reorder_blocks): ...new function.
10754 (pass_branch_target_load_optimize2): Replace declaration with that
10756 (make_pass_branch_target_load_optimize2): ...new function.
10757 (pass_leaf_regs): Replace declaration with that of...
10758 (make_pass_leaf_regs): ...new function.
10759 (pass_split_before_sched2): Replace declaration with that of...
10760 (make_pass_split_before_sched2): ...new function.
10761 (pass_compare_elim_after_reload): Replace declaration with that of...
10762 (make_pass_compare_elim_after_reload): ...new function.
10763 (pass_sched2): Replace declaration with that of...
10764 (make_pass_sched2): ...new function.
10765 (pass_stack_regs): Replace declaration with that of...
10766 (make_pass_stack_regs): ...new function.
10767 (pass_stack_regs_run): Replace declaration with that of...
10768 (make_pass_stack_regs_run): ...new function.
10769 (pass_df_finish): Replace declaration with that of...
10770 (make_pass_df_finish): ...new function.
10771 (pass_compute_alignments): Replace declaration with that of...
10772 (make_pass_compute_alignments): ...new function.
10773 (pass_duplicate_computed_gotos): Replace declaration with that of...
10774 (make_pass_duplicate_computed_gotos): ...new function.
10775 (pass_variable_tracking): Replace declaration with that of...
10776 (make_pass_variable_tracking): ...new function.
10777 (pass_free_cfg): Replace declaration with that of...
10778 (make_pass_free_cfg): ...new function.
10779 (pass_machine_reorg): Replace declaration with that of...
10780 (make_pass_machine_reorg): ...new function.
10781 (pass_cleanup_barriers): Replace declaration with that of...
10782 (make_pass_cleanup_barriers): ...new function.
10783 (pass_delay_slots): Replace declaration with that of...
10784 (make_pass_delay_slots): ...new function.
10785 (pass_split_for_shorten_branches): Replace declaration with that of...
10786 (make_pass_split_for_shorten_branches): ...new function.
10787 (pass_split_before_regstack): Replace declaration with that of...
10788 (make_pass_split_before_regstack): ...new function.
10789 (pass_convert_to_eh_region_ranges): Replace declaration with that
10791 (make_pass_convert_to_eh_region_ranges): ...new function.
10792 (pass_shorten_branches): Replace declaration with that of...
10793 (make_pass_shorten_branches): ...new function.
10794 (pass_set_nothrow_function_flags): Replace declaration with that of...
10795 (make_pass_set_nothrow_function_flags): ...new function.
10796 (pass_dwarf2_frame): Replace declaration with that of...
10797 (make_pass_dwarf2_frame): ...new function.
10798 (pass_final): Replace declaration with that of...
10799 (make_pass_final): ...new function.
10800 (pass_rtl_seqabstr): Replace declaration with that of...
10801 (make_pass_rtl_seqabstr): ...new function.
10802 (pass_release_ssa_names): Replace declaration with that of...
10803 (make_pass_release_ssa_names): ...new function.
10804 (pass_early_inline): Replace declaration with that of...
10805 (make_pass_early_inline): ...new function.
10806 (pass_inline_parameters): Replace declaration with that of...
10807 (make_pass_inline_parameters): ...new function.
10808 (pass_update_address_taken): Replace declaration with that of...
10809 (make_pass_update_address_taken): ...new function.
10810 (pass_convert_switch): Replace declaration with that of...
10811 (make_pass_convert_switch): ...new function.
10812 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10813 to a subclass of simple_ipa_opt_pass along with...
10814 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10815 (make_pass_ipa_tree_profile): ...new function.
10816 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10817 subclass of gimple_opt_pass along with...
10818 (pass_data_sra_early): ...new pass_data instance and...
10819 (make_pass_sra_early): ...new function.
10820 (pass_sra): Convert from a global struct to a subclass of
10821 gimple_opt_pass along with...
10822 (pass_data_sra): ...new pass_data instance and...
10823 (make_pass_sra): ...new function.
10824 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10825 gimple_opt_pass along with...
10826 (pass_data_early_ipa_sra): ...new pass_data instance and...
10827 (make_pass_early_ipa_sra): ...new function.
10828 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10829 subclass of gimple_opt_pass along with...
10830 (pass_data_ccp): ...new pass_data instance and...
10831 (make_pass_ccp): ...new function.
10832 (pass_fold_builtins): Convert from a global struct to a subclass of
10833 gimple_opt_pass along with...
10834 (pass_data_fold_builtins): ...new pass_data instance and...
10835 (make_pass_fold_builtins): ...new function.
10836 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10837 subclass of gimple_opt_pass along with...
10838 (pass_data_copy_prop): ...new pass_data instance and...
10839 (make_pass_copy_prop): ...new function.
10840 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10841 global struct to a subclass of gimple_opt_pass along with...
10842 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10843 (make_pass_rename_ssa_copies): ...new function.
10844 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10845 subclass of gimple_opt_pass along with...
10846 (pass_data_dce): ...new pass_data instance and...
10847 (make_pass_dce): ...new function.
10848 (pass_dce_loop): Convert from a global struct to a subclass of
10849 gimple_opt_pass along with...
10850 (pass_data_dce_loop): ...new pass_data instance and...
10851 (make_pass_dce_loop): ...new function.
10852 (pass_cd_dce): Convert from a global struct to a subclass of
10853 gimple_opt_pass along with...
10854 (pass_data_cd_dce): ...new pass_data instance and...
10855 (make_pass_cd_dce): ...new function.
10856 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10857 subclass of gimple_opt_pass along with...
10858 (pass_data_dominator): ...new pass_data instance and...
10859 (make_pass_dominator): ...new function.
10860 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10861 gimple_opt_pass along with...
10862 (pass_data_phi_only_cprop): ...new pass_data instance and...
10863 (make_pass_phi_only_cprop): ...new function.
10864 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10865 subclass of gimple_opt_pass along with...
10866 (pass_data_dse): ...new pass_data instance and...
10867 (make_pass_dse): ...new function.
10868 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10869 a subclass of gimple_opt_pass along with...
10870 (pass_data_forwprop): ...new pass_data instance and...
10871 (make_pass_forwprop): ...new function.
10872 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10873 struct to a subclass of gimple_opt_pass along with...
10874 (pass_data_tree_ifcombine): ...new pass_data instance and...
10875 (make_pass_tree_ifcombine): ...new function.
10876 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10877 subclass of gimple_opt_pass along with...
10878 (pass_data_ch): ...new pass_data instance and...
10879 (make_pass_ch): ...new function.
10880 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10881 subclass of gimple_opt_pass along with...
10882 (pass_data_tree_loop): ...new pass_data instance and...
10883 (make_pass_tree_loop): ...new function.
10884 (pass_tree_loop_init): Convert from a global struct to a subclass of
10885 gimple_opt_pass along with...
10886 (pass_data_tree_loop_init): ...new pass_data instance and...
10887 (make_pass_tree_loop_init): ...new function.
10888 (pass_lim): Convert from a global struct to a subclass of
10889 gimple_opt_pass along with...
10890 (pass_data_lim): ...new pass_data instance and...
10891 (make_pass_lim): ...new function.
10892 (pass_tree_unswitch): Convert from a global struct to a subclass of
10893 gimple_opt_pass along with...
10894 (pass_data_tree_unswitch): ...new pass_data instance and...
10895 (make_pass_tree_unswitch): ...new function.
10896 (pass_predcom): Convert from a global struct to a subclass of
10897 gimple_opt_pass along with...
10898 (pass_data_predcom): ...new pass_data instance and...
10899 (make_pass_predcom): ...new function.
10900 (pass_vectorize): Convert from a global struct to a subclass of
10901 gimple_opt_pass along with...
10902 (pass_data_vectorize): ...new pass_data instance and...
10903 (make_pass_vectorize): ...new function.
10904 (pass_graphite): Convert from a global struct to a subclass of
10905 gimple_opt_pass along with...
10906 (pass_data_graphite): ...new pass_data instance and...
10907 (make_pass_graphite): ...new function.
10908 (pass_graphite_transforms): Convert from a global struct to a subclass
10909 of gimple_opt_pass along with...
10910 (pass_data_graphite_transforms): ...new pass_data instance and...
10911 (make_pass_graphite_transforms): ...new function.
10912 (pass_check_data_deps): Convert from a global struct to a subclass of
10913 gimple_opt_pass along with...
10914 (pass_data_check_data_deps): ...new pass_data instance and...
10915 (make_pass_check_data_deps): ...new function.
10916 (pass_iv_canon): Convert from a global struct to a subclass of
10917 gimple_opt_pass along with...
10918 (pass_data_iv_canon): ...new pass_data instance and...
10919 (make_pass_iv_canon): ...new function.
10920 (pass_scev_cprop): Convert from a global struct to a subclass of
10921 gimple_opt_pass along with...
10922 (pass_data_scev_cprop): ...new pass_data instance and...
10923 (make_pass_scev_cprop): ...new function.
10924 (pass_record_bounds): Convert from a global struct to a subclass of
10925 gimple_opt_pass along with...
10926 (pass_data_record_bounds): ...new pass_data instance and...
10927 (make_pass_record_bounds): ...new function.
10928 (pass_complete_unroll): Convert from a global struct to a subclass of
10929 gimple_opt_pass along with...
10930 (pass_data_complete_unroll): ...new pass_data instance and...
10931 (make_pass_complete_unroll): ...new function.
10932 (pass_complete_unrolli): Convert from a global struct to a subclass of
10933 gimple_opt_pass along with...
10934 (pass_data_complete_unrolli): ...new pass_data instance and...
10935 (make_pass_complete_unrolli): ...new function.
10936 (pass_parallelize_loops): Convert from a global struct to a subclass
10937 of gimple_opt_pass along with...
10938 (pass_data_parallelize_loops): ...new pass_data instance and...
10939 (make_pass_parallelize_loops): ...new function.
10940 (pass_loop_prefetch): Convert from a global struct to a subclass of
10941 gimple_opt_pass along with...
10942 (pass_data_loop_prefetch): ...new pass_data instance and...
10943 (make_pass_loop_prefetch): ...new function.
10944 (pass_iv_optimize): Convert from a global struct to a subclass of
10945 gimple_opt_pass along with...
10946 (pass_data_iv_optimize): ...new pass_data instance and...
10947 (make_pass_iv_optimize): ...new function.
10948 (pass_tree_loop_done): Convert from a global struct to a subclass of
10949 gimple_opt_pass along with...
10950 (pass_data_tree_loop_done): ...new pass_data instance and...
10951 (make_pass_tree_loop_done): ...new function.
10952 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10953 struct to a subclass of gimple_opt_pass along with...
10954 (pass_data_cse_reciprocals): ...new pass_data instance and...
10955 (make_pass_cse_reciprocals): ...new function.
10956 (pass_cse_sincos): Convert from a global struct to a subclass of
10957 gimple_opt_pass along with...
10958 (pass_data_cse_sincos): ...new pass_data instance and...
10959 (make_pass_cse_sincos): ...new function.
10960 (pass_optimize_bswap): Convert from a global struct to a subclass of
10961 gimple_opt_pass along with...
10962 (pass_data_optimize_bswap): ...new pass_data instance and...
10963 (make_pass_optimize_bswap): ...new function.
10964 (pass_optimize_widening_mul): Convert from a global struct to a
10965 subclass of gimple_opt_pass along with...
10966 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10967 (make_pass_optimize_widening_mul): ...new function.
10968 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10969 subclass of gimple_opt_pass along with...
10970 (pass_data_phiopt): ...new pass_data instance and...
10971 (make_pass_phiopt): ...new function.
10972 (pass_cselim): Convert from a global struct to a subclass of
10973 gimple_opt_pass along with...
10974 (pass_data_cselim): ...new pass_data instance and...
10975 (make_pass_cselim): ...new function.
10976 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10977 subclass of gimple_opt_pass along with...
10978 (pass_data_phiprop): ...new pass_data instance and...
10979 (make_pass_phiprop): ...new function.
10980 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10981 subclass of gimple_opt_pass along with...
10982 (pass_data_pre): ...new pass_data instance and...
10983 (make_pass_pre): ...new function.
10984 (pass_fre): Convert from a global struct to a subclass of
10985 gimple_opt_pass along with...
10986 (pass_data_fre): ...new pass_data instance and...
10987 (make_pass_fre): ...new function.
10988 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10989 subclass of gimple_opt_pass along with...
10990 (pass_data_reassoc): ...new pass_data instance and...
10991 (make_pass_reassoc): ...new function.
10992 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10993 subclass of gimple_opt_pass along with...
10994 (pass_data_sink_code): ...new pass_data instance and...
10995 (make_pass_sink_code): ...new function.
10996 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10997 subclass of gimple_opt_pass along with...
10998 (pass_data_strlen): ...new pass_data instance and...
10999 (make_pass_strlen): ...new function.
11000 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11001 struct to a subclass of gimple_opt_pass along with...
11002 (pass_data_build_alias): ...new pass_data instance and...
11003 (make_pass_build_alias): ...new function.
11004 (pass_build_ealias): Convert from a global struct to a subclass of
11005 gimple_opt_pass along with...
11006 (pass_data_build_ealias): ...new pass_data instance and...
11007 (make_pass_build_ealias): ...new function.
11008 (pass_ipa_pta): Convert from a global struct to a subclass of
11009 simple_ipa_opt_pass along with...
11010 (pass_data_ipa_pta): ...new pass_data instance and...
11011 (make_pass_ipa_pta): ...new function.
11012 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11013 subclass of gimple_opt_pass along with...
11014 (pass_data_uncprop): ...new pass_data instance and...
11015 (make_pass_uncprop): ...new function.
11016 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11017 global struct to a subclass of gimple_opt_pass along with...
11018 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11019 (make_pass_late_warn_uninitialized): ...new function.
11020 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11021 to a subclass of gimple_opt_pass along with...
11022 (pass_data_init_datastructures): ...new pass_data instance and...
11023 (make_pass_init_datastructures): ...new function.
11024 (pass_early_warn_uninitialized): Convert from a global struct to a
11025 subclass of gimple_opt_pass along with...
11026 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11027 (make_pass_early_warn_uninitialized): ...new function.
11028 (pass_update_address_taken): Convert from a global struct to a
11029 subclass of gimple_opt_pass along with...
11030 (pass_data_update_address_taken): ...new pass_data instance and...
11031 (make_pass_update_address_taken): ...new function.
11032 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11033 struct to a subclass of gimple_opt_pass along with...
11034 (pass_data_release_ssa_names): ...new pass_data instance and...
11035 (make_pass_release_ssa_names): ...new function.
11036 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11037 subclass of gimple_opt_pass along with...
11038 (pass_data_stdarg): ...new pass_data instance and...
11039 (make_pass_stdarg): ...new function.
11040 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11041 global struct to a subclass of gimple_opt_pass along with...
11042 (pass_data_convert_switch): ...new pass_data instance and...
11043 (make_pass_convert_switch): ...new function.
11044 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11045 to a subclass of gimple_opt_pass along with...
11046 (pass_data_tail_recursion): ...new pass_data instance and...
11047 (make_pass_tail_recursion): ...new function.
11048 (pass_tail_calls): Convert from a global struct to a subclass of
11049 gimple_opt_pass along with...
11050 (pass_data_tail_calls): ...new pass_data instance and...
11051 (make_pass_tail_calls): ...new function.
11052 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11053 struct to a subclass of gimple_opt_pass along with...
11054 (pass_data_lower_vector): ...new pass_data instance and...
11055 (make_pass_lower_vector): ...new function.
11056 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11057 gimple_opt_pass along with...
11058 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11059 (make_pass_lower_vector_ssa): ...new function.
11060 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11061 to a subclass of gimple_opt_pass along with...
11062 (pass_data_slp_vectorize): ...new pass_data instance and...
11063 (make_pass_slp_vectorize): ...new function.
11064 (pass_ipa_increase_alignment): Convert from a global struct to a
11065 subclass of simple_ipa_opt_pass along with...
11066 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11067 (make_pass_ipa_increase_alignment): ...new function.
11068 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11069 gimple_opt_pass along with...
11070 (pass_data_vrp): ...new pass_data instance and...
11071 (make_pass_vrp): ...new function.
11072 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11073 subclass of simple_ipa_opt_pass along with...
11074 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11075 (make_pass_ipa_free_lang_data): ...new function.
11076 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11077 gimple_opt_pass along with...
11078 (pass_data_tsan): ...new pass_data instance and...
11079 (make_pass_tsan): ...new function.
11080 (pass_tsan_O0): Convert from a global struct to a subclass of
11081 gimple_opt_pass along with...
11082 (pass_data_tsan_O0): ...new pass_data instance and...
11083 (make_pass_tsan_O0): ...new function.
11084 * var-tracking.c (pass_variable_tracking): Convert from a global
11085 struct to a subclass of rtl_opt_pass along with...
11086 (pass_data_variable_tracking): ...new pass_data instance and...
11087 (make_pass_variable_tracking): ...new function.
11088 * web.c (pass_web): Convert from a global struct to a subclass of
11089 rtl_opt_pass along with...
11090 (pass_data_web): ...new pass_data instance and...
11091 (make_pass_web): ...new function.
11092 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11093 declaration with that of...
11094 (make_pass_mode_switch_use): ...new function.
11095 (pass_resolve_sw_modes): Replace declaration with that of...
11096 (make_pass_resolve_sw_modes): ...new function.
11097 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11098 from a global struct to a subclass of rtl_opt_pass along with...
11099 (pass_data_mode_switch_use): ...new pass_data instance and...
11100 (make_pass_mode_switch_use): ...new function.
11101 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11102 from a global struct to a subclass of rtl_opt_pass along with...
11103 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11104 (make_pass_resolve_sw_modes): ...new function.
11105 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11106 struct to a subclass of rtl_opt_pass along with...
11107 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11108 (make_pass_insert_vzeroupper): ...new function.
11109 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11110 global struct to a subclass of rtl_opt_pass along with...
11111 (pass_data_work_around_errata): ...new pass_data instance and...
11112 (make_pass_work_around_errata): ...new function.
11113 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11114 struct to a subclass of rtl_opt_pass along with...
11115 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11116 (make_pass_mips_machine_reorg2): ...new function.
11118 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11120 * passes.c (pass_manager::operator new): New.
11122 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11124 Handwritten part of conversion of passes to C++ classes.
11126 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11127 (toplev.o): Add dep on PASS_MANAGER_H.
11128 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11129 of early local pases to reflect this moving from a global to a
11130 member of gcc::pass_manager.
11131 (cgraph_add_new_function): Likewise.
11132 * lto-cgraph.c (lto_output_node): Update for conversion of
11133 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11134 * passes.c (opt_pass::clone): New.
11135 (opt_pass::gate): New.
11136 (opt_pass::execute): New.
11137 (opt_pass::opt_pass): New.
11138 (pass_manager::execute_early_local_passes): New.
11139 (pass_manager::execute_pass_mode_switching): new.
11140 (finish_optimization_passes): Convert to...
11141 (pass_manager::finish_optimization_passes): ...this.
11142 (finish_optimization_passes): Update for conversion of passes to
11144 (register_dump_files_1): Use has_gate since we cannot portably
11145 check a vtable entry against NULL.
11146 (dump_one_pass): Likewise.
11147 (ipa_write_summaries_2): Likewise.
11148 (ipa_write_optimization_summaries_1): Likewise.
11149 (ipa_read_summaries_1): Likewise.
11150 (ipa_read_optimization_summaries_1): Likewise.
11151 (execute_ipa_stmt_fixups): Likewise.
11152 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11153 pass-creation functions rather than wiring up globals, and
11154 storing the results in fields of pass_manager generated using
11155 pass-instances.def.
11156 (pass_manager::dump_profile_report): Update for conversion of
11157 passes to C++ classes.
11158 (pass_manager::execute_ipa_summary_passes): Likewise.
11159 (execute_one_ipa_transform_pass): Likewise.
11160 (execute_one_pass): Use has_gate and has_execute since we cannot
11161 portably check a vtable entry against NULL.
11162 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11163 (pass_manager): Use pass-instances.def to add fields for the
11164 various pass instances.
11165 * toplev.c (finalize): Update for move of
11166 finish_optimization_passes to a method of gcc::pass_manager.
11167 * toplev.h (finish_optimization_passes): Move to method of class
11169 * tree-pass.h (struct pass_data): New.
11170 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11171 (opt_pass::gate): Convert to virtual function.
11172 (opt_pass::~opt_pass): New.
11173 (opt_pass::clone): New.
11174 (opt_pass::execute): Convert to virtual function.
11175 (opt_pass::opt_pass): New.
11176 (opt_pass::ctxt_): new.
11177 (gimple_opt_pass): Convert to subclass of opt_pass.
11178 (gimple_opt_pass::gimple_opt_pass): New.
11179 (rtl_opt_pass): Convert to subclass of opt_pass.
11180 (rtl_opt_pass::rtl_opt_pass): New.
11181 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11182 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11183 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11184 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11185 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11186 invocation of pass_mode_switching to reflect this moving from a
11187 global to a member of gcc::pass_manager.
11188 (ix86_option_override): Rework how pass_insert_vzeroupper is
11189 added to the pass_manager to reflect autogenerated changes.
11190 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11192 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11194 PR rtl-optimization/57708
11195 * recog.c (peep2_find_free_register): Validate all regs in a
11198 2013-08-05 Jan Hubicka <jh@suse.cz>
11201 * cgraph.c (verify_cgraph_node): Accept local flags from other
11203 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11204 (function_and_variable_visibility): Likewise.
11205 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11207 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11209 * graph.c (init_graph_slim_pretty_print): Remove.
11210 (print_graph_cfg): Do not call it. Use local pretty printer.
11211 (start_graph_dump): Likewise.
11213 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11215 * gimple-pretty-print.c (buffer): Remove.
11216 (initialized): Likewise.
11217 (maybe_init_pretty_print): Likewise.
11218 (print_gimple_stmt): Do not call it. Use non-static local
11219 pretty_printer variable.
11220 (print_gimple_expr): Likewise.
11221 (print_gimple_seq): Likewise.
11222 (gimple_dump_bb): Likewise.
11224 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11226 * asan.c (asan_pp): Remove.
11227 (asan_pp_initialized): Likewise.
11228 (asan_pp_initialize): Likewise.
11229 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11230 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11231 (asan_add_global): Likewise.
11233 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11235 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11236 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11237 * pretty-print.c (pp_formatted_text_data): Likewise.
11238 (pp_indent): Rename from pp_base_indent.
11239 (pp_format): Rename from pp_base_format.
11240 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11241 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11242 (pp_flush): Rename from pp_base_flush.
11243 (pp_set_line_maximum_length): Rename from
11244 pp_base_set_line_maximum_length.
11245 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11246 (pp_set_prefix): Rename from pp_base_set_prefix.
11247 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11248 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11249 (pp_append_text): Rename from pp_base_append_text.
11250 (pp_formatted_text): Rename from pp_base_formatted_text.
11251 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11252 (pp_remaining_character_count_for_line): Rename from
11253 pp_base_remaining_character_count_for_line.
11254 (pp_newline): Rename from pp_base_newline.
11255 (pp_character): Rename from pp_base_character.
11256 (pp_string): Rename from pp_base_string.
11257 (pp_maybe_space): Rename from pp_base_maybe_space.
11258 * asan.c (asan_pp_string): Adjust.
11259 (asan_emit_stack_protection): Likewise.
11260 (asan_add_global): Likewise.
11261 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11262 * tree-mudflap.c (mf_varname_tree): Likewise.
11263 * tree-pretty-print.c (pp_tree_identifier): Rename from
11264 pp_base_tree_identifier.
11265 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11266 Declare as function.
11268 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11270 * pretty-print.h (pp_bar_bar): New.
11271 (pp_ampersand_ampersand): Likewise.
11272 (pp_less_equal): Likewise.
11273 (pp_greater_equal): Likewise.
11274 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11275 printer functions instead of pp_string or operators and punctuators.
11276 (dump_gimple_call): Likewise.
11277 (dump_gimple_omp_for): Likewise.
11278 (dump_gimple_transaction): Likewise.
11279 (dump_gimple_phi): Likewise.
11280 (pp_gimple_stmt_1): Likewise.
11281 * sched-vis.c (print_insn): Likewise.
11282 * tree-mudflap.c (mf_varname_tree): Likewise.
11283 * tree-pretty-print.c (dump_block_node): Likewise.
11284 (dump_generic_node): Likewise.
11286 2013-08-02 Jan Hubicka <jh@suse.cz>
11288 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11290 * lto-streamer-out.c (tree_is_indexable): Results decls and
11291 parm decls are not indexable.
11292 (DFS_write_tree_body): Do not follow args and results.
11293 (hash_tree): Likewise.
11294 (output_functions): Rearrange so struct function is needed
11295 only when real body is output; be able to also ouptut abstract
11296 functions; output DECL_ARGUMENTS and DECL_RESULT.
11297 (lto_output): When not in WPA, ale store abstract functions.
11298 (write_symbol): Do not care about RESULT_DECL.
11299 (output_symbol_p): Handle correctly sbtract decls.
11300 * lto-streamer-in.c (input_function): Rearrange so struct
11301 function can be NULL at entry; allow streaming of
11302 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11303 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11304 sanity check during LTO.
11305 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11306 RESULT_DECl and DECL_ARGUMENTS.
11307 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11310 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11312 * pretty-print.h (pp_underscore): New.
11314 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11315 printer functions instead of pp_character.
11316 (dump_binary_rhs): Likewise.
11317 (dump_ternary_rhs): Likewise.
11318 (dump_gimple_call_args): Likewise.
11319 (pp_points_to_solution): Likewise.
11320 (dump_gimple_call): Likewise.
11321 (dump_gimple_switch): Likewise.
11322 (dump_gimple_cond): Likewise.
11323 (dump_gimple_bind): Likewise.
11324 (dump_gimple_try): Likewise.
11325 (dump_gimple_omp_for): Likewise.
11326 (dump_gimple_omp_continue): Likewise.
11327 (dump_gimple_omp_single): Likewise.
11328 (dump_gimple_omp_sections): Likewise.
11329 (dump_gimple_omp_block): Likewise.
11330 (dump_gimple_omp_critical): Likewise.
11331 (dump_gimple_transaction): Likewise.
11332 (dump_gimple_asm): Likewise.
11333 (dump_gimple_phi): Likewise.
11334 (dump_gimple_omp_parallel): Likewise.
11335 (dump_gimple_omp_task): Likewise.
11336 (dump_gimple_omp_atomic_load): Likewise.
11337 (dump_gimple_omp_atomic_store): Likewise.
11338 (dump_gimple_mem_ops): Likewise.
11339 (pp_gimple_stmt_1): Likewise.
11340 (pp_cfg_jump): Likewise.
11341 (dump_implicit_edges): Likewise.
11342 (gimple_dump_bb_for_graph): Likewise.
11343 * graph.c (draw_cfg_node): Likewise.
11344 * langhooks.c (lhd_print_error_function): Likewise.
11345 * sched-vis.c (print_exp): Likewise.
11346 (print_value): Likewise.
11347 (print_pattern): Likewise.
11348 (print_insn): Likewise.
11349 (rtl_dump_bb_for_graph): Likewise.
11350 * tree-pretty-print.c (dump_function_declaration): Likewise.
11351 (dump_array_domain): Likewise.
11352 (dump_omp_clause): Likewise.
11353 (dump_location): Likewise.
11354 (dump_generic_node): Likewise.
11355 (print_struct_decl): Likewise.
11356 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11358 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11360 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11361 candidate table when replacing a candidate statement.
11362 (replace_rhs_if_not_dup): Likewise.
11363 (replace_one_candidate): Likewise.
11365 2013-08-02 Jan Hubicka <jh@suse.cz>
11366 Martin Liska <marxin.liska@gmail.com>
11368 * cgraphunit.c (add_new_function): Fix logic when adding from
11370 (assemble_thunk): Rename to ...
11371 (expand_thunk); .. this one; export; get it working with
11372 general functions; make produced gimple valid.
11373 * cgraph.h (expand_thunk): Declare.
11375 2013-08-02 Jan Hubicka <jh@suse.cz>
11377 * ipa-cp.c (gather_context_independent_values): Use
11378 ipa_get_param_move_cost.
11379 (get_replacement_map): Remove PARAM; move parameter folding
11381 (create_specialized_node): Update.
11382 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11383 assert that we have gimple body; update move_cost.
11384 (count_formal_params): Assert that we have gimple body.
11385 (ipa_dump_param): New function.
11386 (ipa_alloc_node_params): Break out from ...
11387 (ipa_initialize_node_params): ... here.
11388 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11389 (ipa_write_node_info): Stream move costs.
11390 (ipa_read_node_info): Read move costs.
11391 (ipa_update_after_lto_read): Do not recompute node params.
11392 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11393 (ipa_get_param): Check we are not in WPA.
11394 (ipa_get_param_move_cost): New.
11395 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11396 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11397 parm numbers to be present.
11399 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11401 PR rtl-optimization/58048
11402 * lra-constraints.c (process_alt_operands): Don't check asm
11403 operand on register.
11405 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11407 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11408 the implied StoreLoad barrier for atomic operations if before.
11410 2013-08-02 Jan Hubicka <jh@suse.cz>
11411 Martin Liska <marxin.liska@gmail.com>
11413 * cgraph.c (cgraph_function_body_availability): Do not check
11415 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11416 symtab_node_availability): Declare.
11417 * ipa.c (can_replace_by_local_alias): New.
11418 (function_and_variable_visibility): Use it.
11419 * symtab.c (symtab_for_node_and_aliases,
11420 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11422 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11424 PR rtl-optimization/57963
11425 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11426 (lra_constraints): Use them.
11428 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11430 * config/arm/types.md (define_attr "type"): Add "load_acq"
11432 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11434 (cortex_a53_store1): Likewise.
11436 2013-08-01 Jan Hubicka <jh@suse.cz>
11438 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11439 partitions are not needed.
11441 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11443 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11444 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11445 MAYBE_NON_Q_CLASS_P where appropriate.
11447 2013-08-01 Jan Hubicka <jh@suse.cz>
11449 * cgraph.h (release_function_body): Declare.
11450 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11451 of unused delcarations.
11453 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11455 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11456 RTL form when subtracting a constant.
11458 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11460 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11461 Generate canonical plus rtx with negated immediate instead of minus
11463 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11465 2013-08-01 Jan Hubicka <jh@suse.cz>
11467 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11468 (cgraph_release_function_body): Likewise.
11469 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11470 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11471 to used_as_abstract_origin.
11472 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11473 symbols used as abstract origins.
11474 * cgraphunit.c (analyze_functions): Update.
11475 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11476 used_as_abstract_origin.
11477 * tree-inline.c (tree_function_versioning): Update
11478 used_as_abstract_origin; be ready for DECL_RESULT and
11479 DECL_ARGUMENTS to be NULL.
11481 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11482 for abstract functions.
11483 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11486 2013-08-01 Jan Hubicka <jh@suse.cz>
11488 * profile.c (compute_value_histograms): Fix thinko.
11490 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11492 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11493 aarch-common-protos.h to extra_headers.
11494 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11495 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11496 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11498 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11500 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11501 Include "../arm/types.md". Define "type" attribute for all patterns.
11502 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11505 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11507 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11508 to support power8 load fusion.
11509 (fusion_gpr_mem_load): Likewise.
11511 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11513 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11514 declarations for power8 load fusion.
11515 (emit_fusion_gpr_load): Likewise.
11517 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11518 tuning for power8, turn on fusion mode by default. Turn on sign
11519 extending fusion mode if normal fusion mode is on, and we are at
11521 (fusion_gpr_load_p): New function, return true if we can fuse an
11522 addis instruction with a dependent load to a GPR.
11523 (emit_fusion_gpr_load): Emit the instructions for power8 load
11526 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11527 (VSX load fusion peepholes): New peepholes to fuse together an
11528 addi instruction with a VSX load instruction.
11530 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11531 peepholes to fuse an addis instruction with a load to a GPR base
11532 register. If we are supporting sign extending fusions, convert
11533 sign extending loads to zero extending loads and add an explicit
11536 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11538 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11539 aarch-common-protos.h to extra_headers.
11540 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11541 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11542 (arm_early_store_addr_dep): Likewise.
11543 (arm_no_early_alu_shift_dep): Likewise.
11544 (arm_no_early_alu_shift_value_dep): Likewise.
11545 (arm_no_early_mul_dep): Likewise.
11546 (arm_no_early_store_addr_dep): Likewise.
11547 (arm_mac_accumulator_is_mul_result): Likewise.
11548 (arm_mac_accumulator_is_result): Likewise.
11549 * config/arm/aarch-common.c: ... here. New file.
11550 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11552 (arm_early_store_addr_dep): Likewise.
11553 (arm_no_early_alu_shift_dep): Likewise.
11554 (arm_no_early_alu_shift_value_dep): Likewise.
11555 (arm_no_early_mul_dep): Likewise.
11556 (arm_no_early_store_addr_dep): Likewise.
11557 (arm_mac_accumulator_is_mul_result): Likewise.
11558 (arm_mac_accumulator_is_result): Likewise.
11559 * config/arm/aarch-common-protos.h: ... here. New file.
11560 * config/arm/t-arm (aarch-common.o): Define.
11562 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11564 * config/arm/arm.md: Include new file "types.md".
11565 (define_attr "type"): Move from here to ...
11566 (define_attr "mul32"): Likewise.
11567 (define_attr "mul64"): Likewise.
11568 * config/arm/types.md: ... here. New file.
11570 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11572 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11573 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11575 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11577 * gen-pass-instances.awk: Fix offset of substr().
11579 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11581 * Makefile.in (pass-instances.def): New.
11582 (passes.o): Replace dependency on passes.def with one on
11585 * gen-pass-instances.awk: New.
11587 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11588 rather than passes.def, updating local definition of NEXT_PASS
11589 macro to add an extra NUM parameter (currently unused).
11591 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11593 * Makefile.in (PASS_MANAGER_H): New.
11594 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11595 (passes.o): Likewise.
11596 (statistics.o): Likewise.
11597 (cgraphunit.o): Likewise.
11598 (context.o): Depend on PASS_MANAGER_H.
11600 * pass_manager.h: New.
11602 * cgraphunit.c (cgraph_add_new_function): Update for moves
11603 of globals to fields of pass_manager.
11604 (analyze_function): Likewise.
11605 (expand_function): Likewise.
11606 (ipa_passes): Likewise.
11607 (compile): Likewise.
11609 * context.c (context::context): New.
11610 * context.h (context::context): New.
11611 (context::get_passes): New.
11612 (context::passes_): New.
11614 * lto-cgraph.c (input_node): Update for moves of globals to
11615 fields of pass_manager.
11617 * passes.c (all_passes): Remove, in favor of a field of the
11618 same name within the new class pass_manager.
11619 (all_small_ipa_passes): Likewise.
11620 (all_lowering_passes): Likewise.
11621 (all_regular_ipa_passes): Likewise.
11622 (all_late_ipa_passes): Likewise.
11623 (all_lto_gen_passes): Likewise.
11624 (passes_by_id): Likewise.
11625 (passes_by_id_size): Likewise.
11626 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11627 the new class pass_manager.
11628 (set_pass_for_id): Convert to...
11629 (pass_manager::set_pass_for_id): ...method.
11630 (get_pass_for_id): Convert to...
11631 (pass_manager::get_pass_for_id): ...method.
11632 (register_one_dump_file): Move body of implementation into...
11633 (pass_manager::register_one_dump_file): ...here.
11634 (register_dump_files_1): Convert to...
11635 (pass_manager::register_dump_files_1): ...method.
11636 (register_dump_files): Convert to...
11637 (pass_manager::register_dump_files): ...method.
11638 (create_pass_tab): Update for moves of globals to fields of
11640 (dump_passes): Move body of implementation into...
11641 (pass_manager::dump_passes): ...here.
11642 (register_pass): Move body of implementation into...
11643 (pass_manager::register_pass): ...here.
11644 (init_optimization_passes): Convert into...
11645 (pass_manager::pass_manager): ...constructor for new
11646 pass_manager class, and initialize the pass_lists array.
11647 (check_profile_consistency): Update for moves of globals to
11648 fields of pass_manager.
11649 (dump_profile_report): Move body of implementation into...
11650 (pass_manager::dump_profile_report): ...here.
11651 (ipa_write_summaries_1): Update for moves of pass lists from
11652 being globals to fields of pass_manager.
11653 (ipa_write_optimization_summaries): Likewise.
11654 (ipa_read_summaries): Likewise.
11655 (ipa_read_optimization_summaries): Likewise.
11656 (execute_all_ipa_stmt_fixups): Likewise.
11658 * statistics.c (statistics_fini): Update for moves of globals to
11659 fields of pass_manager.
11661 * toplev.c (general_init): Replace call to
11662 init_optimization_passes with construction of the pass_manager
11665 * tree-pass.h (all_passes): Remove, in favor of a field of the
11666 same name within the new class pass_manager.
11667 (all_small_ipa_passes): Likewise.
11668 (all_lowering_passes): Likewise.
11669 (all_regular_ipa_passes): Likewise.
11670 (all_lto_gen_passes): Likewise.
11671 (all_late_ipa_passes): Likewise.
11672 (passes_by_id): Likewise.
11673 (passes_by_id_size): Likewise.
11674 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11675 the new class pass_manager.
11676 (get_pass_for_id): Remove.
11678 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11680 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11683 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11685 * arm.md (mulhi3): New expand pattern.
11687 2013-07-30 Jan Hubicka <jh@suse.cz>
11688 Martin Liska <marxin.liska@gmail.com>
11690 * profile.c (compute_value_histograms): Do not ICE when
11691 there is mismatch only on some counters.
11693 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11695 PR rtl-optimization/57637
11696 * function.c (move_insn_for_shrink_wrap): Also check the
11697 GEN set of the LIVE problem for the liveness analysis
11698 if it exists, otherwise give up.
11700 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11702 PR tree-optimization/57993
11703 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11704 replaced statement in the candidate table.
11705 (phi_add_costs): Return infinite cost when the hidden basis does
11706 not dominate all phis on which the candidate is dependent.
11707 (replace_one_candidate): Record replaced statement in the
11710 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11712 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11713 (ashlv2si3): New expander.
11714 (*ashlv2si3_i): New define_insn_and_split.
11715 * predicates.md (float_operation): Allow patterns with three
11716 basic sub-patterns.
11718 PR rtl-optimization/58021
11719 * mode-switching.c (create_pre_exit): Always split off preceding
11720 insns if we are not at the basic block head.
11722 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11724 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11725 (UCLIBC_DYNAMIC_LINKER): New macro.
11726 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11728 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11729 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11731 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11732 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11733 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11734 for SF and DF modes. Use ieee_quad_format for TF mode.
11735 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11736 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11737 (mips_option_override): Handle `-mnan=legacy'.
11738 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11739 `-mabs=2008' and `-mnan=2008'.
11740 (OPTION_DEFAULT_SPECS): Add "nan" default.
11741 (ASM_SPEC): Handle `-mnan='.
11742 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11743 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11744 comment accordingly.
11745 (neg<mode>2): Likewise.
11746 * config/mips/mips.opt (mabs, mnan): New options.
11747 * doc/install.texi (Configuration): Document `--with-nan=' option.
11748 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11750 (MIPS Options): Document them.
11751 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11752 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11753 * configure: Regenerate.
11754 * config.in: Regenerate.
11756 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11758 * config/i386/i386.md (float post-reload splitters): Do not check
11759 for subregs of SSE registers.
11761 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11762 H.J. Lu <hongjiu.lu@intel.com>
11766 * config/i386/i386.md (post-reload splitter
11767 to avoid partial SSE reg dependency stalls): New pattern.
11769 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11771 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11773 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11775 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11776 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11778 2013-07-26 Cary Coutant <ccoutant@google.com>
11780 * dwarf2out.c (die_checksum_ordered): Don't include template
11781 instantiations in signature.
11782 (is_template_parameter): New function.
11783 (is_template_instantiation): New function.
11784 (generate_skeleton_bottom_up): Don't include template instantiations
11786 (generate_skeleton): Likewise.
11787 (break_out_comdat_types): Move recursive call to break out nested
11789 (prune_unused_types_mark_generic_parms_dies): Call
11790 is_template_parameter.
11792 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11794 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11795 uses vector registers.
11796 * config/aarch64/iterators.md: Add attributes rtn and vas.
11798 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11799 Richard Earnshaw <richard.earnshaw@arm.com>
11801 * combine.c (simplify_comparison): Re-canonicalize operands
11803 * config/arm/arm.md (movcond_addsi): New splitter.
11805 2013-07-25 Sterling Augustine <saugustine@google.com>
11807 * dwarf2out.c (size_of_pubnames): Move code to...
11808 (include_pubname_in_output): ...here. New.
11809 (want_pubnames): Rearrange.
11810 (output_pubnames): Call include_pubname_in_output. Move assertion.
11812 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11814 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11816 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11819 * doc/extend.texi: Remove obsolete builtins. Fix
11820 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11822 2013-07-25 Jan Hubicka <jh@suse.cz>
11824 * cgraph.c (release_function_body): Break out from ...
11825 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11826 and DECL_ARGUMENTS.
11827 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11828 old_tree in the map.
11829 (create_specialized_node): Update.
11830 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11832 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11833 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11834 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11835 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11838 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11840 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11841 addsi3_carryin_alt2_<optab>): Correct output template.
11843 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11845 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11846 Adjust for arm_restrict_it.
11847 Remove trailing whitespace.
11849 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11851 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11852 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11854 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11856 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11858 PR rtl-optimization/57960
11859 * lra-constraints.c (process_alt_operands): Use the right mode
11860 when checking strict_low.
11862 2013-07-25 Jan Hubicka <jh@suse.cz>
11864 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11865 * cgraph.c (cgraph_remove_node): Do not release function body
11866 when in cgraph streaming.
11867 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11868 in other partitions are not considered reachable; fix handling of
11871 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11873 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11875 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11880 * config/arm/arm.md ("*sibcall_insn): Replace use of
11881 Ss with US. Adjust output for v5 and v4t.
11882 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11884 * config/arm/constraints.md ("Ss"): Rename to US.
11886 2013-07-25 Terry Guo <terry.guo@arm.com>
11888 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11889 shift_add/shift_sub0/shift_sub1 RTXs.
11891 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11892 Anton Blanchard <anton@au1.ibm.com>
11894 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11895 (altivec_vpks<VI_char>ss): Likewise.
11896 (altivec_vpks<VI_char>us): Likewise.
11897 (altivec_vpku<VI_char>us): Likewise.
11898 (altivec_vpku<VI_char>um): Likewise.
11900 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11902 Introduce context class.
11904 * Makefile.in (CONTEXT_H): New.
11905 (OBJS): Add context.o.
11906 (toplev.o): Add CONTEXT_H to dependencies.
11909 * toplev.c (general_init): Create the singleton gcc::context instance.
11915 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11917 PR rtl-optimization/57968
11918 * mode-switching.c (create_pre_exit): Allow instructions that
11919 don't set a return register to need a non-exit mode.
11921 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11922 Anton Blanchard <anton@au1.ibm.com>
11924 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11925 operands to vperm for little endian.
11926 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11927 of lvsl to create the control mask for a vperm for little endian.
11929 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11930 Anton Blanchard <anton@au1.ibm.com>
11932 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11933 two operands for little-endian.
11935 2013-07-23 Steve Ellcey <sellcey@mips.com>
11937 * config/mips/mips.c (mips_case_values_threshold): New.
11938 (TARGET_CASE_VALUES_THRESHOLD): Define.
11940 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11941 Anton Blanchard <anton@au1.ibm.com>
11943 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11944 selection of field for vector splat in little endian mode.
11946 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11948 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11949 expanders to rs6000.md.
11950 (ior<mode>3): Likewise.
11951 (and<mode>3): Likewise.
11952 (one_cmpl<mode>2): Likewise.
11953 (nor<mode>3): Likewise.
11954 (andc<mode>3): Likewise.
11955 (eqv<mode>3): Likewise.
11956 (nand<mode>3): Likewise.
11957 (orc<mode>3): Likewise.
11959 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11962 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11963 to split multi-word logical operations.
11964 (rs6000_split_logical_di): Likewise.
11965 (rs6000_split_logical): Likewise.
11967 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11968 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11969 and allow TImode operations in 32-bit.
11970 (vsx_and<mode>3_64bit): Likewise.
11971 (vsx_ior<mode>3_32bit): Likewise.
11972 (vsx_ior<mode>3_64bit): Likewise.
11973 (vsx_xor<mode>3_32bit): Likewise.
11974 (vsx_xor<mode>3_64bit): Likewise.
11975 (vsx_one_cmpl<mode>2_32bit): Likewise.
11976 (vsx_one_cmpl<mode>2_64bit): Likewise.
11977 (vsx_nor<mode>3_32bit): Likewise.
11978 (vsx_nor<mode>3_64bit): Likewise.
11979 (vsx_andc<mode>3_32bit): Likewise.
11980 (vsx_andc<mode>3_64bit): Likewise.
11981 (vsx_eqv<mode>3_32bit): Likewise.
11982 (vsx_eqv<mode>3_64bit): Likewise.
11983 (vsx_nand<mode>3_32bit): Likewise.
11984 (vsx_nand<mode>3_64bit): Likewise.
11985 (vsx_orc<mode>3_32bit): Likewise.
11986 (vsx_orc<mode>3_64bit): Likewise.
11988 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11989 logical types in GPRs.
11991 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11992 logical insns to rs6000.md, and allow TImode operations in
11994 (altivec_ior<mode>3): Likewise.
11995 (altivec_xor<mode>3): Likewise.
11996 (altivec_one_cmpl<mode>2): Likewise.
11997 (altivec_nor<mode>3): Likewise.
11998 (altivec_andc<mode>3): Likewise.
12000 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12001 attributes for moving the 128-bit logical operations into
12003 (BOOL_REGS_OUTPUT): Likewise.
12004 (BOOL_REGS_OP1): Likewise.
12005 (BOOL_REGS_OP2): Likewise.
12006 (BOOL_REGS_UNARY): Likewise.
12007 (BOOL_REGS_AND_CR0): Likewise.
12008 (one_cmpl<mode>2): Add support for DI logical operations on
12009 32-bit, splitting the operations to 32-bit.
12010 (anddi3): Likewise.
12011 (iordi3): Likewise.
12012 (xordi3): Likewise.
12013 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12014 changes to combine the 32/64-bit code, allow logical operations on
12015 TI mode in 32-bit, and to use similar match_operator patterns like
12016 scalar mode uses. Combine the Altivec and VSX code for logical
12017 operations, and move it here.
12018 (ior<mode>3, 128-bit types): Likewise.
12019 (xor<mode>3, 128-bit types): Likewise.
12020 (one_cmpl<mode>3, 128-bit types): Likewise.
12021 (nor<mode>3, 128-bit types): Likewise.
12022 (andc<mode>3, 128-bit types): Likewise.
12023 (eqv<mode>3, 128-bit types): Likewise.
12024 (nand<mode>3, 128-bit types): Likewise.
12025 (orc<mode>3, 128-bit types): Likewise.
12026 (and<mode>3_internal): Likewise.
12027 (bool<mode>3_internal): Likewise.
12028 (boolc<mode>3_internal1): Likewise.
12029 (boolc<mode>3_internal2): Likewise.
12030 (boolcc<mode>3_internal1): Likewise.
12031 (boolcc<mode>3_internal2): Likewise.
12032 (eqv<mode>3_internal1): Likewise.
12033 (eqv<mode>3_internal2): Likewise.
12034 (one_cmpl1<mode>3_internal): Likewise.
12036 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12038 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12039 Rename flag_stack_usage to flag_stack_usage_info.
12041 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12043 * config/microblaze/sync.md: New file.
12044 * config/microblaze/microblaze.md: Include sync.md
12045 * config/microblaze/microblaze.c: Add print_operand 'y'.
12046 * config/microblaze/constraints.md: Add memory_contraint
12047 'Q' which is a single register.
12049 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12051 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12053 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12055 * reload.c (find_reloads): Exit loop once we find this operand
12056 cannot be reloaded somehow for this alternative.
12058 * reload.c (find_reloads): Exit loop once we find a hard register.
12060 * rtlanal.c (computed_jump_p): Exit loop once we find label
12063 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12065 * cfgloopmanip.c (remove_path): Exit loop after setting
12068 * gensupport.c (subst_dup): Avoid loop if code is not
12069 MATCH_DUP nor MATCH_OP_DUP.
12071 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12073 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12075 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12077 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12078 true for SP_REGNUM if mode == ptr_mode.
12079 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12080 with value R0_REGNUM + 31.
12082 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12084 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12085 pad pointer-typed argument downward.
12087 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12089 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12090 and __ILP32__ when the ILP32 model is in use.
12092 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12094 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12095 (aarch64_load_symref_appropriately): In the case of
12096 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12097 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12098 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12099 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12100 if 'mode' doesn't equal to 'ptr_mode'.
12101 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12102 'vcall_offset'; change to call aarch64_emit_move differently depending
12103 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12104 to calculate the upper bound of 'vcall_offset'.
12105 (aarch64_cannot_force_const_mem): Change to also return true if
12107 (aarch64_legitimize_reload_address): In the case of large
12108 displacements, add new local variable 'xmode' and an assertion
12109 based on it; change to use 'xmode' to generate the new rtx and
12111 (aarch64_asm_trampoline_template): Change to generate the template
12112 differently depending on TARGET_ILP32 or not; change to use
12113 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12114 (aarch64_trampoline_size): Removed.
12115 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12116 and replace immediate literals with it. Change to use 'ptr_mode'
12117 instead of 'DImode' and call convert_memory_address if the mode
12118 of 'fnaddr' doesn't equal to 'ptr_mode'.
12119 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12121 (aarch64_elf_asm_destructor): Likewise.
12122 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12123 on TARGET_ILP32 instead of aarch64_trampoline_size.
12124 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12125 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12126 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12127 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12128 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12129 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12130 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12131 depending on the value of 'mode'.
12132 (add_losym_<mode>): New.
12133 (ldr_got_small_<mode>): New, based on ldr_got_small.
12134 (ldr_got_small): Remove.
12135 (ldr_got_small_sidi): New.
12136 * config/aarch64/iterators.md (P): New.
12137 (PTR): Change to 'ptr_mode' in the condition.
12139 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12141 * config.gcc (aarch64*-*-*): Support --with-abi.
12142 (aarch64*-*-elf): Support --with-multilib-list.
12143 (aarch64*-*-linux*): Likewise.
12144 (supported_defaults): Add abi to aarch64*-*-*.
12145 * configure.ac: Mention AArch64 for --with-multilib-list.
12146 * configure: Re-generated.
12147 * config/aarch64/biarchilp32.h: New file.
12148 * config/aarch64/biarchlp64.h: New file.
12149 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12151 (MULTILIB_DEFAULTS): Ditto.
12152 (DRIVER_SELF_SPECS): Ditto.
12153 (ASM_SPEC): Update to also substitute -mabi.
12154 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12155 file whose name depends on -mabi= and -mbig-endian.
12156 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12158 (POINTER_SIZE): New define.
12159 (POINTERS_EXTEND_UNSIGNED): Ditto.
12160 (enum aarch64_abi_type): New enumeration tag.
12161 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12162 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12163 (TARGET_ILP32): New define.
12164 * config/aarch64/aarch64.opt (mabi): New.
12165 (aarch64_abi): New.
12166 (ilp32, lp64): New values for -mabi.
12167 * config/aarch64/t-aarch64 (comma): New define.
12168 (MULTILIB_OPTIONS): Ditto.
12169 (MULTILIB_DIRNAMES): Ditto.
12170 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12171 * doc/invoke.texi: Document -mabi for AArch64.
12173 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12175 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12177 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12178 Anton Blanchard <anton@au1.ibm.com>
12180 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12181 endianness when selecting field to splat.
12183 2013-07-22 Eric Christopher <echristo@gmail.com>
12185 * dwarf2out.c (die_odr_checksum): New function to use
12186 CHECKSUM_ macros and ULEB128 for DIE tag.
12187 (generate_type_signature): Use.
12189 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12191 * config.gcc (sparc*-*-*): Accept leon3 processor.
12192 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12193 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12194 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12195 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12196 (mfix-ut699): Adjust comment.
12197 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12198 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12199 (CPP_CPU_SPEC): Likewise.
12200 (ASM_CPU_SPEC): Likewise.
12201 * config/sparc/sparc.c (leon3_cost): New constant.
12202 (sparc_option_override): Add leon3 support.
12203 (mem_ref): New function.
12204 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12205 (sparc_do_work_around_errata): Look into the instruction in the delay
12206 slot and adjust accordingly. Add fix for the data cache nullify issues
12207 of the UT699. Change insertion position for the NOP.
12208 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12209 (leon3_load): New reservation.
12210 (leon_store): Bump latency to 2.
12211 (grfpu): New automaton.
12212 (grfpu_alu): New unit.
12213 (grfpu_ds): Likewise.
12214 (leon_fp_alu): Adjust.
12215 (leon_fp_mult): Delete.
12216 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12217 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12218 * config/sparc/sparc.md (cpu): Add leon3.
12219 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12220 (swapsi): Likewise.
12221 (atomic_test_and_set): Likewise.
12222 (ldstub): Likewise.
12224 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12226 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12227 default for R5900 targets.
12228 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12229 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12230 * config/mips/mips.c (mips_option_override): Report an error for
12231 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12232 for -march=r5900 -mhard-float.
12234 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12236 * df-problems.c (can_move_insns_across): Exit loop once we
12237 find a non-fixed, non-global register.
12239 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12242 * omega.c (omega_eliminate_red): Break after setting red_found.
12243 (omega_problem_has_red_equations): Similarly after setting found.
12244 (omega_query_variable): Similarly after setting coupled.
12246 2013-07-22 Marek Polacek <polacek@redhat.com>
12248 * gimplify.c: Don't include gimple.h twice.
12250 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12252 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12253 instead of TARGET_THUMB1.
12254 (Pz): New constraint.
12255 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12257 (compare_negsi_si): Likewise.
12258 (compare_addsi2_op0): Likewise.
12259 (compare_addsi2_op1): Likewise.
12260 (addsi3_carryin_<optab>): Likewise.
12261 (addsi3_carryin_alt2_<optab>): Likewise.
12262 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12263 for arm_restrict_it.
12264 (subsi3_carryin): Likewise.
12265 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12266 (minmax_arithsi): Disable for arm_restrict_it.
12267 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12268 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12269 (satsi_<SAT:code>_shift): Likewise.
12270 (arm_shiftsi3): Add alternative for 16-bit encoding.
12271 (arm32_movhf): Disable for arm_restrict_it.
12272 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12273 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12275 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12277 * config/arm/arm.md (attribute "insn"): Delete.
12278 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12279 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12280 (not_shiftsi): Update for attribute change.
12281 (not_shiftsi_compare0): Likewise.
12282 (not_shiftsi_compare0_scratch): Likewise.
12283 (arm_one_cmplsi2): Likewise.
12284 (thumb1_one_cmplsi2): Likewise.
12285 (notsi_compare0): Likewise.
12286 (notsi_compare0_scratch): Likewise.
12287 (thumb1_movdi_insn): Likewise.
12288 (arm_movsi_insn): Likewise.
12289 (movhi_insn_arch4): Likewise.
12290 (movhi_bytes): Likewise.
12291 (arm_movqi_insn): Likewise.
12292 (thumb1_movqi_insn): Likewise.
12293 (arm32_movhf): Likewise.
12294 (thumb1_movhf): Likewise.
12295 (arm_movsf_soft_insn): Likewise.
12296 (thumb1_movsf_insn): Likewise.
12297 (thumb_movdf_insn): Likewise.
12298 (movsicc_insn): Likewise.
12299 (movsfcc_soft_insn): Likewise.
12300 (and_scc): Likewise.
12301 (cond_move): Likewise.
12302 (if_move_not): Likewise.
12303 (if_not_move): Likewise.
12304 (if_shift_move): Likewise.
12305 (if_move_shift): Likewise.
12306 (if_shift_shift): Likewise.
12307 (if_not_arith): Likewise.
12308 (if_arith_not): Likewise.
12309 (cond_move_not): Likewise.
12310 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12311 (neon_mov<mode>): Likewise.
12312 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12313 (thumb2_movsi_vfp): Likewise.
12314 (movsf_vfp): Likewise.
12315 (thumb2_movsf_vfp): Likewise.
12316 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12318 (cortexa7_older_only): Likewise.
12319 (cortexa7_younger): Likewise.
12320 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12321 (1020alu_shift_op): Likewise.
12322 (1020alu_shift_reg_op): Likewise.
12323 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12324 (alu_shift_op): Likewise.
12325 (alu_shift_reg_op): Likewise.
12326 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12327 (11_alu_shift_op): Likewise.
12328 (11_alu_shift_reg_op): Likewise.
12329 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12330 (9_alu_shift_reg_op): Likewise.
12331 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12333 (cortex_a15_alu_shift): Likewise.
12334 (cortex_a15_alu_shift_reg): Likewise.
12335 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12336 (cortex_a5_alu_shift): Likewise.
12337 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12339 (cortex_a53_alu_shift): Likewise.
12340 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12342 (cortex_a7_alu_reg): Likewise.
12343 (cortex_a7_alu_shift): Likewise.
12344 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12345 (cortex_a8_alu_shift): Likewise.
12346 (cortex_a8_alu_shift_reg): Likewise.
12347 (cortex_a8_mov): Likewise.
12348 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12349 (cortex_a9_dp_shift): Likewise.
12350 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12351 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12352 (cortex_r4_mov): Likewise.
12353 (cortex_r4_alu_shift): Likewise.
12354 (cortex_r4_alu_shift_reg): Likewise.
12355 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12356 (526_alu_shift_op): Likewise.
12357 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12358 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12359 (626te_alu_shift_op): Likewise.
12360 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12361 (726te_alu_op): Likewise.
12362 (726te_alu_shift_op): Likewise.
12363 (726te_alu_shift_reg_op): Likewise.
12364 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12365 (mp626_alu_shift_op): Likewise.
12366 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12367 (pj4_alu_e1_conds): Likewise.
12368 (pj4_alu): Likewise.
12369 (pj4_alu_conds): Likewise.
12370 (pj4_shift): Likewise.
12371 (pj4_shift_conds): Likewise.
12372 (pj4_alu_shift): Likewise.
12373 (pj4_alu_shift_conds): Likewise.
12375 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12377 * config/arm/predicates.md (shiftable_operator_strict_it):
12379 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12380 Disable cond_exec version for arm_restrict_it.
12381 (thumb2_smaxsi3): Convert to generate cond_exec.
12382 (thumb2_sminsi3): Likewise.
12383 (thumb32_umaxsi3): Likewise.
12384 (thumb2_uminsi3): Likewise.
12385 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12386 (thumb2_neg_abssi2): Likewise.
12387 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12388 (thumb2_movsicc_insn): Adjust alternatives.
12389 (thumb2_mov_negscc): Disable for arm_restrict_it.
12390 (thumb2_mov_negscc_strict_it): New pattern.
12391 (thumb2_mov_notscc_strict_it): New pattern.
12392 (thumb2_mov_notscc): Disable for arm_restrict_it.
12393 (thumb2_ior_scc): Likewise.
12394 (thumb2_ior_scc_strict_it): New pattern.
12395 (thumb2_cond_move): Adjust for arm_restrict_it.
12396 (thumb2_cond_arith): Disable for arm_restrict_it.
12397 (thumb2_cond_arith_strict_it): New pattern.
12398 (thumb2_cond_sub): Adjust for arm_restrict_it.
12399 (thumb2_movcond): Likewise.
12400 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12401 (thumb2_zero_extendhisi2_v6): Likewise.
12402 (thumb2_zero_extendqisi2_v6): Likewise.
12403 (orsi_notsi_si): Likewise.
12404 (orsi_not_shiftsi_si): Likewise.
12406 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12408 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12409 instruction sequence is 1 byte shorter.
12411 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12413 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12414 it is not needed after split.
12416 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12419 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12420 second label for nonlocal goto receivers. Don't output pic base labels
12421 unless we're producing PIC; mark that action unreachable().
12422 (ix86_save_reg): If the function contains a nonlocal label, save the
12424 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12425 * config/darwin.c (emitted_pic_label_num): New GTY.
12426 (update_pic_label_number_if_needed): New.
12427 (machopic_output_function_base_name): Adjust for nonlocal receiver
12429 (machopic_should_output_picbase_label): New.
12430 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12431 (nonlocal_goto_receiver): New insn and split.
12433 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12435 * config/aarch64/aarch64-builtins.c
12436 (aarch64_fold_builtin): Fold abs in all modes.
12437 * config/aarch64/aarch64-simd-builtins.def
12438 (abs): Enable for all modes.
12439 * config/aarch64/arm_neon.h
12440 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12441 (vabs_f64): Add missing intrinsic.
12443 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12445 * config/aarch64/arm_neon.h (vabs_s64): New function
12447 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12450 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12451 * config/avr/avr.md (adjust_len): Add `round'.
12452 * config/avr/avr-protos.h (avr_out_round): New prototype.
12453 (avr_out_plus): Add `out_label' argument.
12454 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12455 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12456 Handle the case where `insn' is just a pattern.
12457 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12458 (avr_out_round): New function.
12459 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12461 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12463 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12464 Add check for flag_stack_usage to handle -fstack-usage support
12466 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12468 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12469 interaction for new Power8 flags and VSX.
12471 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12473 PR middle-end/57698
12474 * tree-inline.c (expand_call_inline): Emit errors during
12475 early_inlining only if optimization is not turned on.
12477 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12481 * passes.c (init_optimization_passes): Move the construction of
12482 the pass hierarchy into a new passes.def file.
12484 * Makefile.in (passes.o): Add dependency on passes.def.
12486 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12488 * passes.c (init_optimization_passes): Introduce macros for
12489 constructing the tree of passes (INSERT_PASSES_AFTER,
12490 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12491 TERMINATE_PASS_LIST).
12493 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12494 Wei Mi <wmi@google.com>
12496 PR rtl-optimization/57878
12497 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12499 (reload_pseudo_compare_func): Check nregs first for reload
12502 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12504 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12506 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12508 * read-rtl.c (validate_const_int): Once an invalid character is
12509 seen, quit the loop.
12511 * gengtype.c (write_roots): Similarly once we find the "deletable"
12512 or "if_marked" option.
12514 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12516 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12517 "xtab" and "sat". Move value "clz" from here to ...
12518 (attriubte "type"): ... here.
12519 (satsi_<SAT:code>): Delete "insn" attribute.
12520 (satsi_<SAT:code>_shift): Likewise.
12521 (arm_zero_extendqisi2addsi): Likewise.
12522 (arm_extendqisi2addsi): Likewise.
12523 (clzsi2): Update for attribute changes.
12524 (rbitsi2): Likewise.
12525 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12527 (arm_usatsihi): Likewise.
12528 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12530 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12532 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12533 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12534 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12535 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12536 in alphabetical order.
12537 (attribute "core_cycles"): Update for attribute changes.
12538 (arm_addsi3): Likewise.
12539 (addsi3_compare0): Likewise.
12540 (addsi3_compare0_scratch): Likewise.
12541 (addsi3_compare_op1): Likewise.
12542 (addsi3_compare_op2): Likewise.
12543 (compare_addsi2_op0): Likewise.
12544 (compare_addsi2_op1): Likewise.
12545 (addsi3_carryin_shift_<optab>): Likewise.
12546 (subsi3_carryin_shift): Likewise.
12547 (rsbsi3_carryin_shift): Likewise.
12548 (arm_subsi3_insn): Likewise.
12549 (subsi3_compare0): Likewise.
12550 (subsi3_compare): Likewise.
12551 (arm_andsi3_insn): Likewise.
12552 (thumb1_andsi3_insn): Likewise.
12553 (andsi3_compare0): Likewise.
12554 (andsi3_compare0_scratch): Likewise.
12555 (zeroextractsi_compare0_scratch
12556 (andsi_not_shiftsi_si): Likewise.
12557 (iorsi3_insn): Likewise.
12558 (iorsi3_compare0): Likewise.
12559 (iorsi3_compare0_scratch): Likewise.
12560 (arm_xorsi3): Likewise.
12561 (thumb1_xorsi3_insn): Likewise.
12562 (xorsi3_compare0): Likewise.
12563 (xorsi3_compare0_scratch): Likewise.
12564 (satsi_<SAT:code>_shift): Likewise.
12566 (arm_shiftsi3): Likewise.
12567 (shiftsi3_compare0): Likewise.
12568 (not_shiftsi): Likewise.
12569 (not_shiftsi_compare0): Likewise.
12570 (not_shiftsi_compare0_scratch): Likewise.
12571 (arm_one_cmplsi2): Likewise.
12572 (thumb_one_complsi2): Likewise.
12573 (notsi_compare0): Likewise.
12574 (notsi_compare0_scratch): Likewise.
12575 (thumb1_zero_extendhisi2): Likewise.
12576 (arm_zero_extendhisi2): Likewise.
12577 (arm_zero_extendhisi2_v6): Likewise.
12578 (arm_zero_extendhisi2addsi): Likewise.
12579 (thumb1_zero_extendqisi2): Likewise.
12580 (thumb1_zero_extendqisi2_v6): Likewise.
12581 (arm_zero_extendqisi2): Likewise.
12582 (arm_zero_extendqisi2_v6): Likewise.
12583 (arm_zero_extendqisi2addsi): Likewise.
12584 (thumb1_extendhisi2): Likewise.
12585 (arm_extendhisi2): Likewise.
12586 (arm_extendhisi2_v6): Likewise.
12587 (arm_extendqisi): Likewise.
12588 (arm_extendqisi_v6): Likewise.
12589 (arm_extendqisi2addsi): Likewise.
12590 (thumb1_extendqisi2): Likewise.
12591 (thumb1_movdi_insn): Likewise.
12592 (arm_movsi_insn): Likewise.
12593 (movsi_compare0): Likewise.
12594 (movhi_insn_arch4): Likewise.
12595 (movhi_bytes): Likewise.
12596 (arm_movqi_insn): Likewise.
12597 (thumb1_movqi_insn): Likewise.
12598 (arm32_movhf): Likewise.
12599 (thumb1_movhf): Likewise.
12600 (arm_movsf_soft_insn): Likewise.
12601 (thumb1_movsf_insn): Likewise.
12602 (movdf_soft_insn): Likewise.
12603 (thumb_movdf_insn): Likewise.
12604 (arm_cmpsi_insn): Likewise.
12605 (cmpsi_shiftsi): Likewise.
12606 (cmpsi_shiftsi_swp): Likewise.
12607 (arm_cmpsi_negshiftsi_si): Likewise.
12608 (movsicc_insn): Likewise.
12609 (movsfcc_soft_insn): Likewise.
12610 (arith_shiftsi): Likewise.
12611 (arith_shiftsi_compare0
12612 (arith_shiftsi_compare0_scratch
12613 (sub_shiftsi): Likewise.
12614 (sub_shiftsi_compare0
12615 (sub_shiftsi_compare0_scratch
12616 (and_scc): Likewise.
12617 (cond_move): Likewise.
12618 (if_plus_move): Likewise.
12619 (if_move_plus): Likewise.
12620 (if_move_not): Likewise.
12621 (if_not_move): Likewise.
12622 (if_shift_move): Likewise.
12623 (if_move_shift): Likewise.
12624 (if_shift_shift): Likewise.
12625 (if_not_arith): Likewise.
12626 (if_arith_not): Likewise.
12627 (cond_move_not): Likewise.
12628 (thumb1_ashlsi3): Set type attribute.
12629 (thumb1_ashrsi3): Likewise.
12630 (thumb1_lshrsi3): Likewise.
12631 (thumb1_rotrsi3): Likewise.
12632 (shiftsi3_compare0_scratch): Likewise.
12633 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12634 (neon_mov<mode>): Likewise.
12635 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12637 (thumb2_movsi_insn): Likewise.
12638 (thumb2_cmpsi_neg_shiftsi): Likewise.
12639 (thumb2_extendqisi_v6): Likewise.
12640 (thumb2_zero_extendhisi2_v6): Likewise.
12641 (thumb2_zero_extendqisi2_v6): Likewise.
12642 (thumb2_shiftsi3_short): Likewise.
12643 (thumb2_addsi3_compare0_scratch): Likewise.
12644 (orsi_not_shiftsi_si): Likewise.
12645 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12646 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12648 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12649 (1020alu_shift_op): Likewise.
12650 (1020alu_shift_reg_op): Likewise.
12651 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12652 (alu_shift_op): Likewise.
12653 (alu_shift_reg_op): Likewise.
12654 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12655 (11_alu_shift_op): Likewise.
12656 (11_alu_shift_reg_op): Likewise.
12657 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12658 (9_alu_shift_reg_op): Likewise.
12659 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12661 (cortex_a15_alu_shift): Likewise.
12662 (cortex_a15_alu_shift_reg): Likewise.
12663 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12665 (cortex_a5_alu_shift): Likewise.
12666 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12668 (cortex_a53_alu_shift): Likewise.
12669 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12671 (cortex_a7_alu_reg): Likewise.
12672 (cortex_a7_alu_shift): Likewise.
12673 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12675 (cortex_a8_alu_shift): Likewise.
12676 (cortex_a8_alu_shift_reg): Likewise.
12677 (cortex_a8_mov): Likewise.
12678 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12679 (cortex_a9_dp_shift): Likewise.
12680 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12682 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12684 (cortex_r4_mov): Likewise.
12685 (cortex_r4_alu_shift): Likewise.
12686 (cortex_r4_alu_shift_reg): Likewise.
12687 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12688 (526_alu_shift_op): Likewise.
12689 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12690 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12691 (626te_alu_shift_op): Likewise.
12692 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12693 (726te_alu_op): Likewise.
12694 (726te_alu_shift_op): Likewise.
12695 (726te_alu_shift_reg_op): Likewise.
12696 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12697 (mp626_alu_shift_op): Likewise.
12698 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12699 (pj4_alu_e1_conds): Likewise.
12700 (pj4_alu): Likewise.
12701 (pj4_alu_conds): Likewise.
12702 (pj4_shift): Likewise.
12703 (pj4_shift_conds): Likewise.
12704 (pj4_alu_shift): Likewise.
12705 (pj4_alu_shift_conds): Likewise.
12706 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12708 (cortexa7_older_only): Likewise.
12709 (cortexa7_younger): Likewise.
12711 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12713 * ipa-pure-const.c (generate_summary): Rename to...
12714 (pure_const_generate_summary): ... this.
12716 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12718 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12720 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12723 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12726 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12728 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12729 enabled without -march=zEC12.
12730 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12733 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12735 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12736 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12737 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12738 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12739 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12740 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12741 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12743 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12744 and ISA_HAS_NMADD3_NMSUB3.
12745 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12746 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12747 (nmsub4<mode>, nmsub3<mode>): Likewise.
12748 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12750 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12752 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12753 TARGET_MIPS5400 checking.
12755 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12756 Peter Bergner <bergner@vnet.ibm.com>
12758 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12759 registers in the comment.
12760 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12761 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12762 rather than FIRST_PSEUDO_REGISTERS.
12764 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12766 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12767 enable extra ISA flags with TARGET_HTM.
12769 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12771 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12774 2013-07-15 Cong Hou <congh@google.com>
12776 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12777 in compare function for sorting.
12779 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12781 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12782 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12783 * config/rs6000/rs6000.opt: Add -mhtm option.
12784 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12785 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12786 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12787 __HTM__ if the HTM instructions are available.
12788 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12789 htm_spr_reg_operand): New define_predicates.
12790 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12791 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12793 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12794 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12795 HTM builtin functions.
12796 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12797 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12798 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12799 (rs6000_builtin_mask_calculate): Likewise.
12800 (rs6000_option_override_internal): Likewise.
12801 (bdesc_htm): Add new HTM builtin support.
12802 (htm_spr_num): New function.
12803 (htm_spr_regno): Likewise.
12804 (rs6000_htm_spr_icode): Likewise.
12805 (htm_expand_builtin): Likewise.
12806 (htm_init_builtins): Likewise.
12807 (rs6000_expand_builtin): Add support for HTM builtin functions.
12808 (rs6000_init_builtins): Likewise.
12809 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12811 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12812 (TARGET_HTM, MASK_HTM): Define macros.
12813 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12814 (FIXED_REGISTERS): Likewise.
12815 (CALL_USED_REGISTERS): Likewise.
12816 (CALL_REALLY_USED_REGISTERS): Likewise.
12817 (REG_ALLOC_ORDER): Likewise.
12818 (enum reg_class): Likewise.
12819 (REG_CLASS_NAMES): Likewise.
12820 (REG_CLASS_CONTENTS): Likewise.
12821 (REGISTER_NAMES): Likewise.
12822 (ADDITIONAL_REGISTER_NAMES): Likewise.
12823 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12824 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12825 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12826 * config/rs6000/htm.md: New file.
12827 * config/rs6000/htmintrin.h: New file.
12828 * config/rs6000/htmxlintrin.h: New file.
12830 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12832 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12833 Define SYMBOL_TINY_GOT, update comment.
12834 * config/aarch64/aarch64.c
12835 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12836 (aarch64_expand_mov_immediate): Likewise.
12837 (aarch64_print_operand): Likewise.
12838 (aarch64_classify_symbol): Likewise.
12839 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12840 (ldr_got_tiny): Define.
12842 2013-07-13 Tobias Grosser <tobias@grosser.es>
12844 PR tree-optimization/54094
12845 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12846 scheduling dimension for the parallelism check from the polyhedral
12847 information in the AST.
12848 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12851 2013-07-13 Jason Merrill <jason@redhat.com>
12853 * print-tree.c (debug_vec_tree): Use debug_raw.
12854 (debug_raw (vec<tree, va_gc> &)): New.
12855 (debug_raw (vec<tree, va_gc> *)): New.
12856 * tree.h: Declare them.
12858 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12860 * ifcvt.c (ifcvt_after_combine): New static variable.
12861 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12863 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12864 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12865 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12867 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12869 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12871 2013-07-12 Michael Matz <matz@suse.de>
12873 PR middle-end/55771
12874 * convert.c (convert_to_real): Reject non-float inner types.
12876 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12878 * config/aarch64/aarch64-protos.h
12879 (aarch64_simd_immediate_valid_for_move): Remove.
12880 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12881 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12883 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12885 2013-07-11 Steve Ellcey <sellcey@mips.com>
12887 * config/mips/mips.c (mips_conditional_register_usage): Do not
12888 use t[0-7] registers in MIPS16 mode when optimizing for size.
12890 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12892 * config/i386/i386.c (dispatch_function_versions): Fix array
12893 indexing of function_version_info to match actual_versions.
12895 2013-07-11 Teresa Johnson <tejohnson@google.com>
12897 * vec.h (struct va_gc): Move release out-of-line.
12898 (va_gc::release): Call ggc_free on released vec.
12900 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12902 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12903 Require GOT register as additional operand in UNSPEC.
12904 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12905 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12906 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12907 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12908 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12909 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12910 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12912 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12915 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12916 name seen by assembler/linker rather if available.
12918 2013-07-11 Andreas Schwab <schwab@suse.de>
12920 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12922 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12924 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12926 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12928 * doc/tm.texi.in: Move hook documentation to ....
12929 * target.def: ... here.
12931 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12932 text on @hook line.
12933 * doc/tm.texi: Regenerate.
12935 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12938 * doc/invoke.texi: Document -Wconditionally-supported.
12940 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12943 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12946 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12949 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12950 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12951 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12952 Remove duplicate devices.
12953 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12954 * config/avr/t-multilib: Regenerate.
12955 * config/avr/avr-tables.opt: Regenerate.
12956 * doc/avr-mmcu.texi: Regenerate.
12958 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12961 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12963 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12965 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12966 the cost of MULT when optimizing for size.
12968 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12970 * config/cr16/cr16-protos.h: Don't include target.h.
12972 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12974 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12975 adjust register size for TDmode and TFmode for VSX registers.
12977 2013-07-08 Kai Tietz <ktietz@redhat.com>
12980 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12981 hook_bool_const_tree_true.
12983 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12985 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12986 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12987 * config/s390/s390.md: Define FPR*_REGNUM constants.
12988 Fix FPR2_REGNUM constant (18 -> 17).
12989 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12990 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12991 ("trunc<BFP:mode><DFP_ALL:mode>2")
12992 ("trunc<DFP_ALL:mode><BFP:mode>2")
12993 ("*extend<BFP:mode><DFP_ALL:mode>2")
12994 ("*extend<DFP_ALL:mode><BFP:mode>2")
12995 ("extend<BFP:mode><DFP_ALL:mode>2")
12996 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12999 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13001 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13003 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13005 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13006 and cfun_fpr_bit_p to cfun_fpr_save_p.
13007 (s390_frame_area, s390_register_info, s390_frame_info)
13008 (s390_emit_prologue, s390_emit_epilogue)
13009 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13011 * config/s390/s390.h: Define *_REGNUM macros for floating point
13014 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13016 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13018 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13020 PR rtl-optimization/57786
13021 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13022 and break out of the loop when it is set to false.
13024 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13027 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13028 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13029 (const_int 63)) 0)).
13030 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13031 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13032 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13034 PR rtl-optimization/57829
13035 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13036 mask bits outside of mode are just sign-extension from mode to HWI.
13038 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13040 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13041 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13042 adjust_address instead of change_address to keep info about alignment.
13043 (emit_strmov): Remove.
13044 (emit_memmov): New function.
13045 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13046 (expand_movmem_epilogue): Likewise and return updated rtx for
13048 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13049 destination and source.
13050 (decide_alignment): Refactor, handle vector_loop.
13051 (ix86_expand_movmem): Likewise.
13052 (ix86_expand_setmem): Likewise.
13053 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13055 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13057 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13058 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13060 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13062 * config/i386/sse.md (sse_movlhps): Change alternative 3
13063 of operand 2 to "m".
13065 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13068 * config/i386/sse.md (iptr): New mode attribute.
13069 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13070 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13071 (<sse>_vmmul<mode>3): Ditto.
13072 (<sse>_vmdiv<mode>3): Ditto.
13073 (sse_vmrcpv4sf2): Ditto.
13074 (<sse>_vmsqrt<mode>2): Ditto.
13075 (sse_vmrsqrtv4sf2): Ditto.
13076 (<sse>_vm<code><mode>3): Ditto.
13077 (avx_vmcmp<mode>3): Ditto.
13078 (<sse>_vmmaskcmp<mode>3): Ditto.
13079 (<sse>_comi): Ditto.
13080 (<sse>_ucomi): Ditto.
13081 (*xop_vmfrcz_<mode>): Ditto.
13082 (*fmai_fmadd_<mode>): Ditto.
13083 (*fmai_fmsub_<mode>): Ditto.
13084 (*fmai_fnmadd_<mode>): Ditto.
13085 (*fmai_fnmsub_<mode>): Ditto.
13086 (*fma4i_vmfmadd_<mode>): Ditto.
13087 (*fma4i_vmfmsub_<mode>): Ditto.
13088 (*fma4i_vmfnmadd_<mode>): Ditto.
13089 (*fma4i_vmfnmsub_<mode>): Ditto.
13090 (*xop_vmfrcz_<mode>): Ditto.
13091 (sse_cvtps2pi): Ditto.
13092 (sse_cvttps2pi): Ditto.
13093 (sse_cvtss2si): Ditto.
13094 (sse_cvtss2si_2): Ditto.
13095 (sse_cvtss2siq_2): Ditto.
13096 (sse_cvttss2si): Ditto.
13097 (sse_cvttss2siq): Ditto.
13098 (sse_cvtsd2si): Ditto.
13099 (sse_cvtsd2si_2): Ditto.
13100 (sse_cvtsd2siq_2): Ditto.
13101 (sse_cvttsd2si): Ditto.
13102 (sse_cvttsd2siq): Ditto.
13103 (sse_cvtsd2ss): Ditto.
13104 (sse_cvtss2sd): Ditto.
13105 (avx2_pbroadcast<mode>): Ditto.
13106 (avx2_pbroadcast<mode>_1): Ditto.
13107 (*avx_vperm_broadcast_v4sf): Ditto.
13109 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13110 (sse_movlhps): Ditto.
13111 (sse_storehps): Ditto.
13112 (sse_loadhps): Ditto.
13113 (sse_storelps): Ditto.
13114 (sse_loadlps): Ditto.
13115 (*vec_concatv4sf): Ditto.
13116 (*vec_interleave_highv2df): Ditto.
13117 (*vec_interleave_lowv2df): Ditto.
13118 (*vec_extractv2df_1_sse): Ditto.
13119 (*vec_extractv2df_0_sse): Ditto.
13120 (sse2_storelpd): Ditto.
13121 (sse2_loadlpd): Ditto.
13122 (sse2_movsd): Ditto.
13123 (*vec_concatv4si): Ditto.
13124 (vec_concatv2di): Ditto.
13126 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13127 for Intel asm dialect.
13128 (mmx_punpcklwd): Ditto.
13129 (mmx_punpckldq): Ditto.
13131 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13132 for intel assembler dialect.
13134 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13137 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13138 for BUILT_IN_C{LZ,LRSB}*.
13139 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13140 * tree-vrp.c (extract_range_basic): Handle
13141 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13142 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13143 fall thru to code calling set_value*.
13144 * builtins.c (expand_builtin): Remove *IMAX cases.
13145 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13146 if width is bigger than 2*HWI.
13148 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13150 PR rtl-optimization/55342
13151 * lra-int.h (lra_subreg_reload_pseudos): New.
13152 * lra.c: Add undoing optional reloads to the block diagram.
13153 (lra_subreg_reload_pseudos): New.
13154 (lra_optional_reload_pseudos): Change comments.
13155 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13156 lra_optional_reload_pseudos after undo transformations.
13157 * lra-assigns.c (pseudo_prefix_title): New.
13158 (lra_setup_reg_renumber): Use it.
13159 (spill_for): Ditto. Check subreg reload pseudos too.
13160 (assign_by_spills): Consider subreg reload pseudos too.
13161 * lra-constraints.c (simplify_operand_subreg): Use
13162 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13163 (curr_insn_transform): Recognize and do optional reloads.
13164 (undo_optional_reloads): New.
13165 (lra_undo_inheritance): Call undo_optional_reloads.
13167 2013-07-05 Thomas Quinot <quinot@adacore.com>
13169 * tree-complex.c (expand_complex_operations_1): Fix typo.
13171 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13173 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13174 (tune_params): New member 'const vec_costs'.
13175 * config/aarch64/aarch64.c (generic_vector_cost): New.
13176 (generic_tunings): New member 'generic_vector_cost'.
13177 (aarch64_builtin_vectorization_cost): New.
13178 (aarch64_add_stmt_cost): New.
13179 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13180 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13182 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13185 * config/i386/predicates.md (vsib_address_operand): Disallow
13186 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13188 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13190 PR middle-end/55030
13191 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13192 expand_builtin_setjmp_receiver.
13193 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13194 with NULL for the label parameter.
13195 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13196 the frame-pointer. Adjust comments.
13197 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13198 only if LABEL is non-NULL.
13200 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13202 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13203 (ARM_ABI_AAPCS64): Ditto.
13205 (ARM_DEFAULT_ABI): Ditto.
13207 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13209 * config/aarch64/aarch64-builtins.c
13210 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13211 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13213 * config/aarch64/aarch64-simd.md
13214 (aarch64_ld1<VALL:mode>): New.
13215 (aarch64_st1<VALL:mode>): Likewise.
13216 * config/aarch64/arm_neon.h
13217 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13219 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13221 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13223 (ix86_option_override_internal):Turn on all -mavx target flags by
13224 default as they are dependent on AVX anyway.
13226 2013-07-02 Cary Coutant <ccoutant@google.com>
13228 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13229 deterministic hash.
13230 (loc_checksum_ordered): Likewise.
13231 (hash_loc_operands): Remove inline keyword.
13233 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13235 PR tree-optimization/57741
13236 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13237 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13238 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13239 Allow REAL_CST step_exprs if flag_associative_math.
13240 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13242 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13244 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13246 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13248 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13250 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13252 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13254 (iorsi3_insn): Likewise.
13255 (arm_xorsi3): Likewise.
13257 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13259 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13261 (attribute "type"): ... here, and prefix with "wmmx_".
13262 (attribute "core_cycles"): Update for attribute changes.
13263 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13264 (tbcstv4hi): Likewise.
13265 (tbcstv2si): Likewise.
13266 (iwmmxt_iordi3): Likewise.
13267 (iwmmxt_xordi3): Likewise.
13268 (iwmmxt_anddi3): Likewise.
13269 (iwmmxt_nanddi3): Likewise.
13270 (iwmmxt_arm_movdi): Likewise.
13271 (iwmmxt_movsi_insn): Likewise.
13272 (mov<mode>_internal): Likewise.
13273 (and<mode>3_iwmmxt): Likewise.
13274 (ior<mode>3_iwmmxt): Likewise.
13275 (xor<mode>3_iwmmxt): Likewise.
13276 (add<mode>3_iwmmxt): Likewise.
13277 (ssaddv8qi3): Likewise.
13278 (ssaddv4hi3): Likewise.
13279 (ssaddv2si3): Likewise.
13280 (usaddv8qi3): Likewise.
13281 (usaddv4hi3): Likewise.
13282 (usaddv2si3): Likewise.
13283 (sub<mode>3_iwmmxt): Likewise.
13284 (sssubv8qi3): Likewise.
13285 (sssubv4hi3): Likewise.
13286 (sssubv2si3): Likewise.
13287 (ussubv8qi3): Likewise.
13288 (ussubv4hi3): Likewise.
13289 (ussubv2si3): Likewise.
13290 (mulv4hi3_iwmmxt): Likewise.
13291 (smulv4hi3_highpart): Likewise.
13292 (umulv4hi3_highpart): Likewise.
13293 (iwmmxt_wmacs): Likewise.
13294 (iwmmxt_wmacsz): Likewise.
13295 (iwmmxt_wmacu): Likewise.
13296 (iwmmxt_wmacuz): Likewise.
13297 (iwmmxt_clrdi): Likewise.
13298 (iwmmxt_clrv8qi): Likewise.
13299 (iwmmxt_clr4hi): Likewise.
13300 (iwmmxt_clr2si): Likewise.
13301 (iwmmxt_uavgrndv8qi3): Likewise.
13302 (iwmmxt_uavgrndv4hi3): Likewise.
13303 (iwmmxt_uavgv8qi3): Likewise.
13304 (iwmmxt_uavgv4hi3): Likewise.
13305 (iwmmxt_tinsrb): Likewise.
13306 (iwmmxt_tinsrh): Likewise.
13307 (iwmmxt_tinsrw): Likewise.
13308 (iwmmxt_textrmub): Likewise.
13309 (iwmmxt_textrmsb): Likewise.
13310 (iwmmxt_textrmuh): Likewise.
13311 (iwmmxt_textrmsh): Likewise.
13312 (iwmmxt_textrmw): Likewise.
13313 (iwmxxt_wshufh): Likewise.
13314 (eqv8qi3): Likewise.
13315 (eqv4hi3): Likewise.
13316 (eqv2si3): Likewise.
13317 (gtuv8qi3): Likewise.
13318 (gtuv4hi3): Likewise.
13319 (gtuv2si3): Likewise.
13320 (gtv8qi3): Likewise.
13321 (gtv4hi3): Likewise.
13322 (gtv2si3): Likewise.
13323 (smax<mode>3_iwmmxt): Likewise.
13324 (umax<mode>3_iwmmxt): Likewise.
13325 (smin<mode>3_iwmmxt): Likewise.
13326 (umin<mode>3_iwmmxt): Likewise.
13327 (iwmmxt_wpackhss): Likewise.
13328 (iwmmxt_wpackwss): Likewise.
13329 (iwmmxt_wpackdss): Likewise.
13330 (iwmmxt_wpackhus): Likewise.
13331 (iwmmxt_wpackwus): Likewise.
13332 (iwmmxt_wpackdus): Likewise.
13333 (iwmmxt_wunpckihb): Likewise.
13334 (iwmmxt_wunpckihh): Likewise.
13335 (iwmmxt_wunpckihw): Likewise.
13336 (iwmmxt_wunpckilb): Likewise.
13337 (iwmmxt_wunpckilh): Likewise.
13338 (iwmmxt_wunpckilw): Likewise.
13339 (iwmmxt_wunpckehub): Likewise.
13340 (iwmmxt_wunpckehuh): Likewise.
13341 (iwmmxt_wunpckehuw): Likewise.
13342 (iwmmxt_wunpckehsb): Likewise.
13343 (iwmmxt_wunpckehsh): Likewise.
13344 (iwmmxt_wunpckehsw): Likewise.
13345 (iwmmxt_wunpckelub): Likewise.
13346 (iwmmxt_wunpckeluh): Likewise.
13347 (iwmmxt_wunpckeluw): Likewise.
13348 (iwmmxt_wunpckelsb): Likewise.
13349 (iwmmxt_wunpckelsh): Likewise.
13350 (iwmmxt_wunpckelsw): Likewise.
13351 (ror<mode>3): Likewise.
13352 (ashr<mode>3_iwmmxt): Likewise.
13353 (lshr<mode>3_iwmmxt): Likewise.
13354 (ashl<mode>3_iwmmxt): Likewise.
13355 (ror<mode>3_di): Likewise.
13356 (ashr<mode>3_di): Likewise.
13357 (lshr<mode>3_di): Likewise.
13358 (ashl<mode>3_di): Likewise.
13359 (iwmmxt_wmadds): Likewise.
13360 (iwmmxt_wmaddu): Likewise.
13361 (iwmmxt_tmia): Likewise.
13362 (iwmmxt_tmiaph): Likewise.
13363 (iwmmxt_tmiabb): Likewise.
13364 (iwmmxt_tmiatb): Likewise.
13365 (iwmmxt_tmiabt): Likewise.
13366 (iwmmxt_tmiatt): Likewise.
13367 (iwmmxt_tmovmskb): Likewise.
13368 (iwmmxt_tmovmskh): Likewise.
13369 (iwmmxt_tmovmskw): Likewise.
13370 (iwmmxt_waccb): Likewise.
13371 (iwmmxt_wacch): Likewise.
13372 (iwmmxt_waccw): Likewise.
13373 (iwmmxt_waligni): Likewise.
13374 (iwmmxt_walignr): Likewise.
13375 (iwmmxt_walignr0): Likewise.
13376 (iwmmxt_walignr1): Likewise.
13377 (iwmmxt_walignr2): Likewise.
13378 (iwmmxt_walignr3): Likewise.
13379 (iwmmxt_wsadb): Likewise.
13380 (iwmmxt_wsadh): Likewise.
13381 (iwmmxt_wsadbz): Likewise.
13382 (iwmmxt_wsadhz): Likewise.
13383 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13384 (iwmmxt_wabsdiffb): Likewise.
13385 (iwmmxt_wabsdiffh): Likewise.
13386 (iwmmxt_wabsdiffw): Likewise.
13387 (iwmmxt_waddsubhx): Likewise
13388 (iwmmxt_wsubaddhx): Likewise.
13389 (addc<mode>3): Likewise.
13390 (iwmmxt_avg4): Likewise.
13391 (iwmmxt_avg4r): Likewise.
13392 (iwmmxt_wmaddsx): Likewise.
13393 (iwmmxt_wmaddux): Likewise.
13394 (iwmmxt_wmaddsn): Likewise.
13395 (iwmmxt_wmaddun): Likewise.
13396 (iwmmxt_wmulwsm): Likewise.
13397 (iwmmxt_wmulwum): Likewise.
13398 (iwmmxt_wmulsmr): Likewise.
13399 (iwmmxt_wmulumr): Likewise.
13400 (iwmmxt_wmulwsmr): Likewise.
13401 (iwmmxt_wmulwumr): Likewise.
13402 (iwmmxt_wmulwl): Likewise.
13403 (iwmmxt_wqmulm): Likewise.
13404 (iwmmxt_wqmulwm): Likewise.
13405 (iwmmxt_wqmulmr): Likewise.
13406 (iwmmxt_wqmulwmr): Likewise.
13407 (iwmmxt_waddbhusm): Likewise.
13408 (iwmmxt_waddbhusl): Likewise.
13409 (iwmmxt_wqmiabb): Likewise.
13410 (iwmmxt_wqmiabt): Likewise.
13411 (iwmmxt_wqmiatb): Likewise.
13412 (iwmmxt_wqmiatt): Likewise.
13413 (iwmmxt_wqmiabbn): Likewise.
13414 (iwmmxt_wqmiabtn): Likewise.
13415 (iwmmxt_wqmiatbn): Likewise.
13416 (iwmmxt_wqmiattn): Likewise.
13417 (iwmmxt_wmiabb): Likewise.
13418 (iwmmxt_wmiabt): Likewise.
13419 (iwmmxt_wmiatb): Likewise.
13420 (iwmmxt_wmiatt): Likewise.
13421 (iwmmxt_wmiabbn): Likewise.
13422 (iwmmxt_wmiabtn): Likewise.
13423 (iwmmxt_wmiatbn): Likewise.
13424 (iwmmxt_wmiattn): Likewise.
13425 (iwmmxt_wmiawbb): Likewise.
13426 (iwmmxt_wmiawbt): Likewise.
13427 (iwmmxt_wmiawtb): Likewise.
13428 (iwmmxt_wmiawtt): Likewise.
13429 (iwmmxt_wmiawbbn): Likewise.
13430 (iwmmxt_wmiawbtn): Likewise.
13431 (iwmmxt_wmiawtbn): Likewise.
13432 (iwmmxt_wmiawttn): Likewise.
13433 (iwmmxt_wmerge): Likewise.
13434 (iwmmxt_tandc<mode>3): Likewise.
13435 (iwmmxt_torc<mode>3): Likewise.
13436 (iwmmxt_torvsc<mode>3): Likewise.
13437 (iwmmxt_textrc<mode>3): Likewise.
13438 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13439 (wmmxt_pack): Likewise.
13440 (wmmxt_mult_c1): Likewise.
13441 (wmmxt_mult_c2): Likewise.
13442 (wmmxt_alu_c1): Likewise.
13443 (wmmxt_alu_c2): Likewise.
13444 (wmmxt_alu_c3): Likewise.
13445 (wmmxt_transfer_c1): Likewise.
13446 (wmmxt_transfer_c2): Likewise.
13447 (wmmxt_transfer_c3): Likewise.
13448 (marvell_f_iwmmxt_wstr): Likewise.
13449 (marvell_f_iwmmxt_wldr): Likewise.
13451 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13453 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13455 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13458 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13459 * lra-constraints.c (need_for_split_p): Check call used hard regs
13460 living through calls.
13462 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13463 call used regs for call insn.
13465 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13468 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13469 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13471 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13473 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13475 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13477 * lra-constraints.c (need_for_split_p): Check call used hard regs
13478 living through calls.
13480 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13483 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13484 to tie with any other modes. Eliminate Altivec vector mode tests,
13485 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13486 code, to return 0 if testing MODE2 for a condition, if we've
13487 already tested MODE1 for the same condition.
13489 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13491 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13494 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13496 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13497 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13499 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13501 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13503 (aarch64_symbolic_constant_p): Remove.
13504 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13505 static. Fix line length and white space.
13506 (aarch64_symbolic_constant_p): Remove.
13507 * config/aarch64/predicates.md (aarch64_valid_symref):
13508 Use aarch64_classify_symbol_expression.
13510 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13512 * config/arm/constraints.md (Ts): New constraint.
13513 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13515 (compare_scc): Use "Ts" constraint for operand 0.
13516 (ior_scc_scc): Likewise.
13517 (and_scc_scc): Likewise.
13518 (and_scc_scc_nodom): Likewise.
13519 (ior_scc_scc_cmp): Likewise for operand 7.
13520 (and_scc_scc_cmp): Likewise.
13521 * config/arm/thumb2.md (thumb2_movsi_insn):
13522 Add alternatives for 16-bit encodings.
13523 (thumb2_movhi_insn): Likewise.
13524 (thumb2_movsicc_insn): Likewise.
13525 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13526 (thumb2_negscc): Use "Ts" constraint.
13527 Move mvn instruction outside cond_exec block.
13528 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13529 for 16-bit encodings.
13531 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13533 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13535 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13536 (mulsi3subsi): Likewise.
13537 (mulsidi3adddi): Likewise.
13538 (mulsidi3_v6): Likewise.
13539 (umulsidi3_v6): Likewise.
13540 (umulsidi3adddi_v6): Likewise.
13541 (smulsi3_highpart_v6): Likewise.
13542 (umulsi3_highpart_v6): Likewise.
13543 (mulhisi3tb): Likewise.
13544 (mulhisi3bt): Likewise.
13545 (mulhisi3tt): Likewise.
13546 (maddhisi4): Likewise.
13547 (maddhisi4tb): Likewise.
13548 (maddhisi4tt): Likewise.
13549 (maddhidi4): Likewise.
13550 (maddhidi4tb): Likewise.
13551 (maddhidi4tt): Likewise.
13552 (zeroextractsi_compare0_scratch): Likewise.
13553 (insv_zero): Likewise.
13554 (insv_t2): Likewise.
13555 (anddi_notzesidi_di): Likewise.
13556 (anddi_notsesidi_di): Likewise.
13557 (andsi_notsi_si): Likewise.
13558 (iordi_zesidi_di): Likewise.
13559 (xordi_zesidi_di): Likewise.
13560 (andsi_iorsi3_notsi): Likewise.
13561 (smax_0): Likewise.
13562 (smax_m1): Likewise.
13563 (smin_0): Likewise.
13564 (not_shiftsi): Likewise.
13565 (unaligned_loadsi): Likewise.
13566 (unaligned_loadhis): Likewise.
13567 (unaligned_loadhiu): Likewise.
13568 (unaligned_storesi): Likewise.
13569 (unaligned_storehi): Likewise.
13570 (extv_reg): Likewise.
13571 (extzv_t2): Likewise.
13572 (divsi3): Likewise.
13573 (udivsi3): Likewise.
13574 (arm_zero_extendhisi2addsi): Likewise.
13575 (arm_zero_extendqisi2addsi): Likewise.
13576 (compareqi_eq0): Likewise.
13577 (arm_extendhisi2_v6): Likewise.
13578 (arm_extendqisi2addsi): Likewise.
13579 (arm_movt): Likewise.
13580 (thumb2_ldrd): Likewise.
13581 (thumb2_ldrd_base): Likewise.
13582 (thumb2_ldrd_base_neg): Likewise.
13583 (thumb2_strd): Likewise.
13584 (thumb2_strd_base): Likewise.
13585 (thumb2_strd_base_neg): Likewise.
13586 (arm_negsi2): Add alternative for 16-bit encoding.
13587 (arm_one_cmplsi2): Likewise.
13589 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13591 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13592 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13593 (movdfcc): Likewise.
13594 * config/arm/vfp.md (*thumb2_movsf_vfp):
13595 Disable predication for arm_restrict_it.
13596 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13597 (*thumb2_movdfcc_vfp): Likewise.
13598 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13599 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13600 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13601 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13602 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13603 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13604 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13605 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13606 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13607 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13608 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13609 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13610 Disable predication for arm_restrict_it.
13612 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13614 * config/i386/bmiintrin.h (_bextr_u32): New.
13615 (_bextr_u64): Ditto.
13617 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13619 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13620 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13621 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13622 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13623 * config/mips/n32-elf.h: ...this new file.
13625 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13628 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13629 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13631 2013-06-27 Catherine Moore <clm@codesourcery.com>
13633 * config/mips/mips-tables.opt: Regenerate.
13634 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13635 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13636 * doc/invoke.texi: Add -m14kc.
13638 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13641 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13642 constraints of operand 1 and 2.
13645 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13646 to match RTL canonicalization. Swap predicates and
13647 constraints of operand 1 and 2.
13649 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13651 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13652 Process OP_INOUT regs for splitting too.
13654 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13656 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13657 decl before the loop, initialize to NULL.
13658 (vectorizable_load): Initialize ptr_incr to NULL.
13660 2013-06-27 Martin Jambor <mjambor@suse.cz>
13663 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13664 * ipa-ref.c (ipa_maybe_record_reference): New function.
13665 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13666 * ipa-cp.c (create_specialized_node): Record potential references from
13668 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13670 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13672 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13673 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13674 'mode' to force_reg.
13675 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13676 (aarch64_expand_mov_immediate): Likewise.
13678 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13680 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13681 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13683 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13685 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13686 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13687 (struct machine_function): Add tbegin_p.
13688 (s390_canonicalize_comparison): Fold CC mode compares to
13689 conditional jump if possible.
13690 (s390_emit_jump): Return the emitted jump.
13691 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13692 Handle CCRAWmode compares.
13693 (s390_option_override): Default to -mhtm if available.
13694 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13695 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13696 FPRs instead of df_regs_ever_live_p.
13697 (s390_optimize_nonescaping_tx): New function.
13698 (s390_init_frame_layout): Extend clobbered_regs array to cover
13700 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13701 (s390_expand_tbegin): New function.
13702 (enum s390_builtin): New enum definition.
13703 (code_for_builtin): New array definition.
13704 (s390_init_builtins): New function.
13705 (s390_expand_builtin): New function.
13706 (TARGET_INIT_BUILTINS): Define.
13707 (TARGET_EXPAND_BUILTIN): Define.
13708 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13709 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13710 (s390_alc_comparison): Likewise.
13711 * config/s390/s390-modes.def: Add CCRAWmode.
13712 * config/s390/s390.h (processor_flags): Add PF_TX.
13713 (TARGET_CPU_HTM): Define macro.
13714 (TARGET_HTM): Define macro.
13715 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13716 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13717 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13718 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13719 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13720 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13721 ("*cc_to_int"): Move up.
13722 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13723 constants other than 0.
13724 ("*ccraw_to_int"): New insn and splitter definition.
13725 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13726 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13727 ("tx_assist"): New expander.
13728 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13729 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13730 * config/s390/s390.opt: Add -mhtm option.
13731 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13732 * config/s390/htmxlintrin.h: New file.
13733 * config/s390/htmintrin.h: New file.
13734 * config/s390/s390intrin.h: New file.
13735 * doc/extend.texi: Document htm builtins.
13736 * config.gcc: Add the new header files to extra_headers.
13738 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13740 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13741 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13743 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13744 Pat Haugen <pthaugen@us.ibm.com>
13745 Peter Bergner <bergner@vnet.ibm.com>
13747 * config/rs6000/power8.md: New.
13748 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13749 setting for power8 entry.
13750 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13751 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13752 test for Power4/Power5 only.
13753 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13755 (force_new_group): Adjust comment.
13756 * config/rs6000/rs6000.md: Include power8.md.
13758 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13760 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13761 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13763 (tune_params): Update comment.
13764 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13765 (arm_max_conditional_execute): New function.
13766 (thumb2_final_prescan_insn): Use max_insn_skipped and
13767 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13769 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13771 PR tree-optimization/57705
13772 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13773 SSA_NAME step, provided that it is not defined inside the loop.
13774 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13775 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13777 2013-06-25 Martin Jambor <mjambor@suse.cz>
13779 PR middle-end/57670
13780 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13781 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13783 (ipa_note_param_call): Initialize member_ptr flag.
13784 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13785 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13786 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13787 (ipa_read_indirect_edge_info): Likewise.
13789 2013-06-25 Richard Biener <rguenther@suse.de>
13791 PR middle-end/56977
13792 * passes.c (init_optimization_passes): Move pass_fold_builtins
13793 and pass_dce earlier with -Og.
13795 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13797 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13798 <BIT_FIELD_REF>: Remove trailing TAB.
13799 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13802 2013-06-24 Martin Jambor <mjambor@suse.cz>
13804 PR tree-optimization/57358
13805 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13806 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13807 (ipa_analyze_params_uses): Generate pessimistic info when true.
13809 2013-06-24 Martin Jambor <mjambor@suse.cz>
13811 PR tree-optimization/57539
13812 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13813 global.inlined_to of the new node to it. All callers changed.
13814 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13815 inlining_into, pass it to cgraph_clone_node.
13816 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13817 ipa_free_edge_args_substructures.
13818 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13819 rdesc linked list. Do not assert rdesc edges have inlined caller.
13820 Assert we have found an rdesc in the rdesc list.
13822 2013-06-24 Richard Biener <rguenther@suse.de>
13824 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13825 (pointer_set_lookup): Declare.
13826 (class pointer_map): New template class implementing a
13827 generic pointer to T map.
13828 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13829 pointer_map<T>::contains, pointer_map<T>::insert,
13830 pointer_map<T>::traverse): New functions.
13831 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13832 (pointer_set_lookup): New function.
13833 (pointer_set_contains): Use pointer_set_lookup.
13834 (pointer_set_insert): Likewise.
13835 (insert_aux): Remove.
13836 (struct pointer_map_t): Embed a pointer_set_t.
13837 (pointer_map_create): Adjust.
13838 (pointer_map_destroy): Likewise.
13839 (pointer_map_contains): Likewise.
13840 (pointer_map_insert): Likewise.
13841 (pointer_map_traverse): Likewise.
13842 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13843 pointer_map<unsigned> instead of a pointer_map_t.
13844 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13845 (streamer_tree_cache_lookup): Likewise.
13846 (streamer_tree_cache_create): Likewise.
13847 (streamer_tree_cache_delete): Likewise.
13848 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13849 pointer_map<unsigned> instead of a pointer_map_t.
13850 (lto_init_tree_ref_encoder): Adjust.
13851 (lto_destroy_tree_ref_encoder): Likewise.
13852 * lto-section-out.c (lto_output_decl_index): Likewise.
13853 (lto_record_function_out_decl_state): Likewise.
13854 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13856 2013-06-24 Richard Biener <rguenther@suse.de>
13858 PR tree-optimization/57488
13859 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13861 2013-06-24 Alan Modra <amodra@gmail.com>
13863 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13864 (gen_easy_altivec_constant): Likewise.
13865 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13866 easy_vector_constant_msb): Likewise.
13868 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13871 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13872 add missing return true.
13874 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13877 * config/sh/predicates.md (general_extend_operand): Invoke
13878 general_movsrc_operand for memory operands.
13879 (general_movsrc_operand): Allow reg+reg addressing, do not use
13880 general_operand for memory operands.
13882 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13884 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13885 when current target options does not apply.
13886 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13887 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13888 * config/i386/bmiintrin.h: Pass appropriate target
13889 attributes to header.
13890 * config/i386/mmintrin.h: Ditto.
13891 * config/i386/nmmintrin.h: Ditto.
13892 * config/i386/avx2intrin.h: Ditto.
13893 * config/i386/fxsrintrin.h: Ditto.
13894 * config/i386/tbmintrin.h: Ditto.
13895 * config/i386/xsaveintrin.h: Ditto.
13896 * config/i386/f16cintrin.h: Ditto.
13897 * config/i386/xtestintrin.h: Ditto.
13898 * config/i386/xsaveoptintrin.h: Ditto.
13899 * config/i386/bmi2intrin.h: Ditto.
13900 * config/i386/lzcntintrin.h: Ditto.
13901 * config/i386/smmintrin.h: Ditto.
13902 * config/i386/wmmintrin.h: Ditto.
13903 * config/i386/x86intrin.h: Remove all header include guards.
13904 * config/i386/prfchwintrin.h: Ditto.
13905 * config/i386/pmmintrin.h: Ditto.
13906 * config/i386/tmmintrin.h: Ditto.
13907 * config/i386/xmmintrin.h: Ditto.
13908 * config/i386/popcntintrin.h: Ditto.
13909 * config/i386/rdseedintrin.h: Ditto.
13910 * config/i386/ammintrin.h: Ditto.
13911 * config/i386/emmintrin.h: Ditto.
13912 * config/i386/immintrin.h: Remove all header include guards.
13913 * config/i386/fma4intrin.h: Ditto.
13914 * config/i386/lwpintrin.h: Ditto.
13915 * config/i386/xopintrin.h: Ditto.
13916 * config/i386/ia32intrin.h: Ditto.
13917 * config/i386/avxintrin.h: Ditto.
13918 * config/i386/rtmintrin.h: Ditto.
13919 * config/i386/fmaintrin.h: Ditto.
13920 * config/i386/mm3dnow.h: Ditto.
13922 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13924 * common/config/i386/i386-common.c: Handle LZCNT.
13926 2013-06-22 Andi Kleen <ak@linux.intel.com>
13928 * doc/extend.texi: Use __atomic_store_n instead of
13929 __atomic_store in HLE example.
13931 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13933 * config/sh/sh.c: Remove <cstdlib> workaround.
13935 2013-06-21 Andi Kleen <ak@linux.intel.com>
13937 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13939 2013-06-21 Andi Kleen <ak@linux.intel.com>
13941 * doc/extend.texi: Document that __atomic_clear and
13942 __atomic_test_and_set should only be used with bool.
13944 2013-06-20 Jan Hubicka <jh@suse.cz>
13946 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13947 types_same_for_odr.
13948 * tree.c (decls_same_for_odr): New function.
13949 (same_for_edr): New function.
13950 (types_same_for_odr): New function.
13951 (get_binfo_at_offset): Use it.
13952 * tree.h (types_same_for_odr): Declare.
13954 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13955 Jason Merrill <jason@redhat.com>
13957 * system.h: Include <cstdlib> as well as <stdlib.h>.
13959 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13962 * config/i386/i386.c (construct_container): Report error if
13963 long double is used with disabled x87 float returns.
13965 2013-06-20 Jan Hubicka <jh@suse.cz>
13967 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13969 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13971 PR rtl-optimization/57425
13972 PR rtl-optimization/57569
13973 * alias.c (write_dependence_p): Remove parameters mem_mode and
13974 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13975 Changed all callers.
13976 (canon_anti_dependence): Get comments and semantics in sync.
13977 Add parameter mem_canonicalized. Changed all callers.
13978 * rtl.h (canon_anti_dependence): Update prototype.
13980 2013-06-20 Richard Biener <rguenther@suse.de>
13982 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13983 byte case, inline streamer_read_uchar and defer section
13986 2013-06-20 Richard Biener <rguenther@suse.de>
13988 PR tree-optimization/57584
13989 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13990 SSA names into the expanded expression that take part in
13991 abnormal coalescing.
13993 2013-06-19 Sharad Singhai <singhai@google.com>
13995 * gcov.c (print_usage): Handle new option.
13996 (process_args): Ditto.
13997 (get_gcov_intermediate_filename): New function.
13998 (output_intermediate_file): New function.
13999 (output_gcov_file): New function
14000 (generate_results): Handle new option.
14001 (release_function): Relase demangled name.
14002 (read_graph_file): Handle demangled name.
14003 (output_lines): Ditto.
14004 * doc/gcov.texi: Document gcov intermediate format.
14006 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14009 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14010 (lra_emit_add): Use the functions. Add comment about Y as an
14013 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14016 * collect2.c (collect_atexit): New.
14017 (collect_exit): Delete.
14018 (main): Register collect_atexit with atexit.
14019 (collect_wait): Change collect_exit to exit.
14021 * collect2.h (collect_exit): Delete.
14022 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14024 2013-06-19 Wei Mi <wmi@google.com>
14026 PR rtl-optimization/57518
14027 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14028 if regno is used in paradoxical subreg.
14029 (update_equiv_regs): Check pdx_subregs[regno] before
14030 set a reg to be equivalent with a mem.
14032 2013-06-19 Matthias Klose <doko@ubuntu.com>
14035 * file-find.h (find_a_file): Add a mode parameter.
14036 * file-find.c (find_a_file): Likewise.
14037 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14038 with X_OK for the executables.
14039 * collect2.c (main): Call find_a_file with X_OK.
14041 2013-06-19 Steve Ellcey <sellcey@mips.com>
14044 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14045 Use NEXT_INSN instead of next_real_insn.
14047 2013-06-19 Jan Hubicka <jh@suse.cz>
14049 * cgraph.h (const_value_known_p): Replace by ...
14050 (ctor_for_folding): .. this one.
14051 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14052 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14053 * expr.c (expand_expr_real_1): Likewise.
14054 (string_constant): Likewise.
14055 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14056 * ipa.c (process_references): Likewise.
14057 (symtab_remove_unreachable_nodes): Likewise.
14058 * ipa-inline-analysis.c (param_change_prob): Likewise.
14059 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14060 (get_base_constructor): Likwise.
14061 * varpool.c (varpool_remove_node): Likewise.
14062 (varpool_remove_initializer): LIkewise.
14063 (dump_varpool_node): LIkwise.
14064 (const_value_known_p): Rewrite to ...
14065 (ctor_for_folding): ... this one.
14067 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14070 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14071 PERSONALITY in $PATH derived prefixes.
14073 2013-06-19 Jeff Law <law@redhat.com>
14075 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14078 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14079 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14082 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14084 * config/arm/vfp.md: Move VFP instruction classification documentation
14086 * config/arm/arm.md: ... here. Update instruction classification
14089 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14091 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14092 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14093 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14095 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14098 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14099 with NEXT_INSN. Use tablejump_p to check for jump table data
14102 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14105 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14106 that now in C++ the value is correct per the C++ standards.
14108 2013-06-19 Richard Biener <rguenther@suse.de>
14110 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14111 for global context.
14113 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14116 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14119 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14120 with next_active_insn.
14122 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14124 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14125 functions are inlined during failures to flag an error.
14126 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14127 in early inline pass.
14129 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14131 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14134 2013-06-18 Julian Brown <julian@codesourcery.com>
14136 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14137 Permit virtual register pre-reload if !strict.
14138 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14140 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14142 * config/arm/neon.md (movmisalign<mode>): Use
14143 neon_perm_struct_or_reg_operand instead of
14144 neon_struct_or_register_operand.
14145 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14146 neon_permissive_struct_operand instead of neon_struct_operand.
14147 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14148 neon_vector_mem_operand.
14149 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14150 neon_vector_mem_operand.
14151 (neon_permissive_struct_operand): New.
14152 (neon_struct_or_register_operand): Rename to...
14153 (neon_perm_struct_or_reg_operand): This. Adjust call to
14154 neon_vector_mem_operand.
14156 2013-06-18 Richard Biener <rguenther@suse.de>
14158 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14159 * lto-streamer.h: Include pointer-set.h.
14160 (struct lto_decl_slot): Remove.
14161 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14162 Remove next_index entry.
14163 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14164 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14165 (lto_init_tree_ref_encoder): Adjust.
14166 (lto_destroy_tree_ref_encoder): Likewise.
14167 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14168 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14169 (lto_output_decl_index): Adjust.
14170 (lto_new_out_decl_state): Likewise.
14171 (lto_record_function_out_decl_state): Likewise.
14172 * lto-streamer-out.c (copy_function): Likewise.
14174 2013-06-18 Richard Biener <rguenther@suse.de>
14176 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14177 * cgraphunit.c: Include cfgloop.h.
14178 (init_lowered_empty_function): Initialize the loop tree.
14179 (assemble_thunk): Insert new BBs into loops.
14181 2013-06-18 Richard Biener <rguenther@suse.de>
14183 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14184 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14185 the map from cache entry to cache index optional.
14186 (streamer_tree_cache_replace_tree): Adjust accordingly.
14187 (streamer_tree_cache_append): Likewise.
14188 (streamer_tree_cache_delete): Likewise.
14189 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14190 streamer cache map from cache entry to cache index.
14191 * lto-streamer-out.c (create_output_block): Adjust.
14193 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14195 * config/arm/arm.md (attribute "insn"): Move multiplication and
14196 division attributes to...
14197 (attribute "type"): ... here. Remove mult.
14198 (attribute "mul32"): New attribute.
14199 (attribute "mul64"): Add umaal.
14200 (*arm_mulsi3): Update attributes.
14201 (*arm_mulsi3_v6): Likewise.
14202 (*thumb_mulsi3): Likewise.
14203 (*thumb_mulsi3_v6): Likewise.
14204 (*mulsi3_compare0): Likewise.
14205 (*mulsi3_compare0_v6): Likewise.
14206 (*mulsi_compare0_scratch): Likewise.
14207 (*mulsi_compare0_scratch_v6): Likewise.
14208 (*mulsi3addsi): Likewise.
14209 (*mulsi3addsi_v6): Likewise.
14210 (*mulsi3addsi_compare0): Likewise.
14211 (*mulsi3addsi_compare0_v6): Likewise.
14212 (*mulsi3addsi_compare0_scratch): Likewise.
14213 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14214 (*mulsi3subsi): Likewise.
14215 (*mulsidi3adddi): Likewise.
14216 (*mulsi3addsi_v6): Likewise.
14217 (*mulsidi3adddi_v6): Likewise.
14218 (*mulsidi3_nov6): Likewise.
14219 (*mulsidi3_v6): Likewise.
14220 (*umulsidi3_nov6): Likewise.
14221 (*umulsidi3_v6): Likewise.
14222 (*umulsidi3adddi): Likewise.
14223 (*umulsidi3adddi_v6): Likewise.
14224 (*smulsi3_highpart_nov6): Likewise.
14225 (*smulsi3_highpart_v6): Likewise.
14226 (*umulsi3_highpart_nov6): Likewise.
14227 (*umulsi3_highpart_v6): Likewise.
14228 (mulhisi3): Likewise.
14229 (*mulhisi3tb): Likewise.
14230 (*mulhisi3bt): Likewise.
14231 (*mulhisi3tt): Likewise.
14232 (maddhisi4): Likewise.
14233 (*maddhisi4tb): Likewise.
14234 (*maddhisi4tt): Likewise.
14235 (maddhidi4): Likewise.
14236 (*maddhidi4tb): Likewise.
14237 (*maddhidi4tt): Likewise.
14238 (divsi3): Likewise.
14239 (udivsi3): Likewise.
14240 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14241 (thumb2_mulsi_short_compare0): Likewise.
14242 (thumb2_mulsi_short_compare0_scratch): Likewise.
14243 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14244 (1020mult2): Likewise.
14245 (1020mult3): Likewise.
14246 (1020mult4): Likewise.
14247 (1020mult5): Likewise.
14248 (1020mult6): Likewise.
14249 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14251 (cortex_a15_mult64): Likewise.
14252 (cortex_a15_sdiv): Likewise.
14253 (cortex_a15_udiv): Likewise.
14254 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14260 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14261 (pj4_ir_div): Likewise.
14262 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14263 (11_mult2): Likewise.
14264 (11_mult3): Likewise.
14265 (11_mult4): Likewise.
14266 (11_mult5): Likewise.
14267 (11_mult6): Likewise.
14268 (11_mult7): Likewise.
14269 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14270 (cortex_a8_mla): Likewise.
14271 (cortex_a8_mull): Likewise.
14272 (cortex_a8_smulwy): Likewise.
14273 (cortex_a8_smlald): Likewise.
14274 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14275 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14276 (cortex_r4_mul_3): Likewise.
14277 (cortex_r4_mla_4): Likewise.
14278 (cortex_r4_mla_3): Likewise.
14279 (cortex_r4_smlald): Likewise.
14280 (cortex_r4_mull): Likewise.
14281 (cortex_r4_sdiv): Likewise.
14282 (cortex_r4_udiv): Likewise.
14283 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14284 (cortex_a7_idiv): Likewise.
14285 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14286 (9_mult2): Likewise.
14287 (9_mult3): Likewise.
14288 (9_mult4): Likewise.
14289 (9_mult5): Likewise.
14290 (9_mult6): Likewise.
14291 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14292 (cortex_a53_sdiv): Likewise.
14293 (cortex_a53_udiv): Likewise.
14294 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14295 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14296 (mp626_mult2): Likewise.
14297 (mp626_mult3): Likewise.
14298 (mp626_mult4): Likewise.
14299 * config/arm/fa526.md (526_mult1): Update attribute change.
14300 (526_mult2): Likewise.
14301 * config/arm/arm-generic.md (mult): Update attribute change.
14302 (mult_ldsched_strongarm): Likewise.
14303 (mult_ldsched): Likewise.
14304 (multi_cycle): Likewise.
14305 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14306 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14307 (606te_mult2): Likewise.
14308 (606te_mult3): Likewise.
14309 (606te_mult4): Likewise.
14310 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14311 (cortex_a9_mac16): Likewise.
14312 (cortex_a9_multiply): Likewise.
14313 (cortex_a9_mac): Likewise.
14314 (cortex_a9_multiply_long): Likewise.
14315 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14316 (626te_mult2): Likewise.
14317 (626te_mult3): Likewise.
14318 (626te_mult4): Likewise.
14320 2013-06-18 Richard Biener <rguenther@suse.de>
14323 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14325 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14328 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14329 with next_active_insn.
14331 2013-06-18 Alan Modra <amodra@gmail.com>
14333 * config/rs6000/rs6000.h (enum data_align): New.
14334 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14335 (DATA_ABI_ALIGNMENT): Define.
14336 (CONSTANT_ALIGNMENT): Correct comment.
14337 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14338 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14340 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14342 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14343 ATTRIBUTE_UNUSED marking.
14345 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14347 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14348 alternative and update.
14349 (aarch64_dup_lanedi): Delete.
14350 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14351 * config/aarch64/aarch64-simd-builtins.def: Update.
14353 2013-06-17 Richard Biener <rguenther@suse.de>
14355 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14356 (lto_input_scc): Declare.
14357 (lto_input_tree_1): Likewise.
14358 (struct lto_stats_d): Add num_tree_bodies_output and
14359 num_pickle_refs_output.
14360 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14361 (lto_read_tree_1): Split out from ...
14362 (lto_read_tree): ... this.
14363 (lto_input_scc): New function.
14364 (lto_input_tree_1): Split out from ...
14365 (lto_input_tree): ... this. Handle LTO_tree_scc.
14366 (lto_data_in_create): Create the streamer cache without hashes.
14367 * lto-streamer-out.c (create_output_block): Create the streamer
14368 cache with hashes when not doing WPA.
14369 (lto_write_tree_1): Split out from ...
14370 (lto_write_tree): ... this.
14371 (get_symbol_initial_value): New function.
14372 (lto_output_tree_1): Split out from ...
14373 (lto_output_tree): ... this. Write trees as series of SCCs
14374 using a DFS walk via DFS_write_tree.
14375 (struct sccs, struct scc_entry): New types.
14376 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14377 (DFS_write_tree_body): New function.
14378 (DFS_write_tree): Likewise.
14379 (hash_tree): Likewise.
14380 (scc_entry_compare): Likewise.
14381 (hash_scc): Likewise.
14382 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14383 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14384 TREE_CHAIN as regular reference.
14385 (streamer_read_integer_cst): Remove.
14386 (streamer_get_pickled_tree): Adjust.
14387 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14388 of DECL_EXTERNALs in BLOCK_VARS for now.
14389 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14391 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14392 Add hash value argument and record that if hashes are recorded
14394 (streamer_tree_cache_insert_1): Adjust.
14395 (streamer_tree_cache_insert): Likewise.
14396 (streamer_tree_cache_insert_at): Rename to ...
14397 (streamer_tree_cache_replace_tree): ... this and adjust.
14398 (streamer_tree_cache_append): Adjust.
14399 (record_common_node): Likewise.
14400 (streamer_tree_cache_create): Add argument whether to
14401 record hash values together with trees.
14402 (streamer_tree_cache_delete): Adjust.
14403 * tree-streamer.h (struct streamer_tree_cache_d): Add
14405 (streamer_read_integer_cst): Remove.
14406 (streamer_tree_cache_insert): Adjust.
14407 (streamer_tree_cache_append): Likewise.
14408 (streamer_tree_cache_insert_at): Rename to ...
14409 (streamer_tree_cache_replace_tree): ... this and adjust.
14410 (streamer_tree_cache_create): Add argument whether to record hashes.
14411 (streamer_tree_cache_get): Rename to ...
14412 (streamer_tree_cache_get_tree): ... this.
14413 (streamer_tree_cache_get_hash): New function.
14414 * tree.c (cache_integer_cst): New function.
14415 * tree.h (cache_integer_cst): Declare.
14416 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14417 * lto-symtab.c (lto_varpool_replace_node): Only release
14418 DECL_INITIAL of non-prevailing decls.
14419 * varpool.c (varpool_remove_initializer): Do not release
14420 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14422 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14424 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14425 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14426 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14427 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14428 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14429 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14430 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14431 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14432 instead of TARGET_64BIT.
14433 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14434 Require ISA_HAS_<D>DIV.
14436 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14438 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14439 (mips*-*-linux*): Move default with_llsc setting to where other
14441 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14443 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14444 Likewise. Remove default with_tune setting. Move default float
14445 setting to its own block. Handle with_llsc in the same block as above.
14447 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14449 PR rtl-optimization/57425
14450 PR rtl-optimization/57569
14451 * alias.c (write_dependence_p): Add new parameters mem_mode,
14452 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14453 Changed all callers.
14454 (canon_anti_dependence): New function.
14455 * cse.c (check_dependence): Use canon_anti_dependence.
14456 * cselib.c (cselib_invalidate_mem): Likewise.
14457 * rtl.h (canon_anti_dependence): Declare.
14459 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14461 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14462 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14463 ".set mips3" for 64-bit targets.
14465 2013-06-15 Dehao Chen <dehao@google.com>
14467 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14468 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14469 (gimple_check_call_args): Likewise.
14470 * value-prof.c (check_ic_target): Likewise.
14471 * ipa-inline.c (early_inliner): Likewise.
14472 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14473 * cgraph.c (cgraph_create_edge_1): Likewise.
14474 (cgraph_make_edge_direct): Likewise.
14476 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14479 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14480 rs6000_output_move_128bit to handle emitting quad memory
14481 operations. Set attribute length to 8 bytes.
14483 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14485 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14487 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14488 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14489 (aarch64_<su>mlsl<mode>): Likewise.
14491 2013-06-14 Mike Stump <mikestump@comcast.net>
14493 * Makefile.in (TARGET_H): Add insn-codes.h.
14495 2013-06-14 Alan Modra <amodra@gmail.com>
14497 PR middle-end/57134
14498 PR middle-end/57586
14499 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14500 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14501 bitfield expansion when EXPAND_MEMORY.
14502 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14504 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14506 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14507 test for clearing quad memory on 32-bit later.
14509 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14511 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14512 (fold_negate_expr): Likewise.
14513 (fold_real_zero_addition_p): Handle vectors.
14514 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14516 2013-06-14 Alan Modra <amodra@gmail.com>
14518 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14520 2013-06-13 Jan Hubicka <jh@suse.cz>
14522 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14523 Local comdats are not externally visible.
14524 * symtab.c (dump_symtab_base): Dump externally visible.
14525 (verify_symtab_base): Verify back links in the symtab hash.
14527 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14529 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14530 CONVERT_EXPR as equal nodes.
14532 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14534 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14536 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14538 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14539 Generalize to complex and vector.
14540 * tree.c (build_all_ones_cst): New function.
14541 * tree.h (build_all_ones_cst): Declare it.
14543 2013-06-13 Alan Modra <amodra@gmail.com>
14545 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14546 * config/rs6000/rs6000.md (signbittf2): New insn.
14547 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14548 (abstf2_internal, cmptf_internal2): Likewise.
14549 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14551 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14552 Pat Haugen <pthaugen@us.ibm.com>
14553 Peter Bergner <bergner@vnet.ibm.com>
14555 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14556 power8 byte, half-word, and quad-word atomic instructions.
14557 (emit_store_conditional): Likewise.
14558 (rs6000_expand_atomic_compare_and_swap): Likewise.
14559 (rs6000_expand_atomic_op): Likewise.
14561 * config/rs6000/sync.md (larx): Add new modes for power8.
14563 (AINT): New mode iterator to include TImode as well as normal
14564 integer modes on power8.
14565 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14566 that VSX registers are not considered. Use AINT mode iterator
14567 instead of INT1 to allow inclusion of quad word atomic operations
14569 (load_locked<mode>): Likewise.
14570 (store_conditional<mode>): Likewise.
14571 (atomic_compare_and_swap<mode>): Likewise.
14572 (atomic_exchange<mode>): Likewise.
14573 (atomic_nand<mode>): Likewise.
14574 (atomic_fetch_<fetchop_name><mode>): Likewise.
14575 (atomic_nand_fetch<mode>): Likewise.
14576 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14578 (ATOMIC): On power8, add QImode, HImode modes.
14579 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14580 modes that promote to SImode.
14581 (load_lockedti): Convert TImode arguments to PTImode, so that we
14582 get a guaranteed even/odd register pair.
14583 (load_lockedpti): Likewise.
14584 (store_conditionalti): Likewise.
14585 (store_conditionalpti): Likewise.
14587 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14588 atomic load/store instructions.
14591 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14593 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14595 (insn_count): New attribute, with most cases extracted from...
14596 (length): ...here. Redefine most cases in terms of insn_count.
14597 (single_insn): Delete.
14598 (can_delay): Use insn_count to check for single instructions.
14599 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14600 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14601 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14602 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14603 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14604 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14605 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14606 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14607 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14608 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14609 rather than "length".
14610 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14611 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14612 Use "insn_count" rather than "length".
14613 * config/mips/mips-dsp.md
14614 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14615 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14618 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14620 PR tree-optimization/57361
14621 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14623 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14625 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14627 (aarch64_simd_combine<mode>): New instruction expansion.
14628 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14629 function prototype.
14630 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14631 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14633 2013-06-12 Jan Hubicka <jh@suse.cz>
14635 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14636 decl has when in streaming stage.
14637 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14638 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14640 2013-06-12 Roland Stigge <stigge@antcom.de>
14643 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14645 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14647 PR tree-optimization/57537
14648 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14649 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14651 2013-06-12 Richard Biener <rguenther@suse.de>
14653 * data-streamer.h (streamer_write_char_stream): CSE
14654 obs->current_pointer.
14655 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14656 streamer_write_char_stream manually and optimize the resulting loop.
14657 (streamer_write_hwi_stream): Likewise.
14659 2013-06-12 Jan Hubicka <jh@suse.cz>
14661 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14662 * cgraph.h (varpool_create_empty_node): Declare.
14663 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14665 * symtab.c (symtab_unregister_node): Be lax about missin entries
14667 (symtab_get_node): Update comment.
14668 * varpool.c (varpool_create_empty_node): Break out from ...
14669 (varpool_node_for_decl): ... here.
14670 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14672 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14674 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14675 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14676 part. Use straight-line flow at the end.
14677 <COMPONENT_REF>: Remove superfluous else.
14678 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14680 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14683 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14684 target hook even for !TREE_PUBLIC decls. If no resolution info
14685 is available, return false for common and external decls.
14687 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14689 * config/rl78/constraints.md (U): New constraint.
14690 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14693 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14696 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14697 to allow returning address to AT_PLATFORM name.
14699 2013-06-11 Jan Hubicka <jh@suse.cz>
14701 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14702 * cgraph.h (symtab_node_base): Add weakref flag.
14703 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14704 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14705 (output_weakrefs): Use weakref flag.
14706 * fold-const.c (simple_operand_p): Handle WEAK.
14707 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14708 * ipa.c (varpool_externally_visible_p): Drop weakref.
14709 (function_and_variable_visibility): Update comment; fix weakref
14710 sanity checks; do not clear DECL_WEAK on them.
14711 * lto-cgraph.c (lto_output_node): update.
14712 (lto_output_varpool_node): Update.
14713 (input_overwrite_node): Update.
14714 (input_node): Update.
14715 (input_varpool_node): Update.
14716 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14717 (lto_symtab_merge_symbols): Add sanity check.
14718 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14719 * passes.c (rest_of_decl_compilation): Set static flag, too.
14720 * symtab.c (dump_symtab_base): Dump weakref.
14721 (verify_symtab_base): Sanity check weakrefs.
14722 (symtab_make_decl_local): Remove duplicated code.
14723 (symtab_alias_ultimate_target): Simplify.
14724 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14726 2013-06-11 Tom de Vries <tom@codesourcery.com>
14728 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14729 sequence_vect == NULL.
14731 2013-06-11 DJ Delorie <dj@redhat.com>
14733 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14734 (rl78_unwind_word_mode): New.
14736 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14738 * final.c (debug_prefix_maps): Make static.
14740 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14742 * function.c (initial_trampoline): Remove stray copy.
14744 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14746 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14748 2013-06-11 Martin Jambor <mjambor@suse.cz>
14750 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14751 within bounds at the beginning of the function.
14753 2013-06-11 Alan Modra <amodra@gmail.com>
14755 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14757 (get_named_section): Don't NULL !DECL_P decl.
14759 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14761 * doc/invoke.texi (core-avx2): Document.
14763 (atom): Updated with MOVBE.
14765 2013-06-11 Richard Biener <rguenther@suse.de>
14767 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14769 2013-06-11 Anton Blanchard <anton@samba.org>
14771 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14772 correct shift value in little-endian mode.
14774 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14777 * varasm.c (get_variable_align): Move #endif to the right place.
14779 2013-06-10 Cary Coutant <ccoutant@google.com>
14781 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14782 for hash so that hash table traversal order is deterministic.
14784 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14785 Pat Haugen <pthaugen@us.ibm.com>
14786 Peter Bergner <bergner@vnet.ibm.com>
14788 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14789 of vectors in GPRS if they are direct moves or quad word load or
14792 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14794 (direct_move_p): Likewise.
14795 (quad_load_store_p): Likewise.
14797 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14798 classes into bins based on the physical register type.
14799 (reg_class_to_reg_type): Likewise.
14800 (IS_STD_REG_TYPE): Likewise.
14801 (IS_FP_VECT_REG_TYPE): Likewise.
14802 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14803 use direct move instructions.
14804 (reload_gpr_vsx): Likewise.
14805 (reload_vsx_gpr): Likewise.
14806 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14807 information that is a simplification of register classes. Also
14808 precalculate direct move reload helpers.
14809 (direct_move_p): New function to return true if the operation can
14810 be done as a direct move instruciton.
14811 (quad_load_store_p): New function to return true if the operation
14812 is a quad memory operation.
14813 (rs6000_legitimize_address): If quad memory, only allow register
14814 indirect for TImode addresses.
14815 (rs6000_legitimate_address_p): Likewise.
14816 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14817 (rs6000_reload_register_type): Likewise.
14818 (register_to_reg_type): Return register type.
14819 (rs6000_secondary_reload_simple_move): New helper function for
14820 secondary reload and secondary memory needed to identify anything
14821 that is a simple move, and does not need reloading.
14822 (rs6000_secondary_reload_direct_move): New helper function for
14823 secondary reload to identify cases that can be done with several
14824 instructions via the direct move instructions.
14825 (rs6000_secondary_reload_move): New helper function for secondary
14826 reload to identify moves between register types that can be done.
14827 (rs6000_secondary_reload): Add support for quad memory operations
14828 and for direct move.
14829 (rs6000_secondary_memory_needed): Likewise.
14830 (rs6000_debug_secondary_memory_needed): Change argument names.
14831 (rs6000_output_move_128bit): New function to return the move to
14832 use for 128-bit moves, including knowing about the various
14833 limitations of quad memory operations.
14835 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14836 memory operations. call rs6000_output_move_128bit for the actual
14837 instruciton(s) to generate.
14838 (vsx_movti_64bit): Likewise.
14840 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14841 (UNSPEC_P8V_MTVSRWZ): Likewise.
14842 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14843 (UNSPEC_P8V_MTVSRD): Likewise.
14844 (UNSPEC_P8V_XXPERMDI): Likewise.
14845 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14846 (UNSPEC_FUSION_GPR): Likewise.
14847 (FMOVE128_GPR): New iterator for direct move.
14848 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14849 (f32_sv): Likewise.
14850 (f32_dm): Likewise.
14851 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14852 loads and direct move instructions.
14853 (zero_extendsidi2_lfiwzx): Likewise.
14854 (extendsidi2_lfiwax): Likewise.
14855 (extendsidi2_nocell): Likewise.
14856 (floatsi<mode>2_lfiwax): Likewise.
14857 (lfiwax): Likewise.
14858 (floatunssi<mode>2_lfiwzx): Likewise.
14859 (lfiwzx): Likewise.
14860 (fix_trunc<mode>_stfiwx): Likewise.
14861 (fixuns_trunc<mode>_stfiwx): Likewise.
14862 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14863 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14864 (parity<mode>2_cmpb): Set length/type attr.
14865 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14866 for 'mr.' to fast_compare.
14867 (bpermd_<mode>): Change type attr to popcnt.
14868 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14869 (p8_mtvsrwz_1): Likewise.
14870 (p8_mtvsrwz_2): Likewise.
14871 (reload_fpr_from_gpr<mode>): Likewise.
14872 (p8_mtvsrd_1): Likewise.
14873 (p8_mtvsrd_2): Likewise.
14874 (p8_xxpermdi_<mode>): Likewise.
14875 (reload_vsx_from_gpr<mode>): Likewise.
14876 (reload_vsx_from_gprsf): Likewise.
14877 (p8_mfvsrd_3_<mode>): LIkewise.
14878 (reload_gpr_from_vsx<mode>): Likewise.
14879 (reload_gpr_from_vsxsf): Likewise.
14880 (p8_mfvsrd_4_disf): Likewise.
14881 (multi-word GPR splits): Do not split direct moves or quad memory
14884 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14886 * tree-into-ssa.c (interesting_blocks): Make static.
14888 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14891 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14892 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14893 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14894 (get_variable_align): New function.
14895 (get_variable_section, emit_bss, emit_common,
14896 assemble_variable_contents, place_block_symbol): Use
14897 get_variable_align instead of DECL_ALIGN.
14898 (assemble_noswitch_variable): Add align argument, use it
14899 instead of DECL_ALIGN.
14900 (assemble_variable): Adjust caller. Use get_variable_align
14901 instead of DECL_ALIGN.
14902 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14904 (DATA_ABI_ALIGNMENT): Define.
14905 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14906 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14907 opt is false, only return the psABI mandated alignment increase.
14908 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14909 (DATA_ABI_ALIGNMENT): ... this.
14910 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14911 (DATA_ABI_ALIGNMENT): ... this.
14912 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14913 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14914 (DATA_ABI_ALIGNMENT): ... this.
14915 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14916 * doc/tm.texi: Regenerated.
14918 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14920 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14921 cmp_code to construct REG_EQUAL note.
14923 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14926 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14927 that operands[2] doesn't overlap with operands[0].
14929 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14930 Jan Hubicka <jh@suse.cz>
14932 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14933 hack to mark symbols as used.
14935 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14937 PR rtl-optimization/57559
14938 * lra-constraints.c (process_alt_operands): Don't discourage
14939 memory with known offset for offsetable memory constraint.
14940 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14942 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14944 * varasm.c (struct oc_local_state): Reorder fields.
14945 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14946 and adjust accordingly.
14947 (output_constructor): Reorder initialization code and adjust call to
14948 output_constructor_bitfield.
14950 2013-06-07 Jan Hubicka <jh@suse.cz>
14952 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14954 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14956 * tree-object-size.c (unknown): Make const.
14958 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14960 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14961 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14962 for last alternative in the cpu_facility attribute.
14964 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14967 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14968 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14969 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14970 * config/arm/constraints.md (Dg): New constraint.
14971 * config/arm/neon.md (xordi3_neon): Remove.
14972 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14973 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14975 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14977 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14978 Clean up alternatives.
14980 2013-06-07 Alan Modra <amodra@gmail.com>
14982 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14985 2013-06-07 Alan Modra <amodra@gmail.com>
14987 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14989 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14991 * config/arm/constraints.md (Df): New constraint.
14992 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14993 Correct length attribute for last two alternatives.
14995 2013-06-07 Alan Modra <amodra@gmail.com>
14997 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14998 override user -mfp-in-toc.
14999 (offsettable_ok_by_alignment): Consider just the current access
15000 rather than the whole object, unless BLKmode. Handle
15001 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15002 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15003 for -mcmodel=medium.
15004 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15005 override user -mfp-in-toc or -msum-in-toc. Default to
15006 -mno-fp-in-toc for -mcmodel=medium.
15008 2013-06-06 DJ Delorie <dj@redhat.com>
15010 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15011 TARGET_VALID_POINTER_MODE.
15013 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15014 Pat Haugen <pthaugen@us.ibm.com>
15015 Peter Bergner <bergner@vnet.ibm.com>
15017 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15018 Document new power8 builtins.
15020 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15021 condition code register, to allow 128-bit logical operations to be
15022 done in the VSX or GPR registers.
15023 (nor<mode>3): Use the canonical form for nor.
15024 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15025 vclz*, and vpopcnt* vector instructions.
15026 (nand<mode>3): Likewise.
15027 (orc<mode>3): Likewise.
15028 (clz<mode>2): LIkewise.
15029 (popcount<mode>2): Likewise.
15031 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15032 that only the GPRs are recognized.
15034 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15035 support for new power8 builtins.
15037 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15039 (xscvdpspn): Likewise.
15045 (vpopcnt): Likewise.
15046 (vpopcntb): Likewise.
15047 (vpopcnth): Likewise.
15048 (vpopcntw): Likewise.
15049 (vpopcntd): Likewise.
15051 (vmrgew): Likewise.
15052 (vmrgow): Likewise.
15054 (eqv_v16qi3): Likewise.
15055 (eqv_v8hi3): Likewise.
15056 (eqv_v4si3): Likewise.
15057 (eqv_v2di3): Likewise.
15058 (eqv_v4sf3): Likewise.
15059 (eqv_v2df3): Likewise.
15061 (nand_v16qi3): Likewise.
15062 (nand_v8hi3): Likewise.
15063 (nand_v4si3): Likewise.
15064 (nand_v2di3): Likewise.
15065 (nand_v4sf3): Likewise.
15066 (nand_v2df3): Likewise.
15068 (orc_v16qi3): Likewise.
15069 (orc_v8hi3): Likewise.
15070 (orc_v4si3): Likewise.
15071 (orc_v2di3): Likewise.
15072 (orc_v4sf3): Likewise.
15073 (orc_v2df3): Likewise.
15075 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15076 allow power8 quad mode in 64-bit.
15077 (rs6000_builtin_vectorized_function): Add support to vectorize
15078 ISA 2.07 count leading zeros, population count builtins.
15079 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15080 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15081 (builtin_function_type): Add vgbbd builtin function which takes an
15083 (altivec_expand_vec_perm_const): Add support for new power8 merge
15086 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15087 that does not include TImdoe for use with 32-bit.
15088 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15090 (UNSPEC_VSX_CVDPSPN): Likewise.
15091 (vsx_xscvdpspn): Likewise.
15092 (vsx_xscvspdpn): Likewise.
15093 (vsx_xscvdpspn_scalar): Likewise.
15094 (vsx_xscvspdpn_directmove): Likewise.
15095 (vsx_and<mode>3): Split logical operations into 32-bit and
15096 64-bit. Add support to do logical operations on TImode as well as
15097 VSX vector types. Allow logical operations to be done in either
15098 VSX registers or in general purpose registers in 64-bit mode. Add
15099 splitters if GPRs were used. For AND, add clobber of CCmode to
15100 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15102 (vsx_and<mode>3_32bit): Likewise.
15103 (vsx_and<mode>3_64bit): Likewise.
15104 (vsx_ior<mode>3): Likewise.
15105 (vsx_ior<mode>3_32bit): Likewise.
15106 (vsx_ior<mode>3_64bit): Likewise.
15107 (vsx_xor<mode>3): Likewise.
15108 (vsx_xor<mode>3_32bit): Likewise.
15109 (vsx_xor<mode>3_64bit): Likewise.
15110 (vsx_one_cmpl<mode>2): Likewise.
15111 (vsx_one_cmpl<mode>2_32bit): Likewise.
15112 (vsx_one_cmpl<mode>2_64bit): Likewise.
15113 (vsx_nor<mode>3): Likewise.
15114 (vsx_nor<mode>3_32bit): Likewise.
15115 (vsx_nor<mode>3_64bit): Likewise.
15116 (vsx_andc<mode>3): Likewise.
15117 (vsx_andc<mode>3_32bit): Likewise.
15118 (vsx_andc<mode>3_64bit): Likewise.
15119 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15120 and xxlorc instructions.
15121 (vsx_eqv<mode>3_64bit): Likewise.
15122 (vsx_nand<mode>3_32bit): Likewise.
15123 (vsx_nand<mode>3_64bit): Likewise.
15124 (vsx_orc<mode>3_32bit): Likewise.
15125 (vsx_orc<mode>3_64bit): Likewise.
15127 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15129 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15131 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15132 (p8_vmrgow): Likewise.
15133 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15134 GPRs to be split under VSX.
15135 (p8v_clz<mode>2): Add power8 count leading zero support.
15136 (p8v_popcount<mode>2): Add power8 population count support.
15137 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15140 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15143 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15145 (vec_nand): Likewise.
15146 (vec_vclz): Likewise.
15147 (vec_vclzb): Likewise.
15148 (vec_vclzd): Likewise.
15149 (vec_vclzh): Likewise.
15150 (vec_vclzw): Likewise.
15151 (vec_vgbbd): Likewise.
15152 (vec_vmrgew): Likewise.
15153 (vec_vmrgow): Likewise.
15154 (vec_vpopcnt): Likewise.
15155 (vec_vpopcntb): Likewise.
15156 (vec_vpopcntd): Likewise.
15157 (vec_vpopcnth): Likewise.
15158 (vec_vpopcntw): Likewise.
15160 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15162 PR rtl-optimization/57468
15163 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15166 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15168 PR rtl-optimization/57459
15169 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15170 type when setting live regs.
15172 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15174 * config/s390/s390.opt (mlra): New option.
15175 * config/s390/s390.c (s390_decompose_address): Check displacement
15176 for all registers for LRA.
15177 (s390_secondary_reload): Don't used secondary reloads for LRA.
15178 (s390_lra_p): New function.
15179 (TARGET_LRA_P): Define.
15180 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15181 of attribute cpu_facility to zarch for the last alternative.
15182 (*cmpmem_short): Ditto.
15184 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15186 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15187 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15188 (arm_expand_prologue): Likewise.
15190 2013-06-06 Teresa Johnson <tejohnson@google.com>
15193 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15194 as this is now done by redirect_edge_and_branch_force.
15195 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15196 barriers, and fix interaction with splitting.
15197 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15198 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15199 reflect changes made in the routine.
15200 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15201 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15202 since this is called in cfglayout mode, and replace partition fixup
15203 with assert as that is now done by force_nonfallthru_and_redirect.
15204 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15205 already be marked with region crossing note.
15206 (insert_section_boundary_note): Make non-static, gate on flag
15207 has_bb_partition, rewrite to also check for multiple partitions.
15208 (rest_of_handle_reorder_blocks): Remove call to
15209 insert_section_boundary_note, now done later during free_cfg.
15210 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15211 * bb-reorder.h (insert_section_boundary_note): Declare.
15212 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15213 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15214 invoke insert_section_boundary_note.
15215 (try_redirect_by_replacing_jump): Remove unnecessary
15216 check for region crossing note.
15217 (fixup_partition_crossing): New function.
15218 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15219 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15220 in non-cfglayout mode.
15221 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15222 remove old code that tried to do this. Emit barrier correctly
15223 when we are in cfglayout mode.
15224 (last_bb_in_partition): New function.
15225 (rtl_split_edge): Correctly fixup partition boundaries.
15226 (commit_one_edge_insertion): Remove old code that tried to
15227 fixup region crossing edge since this is now handled in
15228 split_block, and set up insertion point correctly since
15229 block may now end in a jump.
15230 (verify_hot_cold_block_grouping): Guard against checking when not in
15231 linearized RTL mode.
15232 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15234 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15235 rtl_verify_flow_info, so not called in cfglayout mode.
15236 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15237 (fixup_reorder_chain): Remove old code that attempted to fixup region
15238 crossing note as this is now handled in force_nonfallthru_and_redirect.
15239 (duplicate_insn_chain): Don't duplicate switch section notes.
15240 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15242 * basic-block.h (emit_barrier_after_bb): Declare.
15244 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15246 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15247 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15248 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15250 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15252 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15254 * config/arm/ldmstm.md: Regenerate.
15256 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15258 * config/arm/sync.md (atomic_loaddi_1):
15259 Disable predication for arm_restrict_it.
15260 (arm_load_exclusive<mode>): Likewise.
15261 (arm_load_exclusivesi): Likewise.
15262 (arm_load_exclusivedi): Likewise.
15263 (arm_load_acquire_exclusive<mode>): Likewise.
15264 (arm_load_acquire_exclusivesi): Likewise.
15265 (arm_load_acquire_exclusivedi): Likewise.
15266 (arm_store_exclusive<mode>): Likewise.
15267 (arm_store_exclusive<mode>): Likewise.
15268 (arm_store_release_exclusivedi): Likewise.
15269 (arm_store_release_exclusive<mode>): Likewise.
15271 2013-06-06 Richard Biener <rguenther@suse.de>
15273 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15275 (lto_tag_is_tree_code_p): Adjust.
15276 (lto_tag_is_gimple_code_p): Likewise.
15277 (lto_gimple_code_to_tag): Likewise.
15278 (lto_tag_to_gimple_code): Likewise.
15279 (lto_tree_code_to_tag): Likewise.
15280 (lto_tag_to_tree_code): Likewise.
15281 * data-streamer.h (streamer_write_hwi_in_range): Use
15282 uhwi streaming to stream the normalized range.
15283 (streamer_read_hwi_in_range): Likewise.
15285 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15287 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15288 (predicable_short_it): Likewise.
15289 (predicated): Likewise.
15290 (enabled): Handle above.
15291 (define_cond_exec): Set predicated attribute to yes.
15293 2013-06-05 Mike Stump <mikestump@comcast.net>
15295 * gdbinit.in (__FUNCTION__): Add.
15297 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15299 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15300 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15302 2013-06-05 Jan Hubicka <jh@suse.cz>
15304 * varasm.c (mark_decl_referenced): Revert the removal until targets
15307 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15309 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15310 instead of mark_decl_referenced.
15312 2013-06-05 Jan Hubicka <jh@suse.cz>
15314 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15315 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15316 and symtab_used_from_object_file_p.
15317 (cgraph_make_node_local_1): Clear forced_by_abi.
15318 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15319 * cgraph.h (symtab_node_base): Add forced_by_abi.
15320 (decide_is_variable_needed): Remove.
15321 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15322 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15323 (decide_is_symbol_needed): ... this one; handle symbols in general;
15324 always analyze virtuals; honnor forced_by_abi.
15325 (cgraph_finalize_function): Update.
15326 (varpool_finalize_decl): Update.
15327 (symbol_defined_and_needed): Remove.
15328 (analyze_functions): Update.
15329 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15330 output_refs, input_overwrite_node): Handle forced_by_abi.
15331 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15332 (address_taken_from_non_vtable_p): ... this one.
15333 (comdat_can_be_unshared_p_1): New function.
15334 (cgraph_comdat_can_be_unshared_p): Rename to ...
15335 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15336 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15337 (function_and_variable_visibility): Clear forced_by_abi as needed.
15338 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15339 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15340 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15341 * varpool.c (decide_is_variable_needed): Remove.
15343 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15345 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15346 (arm_option_override): Override arm_restrict_it where appropriate.
15347 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15348 * config/arm/arm.opt (mrestrict-it): New command-line option.
15349 * doc/invoke.texi: Document -mrestrict-it.
15351 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15353 * tsan.c (tsan_atomic_table): Make const.
15355 2013-06-05 Richard Biener <rguenther@suse.de>
15357 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15358 index associated with the tree we are supposed to replace.
15359 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15360 TREE_ASM_WRITTEN as zero for everything but SSA names.
15362 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15364 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15366 2013-06-04 Jan Hubicka <jh@suse.cz>
15368 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15369 (input_node, input_varpool_node): Handle correctly external same
15371 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15372 nodes at ltrans stage.
15374 2013-06-04 Jan Hubicka <jh@suse.cz>
15376 * ipa-inline.c (update_caller_keys): Fix availability test.
15377 (update_callee_keys): Likewise.
15378 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15379 to follow ELF standard.
15381 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15383 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15384 (mips64r5900el-*-elf*): New configurations.
15385 * config/mips/mips-cpus.def (r5900): New processor.
15386 * config/mips/mips-tables.opt: Regenerate.
15387 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15388 (mips_issue_rate): Handle PROCESSOR_R5900.
15389 (mips_reorg_process_insns): Force reorder mode for the R5900.
15390 * config/mips/mips.h (TARGET_MIPS5900): Define.
15391 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15393 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15395 * config/mips/mips.md (processor): Add r5900.
15396 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15398 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15400 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15401 into function to generate MOVI instruction.
15402 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15403 (aarch64_preferred_simd_mode): Turn into wrapper.
15404 (aarch64_output_scalar_simd_mov_immediate): New function.
15405 * config/aarch64/aarch64-protos.h: Add prototype for above.
15407 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15409 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15410 element_char member.
15411 (sizetochar): Return signed char.
15412 (aarch64_simd_valid_immediate): Remove elchar and other
15413 unnecessary variables.
15414 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15415 Calculate element_char as required.
15416 * config/aarch64/aarch64-protos.h: Update and move prototype
15417 for aarch64_output_simd_mov_immediate.
15418 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15421 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15423 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15424 information completed by aarch64_simd_valid_immediate.
15425 (aarch64_legitimate_constant_p): Update arguments.
15426 (aarch64_simd_valid_immediate): Work with struct rather than many
15428 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15429 (aarch64_simd_make_constant): Update arguments.
15430 (aarch64_output_simd_mov_immediate): Work with struct rather than
15431 many pointers. Output immediate directly rather than as operand.
15432 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15434 * config/aarch64/constraints.md (Dn): Update arguments.
15436 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15438 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15440 (aarch64_simd_immediate_valid_for_move): Remove.
15441 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15442 (aarch64_simd_make_constant): Update call.
15443 (aarch64_output_simd_mov_immediate): Update call.
15444 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15446 * config/aarch64/constraints.md (Dn): Update call.
15448 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15450 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15451 return type to bool for prototype.
15452 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15453 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15454 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15456 2013-06-04 Catherine Moore <clm@codesourcery.com>
15458 * config/mips/mips.opt (meva): New.
15459 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15460 (ASM_SPEC): Handle -meva.
15461 * doc/invoke.texi (meva): Document.
15463 2013-06-04 Alan Modra <amodra@gmail.com>
15465 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15468 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15470 * rtl.def: Add extra fourth optional field to define_cond_exec.
15471 * gensupport.c (process_one_cond_exec): Process attributes from
15473 * doc/md.texi: Document fourth field in define_cond_exec.
15475 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15477 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15478 out the processing order as in store_bit_field_1.
15480 2013-06-04 Jan Hubicka <jh@suse.cz>
15482 PR middle-end/57500
15483 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15484 non-VAR_DECL node if it does not exist yet.
15486 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15488 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15489 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15490 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15491 target_cpu_default setting.
15493 2013-06-03 Teresa Johnson <tejohnson@google.com>
15495 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15496 default to -fopt-info=optimized instead of all.
15497 * doc/invoke.texi: Ditto.
15498 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15499 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15500 (execute_vect_slp): Emit BB vectorization success under
15501 MSG_OPTIMIZED_LOCATIONS.
15502 * tree-vect-slp.c (vect_slp_transform_bb): Change
15503 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15504 * tree-vect-loop.c (vect_transform_loop): Ditto.
15506 2013-06-03 Jason Merrill <jason@redhat.com>
15509 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15510 Use TARGET_EXPR for C++.
15512 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15514 PR rtl-optimization/57268
15515 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15516 if DEBUG_INSN_P (insn).
15519 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15521 PR rtl-optimization/57268
15522 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15523 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15525 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15527 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15528 (ix86_avoid_lea_for_addr): Likewise.
15529 (exact_dependency_1): Likewise.
15530 (ix86_adjust_cost): Likewise.
15531 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15533 (do_reorder_for_imul): Fix typo, formatting and
15534 !reload_completed check removed.
15535 (ix86_sched_reorder): Fix typo and formatting.
15536 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15539 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15541 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15543 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15545 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15546 <STRING_CST>: Likewise.
15547 <VECTOR_CST>: Likewise.
15549 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15550 Mikael Morin <mikael@gcc.gnu.org>
15552 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15553 * config.in: Regenerated.
15554 * configure: Regenerated.
15556 2013-06-01 Jan Hubicka <jh@suse.cz>
15558 PR middle-end/57366
15559 * cgraphunit.c (compile): When weakref is not supported,
15560 set up transparent aliases before final output pass.
15561 * varasm.c (assemble_alias): Do not try to do it here.
15563 2013-06-01 Jan Hubicka <jh@suse.cz>
15565 PR middle-end/57467
15566 * passes.c (for_per_function): Skip unanalyzed functions.
15568 2013-06-01 Jan Hubicka <jh@suse.cz>
15570 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15571 (lto_symtab_merge_symbols_1): ... this one.
15572 (lto_symtab_merge_cgraph_nodes): Rename to ...
15573 (lto_symtab_merge_symbols): ... this one; simplify.
15574 * cgraph.c (same_body_aliases_done): Rename to ...
15575 (cpp_implicit_aliases_done): ... this one.
15576 (cgraph_create_function_alias): Update.
15577 (cgraph_same_body_alias): Update.
15578 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15579 (verify_edge_corresponds_to_fndecl): Simplify.
15580 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15581 (cgraph_node): Remove same_body_alias.
15582 (varpool_node): Remove alias_of and extra_name_alias.
15583 (same_body_aliases_done): Rename to ..
15584 (cpp_implicit_aliases_done): ... this one.
15585 (symtab_alias_ultimate_target): Add default parameter.
15586 (symtab_resolve_alias): New function.
15587 (fixup_same_cpp_alias_visibility): Declare.
15588 (cgraph_function_node): Add default parameter.
15589 (cgraph_node_asm_name): Likewise.
15590 (cgraph_function_or_thunk_node): Add default parameter; do
15591 not ICE when it is NULL.
15592 (varpool_variable_node): Likewise.
15593 * tree-emutls.c (create_emultls_var): Update.
15594 (ipa_lower_emutls): Update.
15595 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15596 (cgraph_reset_node): Reset alias info.
15597 (cgraph_finalize_function): Update.
15598 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15599 (analyze_function): Simplify.
15600 (cgraph_process_same_body_aliases): Simplify.
15601 (analyze_functions): Fixup same body aliases.
15602 (handle_alias_pairs): Simplify.
15603 (assemble_thunk): Update.
15604 (assemble_thunks_and_aliases): Update.
15605 (output_weakrefs): Rewrite.
15606 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15607 (lto_output_varpool_node): Likewise.
15608 (compute_ltrans_boundary): Remve assert.
15609 (get_alias_symbol): New functoin.
15610 (input_node): Rewrite alias handling.
15611 (input_varpool_node): Likewise.
15612 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15613 * ipa.c (process_references): Handle weakrefs correctly.
15614 (symtab_remove_unreachable_nodes): Likewise.
15615 * trans-mem.c (get_cg_data): Update.
15616 (ipa_tm_create_version_alias): Update.
15617 (ipa_tm_execute): Update.
15618 * symtab.c (dump_symtab_base): Dump aliases.
15619 (verify_symtab_base): Verify aliases.
15620 (symtab_node_availability): New function.
15621 (symtab_alias_ultimate_target): Simplify.
15622 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15623 handle all the fixup cases.
15624 (symtab_resolve_alias): New function.
15625 * passes.c (ipa_write_summaries): Handle weakrefs.
15626 * varpool.c (varpool_analyze_node): Simplify.
15627 (assemble_aliases): Update.
15628 (varpool_create_variable_alias): Simplify.
15629 (varpool_extra_name_alias): Simplify.
15630 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15631 (lto_symtab_merge_symbols): ... this one.
15633 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15636 PR rtl-optimization/57268
15637 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15638 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15640 2013-06-01 Tobias Burnus <burnus@net-b.de>
15642 Partially reverted:
15643 2013-05-31 Tobias Burnus <burnus@net-b.de>
15645 PR middle-end/57073
15646 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15649 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15651 PR rtl-optimization/57268
15652 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15653 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15655 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15657 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15658 unordered comparison operators when -fno-trapping-math is in effect
15660 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15661 and implement unordered comparison operators properly on the e500.
15663 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15665 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15666 for constant scalar integers.
15667 (simplify_relational_operation_1): Likewise.
15669 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15671 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15672 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15675 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15676 Igor Zamyatin <igor.zamyatin@intel.com>
15678 Silvermont (SLM) architecture performance tuning.
15679 * config/i386/i386.h (enum ix86_tune_indices): Add
15680 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15681 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15683 * config/i386/i386.c (initial_ix86_tune_features)
15684 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15685 (ix86_lea_outperforms): Handle Silvermont tuning.
15686 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15688 (ix86_use_lea_for_mov): Likewise.
15689 (ix86_avoid_lea_for_addr): Likewise.
15690 (ix86_lea_for_add_ok): Likewise.
15691 (exact_dependency_1): New function.
15692 (exact_store_load_dependency): Likewise.
15693 (ix86_adjust_cost): Handle Silvermont tuning.
15694 (do_reoder_for_imul): Likewise.
15695 (swap_top_of_ready_list): New function.
15696 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15698 * config/i386/i386.md (peepholes that split memory operand in fp
15701 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15703 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15704 Remove un-necessary braces.
15706 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15708 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15709 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15711 2013-05-31 Tobias Burnus <burnus@net-b.de>
15713 PR middle-end/57073
15714 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15716 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15719 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15720 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15721 * config/arm/neon.md (iordi3_neon): Remove.
15722 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15723 * config/arm/predicates.md (imm_for_neon_logic_operand):
15724 Move to earlier in the file.
15725 (neon_logic_op2): Likewise.
15726 (arm_iordi_operand_neon): New predicate.
15728 2013-05-31 Richard Biener <rguenther@suse.de>
15730 PR tree-optimization/57478
15731 PR tree-optimization/57453
15732 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15735 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15737 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15738 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15740 2013-05-30 Tobias Burnus <burnus@net-b.de>
15741 Thomas Koenig <tkoenig@gcc.gnu.org>
15743 PR middle-end/57073
15744 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15745 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15747 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15749 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15751 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15753 * target.def (register_usage_leveling_p): New hook.
15754 * targhooks.c (default_register_usage_leveling_p): New.
15755 * targhooks.h (default_register_usage_leveling_p): New prototype.
15756 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15757 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15758 * doc/tm.texi: Update.
15759 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15761 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15763 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15764 (*insv_reg<mode>): New define_insn.
15766 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15768 PR rtl-optimization/57439
15769 * postreload.c (move2add_valid_value_p): Check that we have
15770 a zero subreg_regno_offset when accessing the register in
15771 the requested mode.
15773 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15774 Igor Zamyatin <igor.zamyatin@intel.com>
15776 Silvermont (SLM) architecture pipeline model, tuning and
15778 * config.gcc: Add slm config options and target.
15780 * config/i386/slm.md: New.
15782 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15784 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15786 (ix86_target_macros_internal): Likewise.
15788 * config/i386/i386.c (slm_cost): New cost.
15789 (m_SLM): New macro flag.
15790 (initial_ix86_tune_features): Set m_SLM.
15791 (x86_accumulate_outgoing_args): Likewise.
15792 (x86_arch_always_fancy_math_387): Likewise.
15793 (processor_target_table): Add slm cost.
15794 (cpu_names): Add slm cpu name.
15795 (x86_option_override_internal): Set SLM ISA.
15796 (ix86_issue_rate): New case PROCESSOR_SLM.
15797 (ia32_multipass_dfa_lookahead): Likewise.
15798 (fold_builtin_cpu): Add slm.
15800 * config/i386/i386.h (TARGET_SLM): New target macro.
15801 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15802 (processor_type): Add PROCESSOR_SLM.
15804 * config/i386/i386.md (cpu): Add new value "slm".
15805 (slm.md): Include slm.md.
15807 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15808 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15810 * config/arm/arm-protos.h: Add and update function protos.
15811 * config/arm/arm.c (use_simple_return_p): New added.
15812 (thumb2_expand_return): Check simple_return flag.
15813 * config/arm/arm.md: Add simple_return and conditional simple_return.
15814 * config/arm/iterators.md: Add iterator for return and simple_return.
15816 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15818 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15819 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15820 (arm_emit_vfp_multi_reg_pop): Likewise.
15821 (thumb2_emit_ldrd_pop): Likewise.
15822 (arm_expand_epilogue): Add misc REG_CFA notes.
15823 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15825 2013-05-29 Lawrence Crowl <crowl@google.com>
15827 * config/arm/t-arm: Update for below.
15829 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15830 Change type to hash_table. Update dependent calls and types.
15832 * config/i386/t-cygming: Update for below.
15834 * config/i386/t-interix: Update for below.
15836 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15837 Change type to hash_table. Update dependent calls and types.
15838 (i386_find_on_wrapper_list::wrappers): Likewise.
15840 * config/ia64/t-ia64: Update for below.
15842 * config/ia64/ia64.c (bundle_state_table):
15843 Change type to hash_table. Update dependent calls and types.
15845 * config/mips/mips.c (mips_reorg_process_insns::htab):
15846 Change type to hash_table. Update dependent calls and types.
15848 * config/sol2.c (solaris_comdat_htab):
15849 Change type to hash_table. Update dependent calls and types.
15851 * config/t-sol2: Update for above.
15853 2013-05-29 Teresa Johnson <tejohnson@google.com>
15855 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15856 functions are not yet marked as defined.
15858 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15859 Pat Haugen <pthaugen@us.ibm.com>
15860 Peter Bergner <bergner@vnet.ibm.com>
15862 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15866 (vrotl<mode>3): Likewise.
15867 (vashl<mode>3): Likewise.
15868 (vlshr<mode>3): Likewise.
15869 (vashr<mode>3): Likewise.
15871 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15872 support for power8 V2DI builtins.
15874 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15875 power8 V2DI builtins.
15876 (vupkhsw): Likewise.
15877 (vupklsw): Likewise.
15878 (vaddudm): Likewise.
15879 (vminsd): Likewise.
15880 (vmaxsd): Likewise.
15881 (vminud): Likewise.
15882 (vmaxud): Likewise.
15883 (vpkudum): Likewise.
15884 (vpksdss): Likewise.
15885 (vpkudus): Likewise.
15886 (vpksdus): Likewise.
15891 (vsubudm): Likewise.
15892 (vcmpequd): Likewise.
15893 (vcmpgtsd): Likewise.
15894 (vcmpgtud): Likewise.
15895 (vcmpequd_p): Likewise.
15896 (vcmpgtsd_p): Likewise.
15897 (vcmpgtud_p): Likewise.
15898 (vupkhsw): Likewise.
15899 (vupklsw): Likewise.
15900 (vaddudm): Likewise.
15901 (vmaxsd): Likewise.
15902 (vmaxud): Likewise.
15903 (vminsd): Likewise.
15904 (vminud): Likewise.
15905 (vpksdss): Likewise.
15906 (vpksdus): Likewise.
15907 (vpkudum): Likewise.
15908 (vpkudus): Likewise.
15913 (vsubudm): Likewise.
15915 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15916 support for power8 V2DI instructions.
15918 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15919 power8 V2DI instructions. Combine pack and unpack insns to use an
15920 iterator for each mode. Check whether a particular mode supports
15921 Altivec instructions instead of just checking TARGET_ALTIVEC.
15922 (UNSPEC_VPKUWUM): Likewise.
15923 (UNSPEC_VPKSHSS): Likewise.
15924 (UNSPEC_VPKSWSS): Likewise.
15925 (UNSPEC_VPKUHUS): Likewise.
15926 (UNSPEC_VPKSHUS): Likewise.
15927 (UNSPEC_VPKUWUS): Likewise.
15928 (UNSPEC_VPKSWUS): Likewise.
15929 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15930 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15931 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15932 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15933 (UNSPEC_VUPKHSB): Likewise.
15934 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15935 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15936 (UNSPEC_VUPKHSH): Likewise.
15937 (UNSPEC_VUPKLSB): Likewise.
15938 (UNSPEC_VUPKLSH): Likewise.
15940 (VI_char): Likewise.
15941 (VI_scalar): Likewise.
15942 (VI_unit): Likewise.
15944 (VP_small): Likewise.
15945 (VP_small_lc): Likewise.
15946 (VU_char): Likewise.
15947 (add<mode>3): Likewise.
15948 (altivec_vaddcuw): Likewise.
15949 (altivec_vaddu<VI_char>s): Likewise.
15950 (altivec_vadds<VI_char>s): Likewise.
15951 (sub<mode>3): Likewise.
15952 (altivec_vsubcuw): Likewise.
15953 (altivec_vsubu<VI_char>s): Likewise.
15954 (altivec_vsubs<VI_char>s): Likewise.
15955 (altivec_vavgs<VI_char>): Likewise.
15956 (altivec_vcmpbfp): Likewise.
15957 (altivec_eq<mode>): Likewise.
15958 (altivec_gt<mode>): Likewise.
15959 (altivec_gtu<mode>): Likewise.
15960 (umax<mode>3): Likewise.
15961 (smax<mode>3): Likewise.
15962 (umin<mode>3): Likewise.
15963 (smin<mode>3): Likewise.
15964 (altivec_vpkuhum): Likewise.
15965 (altivec_vpkuwum): Likewise.
15966 (altivec_vpkshss): Likewise.
15967 (altivec_vpkswss): Likewise.
15968 (altivec_vpkuhus): Likewise.
15969 (altivec_vpkshus): Likewise.
15970 (altivec_vpkuwus): Likewise.
15971 (altivec_vpkswus): Likewise.
15972 (altivec_vpks<VI_char>ss): Likewise.
15973 (altivec_vpks<VI_char>us): Likewise.
15974 (altivec_vpku<VI_char>us): Likewise.
15975 (altivec_vpku<VI_char>um): Likewise.
15976 (altivec_vrl<VI_char>): Likewise.
15977 (altivec_vsl<VI_char>): Likewise.
15978 (altivec_vsr<VI_char>): Likewise.
15979 (altivec_vsra<VI_char>): Likewise.
15980 (altivec_vsldoi_<mode>): Likewise.
15981 (altivec_vupkhsb): Likewise.
15982 (altivec_vupkhs<VU_char>): Likewise.
15983 (altivec_vupkls<VU_char>): Likewise.
15984 (altivec_vupkhsh): Likewise.
15985 (altivec_vupklsb): Likewise.
15986 (altivec_vupklsh): Likewise.
15987 (altivec_vcmpequ<VI_char>_p): Likewise.
15988 (altivec_vcmpgts<VI_char>_p): Likewise.
15989 (altivec_vcmpgtu<VI_char>_p): Likewise.
15990 (abs<mode>2): Likewise.
15991 (vec_unpacks_hi_v16qi): Likewise.
15992 (vec_unpacks_hi_v8hi): Likewise.
15993 (vec_unpacks_lo_v16qi): Likewise.
15994 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15995 (vec_unpacks_lo_v8hi): Likewise.
15996 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15997 (vec_pack_trunc_v8h): Likewise.
15998 (vec_pack_trunc_v4si): Likewise.
15999 (vec_pack_trunc_<mode>): Likewise.
16001 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16003 (vec_vmaxsd): Likewise.
16004 (vec_vmaxud): Likewise.
16005 (vec_vminsd): Likewise.
16006 (vec_vminud): Likewise.
16007 (vec_vpksdss): Likewise.
16008 (vec_vpksdus): Likewise.
16009 (vec_vpkudum): Likewise.
16010 (vec_vpkudus): Likewise.
16011 (vec_vrld): Likewise.
16012 (vec_vsld): Likewise.
16013 (vec_vsrad): Likewise.
16014 (vec_vsrd): Likewise.
16015 (vec_vsubudm): Likewise.
16016 (vec_vupkhsw): Likewise.
16017 (vec_vupklsw): Likewise.
16019 2013-05-29 Jan Hubicka <jh@suse.cz>
16021 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16022 flags; reorder rest of fields in more consistent way.
16023 (varpool_node): Remove analyzed, finalized and alias.
16024 (cgraph_ndoe): Likewise.
16025 (symtab_alias_ultimate_target): New function.
16026 (cgraph_function_node): Move offline.
16027 (cgraph_reset_node): Declare.
16028 (cgraph_comdat_can_be_unshared_p): Remove.
16029 (varpool_remove_initializer): Declare.
16030 (varpool_first_defined_variable, varpool_next_defined_variable
16031 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16032 (cgraph_function_with_gimple_body_p): Update.
16033 (varpool_all_refs_explicit_p): Update.
16034 (symtab_alias_target): New function.
16035 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16036 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16037 (cgraph_function_or_thunk_node): Simplify using
16038 symtab_alias_ultimate_target.
16039 (varpool_variable_node): Likewise.
16040 * cgraph.c (cgraph_create_function_alias): Update.
16041 (cgraph_add_thunk): Update.
16042 (cgraph_remove_node): Update.
16043 (dump_cgraph_node): Do not dump removed flags.
16044 (cgraph_function_body_availability): Update.
16045 (cgraph_propagate_frequency): Update.
16046 (verify_cgraph_node): Check sanity of local flag.
16047 (cgraph_function_node): Move here from cgraph.h; revamp for
16048 cgraph_function_or_thunk_node.
16049 * lto-symtab.c (lto_varpool_replace_node): Update.
16050 (lto_symtab_resolve_can_prevail_p): Update.
16051 (lto_symtab_merge_cgraph_nodes): Update.
16052 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16053 propagate_constants_accross_call, devirtualization_time_bonus,
16054 ipcp_propagate_stage): Update.
16055 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16056 * ipa-inline-transform.c (clone_inlined_nodes,
16057 preserve_function_body_p): Update.
16058 * ipa-reference.c (propagate): Update.
16059 (write_node_summary_p): Update.
16060 * toplev.c (wrapup_global_declaration_2): Update.
16061 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16062 (analyze_function) ... this one.
16063 (cgraph_process_new_functions): Update.
16064 (cgraph_reset_node): Export.
16065 (cgraph_finalize_function): Update.
16066 (cgraph_add_new_function): Update.
16067 (process_function_and_variable_attributes): Update.
16068 (varpool_finalize_decl): Update.
16069 (symbol_finalized): Remove.
16070 (symbol_finalized_and_needed): Rename to ...
16071 (symbol_defined_and_needed): ... update.
16072 (cgraph_analyze_functions): Update.
16073 (handle_alias_pairs): Update.
16074 (mark_functions_to_output): Update.
16075 (assemble_thunk): Update.
16076 (output_in_order): Update.
16077 (output_weakrefs): Update.
16078 (finalize_compilation_unit): Update.
16079 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16080 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16081 input_node, input_varpool_node): Update.
16082 * dbxout.c (dbxout_expand_expr): Update.
16083 * cgraphclones.c (cgraph_clone_node): Update.
16084 (cgraph_copy_node_for_versioning): Update.
16085 (cgraph_materialize_clone): Update.
16086 (cgraph_materialize_all_clones): Update.
16087 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16088 propagate_pure_const, propagate_nothrow): Update.
16089 * lto-streamer-out.c (lto_output, write_symbol): Update.
16090 * ipa-utils.c (ipa_reverse_postorder): Update.
16091 * ipa-inline.c (can_inline_edge_p): Update.
16092 (update_caller_keys, ipa_inline): Update.
16093 * dwarf2out.c (reference_to_unused,
16094 premark_types_used_by_global_vars_helper): Update.
16095 * tree-eh.c (tree_could_trap_p): Update.
16096 * ipa-split.c (consider_split, execute_split_functions): Update.
16097 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16098 has_addr_references_p): Update; move ahead in file for better
16100 (process_references): Simplify.
16101 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16102 bodies are removed.
16103 (cgraph_comdat_can_be_unshared_p): Make static.
16104 (cgraph_externally_visible_p): Update.
16105 (varpool_externally_visible_p): Update.
16106 (function_and_variable_visibility): Update.
16107 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16108 ipa_tm_mark_force_output_node): Update.
16109 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16110 estimate_edge_devirt_benefit, inline_generate_summary,
16111 inline_write_summary): Update.
16112 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16113 * ipa-prop.c (ipa_compute_jump_functions): Update.
16114 (ipa_print_node_params, ipa_prop_read_section,
16115 ipa_update_after_lto_read, read_replacements_section): Update.
16116 * varasm.c (mark_decl_referenced): Update.
16117 (assemble_alias, dump_tm_clone_pairs): Update.
16118 * tree-inline.c (copy_bb): Update.
16119 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16121 * symtab.c (dump_symtab_base): Print new flags.
16122 (verify_symtab_base): Verify new flags.
16123 (symtab_alias_ultimate_target): New function.
16124 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16125 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16127 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16129 * i386.c (ix86_get_function_versions_dispatcher,
16130 ix86_generate_version_dispatcher_body): Update.
16131 (fold_builtin_cpu): Use varpool_add_new_variable.
16132 * varpool.c (varpool_remove_initializer): Break out from ...
16133 (varpool_remove_node): ... this one.
16134 (dump_varpool_node, varpool_node_for_asm,
16135 cgraph_variable_initializer_availability, varpool_analyze_node,
16136 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16137 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16140 2013-05-29 Jan Hubicka <jh@suse.cz>
16142 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16144 2013-05-29 Easwaran Raman <eraman@google.com>
16146 PR tree-optimization/57442
16147 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16148 when control exits the main loop.
16150 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16152 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16154 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16155 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16156 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16159 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16161 PR tree-optimization/57441
16162 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16163 Don't limit size of incr_vec to number of candidates.
16165 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16167 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16168 and mips16 directories.
16169 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16170 (MULTILIB_DIRNAMES): Ditto.
16171 (MULTILIB_EXCEPTIONS): Add new exceptions.
16172 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16173 (MULTILIB_DIRNAMES): Ditto.
16174 (MULTILIB_EXCEPTIONS): Add new exceptions.
16176 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16177 Marcus Shawcroft <marcus.shawcroft@arm.com>
16179 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16180 SYMBOL_TINY_ABSOLUTE.
16181 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16182 SYMBOL_TINY_ABSOLUTE.
16183 (aarch64_expand_mov_immediate): Likewise.
16184 (aarch64_classify_symbol): Likewise.
16185 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16186 Permit SYMBOL_TINY_ABSOLUTE.
16187 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16189 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16190 Marcus Shawcroft <marcus.shawcroft@arm.com>
16192 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16193 Refactor if/switch. Replace gcc_assert with if.
16195 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16197 * config/i386/i386.c (initial_ix86_tune_features): Enable
16198 FP Reassociation for AMD bdver1 and bdver2.
16200 2013-05-29 Martin Jambor <mjambor@suse.cz>
16202 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16203 and IMAGPART_EXPR do not occur within other handled_components.
16205 2013-05-29 Richard Biener <rguenther@suse.de>
16207 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16208 access on whether the use is in the BB we currently try to
16210 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16211 vectorize to vect_bb_slp_scalar_cost.
16213 2013-05-29 Richard Biener <rguenther@suse.de>
16215 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16216 computing scalar cost offsetted by stmts that are kept live
16218 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16219 for computation of scalar cost.
16221 2013-05-28 Steve Ellcey <sellcey@mips.com>
16223 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16225 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16227 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16228 array notation built-in reduction functions.
16229 * doc/passes.texi (Passes): Added documentation about changes done
16231 * doc/invoke.texi (C Dialect Options): Added documentation about
16232 the -fcilkplus flag.
16233 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16234 (BUILTINS_DEF): Depend on cilkplus.def.
16235 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16236 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16237 * cilkplus.def: New file.
16239 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16241 PR rtl-optimization/57439
16242 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16244 2013-05-28 Easwaran Raman <eraman@google.com>
16246 PR tree-optimization/57337
16247 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16248 (find_insert_point): Correctly identify the insertion point
16249 when two statements with the same UID is compared.
16251 2013-05-28 Richard Biener <rguenther@suse.de>
16253 PR tree-optimization/56787
16254 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16255 from the list of data references.
16256 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16258 (vect_analyze_loop_operations): Likewise.
16259 (vect_transform_loop): Remove clobbers.
16261 2013-05-28 Martin Jambor <mjambor@suse.cz>
16263 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16264 and REALPART_EXPRs have scalar type.
16266 2013-05-28 Richard Biener <rguenther@suse.de>
16268 PR tree-optimization/57411
16269 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16271 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16272 virtual operand propagation.
16274 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16276 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16277 destination register for bmasksi_vis.
16278 (vector_init_bshuffle): Likewise.
16279 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16281 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16283 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16284 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16285 mode if the instruction isn't available in the original mode.
16286 * config/sparc/sparc.opt (mfix-ut699): New option.
16287 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16288 (divdf3): Turn into expander.
16289 (divdf3_nofix): New insn.
16290 (divdf3_fix): Likewise.
16291 (divsf3): Disable if -mfix-ut699.
16292 (sqrtdf2): Turn into expander.
16293 (sqrtdf2_nofix): New insn.
16294 (sqrtdf2_fix): Likewise.
16295 (sqrtsf2): Disable if -mfix-ut699.
16297 2013-05-27 Richard Biener <rguenther@suse.de>
16299 PR middle-end/57412
16300 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16301 block for the new loop.
16303 2013-05-27 Richard Biener <rguenther@suse.de>
16305 PR tree-optimization/57343
16306 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16307 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16308 (number_of_iterations_cond): Do not build the folded tree.
16310 2013-05-27 Richard Biener <rguenther@suse.de>
16313 PR middle-end/57381
16314 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16315 OEP_CONSTANT_ADDRESS_OF retained.
16317 PR tree-optimization/57417
16318 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16319 for unchanged base.
16320 (set_ssa_val_to): Compare addresses using
16321 get_addr_base_and_unit_offset.
16323 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16325 PR rtl-optimization/56833
16326 * postreload.c (move2add_record_mode): New function.
16327 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16328 (move2add_use_add2_insn): Use move2add_record_sym_value.
16329 (move2add_use_add3_insn): Likewise.
16330 (reload_cse_move2add): Use move2add_valid_value_p and
16331 move2add_record_mode. Invalidate call-clobbered and REG_INC
16332 affected regs by setting reg_mode to VOIDmode.
16333 (move2add_note_store): Don't pretend the inside of a SUBREG is
16334 the actual destination. Invalidate single/leading registers by
16335 setting reg_mode to VOIDmode.
16336 Use move2add_record_sym_value, move2add_valid_value_p and
16337 move2add_record_mode.
16339 2013-05-27 Richard Biener <rguenther@suse.de>
16341 PR tree-optimization/57396
16342 * tree-affine.c (double_int_constant_multiple_p): Properly
16343 return false for val == 0 and div != 0.
16345 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16347 * config/mips/mips.h: Use #elif in preprocessor conditions.
16349 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16352 * config/mips/constraints.md (kl): New constraint.
16353 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16354 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16355 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16356 from LO for MIPS16.
16357 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16359 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16362 * config/mips/mips.c (mips_can_inline_p): New function.
16363 (TARGET_CAN_INLINE_P): Define.
16365 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16367 * sched-int.h (ds_t, dw_t): Make unsigned int.
16368 Fix documentation that describes how all the ds_t bits are used.
16369 Reserve the last bit for delayed-branch scheduling.
16370 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16371 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16372 (gen_dep_weak_1): Remove prototype.
16373 * sched-deps.c (get_dep_weak_1): Make static.
16374 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16375 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16376 * doc/tm.texi: Regenerate.
16377 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16379 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16382 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16384 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16385 Sandra Loosemore <sandra@codesourcery.com>
16387 * config.gcc (powerpc-*): Allow native for with-cpu.
16389 2013-05-24 Jeff Law <law@redhat.com>
16391 PR tree-optimization/57124
16392 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16393 conversion feeding a condition if the range has an overflow
16394 if -fstrict-overflow. Add warnings for when we do make the
16397 2013-05-24 Dehao Chen <dehao@google.com>
16399 * tree-cfg.c (locus_discrim_map): Fix the typo.
16400 (locus_discrim_hasher): Likewise.
16401 (locus_discrim_hasher::hash): Likewise.
16402 (locus_discrim_hasher::equal): Likewise.
16404 2013-05-24 Martin Jambor <mjambor@suse.cz>
16406 PR tree-optimization/57294
16407 * cgraph.h (ipa_record_stmt_references): Declare.
16408 * cgraphbuild.c (ipa_record_stmt_references): New function.
16409 (build_cgraph_edges): Use ipa_record_stmt_references.
16410 (rebuild_cgraph_edges): Likewise.
16411 (cgraph_rebuild_references): Likewise.
16412 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16413 associated with the old statement and build references from the
16414 newly built statements.
16415 * ipa-ref.c (ipa_remove_stmt_references): New function.
16416 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16418 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16420 * lra-constraints.c (emit_spill_move): Use smaller mode for
16422 (check_and_process_move): Consider mem-reg moves for secondary
16424 (curr_insn_transform): Don't lose insns emitted before for
16425 secondary memory moves.
16426 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16427 reg set up in the current insn.
16429 2013-05-24 Dehao Chen <dehao@google.com>
16431 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16433 (locus_descrim_hasher::equal): Likewise.
16434 (build_gimple_cfg): New discriminator assignment algorithm.
16435 (make_edges): Likewise.
16436 (next_discriminator_for_locus): Likewise.
16437 (same_line_p): Likewise.
16438 (assign_discriminators): Likewise.
16439 (make_cond_expr_edges): Likewise.
16440 (make_gimple_switch_edges): Likewise.
16441 (make_goto_expr_edges): Likewise.
16442 (make_gimple_asm_edges): Likewise.
16444 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16446 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16447 X format specifier to only display bottom 16 bits.
16448 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16449 immediate to match for operand 2, since it will be masked.
16451 2013-05-24 Richard Biener <rguenther@suse.de>
16453 PR tree-optimization/57287
16454 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16455 all SSA names that occur in abnormal PHIs.
16457 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16460 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16461 that index is not negative.
16463 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16465 PR rtl-optimization/55177
16466 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16467 (simplify_byte_swapping_operation): New.
16468 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16469 (simplify_relational_operation_1): Deal with BSWAP.
16471 2013-05-23 Richard Henderson <rth@redhat.com>
16474 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16475 (ix86_reorg): Call it.
16477 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16480 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16481 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16482 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16484 2013-05-23 Christian Bruel <christian.bruel@st.com>
16487 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16489 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16490 Marcus Shawcroft <marcus.shawcroft@arm.com>
16492 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16493 * config/aarch64/constraints.md (Usa): Remove.
16494 * doc/md.texi (AArch64 Usa): Remove.
16496 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16497 Marcus Shawcroft <marcus.shawcroft@arm.com>
16499 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16500 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16501 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16502 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16504 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16506 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16507 instruction (AdvSIMD).
16508 * config/aarch64/aarch64-builtins.c
16509 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16510 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16512 2013-05-23 Martin Jambor <mjambor@suse.cz>
16514 PR middle-end/57347
16515 * tree.h (contains_bitfld_component_ref_p): Declare.
16516 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16517 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16519 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16520 not access a bit-field. Assert all final offsets are byte-aligned.
16522 2013-05-23 Richard Biener <rguenther@suse.de>
16524 PR tree-optimization/57380
16525 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16526 least one invariant or re-used load.
16527 * passes.c (init_optimization_passes): Move pass_phiprop before
16530 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16532 * config/aarch64/aarch64-simd.md
16533 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16535 2013-05-23 Richard Biener <rguenther@suse.de>
16537 PR middle-end/57381
16538 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16539 OEP_CONSTANT_ADDRESS_OF retained.
16541 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16543 PR middle-end/57344
16544 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16545 don't lower unit. Handle unit not being always BITS_PER_WORD.
16547 2013-05-23 Richard Biener <rguenther@suse.de>
16549 PR rtl-optimization/57341
16550 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16551 instead of true_dependence.
16553 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16555 * bb-reorder.c (branch_threshold): Make const.
16556 (exec_threshold): Ditto.
16558 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16559 Pat Haugen <pthaugen@us.ibm.com>
16560 Peter Bergner <bergner@vnet.ibm.com>
16562 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16563 documentation for the power8 crypto builtins.
16565 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16567 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16568 macros for defining power8 builtin functions.
16569 (BU_P8V_AV_2): Likewise.
16570 (BU_P8V_AV_P): Likewise.
16571 (BU_P8V_VSX_1): Likewise.
16572 (BU_P8V_OVERLOAD_1): Likewise.
16573 (BU_P8V_OVERLOAD_2): Likewise.
16574 (BU_CRYPTO_1): Likewise.
16575 (BU_CRYPTO_2): Likewise.
16576 (BU_CRYPTO_3): Likewise.
16577 (BU_CRYPTO_OVERLOAD_1): Likewise.
16578 (BU_CRYPTO_OVERLOAD_2): Likewise.
16579 (XSCVSPDP): Fix typo, point to the correct instruction.
16580 (VCIPHER): Add power8 crypto builtins.
16581 (VCIPHERLAST): Likewise.
16582 (VNCIPHER): Likewise.
16583 (VNCIPHERLAST): Likewise.
16584 (VPMSUMB): Likewise.
16585 (VPMSUMH): Likewise.
16586 (VPMSUMW): Likewise.
16587 (VPERMXOR_V2DI): Likewise.
16588 (VPERMXOR_V4SI: Likewise.
16589 (VPERMXOR_V8HI: Likewise.
16590 (VPERMXOR_V16QI: Likewise.
16591 (VSHASIGMAW): Likewise.
16592 (VSHASIGMAD): Likewise.
16593 (VPMSUM): Likewise.
16594 (VPERMXOR): Likewise.
16595 (VSHASIGMA): Likewise.
16597 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16598 __CRYPTO__ if the crypto instructions are available.
16599 (altivec_overloaded_builtins): Add support for overloaded power8
16602 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16603 support for power8 crypto builtins.
16604 (builtin_function_type): Likewise.
16605 (altivec_init_builtins): Add support for builtins that take vector
16606 long long (V2DI) arguments.
16608 * config/rs6000/crypto.md: New file, define power8 crypto
16611 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16612 Pat Haugen <pthaugen@us.ibm.com>
16613 Peter Bergner <bergner@vnet.ibm.com>
16615 * doc/invoke.texi (Option Summary): Add power8 options.
16616 (RS/6000 and PowerPC Options): Likewise.
16618 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16619 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16620 wm, wn, wr documentation.
16622 * config/rs6000/constraints.md (wm): New constraint for VSX
16623 registers if direct move instructions are enabled.
16624 (wn): New constraint for no registers.
16625 (wq): New constraint for quad word even GPR registers.
16626 (wr): New constraint if 64-bit instructions are enabled.
16627 (wv): New constraint if power8 vector instructions are enabled.
16628 (wQ): New constraint for quad word memory locations.
16630 * config/rs6000/predicates.md (const_0_to_15_operand): New
16631 constraint for 0..15 for crypto instructions.
16632 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16633 as GPR and floating point registers.
16634 (int_reg_operand): New predicate to match only GPR registers.
16635 (base_reg_operand): New predicate to match base registers.
16636 (quad_int_reg_operand): New predicate to match even GPR registers
16637 for quad memory operations.
16638 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16639 operations in both GPR and VSX registers.
16640 (quad_memory_operand): New predicate for quad memory operations.
16641 (reg_or_indexed_operand): New predicate for direct move support.
16643 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16644 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16645 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16646 (POWERPC_MASKS): Add power8 options.
16647 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16650 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16651 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16653 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16654 (-mpower8-fusion): New power8 options.
16655 (-mpower8-fusion-sign): Likewise.
16656 (-mpower8-vector): Likewise.
16657 (-mcrypto): Likewise.
16658 (-mdirect-move): Likewise.
16659 (-mquad-memory): Likewise.
16661 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16662 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16664 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16665 (rs6000_debug_vector_unit): Add p8_vector.
16666 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16667 definitions. Also print fusion state.
16668 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16669 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16670 (rs6000_option_override_internal): Add support for power8.
16671 (rs6000_common_init_builtins): Add debugging for skipped builtins
16672 if -mdebug=builtin.
16673 (rs6000_adjust_cost): Add power8 support.
16674 (rs6000_issue_rate): Likewise.
16675 (insn_must_be_first_in_group): Likewise.
16676 (insn_must_be_last_in_group): Likewise.
16677 (force_new_group): Likewise.
16678 (rs6000_register_move_cost): Likewise.
16679 (rs6000_opt_masks): Likewise.
16681 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16682 power8 capable assembler, default to power7 options.
16683 (TARGET_DIRECT_MOVE): Likewise.
16684 (TARGET_CRYPTO): Likewise.
16685 (TARGET_P8_VECTOR): Likewise.
16686 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16687 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16688 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16689 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16690 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16691 (TARGET_XSCVDPSPN): Likewise.
16692 (TARGET_XSCVSPDPN): Likewsie.
16693 (TARGET_SYNC_HI_QI): Likewise.
16694 (TARGET_SYNC_TI): Likewise.
16695 (MASK_CRYPTO): Likewise.
16696 (MASK_DIRECT_MOVE): Likewise.
16697 (MASK_P8_FUSION): Likewise.
16698 (MASK_P8_VECTOR): Likewise.
16699 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16700 temporary used by some of the direct move instructions to get two FP
16701 temporary registers does not force creation of a stack frame.
16702 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16703 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16704 that any VSX registers are tieable, even if they are also an
16705 Altivec vector mode.
16706 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16707 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16708 (RS6000_BTM_CRYPTO): Likewise.
16709 (RS6000_BTM_COMMON): Likewise.
16711 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16712 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16713 (enum rs6000_vector): Add power8 vector support.
16715 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16719 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16720 (any_sibcall_could_use_r3): this and handle indirect calls.
16721 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16723 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16725 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16727 2013-05-22 Richard Biener <rguenther@suse.de>
16729 PR middle-end/57349
16730 * profile.c (branch_prob): Do not split blocks that are
16731 abnormally receiving from ECF_RETURNS_TWICE functions.
16733 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16735 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16736 address mode. Move pointer mode initialization to the same place.
16738 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16740 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16741 while it has any effect.
16743 2013-05-21 Easwaran Raman <eraman@google.com>
16745 PR tree-optimization/57322
16746 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16747 UID of the statement added to the BB to be 1.
16749 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16751 PR tree-optimization/57331
16752 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16753 of conversion from pointer type to integral type with integer.
16755 2013-05-21 Martin Jambor <mjambor@suse.cz>
16758 * ipa-prop.c (ipa_read_node_info): Process param_used and
16759 controlled_uses in the same order as when writing.
16761 2013-05-21 Magnus Granberg <baldrick@free.fr>
16764 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16766 2013-05-21 Richard Biener <rguenther@suse.de>
16768 PR tree-optimization/57318
16769 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16770 estimate stmts with side-effects as likely eliminated.
16772 2013-05-21 Richard Biener <rguenther@suse.de>
16774 PR tree-optimization/57330
16775 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16776 preserve the call stmts fntype.
16778 2013-05-21 Richard Biener <rguenther@suse.de>
16780 PR tree-optimization/57303
16781 * tree-ssa-sink.c (statement_sink_location): Improve killing
16782 stmt detection and properly handle self-assignments.
16784 2013-05-21 Christian Bruel <christian.bruel@st.com>
16786 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16787 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16788 registers. Set register size out of the PARALLEL loop.
16790 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16793 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16794 (*fmasf4, *fmasf4_media): New insns.
16796 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16798 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16799 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16800 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16801 (mips_idiv_insns): Update the comments to say that the returned
16802 instruction counts are in units of BASE_INSN_LENGTH.
16803 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16804 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16805 using 2 rather than 4 as the length of indirect MIPS16 and
16806 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16807 length of a NOP. Don't divide MIPS16 lengths by 2.
16808 (mips16_split_long_branches): Assume a branch is long if the
16809 length is greater than 4 rather than 8.
16810 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16811 rather than multiplying them by 2. Multiply instruction counts
16812 by BASE_INSN_LENGTH rather than 4.
16813 (*jump_mips16, tls_get_tp_mips16_<mode>)
16814 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16816 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16818 * config/mips/mips.md (extended_mips16): Remove branch case.
16819 (length): Remove duplicated extended_mips16 test.
16821 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16823 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16825 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16827 * recog.h (Recog_data): Rename to...
16828 (recog_data_d): ...this.
16829 (recog_data): Update accordingly.
16830 * recog.c (recog_data): Likewise.
16831 * reload.c (save_recog_data): Likewise.
16832 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16833 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16835 2013-05-17 Julian Brown <julian@codesourcery.com>
16837 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16838 found in a REG_EQUAL note, invalidate it.
16840 2013-05-17 Easwaran Raman <eraman@google.com>
16842 * tree-ssa-reassoc.c (find_insert_point): New function.
16843 (insert_stmt_after): Likewise.
16844 (get_def_stmt): Likewise.
16845 (ensure_ops_are_available): Likewise.
16846 (not_dominated_by): Likewise.
16847 (rewrite_expr_tree): Do not move statements beyond what is
16848 necessary. Remove call to swap_ops_for_binary_stmt...
16849 (reassociate_bb): ... and move it here.
16850 (build_and_add_sum): Assign UIDs for new statements.
16851 (linearize_expr): Likewise.
16852 (do_reassoc): Renumber gimple statement UIDs.
16854 2013-05-17 Jan Hubicka <jh@suse.cz>
16856 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16858 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16859 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16861 (output_weakrefs): Update.
16863 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16864 Martin Jambor <mjambor@suse.cz>
16866 PR middle-end/57276
16867 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16868 value that corresponds to the given aggval is found in values vector.
16870 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16872 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16873 sse, sse2, sse3, ssse3 and sse4a flags to options.
16875 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16877 * gengtype-state.c: (s_expr_writer): New class, to handle
16878 prettifying of output layout of s-expressions.
16879 (state_writer): New class, to write out gtype.state.
16880 (state_written_type_count): Move this variable into member data of
16882 (s_expr_writer::s_expr_writer): New code: constructor for new class
16883 (state_writer::state_writer(): ditto
16884 (s_expr_writer::write_new_line): New function
16885 (s_expr_writer::write_any_indent): ditto
16886 (s_expr_writer::begin_s_expr): ditto
16887 (s_expr_writer::end_s_expr): ditto
16888 (write_state_fileloc): convert to method of state_writer...
16889 (state_writer:: write_state_fileloc): ...and use methods of
16890 s_expr_writer to write indentation into the gtype.state output file
16891 to visually represent the hierarchical structure of the list
16893 (write_state_fields): ditto, renaming to...
16894 (state_writer::write_state_fields)
16895 (write_state_a_string): ditto, renaming to...
16896 (state_writer::write_state_a_string)
16897 (write_state_string_option): ditto, renaming to...
16898 (state_writer::write_state_string_option)
16899 (write_state_type_option): ditto, renaming to...
16900 (state_writer::write_state_type_option)
16901 (write_state_nested_option): ditto, renaming to...
16902 (state_writer::write_state_nested_option)
16903 (write_state_option): ditto, renaming to...
16904 (state_writer::write_state_option)
16905 (write_state_options): ditto, renaming to...
16906 (state_writer::write_state_options)
16907 (write_state_lang_bitmap): ditto, renaming to...
16908 (state_writer::write_state_lang_bitmap)
16909 (write_state_version): ditto, renaming to...
16910 (state_writer::write_state_version)
16911 (write_state_scalar_type): ditto, renaming to...
16912 (state_writer::write_state_scalar_type)
16913 (write_state_string_type): ditto, renaming to...
16914 (state_writer::write_state_string_type)
16915 (write_state_undefined_type): ditto, renaming to...
16916 (state_writer::write_state_undefined_type)
16917 (write_state_struct_union_type): ditto, renaming to...
16918 (state_writer::write_state_struct_union_type)
16919 (write_state_struct_type): ditto, renaming to...
16920 (state_writer::write_state_struct_type)
16921 (write_state_user_struct_type): ditto, renaming to...
16922 (state_writer::write_state_user_struct_type)
16923 (write_state_lang_struct_type): ditto, renaming to...
16924 (state_writer::write_state_lang_struct_type)
16925 (write_state_param_struct_type): ditto, renaming to...
16926 (state_writer::write_state_param_struct_type)
16927 (write_state_pointer_type): ditto, renaming to...
16928 (state_writer::write_state_pointer_type)
16929 (write_state_array_type): ditto, renaming to...
16930 (state_writer::write_state_array_type)
16931 (write_state_gc_used): ditto, renaming to...
16932 (state_writer::write_state_gc_used)
16933 (write_state_common_type_content): ditto, renaming to...
16934 (state_writer::write_state_common_type_content)
16935 (write_state_type): ditto, renaming to...
16936 (state_writer::write_state_type)
16937 (write_state_pair_list): ditto, renaming to...
16938 (state_writer::write_state_pair_list)
16939 (write_state_pair): ditto, renaming to...
16940 (state_writer::write_state_pair)
16941 (write_state_typedefs): ditto, renaming to...
16942 (state_writer::write_state_typedefs)
16943 (write_state_structures): ditto, renaming to...
16944 (state_writer::write_state_structures)
16945 (write_state_param_structs): ditto, renaming to...
16946 (state_writer::write_state_param_structs)
16947 (write_state_variables): ditto, renaming to...
16948 (state_writer::write_state_variables)
16949 (write_state_srcdir): ditto, renaming to...
16950 (state_writer::write_state_srcdir)
16951 (write_state_files_list): ditto, renaming to...
16952 (state_writer::write_state_files_list)
16953 (write_state_languages): ditto, renaming to...
16954 (state_writer::write_state_languages)
16955 (write_state): create a state_writer instance and use it when
16956 writing out the state file
16958 2013-05-17 Mike Stump <mikestump@comcast.net>
16960 PR rtl-optimization/57304
16961 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16962 accessing DF_REF_REAL_LOC.
16964 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16966 PR rtl-optimization/57281
16967 PR rtl-optimization/57300
16968 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16969 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16970 what the other splitter did if the registers are dead.
16972 2013-05-17 Richard Biener <rguenther@suse.de>
16974 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16977 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16979 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16982 2013-05-17 Marek Polacek <polacek@redhat.com>
16984 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16985 length when doing non-zero store of storing '\0' to '\0'.
16987 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16989 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16990 vect_external_def oprnd1 with loop_vinfo, try to emit
16991 optional cast, negation and and stmts on the loop preheader
16992 edge instead of into the pattern def seq.
16994 PR tree-optimization/57051
16995 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16996 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16998 2013-05-16 Nick Clifton <nickc@redhat.com>
17000 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17001 (rl78_is_naked_func): New function.
17002 (rl78_expand_prologue): Skip prologue generation for naked functions.
17003 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17004 * doc/extend.texi (naked): Add RL78 to the list of processors
17005 that supports this attribute.
17007 2013-05-16 Jeff Law <law@redhat.com>
17009 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17011 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17013 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17014 cache parameters using detect_caches_amd also for CYRIX,
17015 NSC and TM2 signatures.
17017 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17018 Dzianis Kahanovich <mahatma@eu.by>
17022 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17023 VIA/Centaur processors and determine their cache parameters
17024 using detect_caches_amd.
17026 2013-05-16 Teresa Johnson <tejohnson@google.com>
17028 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17029 (rtl_verify_edges): New function.
17030 (rtl_verify_bb_insns): Ditto.
17031 (rtl_verify_bb_pointers): Ditto.
17032 (rtl_verify_bb_insn_chain): Ditto.
17033 (rtl_verify_fallthru): Ditto.
17034 (rtl_verify_bb_layout): Ditto.
17035 (rtl_verify_flow_info_1): Outline checks into new functions.
17036 (rtl_verify_flow_info): Ditto.
17038 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17040 * cfghooks.c (copy_bbs): Add update_dominance argument.
17041 * cfghooks.h (copy_bbs): Update prototype.
17042 * tree-cfg.c (gimple_duplicate_sese_region):
17043 Add update_dominance argument.
17044 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17045 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17046 gimple_duplicate_sese_region call.
17047 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17048 Update copy_bbs call.
17049 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17050 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17052 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17054 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17055 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17056 vect_recog_rotate_pattern.
17057 (vect_recog_rotate_pattern): New function.
17059 2013-05-16 Jason Merrill <jason@redhat.com>
17061 * Makefile.in (LLINKER): New variable.
17062 (mostlyclean): Remove link mutex.
17063 * configure.ac: Handle --enable-link-mutex.
17064 * lock-and-run.sh: New script.
17066 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17069 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17072 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17074 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17076 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17078 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17079 * config/arm/arm.c (next_consecutive_mem): New function.
17080 (gen_movmem_ldrd_strd): Likewise.
17081 * config/arm/arm.md (movmemqi): Update condition and code.
17082 (unaligned_loaddi, unaligned_storedi): New patterns.
17084 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17086 * config.gcc: Obsolete *-*-solaris2.9*.
17087 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17089 2013-05-16 Richard Biener <rguenther@suse.de>
17091 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17092 earlier, after GRAPHITE transforms and IV canonicalization.
17094 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17096 * omp-low.c (extract_omp_for_data): For collapsed loops,
17097 if at least one of the loops is known at compile time to
17098 iterate zero times, set count to 0.
17099 (expand_omp_regimplify_p): New function.
17100 (expand_omp_for_generic): For collapsed loops, if at least
17101 one of the loops isn't known to iterate at least once,
17102 add runtime check with setting count to 0.
17103 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17104 For unsigned types if it isn't known at compile time that
17105 the loop will iterate at least once, add runtime check to bypass
17106 the whole loop if initial condition isn't true.
17108 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17110 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17112 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17114 PR middle-end/57286
17115 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17116 transformations to avoid an infinite loop.
17118 2013-05-16 Marek Polacek <polacek@redhat.com>
17120 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17122 2013-05-15 Leif Ekblad <leif@rdos.net>
17124 * config/i386/i386.c (ix86_decompose_address): Use
17125 DEFAULT_TLS_SEG_REG to access TLS segment register.
17126 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17127 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17128 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17130 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17133 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17134 sibling calls to functions that would normally be lazily bound,
17135 unless $gp is call-clobbered.
17137 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17139 * config/i386/i386.c (ix86_option_override_internal): Update
17140 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17141 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17142 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17143 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17145 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17147 2013-05-15 Andreas Schwab <schwab@suse.de>
17149 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17150 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17153 2013-05-15 Teresa Johnson <tejohnson@google.com>
17155 * loop-unroll.c (report_unroll_peel): Check decision before
17156 emitting unroll/peel message.
17158 2013-05-15 Teresa Johnson <tejohnson@google.com>
17160 * function.h (has_bb_partition): New rtl_data flag.
17161 (bb_reorder_complete): Ditto.
17162 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17163 instead of flag_reorder_blocks_and_partition.
17164 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17165 with some enhancements.
17166 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17167 * bb-reorder.c (connect_traces): Check for has_bb_partition
17168 instead of flag_reorder_blocks_and_partition.
17169 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17170 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17171 verify_hot_cold_block_grouping.
17172 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17174 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17177 * config/arm/predicates.md (call_insn_operand): New predicate.
17178 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17179 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17180 if insn is not a tail call.
17181 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17183 * config/arm/arm.h (enum reg_class): New caller save register class.
17184 (REG_CLASS_NAMES): Likewise.
17185 (REG_CLASS_CONTENTS): Likewise.
17186 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17189 2013-05-15 Richard Biener <rguenther@suse.de>
17191 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17192 of MSG_OPTIMIZED_LOCATIONS.
17193 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17194 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17196 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17197 of MSG_OPTIMIZED_LOCATIONS.
17198 (execute_vect_slp): Likewise.
17199 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17200 (vect_create_cond_for_alias_checks): Likewise.
17201 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17202 (vect_recog_widen_mult_pattern): Likewise.
17203 (vect_recog_widen_sum_pattern): Likewise.
17204 (vect_recog_over_widening_pattern): Likewise.
17205 (vect_recog_widen_shift_pattern): Likewise.
17206 (vect_recog_vector_vector_shift_pattern): Likewise.
17207 (vect_recog_divmod_pattern): Likewise.
17208 (vect_recog_mixed_size_cond_pattern): Likewise.
17209 (vect_recog_bool_pattern): Likewise.
17210 (vect_pattern_recog_1): Likewise.
17212 2013-05-15 Martin Jambor <mjambor@suse.cz>
17214 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17215 non-functions to builtin_unreachable.
17216 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17217 correct when new direct edges were discovered.
17219 2013-05-15 Martin Jambor <mjambor@suse.cz>
17221 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17222 header, print symbol order instead of node uid, print more information
17223 about indirect edge targets.
17224 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17226 (ipa_make_edge_direct_to_target): Likewise.
17227 (remove_described_reference): Likewise.
17228 (propagate_controlled_uses): Likewise.
17229 (ipa_print_node_params): Also print symbol order.
17230 (ipcp_transform_function): Print symbol order instead of node uids.
17231 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17232 (cgraph_get_create_real_symbol_node): Likewise.
17233 * ipa-cp.c (print_lattice): Likewise.
17234 (print_all_lattices): Likewise.
17235 (determine_versionability): Likewise.
17236 (initialize_node_lattices): Likewise.
17237 (estimate_local_effects): Likewise.
17238 (update_profiling_info): Likewise.
17239 (create_specialized_node): Likewise.
17240 (perhaps_add_new_callers): Likewise.
17241 (decide_about_value): Likewise.
17242 (decide_whether_version_node): Likewise.
17243 (identify_dead_nodes): Likewise.
17244 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17245 (dump_inline_summary): Likewise.
17246 (estimate_node_size_and_time): Likewise.
17247 (inline_analyze_function): Likewise.
17248 * ipa-inline.c (report_inline_failed_reason): Likewise.
17249 (want_early_inline_function_p): Likewise.
17250 (edge_badness): Likewise.
17251 (update_edge_key): Likewise.
17252 (inline_small_functions): Likewise. Add dumping of order to two other
17254 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17255 instead of node uids.
17256 (propagate_pure_const): Likewise.
17257 (propagate_pure_const): Likewise.
17258 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17259 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17260 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17262 * tree-pretty-print.c (dump_function_header): Likewise.
17263 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17264 Print symbol order instead of node uids.
17266 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17268 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17269 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17271 2013-05-15 Richard Biener <rguenther@suse.de>
17273 PR tree-optimization/57275
17274 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17275 return value for fail to do runtime alias checks for gather loads.
17277 2013-05-15 Jan Hubicka <jh@suse.cz>
17281 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17282 weakrefs are not external.
17283 (lto_symtab_merge_decls): Fix thinko when dealing with
17284 non-lto_symtab decls.
17285 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17286 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17287 * varpool.c (dump_varpool_node): Dump more flags.
17289 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17291 * config/i386/i386.c (processor_alias_table): Add instruction
17292 FSGSBASE for AMD bdver3 architecture.
17294 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17296 * tree.c (warn_deprecated_use): Print file:line using locus color.
17297 * diagnostic.c (diagnostic_report_current_module): Print file:line
17298 and file:line:column using locus color.
17300 2013-05-14 Mike Stump <mikestump@comcast.net>
17302 * gdbinit.in: Add __null.
17304 2013-05-14 Mike Stump <mikestump@comcast.net>
17306 * recog.h: Rename struct recog_data to Recog_data.
17307 * recog.c: Likewise.
17308 * reload.c (can_reload_into): Likewise.
17309 * config/picochip/picochip.c: Likewise.
17311 2013-05-14 Mike Stump <mikestump@comcast.net>
17313 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17315 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17317 * resource.h (struct resources): Remove unch_memory member.
17318 (CLEAR_RESOURCE): Don't clear unch_memory.
17319 * resource.c (mark_referenced_resources): Don't set it.
17320 (mark_set_resources): Likewise.
17321 (mark_target_live_regs): Don't clear it.
17322 (init_resource_info): Likewise.
17323 * reorg.c (resource_conflicts_p): Don't compare it.
17324 (redundant_insn): Don't set it.
17326 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17328 * emit-rtl.c (next_label): Remove unused function.
17329 (skip_consecutive_labels, link_cc0_insns): Move to ...
17330 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17331 only place where these functions are used, and make them static.
17333 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17335 * fold-const.c (fold_negate_expr): Handle vectors.
17336 (fold_truth_not_expr): Make it static.
17337 (fold_invert_truthvalue): New static function.
17338 (invert_truthvalue_loc): Handle vectors. Do not call
17339 fold_truth_not_expr directly.
17340 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17341 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17342 (fold_comparison): Handle vector constants.
17343 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17344 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17345 * tree.h (fold_truth_not_expr): Remove declaration.
17347 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17349 * config/aarch64/aarch64-simd.md
17350 (aarch64_vcond_internal<mode>): Rename to...
17351 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17352 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17353 float modes. Clarify all iterator modes.
17354 (vcond<mode><mode>): Use new name for vcond expanders.
17355 (vcond<v_cmp_result><mode>): Likewise.
17356 (vcondu<mode><mode>: Likewise.
17357 * config/aarch64/iterators.md (VDQF_COND): New.
17359 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17362 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17363 variable for the shift amount. Check that we shift by non-negative
17366 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17369 * config/arm/arm.h (EPILOGUE_USES): Only return true
17370 for LR_REGNUM after epilogue_completed.
17372 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17374 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17375 is error_mark_node.
17377 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17380 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17381 and Solaris 11+/x86 with gld.
17382 * configure: Regenerate.
17384 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17386 * expmed.c (expand_shift_1): Canonicalize rotates by
17387 constant bitsize / 2 to bitsize - 1.
17388 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17389 case ROTATERT>: Likewise.
17392 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17394 * config/i386/i386.md (rotateinv): New code attr.
17395 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17396 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17397 roll $31, %eax, etc.
17399 2013-05-14 Richard Biener <rguenther@suse.de>
17401 PR middle-end/57235
17402 * tree-eh.c (sink_clobbers): Give up for successors with
17403 multiple predecessors and no virtual uses.
17405 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17407 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17408 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17410 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17412 PR middle-end/57251
17413 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17414 the case when both op0 and op1 have VOIDmode.
17416 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17418 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17419 in multiply-accumulate mode.
17421 2013-05-13 Guozhi Wei <carrot@google.com>
17423 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17425 2013-05-13 Kai Tietz <ktietz@redhat.com>
17428 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17429 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17430 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17431 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17432 (ix86_expand_prologue): Likewise.
17433 (ix86_expand_split_stack_prologue): Likewise.
17434 (legitimate_pic_address_disp_p): Likewise.
17435 (legitimize_pic_address): Likewise.
17436 (legitimize_tls_address): Likewise.
17437 (legitimize_pe_coff_symbol): Likewise.
17438 (output_pic_addr_const): Likewise.
17439 (construct_plt_address): Likewise.
17440 (ix86_expand_call): Likewise.
17441 (x86_output_mi_thunk): Likewise.
17442 (x86_function_profiler): Likewise.
17444 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17446 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17447 similar switch cases.
17448 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17449 (aarch64_simd_mov_to_<mode>low): Delete.
17450 (aarch64_simd_mov_to_<mode>high): Delete.
17451 (move_lo_quad_<mode>): Add w<-r alternative.
17452 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17453 (aarch64_simd_mov_from_*): Update type attribute.
17454 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17457 2013-05-13 Jan Hubicka <jh@suse.cz>
17459 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17460 * config/i386/i386.c (ix86_compute_frame_layout,
17461 ix86_expand_epilogue, emit_i387_cw_initialization,
17462 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17463 ix86_local_alignment): Fix use of size/speed predicates.
17465 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17467 PR tree-optimization/45216
17468 PR tree-optimization/57157
17469 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17470 the (-Y) & (B - 1) variant if OP is |.
17471 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17472 return shifted. Use (-op1) & (prec - 1) as other_amount
17473 instead of prec - op1.
17475 2013-05-13 Martin Jambor <mjambor@suse.cz>
17477 PR middle-end/42371
17478 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17479 (ipa_constant_data): New type.
17480 (ipa_jump_func): Use ipa_constant_data to hold information about
17481 constant jump functions.
17482 (ipa_get_jf_constant): Adjust to jump function type changes.
17483 (ipa_get_jf_constant_rdesc): New function.
17484 (ipa_param_descriptor): New field controlled_uses.
17485 (ipa_get_controlled_uses): New function.
17486 (ipa_set_controlled_uses): Likewise.
17487 * ipa-ref.h (ipa_find_reference): Declare.
17488 * ipa-prop.c (ipa_cst_ref_desc): New type.
17489 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17491 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17492 New parameter cs. Adjust all callers.
17493 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17494 (remove_described_reference): New function.
17495 (jfunc_rdesc_usable): Likewise.
17496 (try_make_edge_direct_simple_call): Decrement controlled use count,
17497 attempt to remove reference if it hits zero.
17498 (combine_controlled_uses_counters): New function.
17499 (propagate_controlled_uses): Likewise.
17500 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17501 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17502 (ipa_print_node_params): Print described use counter.
17503 (ipa_write_jump_function): Adjust to jump function type changes.
17504 (ipa_read_jump_function): New parameter CS, pass it to
17505 ipa_set_jf_constant. Adjust caller.
17506 (ipa_write_node_info): Stream controlled use count
17507 (ipa_read_node_info): Likewise.
17508 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17510 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17511 count. Remove cloning-added reference if it reaches zero.
17512 * ipa-ref.c (ipa_find_reference): New function.
17514 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17516 * config/i386/i386.c (processor_target_table): Modified default
17517 alignment values for AMD BD and BT architectures.
17519 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17521 * tree-vect-generic.c (uniform_vector_p): Move ...
17522 * tree.c (uniform_vector_p): ... here.
17523 * tree.h (uniform_vector_p): Declare it.
17524 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17527 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17529 PR tree-optimization/57230
17530 * tree-ssa-strlen.c (handle_char_store): Record length for
17531 array store from STRING_CST.
17533 PR tree-optimization/57230
17534 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17537 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17539 * config/epiphany/epiphany.c (epiphany_init): Check size of
17540 NUM_MODES_FOR_MODE_SWITCHING.
17541 (epiphany_expand_prologue):
17542 Remove CONFIG_REGNUM initial value handling code.
17543 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17544 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17545 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17546 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17547 Don't return 1 for FP_MODE_NONE.
17548 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17549 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17550 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17551 * config/epiphany/epiphany.md (save_config): New pattern.
17553 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17555 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17557 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17559 * config/i386/i386.md (memory): Handle sseishft1.
17560 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17561 (*vec_extractv2di_1): Ditto.
17563 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17565 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17568 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17570 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17571 Add mthumb/march=armv7-a multilib.
17572 Add mthumb/march=armv7-r multilib.
17573 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17575 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17577 * config/v850/t-rtems: Add more multilibs.
17579 2013-05-10 Richard Biener <rguenther@suse.de>
17581 PR tree-optimization/57214
17582 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17583 not propagate from SSA names that occur in abnormal PHI nodes.
17585 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17587 * stor-layout.c (element_precision): New function.
17588 * machmode.h (element_precision): Declare it.
17589 * tree.c (build_minus_one_cst): New function.
17590 (element_precision): Likewise.
17591 * tree.h (build_minus_one_cst): Declare new function.
17592 (element_precision): Likewise.
17593 * fold-const.c (operand_equal_p): Use element_precision.
17594 (fold_binary_loc): Handle vector types.
17595 * convert.c (convert_to_integer): Use element_precision.
17596 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17599 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17601 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17602 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17603 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17604 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17605 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17606 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17607 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17608 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17609 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17610 (Uuw8): New constraints.
17611 (Usb4): Move into alphabetical order.
17612 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17613 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17614 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17615 previously unnamed patterns.
17616 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17617 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17618 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17619 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17620 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17621 of specific lengths.
17623 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17625 * config/i386/i386.md (rotateinv): New code attr.
17626 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17627 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17628 roll $31, %eax, etc.
17630 PR tree-optimization/45216
17631 PR tree-optimization/57157
17632 * tree-ssa-forwprop.c (simplify_rotate): New function.
17633 (ssa_forward_propagate_and_combine): Call it.
17635 2013-05-10 Richard Biener <rguenther@suse.de>
17637 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17638 disable peeling when we version for aliasing.
17639 (vector_alignment_reachable_p): Honor explicit user alignment.
17640 (vect_supportable_dr_alignment): Likewise.
17641 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17642 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17643 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17644 then peeling to arrange for the cost-model check to come first.
17646 2013-05-10 Alan Modra <amodra@gmail.com>
17648 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17649 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17650 * configure: Regenerate.
17652 2013-05-10 Alan Modra <amodra@gmail.com>
17655 * varasm.c (default_elf_select_section): Move !DECL_P check..
17656 (get_named_section): ..to here before calling get_section_name.
17658 (default_section_type_flags): Add DECL_P check.
17659 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17660 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17662 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17664 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17665 When using gen_stack_adjust_str with a register offset, add a
17666 REG_FRAME_RELATED_EXPR note.
17668 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17670 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17671 (*vec_extractv4si_zext_mem): Ditto.
17672 (*vec_extractv2di): Add 0->x and x->x alternatives.
17673 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17674 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17676 2013-05-09 Jason Merrill <jason@redhat.com>
17678 N3639 C++1y VLA support
17679 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17682 * tree.c (build_constructor_va): New.
17683 * tree.h: Declare it.
17685 2013-05-09 Martin Jambor <mjambor@suse.cz>
17688 * gimple-fold.c (canonicalize_constructor_val): Call
17689 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17691 2013-05-09 Jan Hubicka <jh@suse.cz>
17692 Richard Biener <rguenther@suse.de>
17695 * symtab.c (symtab_make_decl_local): Do not add private names.
17697 2013-05-09 Jan Hubicka <jh@suse.cz>
17700 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17701 (unlink_from_assembler_name_hash): Likewise.
17702 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17703 symtab_unregister_node, symtab_initialize_asm_name_hash,
17704 change_decl_assembler_name): Update.
17706 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17708 * config/aarch64/aarch64.md: New movtf split.
17709 (*movtf_aarch64): Update.
17710 (aarch64_movdi_tilow): Handle TF modes and rename to
17711 aarch64_movdi_<mode>low.
17712 (aarch64_movdi_tihigh): Handle TF modes and rename to
17713 aarch64_movdi_<mode>high
17714 (aarch64_movtihigh_di): Handle TF modes and rename to
17715 aarch64_mov<mode>high_di
17716 (aarch64_movtilow_di): Handle TF modes and rename to
17717 aarch64_mov<mode>low_di
17718 (aarch64_movtilow_tilow): Remove spurious whitespace.
17719 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17721 (aarch64_print_operand): Update.
17723 2013-05-09 Alan Modra <amodra@gmail.com>
17725 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17727 * configure: Regenerate.
17729 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17731 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17732 splitter preparation statements.
17733 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17734 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17737 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17739 * gimple-ssa-strength-reduction.c (count_candidates): Change
17740 return value to int.
17741 (analyze_candidates_and_replace): Change type of length to int.
17743 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17745 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17746 (*vec_extract<mode>): Use VI12_128 mode iterator.
17747 (*vec_extract<mode>_mem): Ditto.
17748 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17751 2013-05-08 Diego Novillo <dnovillo@google.com>
17756 2012-08-17 Diego Novillo <dnovillo@google.com>
17759 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17760 * config.in: Regenerate.
17761 * configure: Regenerate.
17762 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17764 2013-05-08 Jan Hubicka <jh@suse.cz>
17767 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17768 * cgraph.h (symtab_node_base): Add unique_name.
17769 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17770 input_overwrite_node, input_varpool_node): Stream unique_name.
17771 * cgraphclones.c (cgraph_create_virtual_clone,
17772 cgraph_function_versioning): Set unique_name.
17773 * ipa.c (function_and_variable_visibility): Set unique_name.
17775 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17777 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17778 (alloc_cand_and_find_basis): Restrict conditional candidate
17779 processing to CAND_MULTs.
17781 2013-05-08 Jan Hubicka <jh@suse.cz>
17784 lto-symtab.c (lto_symtab_symbol_p): New function.
17785 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17786 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17787 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17788 Skip static symbols.
17790 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17792 PR tree-optimization/57200
17793 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17794 Only call inform if the preceding warning_at returns true.
17796 2013-05-07 Han Shen <shenhan@google.com>
17798 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17799 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17800 * common.opt (fstack-protector-strong): New option.
17801 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17802 * doc/invoke.texi (Optimization Options): Document
17803 "-fstack-protector-strong".
17804 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17806 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17808 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17810 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17812 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17813 it is not lhs of the insn.
17814 (setup_reg_equiv): Remove insn having equiv memory which it is not
17816 * lra-constraints.c (process_address): Try to improve generation
17817 code for address base + disp.
17818 (lra_constraints): Make correct the code for checking insn setting
17819 up backward equivalence. Remove insn only if it is in the init
17821 * lra-eliminations.c (update_reg_eliminate): Change return value.
17822 (lra_eliminate): Use the result.
17824 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17826 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17827 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17828 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17829 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17830 register target operands.
17831 (*vec_extractv8hi_sse2): New pattern.
17832 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17833 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17834 (*vec_extract<mode>_mem): New insn and split pattern.
17836 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17838 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17839 (TARGET_ASAN_SHADOW_OFFSET): Define.
17840 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17841 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17843 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17845 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17846 (incr_vec_index): Return -1 if increment not found.
17847 (create_add_on_incoming_edge): Assert if increment not found.
17848 (record_increment): Limit number of increments recorded.
17849 (all_phi_incrs_profitable): Return false if an increment not found.
17850 (replace_profitable_candidates): Don't process increments that were
17852 (analyze_candidates_and_replace): Limit size of incr_vec.
17854 2013-05-07 Richard Biener <rguenther@suse.de>
17856 * calls.c (special_function_p): setjmp-like functions are leaf.
17857 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17858 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17860 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17862 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17863 (aarch64_simd_mov<mode>): New expander.
17864 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17865 (aarch64_simd_mov_to_<mode>high): Likewise.
17866 (aarch64_simd_mov_from_<mode>low): Likewise.
17867 (aarch64_simd_mov_from_<mode>high): Likewise.
17868 (aarch64_dup_lane<mode>): Update.
17869 (aarch64_dup_lanedi): New instruction pattern.
17870 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17871 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17873 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17875 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17876 (replace_mult_candidate): Remove unnecessary argument; remove
17877 unnecessary parameter from call to introduce_cast_before_cand.
17878 (replace_unconditional_candidate): Remove unnecessary parameter
17879 from call to replace_mult_candidate.
17880 (replace_conditional_candidate): Likewise.
17881 (insert_initializers): Use make_temp_ssa_name.
17882 (introduce_cast_before_cand): Remove unnecessary argument; use
17883 make_temp_ssa_name.
17884 (replace_one_candidate): Remove unnecessary argument; remove
17885 unnecessary parameter from calls to introduce_cast_before_cand.
17886 (replace_profitable_candidates): Remove unnecessary parameters
17887 from calls to replace_one_candidate.
17889 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17891 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17892 phi def as possibly hiding a basis for a CAND_ADD whose operands
17893 have been commuted in the analysis.
17894 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17896 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17898 * config/aarch64/aarch64.md
17899 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17900 shift value between 0-4.
17902 2013-05-07 Richard Biener <rguenther@suse.de>
17904 * double-int.h (rshift): New overload.
17905 * double-int.c (rshift): New function.
17906 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17907 (create_reference_ops_from_ref): Remove.
17908 (vn_reference_insert): Use shared ops for constructing the
17909 reference and copy it.
17911 2013-05-07 Richard Biener <rguenther@suse.de>
17913 PR middle-end/57190
17914 * tree-eh.c (sink_clobbers): Properly propagate
17915 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17917 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17919 PR tree-optimization/57149
17920 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17921 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17922 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17923 uninit_undefined_value_p instead of ssa_undefined_value_p.
17926 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17927 for modifier == EXPAND_INITIALIZER.
17929 2013-05-07 Anton Blanchard <anton@samba.org>
17931 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17932 for powerpc64 little endian.
17933 * configure: Regenerate.
17935 2013-05-06 Graham Stott <grahams@btinternet.com>
17937 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17938 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17939 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17940 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17942 2013-05-06 Graham Stott <grahams@btinternet.com>
17944 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17945 codes which allow non-lvalues.
17947 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17949 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17950 components are all 1s.
17951 (integer_minus_onep): New function.
17952 * tree.h (integer_minus_onep): Declare it.
17953 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17954 integer_minus_onep instead of integer_all_onesp.
17956 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17959 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17960 variations of these patterns.
17962 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17964 * config/i386/i386.md (isa): Add x64_sse4 member.
17965 (enabled): Handle x64_sse4.
17966 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17967 instruction for 64bit SSE4_1 targets. Update insn attributes.
17968 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17969 instruction for SSE4_1 targets. Update insn attributes.
17970 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17971 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17972 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17974 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17975 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17976 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17978 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17981 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17983 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17985 * final.c (do_assembler_dialects): Don't handle curly braces and
17986 vertical bar escaped by % as dialect delimiters.
17987 (output_asm_insn): Print curly braces and vertical bar if escaped
17988 by % and ASSEMBLER_DIALECT defined.
17989 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17990 * doc/tm.texi: Regenerated.
17992 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17994 * config/mips/mips.c: Include tree-pass.h.
17995 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17996 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17997 (pass_mips_machine_reorg2): New machine specific pass.
17998 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17999 (mips_option_override): Register the new pass.
18000 * rtl.h (cleanup_barriers): Remove prototype.
18001 (dbr_schedule): Likewise.
18002 * jump.c (cleanup_barriers): Make static.
18003 * reorg.c (dbr_schedule): Likewise.
18005 2013-05-06 Richard Biener <rguenther@suse.de>
18007 PR tree-optimization/57185
18008 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18009 SSA names properly.
18011 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18014 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18015 between "first_arg" and "insn", not "last" and "insn".
18017 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18019 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18020 (find_candidates_in_block): Re-enable slsr_process_phi.
18021 (create_phi_basis): Fix double counting of candidate adjustment.
18023 2013-05-06 Richard Biener <rguenther@suse.de>
18025 PR middle-end/57147
18026 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18027 the edge is also fallthru, preserve it and just clear the
18029 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18030 also complex, preserve that and just clear the fallthru flag.
18031 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18032 update virtual operands.
18034 2013-05-06 Alan Modra <amodra@gmail.com>
18036 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18037 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18038 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18039 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18040 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18041 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18043 2013-05-06 Alan Modra <amodra@gmail.com>
18045 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18046 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18047 (DEFAULT_ASM_ENDIAN): Define.
18048 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18049 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18050 Update -K PIC clause from sysv4.h.
18051 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18052 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18054 2013-05-06 Alan Modra <amodra@gmail.com>
18056 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18057 twice for little-endian.
18058 (ashrdi3_no_power, ashrdi3): Support little-endian.
18060 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18063 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18064 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18066 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18067 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18068 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18071 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18072 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18074 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18075 * config/pa/pa.opt: Make mbig-switch a no-op.
18076 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18077 (CASE_VECTOR_MODE): Always return SImode.
18078 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18079 for the !TARGET_BIG_SWITCH case.
18080 * config/pa/pa-linux.h: Likewise.
18081 * config/pa/pa-openbsd.h: Likewise.
18082 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18083 * config/pa/pa.md (short_jump): Remove define_insn.
18084 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18085 (casesi0): Remove define_insn.
18086 (type): Remove btable_branch.
18087 (pa_combine_type): Likewise.
18088 (in_nullified_branch_delay): Likewise.
18089 (in_call_delay): Likewise.
18090 (define_delay): Likewise.
18091 (define_insn_reservation "Z3"): Likewise.
18092 (define_insn_reservation "Z4"): Likewise.
18093 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18094 (pa_adjust_insn_length): Remove adjustment for btable branches.
18095 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18098 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18100 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18101 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18102 Add m->r,x alternatives.
18103 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18104 splitters using SWI48x mode iterator.
18105 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18106 TARGET_64BIT. Add m->x alternative.
18107 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18108 Add o->x alternative. Enable for TARGET_SSE.
18109 (sse_storeq): Remove expander.
18110 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18111 with memory input operand.
18112 (*vec_extractv2di_1 splitter): New.
18113 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18114 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18116 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18118 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18119 (INT_LOWPART): Delete.
18120 (extract_MB): Adjust.
18121 (extract_ME): Adjust.
18122 (print_operand): Adjust.
18124 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18126 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18127 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18128 (reg_or_logical_cint_operand, easy_fp_constant,
18129 logical_const_operand): Delete "CONST_DOUBLE" case.
18130 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18131 "HOST_BITS_PER_WIDE_INT == 64" test.
18132 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18133 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18134 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18135 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18137 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18138 CONST_DOUBLE DImode/VOIDmode case.
18139 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18140 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18141 CONST_DOUBLE VOIDmode case.
18142 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18143 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18144 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18145 Delete CONST_DOUBLE case.
18146 (splitters for mov FMOVE64 const_double): Delete
18147 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18148 "HOST_BITS_PER_WIDE_INT >= 64" test.
18149 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18151 (mov DI const_double): Delete.
18153 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18155 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18156 on op shows all bits zero in mode of a lowpart subreg, return zero.
18158 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18161 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18162 to save TFmode registers and DImode to save TImode registers for
18163 caller save operations.
18164 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18165 mark being partially clobbered since they only use the first
18168 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18169 and TDmode only use the upper 64-bits of each VSX register.
18171 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18173 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18174 (find_candidates_in_block): Disable slsr_process_phi.
18176 2013-05-03 Guozhi Wei <carrot@google.com>
18178 * coverage.c (coverage_obj_init): Move the construction of gcov
18180 (build_init_ctor): ... here.
18182 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18184 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18185 (slsr_cand_d): Redefine def_phi.
18186 (stride_status, phi_adjust_status, count_phis_status): New enums.
18187 (find_phi_def): New.
18188 (find_basis_for_base_expr): New.
18189 (find_basis_for_candidate): Handle hidden bases.
18190 (alloc_cand_and_find_basis): Handle phi candidates.
18191 (slsr_process_phi): New.
18192 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18193 (create_mul_imm_cand): Likewise.
18194 (create_add_ssa_cand): Exclude phi base candidates.
18195 (create_add_imm_cand): Likewise.
18196 (slsr_process_cast): Likewise.
18197 (slsr_process_copy): Likewise.
18198 (find_candidates_in_block): Handle phi candidates.
18199 (dump_candidate): Likewise.
18200 (unconditional_cands): Delete.
18201 (unconditional_cands_with_known_stride_p): Delete.
18202 (phi_dependent_cand_p): New.
18203 (cand_increment): Handle phi-dependent candidates.
18204 (replace_dependent): Delete.
18205 (replace_mult_candidate): New.
18206 (replace_unconditional_candidate): New.
18207 (incr_vec_index): Move to avoid forward reference.
18208 (create_add_on_incoming_edge): New.
18209 (create_phi_basis): New.
18210 (replace_dependents): Delete.
18211 (replace_conditional_candidate): New.
18212 (phi_add_costs): New.
18213 (replace_uncond_cands_and_profitable_phis): New.
18214 (record_increment): Handle phi adjustments.
18215 (record_phi_increments): New.
18216 (record_increments): Handle phi adjustments.
18217 (phi_incr_cost): New.
18218 (lowest_cost_path): Handle phis.
18219 (total_savings): Likewise.
18220 (analyze_increments): Likewise.
18221 (ncd_with_phi): New.
18222 (ncd_of_cand_and_phis): New.
18223 (nearest_common_dominator_for_cands): Handle phi increments.
18224 (all_phi_incrs_profitable): New.
18225 (replace_profitable_candidates): Handle phi-dependent candidates.
18226 (analyze_candidates_and_replace): Likewise.
18228 2013-05-03 Teresa Johnson <tejohnson@google.com>
18231 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18232 do not exceed REG_BR_PROB_BASE.
18234 2013-05-03 Jeff Law <law@redhat.com>
18236 PR tree-optimization/57144
18237 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18238 operand of the condition will bit into the new type when eliminating
18239 a cast feeding a condition.
18241 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18243 PR rtl-optimization/57130
18244 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18245 of COMPARE as in_code to the recursive call if needed.
18247 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18249 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18250 (enabled): Handle new members.
18251 * config/i386/sse.md (*vec_concatv2si): Merge from
18252 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18253 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18255 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18257 PR tree-optimization/57027
18258 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18259 for fnms opportunity, check we got the prerequisite kind
18260 of tree / gimple before using accessor functions.
18262 2013-05-03 Richard Biener <rguenther@suse.de>
18264 * double-int.h (lshift): New overload without precision
18265 and arith argument.
18266 (operator *=, operator +=, operator -=): Move ...
18267 * double-int.c (operator *=, operator +=, operator -=): ... here
18268 and implement more efficiently.
18269 (mul_double_with_sign): Remove.
18270 (lshift_double): Adjust to take unsinged shift argument, push
18271 dispatching code to callers.
18272 (mul_double_wide_with_sign): Add early out for callers that
18273 are not interested in high parts or overflow.
18274 (lshift): New function.
18275 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18276 dispatch code here.
18277 (lrotate, rrotate): Use logical shifts.
18278 * expr.c (get_inner_reference): Use lshift.
18279 * fixed-value.c (do_fixed_divide): Likewise.
18280 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18281 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18282 (indirect_refs_may_alias_p): Likewise.
18283 (stmt_kills_ref_p_1): Likewise.
18285 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18287 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18289 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18291 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18292 scalar form of FABD instruction.
18294 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18296 * lra-constraints.c (process_alt_operands): Add checking alt
18297 number to choose the best alternative.
18299 2013-05-02 Richard Biener <rguenther@suse.de>
18301 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18302 bitmap and its handling.
18303 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18305 2013-05-02 Richard Biener <rguenther@suse.de>
18307 PR middle-end/57140
18308 * tree-inline.c (copy_loops): Properly handle removed loops.
18309 (copy_cfg_body): Mark destination loops for fixup if source
18310 loops needed fixup.
18312 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18315 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18316 generating simple_return for naked functions.
18318 2013-05-02 Martin Jambor <mjambor@suse.cz>
18320 PR middle-end/56988
18321 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18322 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18324 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18325 ipa_agg_replacement_value structures.
18326 (known_aggs_to_agg_replacement_list): Likewise.
18327 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18328 (read_agg_replacement_chain): Likewise.
18329 (ipcp_transform_function): Also check that by_ref flags match.
18331 2013-05-02 Richard Biener <rguenther@suse.de>
18333 * graphds.h (struct graph): Add obstack member.
18334 * graphds.c (new_graph): Initialize obstack and allocate
18336 (add_edge): Allocate edge from the obstack.
18337 (free_graph): Free the obstack instead of all edges and vertices.
18339 2013-05-02 Teresa Johnson <tejohnson@google.com>
18341 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18343 * cfg.c (update_bb_profile_for_threading): Ditto.
18344 * tree-inline.c (copy_bb): Ditto.
18345 (copy_edges_for_bb): Ditto.
18346 (initialize_cfun): Ditto.
18347 (copy_cfg_body): Ditto.
18348 (expand_call_inline): Ditto.
18349 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18350 (estimate_node_size_and_time): Ditto.
18351 (inline_merge_summary): Ditto.
18352 * cgraphclones.c (cgraph_clone_edge): Ditto.
18353 (cgraph_clone_node): Ditto.
18354 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18355 (compute_trg_info): Ditto.
18357 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18359 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18360 S reg when fp attribute set.
18361 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18363 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18365 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18367 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18368 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18369 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18371 2013-05-02 Richard Biener <rguenther@suse.de>
18373 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18374 (struct instantiate_cache_entry): New type.
18375 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18376 (struct instantiate_cache_type): New type.
18377 (set_instantiated_value, get_instantiated_value): Remove.
18378 (get_instantiated_value_entry): New function.
18379 (instantiate_scev_name): Use the new cache and adjust.
18380 (instantiate_scev_poly): Adjust.
18381 (instantiate_scev_binary): Likewise.
18382 (instantiate_array_ref): Likewise.
18383 (instantiate_scev_convert): Likewise.
18384 (instantiate_scev_not): Likewise.
18385 (instantiate_scev_3): Likewise.
18386 (instantiate_scev_2): Likewise.
18387 (instantiate_scev_r): Likewise.
18388 (instantiate_scev): Likewise.
18389 (resolve_mixers): Likewise.
18391 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18394 * lra-constraints.c (best_small_class_operands_num): Remove.
18395 (process_alt_operands): Remove small_class_operands_num. Take
18396 small classes operands into losers and only if the operand is not
18397 matched. Modify debugging output.
18398 (curr_insn_transform): Remove best_small_class_operands_num.
18401 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18403 * config/aarch64/aarch64-builtins.c
18404 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18405 * config/aarch64/aarch64-simd-builtins.def
18406 (reduc_splus_): Add new modes.
18407 (reduc_uplus_): New.
18408 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18409 (reduc_uplus_v4sf): Likewise.
18410 (reduc_splus_v4sf): Likewise.
18411 (aarch64_addv<mode>): Likewise.
18412 (reduc_uplus_<mode>): Likewise.
18413 (reduc_splus_<mode>): Likewise.
18414 (aarch64_addvv2di): Likewise.
18415 (reduc_uplus_v2di): Likewise.
18416 (reduc_splus_v2di): Likewise.
18417 (aarch64_addvv2si): Likewise.
18418 (reduc_uplus_v2si): Likewise.
18419 (reduc_splus_v2si): Likewise.
18420 (reduc_<sur>plus_<mode>): New.
18421 (reduc_<sur>plus_v2di): Likewise.
18422 (reduc_<sur>plus_v2si): Likewise.
18423 (reduc_<sur>plus_v4sf): Likewise.
18424 (aarch64_addpv4sf): Likewise.
18425 * config/aarch64/arm_neon.h
18426 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18427 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18428 add UNSPEC_SADDV, UNSPEC_UADDV.
18430 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18432 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18434 * config/aarch64/arm_neon.h
18435 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18437 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18439 * config/aarch64/aarch64-builtins
18440 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18442 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18444 * config/aarch64/aarch64-simd-builtins.def
18445 (reduc_smax_): New.
18446 (reduc_smin_): Likewise.
18447 (reduc_umax_): Likewise.
18448 (reduc_umin_): Likewise.
18449 (reduc_smax_nan_): Likewise.
18450 (reduc_smin_nan_): Likewise.
18453 (smax): Update for V2SF, V4SF and V2DF modes.
18456 (smin_nan): Likewise.
18457 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18458 (<su><maxmin><mode>3): ...This, refactor.
18459 (s<maxmin><mode>3): New.
18460 (<maxmin_uns><mode>3): Likewise.
18461 (reduc_<maxmin_uns>_<mode>): Refactor.
18462 (reduc_<maxmin_uns>_v4sf): Likewise.
18463 (reduc_<maxmin_uns>_v2si): Likewise.
18464 (aarch64_<fmaxmin><mode>: Remove.
18465 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18467 (vmin<q>_f<32,64>): Likewise.
18468 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18470 (su): Add mappings for smax, smin, umax, umin.
18472 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18473 (FMAXMIN): Rename as...
18474 (FMAXMIN_UNS): ...This.
18476 (fmaxminv): Likewise.
18477 (fmaxmin): Likewise.
18479 (maxmin_uns_op): Likewise.
18481 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18483 * config/aarch64/arm_neon.h
18484 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18485 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18486 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18488 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18490 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18491 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18493 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18495 * config/aarch64/aarch64-simd.md
18496 (vcond<mode>_internal): Handle special cases for constant masks.
18497 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18498 (vcondu<mode><mode>): Likewise.
18499 (vcond<v_cmp_result><mode>): New.
18501 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18503 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18504 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18505 * config/aarch64/aarch64-simd-builtins.def
18506 (cmeq): Update to BUILTIN_VALLDI.
18511 * config/aarch64/arm_neon.h
18512 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18513 to builtins or C as appropriate.
18515 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18517 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18519 (cmhi): Rename to...
18521 * config/aarch64/aarch64-simd.md
18522 (simd_mode): Add SF.
18523 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18524 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18525 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18526 (cstore<mode>_neg): ...This.
18527 * config/aarch64/iterators.md
18529 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18530 (COMPARISONS): New.
18531 (UCOMPARISONS): Likewise.
18532 (optab): Add missing comparisons.
18538 (VCMP_S): Likewise.
18539 (VCMP_U): Likewise.
18540 (V_cmp_result): Add DF, SF modes.
18541 (v_cmp_result): Likewise.
18543 (vmtype): Likewise.
18544 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18546 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18548 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18549 define_insn to define_insn_and_split.
18550 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18551 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18552 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18553 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18554 (thumb2_negscc): Likewise.
18556 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18558 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18560 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18562 * config/arm/thumb2.md: Remove trailing whitespaces.
18564 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18566 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18567 Use gen_int_mode rather than GEN_INT.
18569 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18571 * value-prof.c (stream_in_histogram_value): Remove the strayed
18574 2013-04-30 Richard Biener <rguenther@suse.de>
18576 PR middle-end/57122
18577 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18579 2013-04-30 Richard Biener <rguenther@suse.de>
18581 PR middle-end/57107
18582 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18584 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18586 PR rtl-optimization/56957
18587 PR rtl-optimization/57105
18588 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18589 variable. Use just INSN_UID for determining whether an insn
18590 should be only disconnected from the insn stream.
18591 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18593 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18595 PR tree-optimization/57104
18596 * tsan.c (instrument_expr): Don't instrument accesses to
18597 DECL_HARD_REGISTER VAR_DECLs.
18599 2013-04-30 Richard Biener <rguenther@suse.de>
18601 * function.h (loops_for_fn): New inline function.
18602 (set_loops_for_fn): Likewise.
18603 * cfgloop.h (place_new_loop): Add struct function parameter.
18604 (get_loop): Likewise.
18605 (get_loops): Likewise.
18606 (number_of_loops): Likewise.
18607 (fel_next): Adjust.
18608 (fel_init): Likewise.
18609 * cfg.c (get_loop_copy): Adjust.
18610 * cfgloop.c (flow_loops_dump): Likewise.
18611 (record_loop_exits): Likewise.
18612 (verify_loop_structure): Likewise.
18613 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18614 (estimate_reg_pressure_cost): Likewise.
18615 (mark_loop_exit_edges): Likewise.
18616 * cfgloopmanip.c (place_new_loop): Likewise.
18617 (add_loop): Likewise.
18618 (duplicate_loop): Likewise.
18619 * graph.c (draw_cfg_nodes): Likewise.
18620 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18621 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18622 (extract_affine_chrec): Likewise.
18623 (build_scop_iteration_domain): Likewise.
18624 * graphite.c (graphite_initialize): Likewise.
18625 * ira-build.c (create_loop_tree_nodes): Likewise.
18626 (more_one_region_p): Likewise.
18627 (rebuild_regno_allocno_maps): Likewise.
18628 (mark_loops_for_removal): Likewise.
18629 (mark_all_loops_for_removal): Likewise.
18630 (remove_unnecessary_regions): Likewise.
18631 (ira_build): Likewise.
18632 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18633 * loop-init.c (fix_loop_structure): Likewise.
18634 (gate_rtl_move_loop_invariants): Likewise.
18635 (gate_rtl_unswitch): Likewise.
18636 (gate_rtl_unroll_and_peel_loops): Likewise.
18637 (rtl_doloop): Likewise.
18638 * lto-streamer-in.c (input_cfg): Likewise.
18639 * lto-streamer-out.c (output_cfg): Likewise.
18640 * modulo-sched.c (sms_schedule): Likewise.
18641 * predict.c (tree_estimate_probability): Likewise.
18642 (tree_estimate_probability_driver): Likewise.
18643 (estimate_loops): Likewise.
18644 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18645 (move_sese_region_to_fn): Likewise.
18646 (debug_loop_num): Likewise.
18647 * tree-chrec.c (chrec_evaluate): Likewise.
18648 (hide_evolution_in_other_loops_than_loop): Likewise.
18649 (chrec_component_in_loop_num): Likewise.
18650 (reset_evolution_in_loop): Likewise.
18651 (evolution_function_is_invariant_rec_p): Likewise.
18652 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18653 * tree-inline.c (copy_loops): Likewise.
18654 (copy_cfg_body): Likewise.
18655 (tree_function_versioning): Likewise.
18656 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18657 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18659 (add_to_evolution_1): Likewise.
18660 (scev_const_prop): Likewise.
18661 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18662 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18663 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18664 (tree_ssa_lim_initialize): Likewise.
18665 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18666 (verify_loop_closed_ssa): Likewise.
18667 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18668 (tree_ssa_loop_im): Likewise.
18669 (tree_ssa_loop_unswitch): Likewise.
18670 (tree_vectorize): Likewise.
18671 (check_data_deps): Likewise.
18672 (tree_ssa_loop_ivcanon): Likewise.
18673 (tree_ssa_loop_bounds): Likewise.
18674 (tree_complete_unroll): Likewise.
18675 (tree_complete_unroll_inner): Likewise.
18676 (tree_parallelize_loops): Likewise.
18677 (tree_ssa_loop_prefetch): Likewise.
18678 (tree_ssa_loop_ivopts): Likewise.
18679 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18680 * tree-vectorizer.c (vectorize_loops): Likewise.
18682 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18684 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18685 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18686 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18687 with EABI_LINK_SPEC.
18689 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18692 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18695 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18698 * lra-constraints.c (process_alt_operands): Discourage a bit more
18699 using memory for pseudos. Print cost dump for alternatives.
18700 Modify cost values for conflicts with early clobbers.
18701 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18703 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18706 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18708 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18710 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18711 from/to S register.
18712 (movdi_aarch64): Support LDR/STR from/to D register.
18714 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18716 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18717 or higher by default.
18719 2013-04-29 Richard Biener <rguenther@suse.de>
18721 PR middle-end/57075
18722 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18723 even if not adding abnormal edges for calls that can make
18726 2013-04-29 Richard Biener <rguenther@suse.de>
18728 PR middle-end/57103
18729 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18731 (move_stmt_r): Remove redundant checking.
18733 2013-04-29 Teresa Johnson <tejohnson@google.com>
18736 * basic-block.h (apply_scale): New function.
18737 (apply_probability): Use apply_scale.
18738 * gimple-streamer-in.c (input_bb): Ditto.
18739 * lto-streamer-in.c (input_cfg): Ditto.
18740 * lto-cgraph.c (merge_profile_summaries): Ditto.
18741 * tree-optimize.c (execute_fixup_cfg): Ditto.
18742 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18743 (copy_edges_for_bb): Ditto.
18744 (copy_cfg_body): Ditto.
18746 2013-04-29 Tom de Vries <tom@codesourcery.com>
18748 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18749 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18750 (tail_merge_optimize): Handle current_loops == NULL.
18752 2013-04-26 Jeff Law <law@redhat.com>
18754 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18755 (simplify_cond_using_ranges): Generalize code to simplify
18756 COND_EXPRs where one argument is a constant and the other
18757 is an SSA_NAME created by an integral type conversion.
18759 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18761 * config/arm/arm.md (store_minmaxsi): Use only when
18762 optimize_insn_for_size_p.
18764 2013-04-29 Christian Bruel <christian.bruel@st.com>
18767 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18769 2013-04-29 Richard Biener <rguenther@suse.de>
18771 PR middle-end/57089
18772 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18773 loop tree make sure to schedule a fixup for the child as well.
18774 (expand_omp_for_generic): Properly add loops.
18775 (expand_omp_for_static_nochunk): Likewise.
18776 (expand_omp_for_static_chunk): Likewise.
18777 (expand_omp_for): For the degenerate case fixup loops.
18778 (expand_omp_sections): Fix default bb placement in loops.
18779 (expand_omp_atomic_pipeline): Properly add loops.
18781 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18783 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18785 2013-04-29 Tom de Vries <tom@codesourcery.com>
18787 * tree-ssa-tail-merge.c: Update header comment.
18789 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18791 * config/aarch64/arm_neon.h
18792 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18793 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18794 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18795 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18796 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18797 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18798 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18799 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18801 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18803 * config/aarch64/aarch64-simd.md
18804 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18805 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18806 fix_trunc, fixuns_trunc.
18807 (ftrunc<VDQF:mode>2): New.
18808 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18809 (fix_trunc_optab): New.
18811 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18813 * config/aarch64/aarch64-builtins.c
18814 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18815 iceilf, lround, iroundf.
18817 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18820 * config/i386/i386.h (enum ix86_tune_indices)
18821 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18822 New, split from X86_TUNE_INTER_UNIT_MOVES.
18823 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18824 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18825 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18826 (TARGET_INTER_UNIT_MOVES): Remove.
18827 * config/i386/i386.c (initial_ix86_tune_features): Update.
18828 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18829 (ix86_expand_convert_uns_didf_sse): Use
18830 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18831 (ix86_expand_vector_init_one_nonzero): Ditto.
18832 (ix86_expand_vector_init_interleave): Ditto.
18833 (inline_secondary_memory_needed): Return true for moves from SSE class
18834 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18835 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18836 * config/i386/constraints.md (Yi, Ym): Depend on
18837 TARGET_INTER_UNIT_MOVES_TO_VEC.
18838 (Yj, Yn): New constraints.
18839 * config/i386/i386.md (*movdi_internal): Change constraints of
18840 operand 1 from Yi to Yj and from Ym to Yn.
18841 (*movsi_internal): Ditto.
18842 (*movdf_internal): Ditto.
18843 (*movsf_internal): Ditto.
18844 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18845 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18846 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18847 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18848 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18849 * config/i386/sse.md (movdi_to_sse): Ditto.
18850 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18851 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18852 TARGET_INTER_UNIT_MOVES.
18853 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18854 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18855 instead of TARGET_INTER_UNIT_MOVES.
18856 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18857 operand 1 from Yi to Yj and from Ym to Yn.
18859 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18861 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18862 (float_truncate_hi_): Likewise.
18863 (float_extend_lo_): Likewise.
18864 (float_truncate_lo_): Likewise.
18865 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18866 (aarch64_float_extend_lo_v2df): Likewise.
18867 (vec_unpacks_hi_v4sf): Likewise.
18868 (aarch64_float_truncate_lo_v2sf): Likewise.
18869 (aarch64_float_truncate_hi_v4sf): Likewise.
18870 (vec_pack_trunc_v2df): Likewise.
18871 (vec_pack_trunc_df): Likewise.
18873 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18875 * config/aarch64/aarch64-builtins.c
18876 (aarch64_fold_builtin): Fold float conversions.
18877 * config/aarch64/aarch64-simd-builtins.def
18878 (floatv2si, floatv4si, floatv2di): New.
18879 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18880 * config/aarch64/aarch64-simd.md
18881 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18882 * config/aarch64/iterators.md (FLOATUORS): New.
18883 (optab): Add float, floatuns.
18884 (su_optab): Likewise.
18886 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18888 * config/aarch64/aarch64-builtins.c
18889 (aarch64_builtin_vectorized_function): Use new names for
18891 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18892 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18893 (fcvtzu): Split as...
18894 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18895 (fcvtas): Split as...
18896 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18897 (fcvtau): Split as...
18898 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18899 (fcvtps): Split as...
18900 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18901 (fcvtpu): Split as...
18902 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18903 (fcvtms): Split as...
18904 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18905 (fcvtmu): Split as...
18906 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18907 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18908 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18909 (lfrintnusf, lfrintnudf): Likewise.
18910 * config/aarch64/aarch64-simd.md
18911 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18913 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18914 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18915 (fcvt_pattern): Likewise.
18917 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18919 * config/aarch64/aarch64-simd.md
18920 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18921 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18923 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18925 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18926 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18927 (vrnd<a,m,n,p>_f32): Implement using builtins.
18928 (vrnd<i,x><q>_f<32, 64>): New.
18930 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18932 * config/aarch64/aarch64-builtins.c
18933 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18934 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18935 (frintz): Rename to...
18937 (frintp): Rename to...
18939 (frintm): Rename to...
18941 (frinti): Rename to...
18942 (nearbyint): ...this.
18943 (frintx): Rename to...
18945 (frinta): Rename to...
18947 * config/aarch64/aarch64-simd.md
18948 (aarch64_frint<frint_suffix><mode>): Delete.
18949 (<frint_pattern><mode>2): Convert to insn.
18950 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18951 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18952 (frint_pattern): Likewise.
18953 (frint_suffix): Likewise.
18955 2013-04-29 Richard Biener <rguenther@suse.de>
18957 PR tree-optimization/57081
18958 * loop-init.c: Include tree-flow.h.
18959 (loop_optimizer_finalize): Free number of iteration estimates.
18960 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18962 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18964 PR tree-optimization/57083
18965 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18966 non-singleton shift count range, zero extend low_bound for uns case.
18968 * config/i386/predicates.md (general_vector_operand): New predicate.
18969 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18970 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18971 if they aren't nonimmediate operands. If their original values
18972 satisfy const_vector_equal_evenodd_p, don't shift them.
18973 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18974 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18975 if not nonimmediate_operand.
18976 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18977 instead of register_operand.
18978 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18980 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18982 * stor-layout.c (finalize_size_functions): Allocate a structure and
18983 reset cfun before dumping the functions.
18985 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18987 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18990 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18991 use xop_pmacsdqh if uns_p.
18992 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18993 the immediate rotate count.
18995 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18997 * rtl.h (struct rtx_def): Add comment for field jump.
18998 (LRA_SUBREG_P): New macro.
18999 * recog.c (register_operand): Check LRA_SUBREG_P.
19000 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19002 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19003 (lra_final_code_change): Skip subreg change for operators.
19004 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19005 if there are no operand changes.
19006 * lra-constraints.c (curr_insn_set): New.
19007 (match_reload): Set LRA_SUBREG_P.
19008 (emit_spill_move): Ditto.
19009 (check_and_process_move): Use curr_insn_set. Process only single
19010 set insns. Don't initialize sec_mem_p and change_p.
19011 (simplify_operand_subreg): Use LRA_SUBREG_P.
19012 (reg_in_class_p): New function.
19013 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19014 of #ifdef. Add code to remove cycling.
19015 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19016 non-null disp. Reload inner instead of disp when base and index
19017 are null. Try to put lo_sum into register.
19018 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19019 (check_and_process_move): Move code for move cost check to
19020 simple_move_p. Remove equiv_substitution.
19021 (simple_move_p): New function.
19022 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19023 curr_insn_set. Call check_and_process_move only for single set
19024 insns. Use the new function. Move call of check_and_process_move
19025 after operand equiv substitution and address process.
19027 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19030 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19031 with nonlocal goto receivers or returns twice calls, ignore
19032 unininitialized values from abnormal edges to nl goto receiver
19033 or returns twice call.
19035 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19037 PR tree-optimization/57051
19038 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19039 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19042 2013-04-26 Richard Biener <rguenther@suse.de>
19044 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19045 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19046 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19047 (fixup_loop_arrays_after_move): New function.
19048 (move_sese_region_to_fn): Properly outline the loop tree parts
19049 of the SESE region.
19051 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19053 * config/i386/i386.md (type, unit): Fix long lines.
19055 2013-04-26 Richard Biener <rguenther@suse.de>
19057 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19058 (lto-streamer-out.o): Likewise.
19059 * cfgloop.c (init_loops_structure): Export, add struct function
19060 argument and adjust.
19061 (flow_loops_find): Adjust.
19062 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19063 (init_loops_structure): Declare.
19064 * lto-streamer-in.c: Include cfgloop.h.
19065 (input_cfg): Input the loop tree.
19066 * lto-streamer-out.c: Include cfgloop.h.
19067 (output_cfg): Output the loop tree.
19068 (output_struct_function_base): Do not drop PROP_loops.
19070 2013-03-26 Richard Biener <rguenther@suse.de>
19072 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19073 (pass_build_cfg): Provide PROP_loops.
19074 (move_sese_region_to_fn): Remove loops that are outlined into fn
19076 * tree-inline.c: Include cfgloop.h.
19077 (initialize_cfun): Do not drop PROP_loops.
19078 (copy_loops): New function.
19079 (copy_cfg_body): Copy loop structure.
19080 (tree_function_versioning): Initialize destination loop tree.
19081 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19082 (pass_parallelize_loops): Do IL verification.
19083 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19084 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19085 the CFG make sure we fixup loops as well.
19086 * tree-ssa-tail-merge.c: Include cfgloop.h.
19087 (replace_block_by): When merging loop latches mark loops for fixup.
19088 * lto-streamer-out.c (output_struct_function_base): Drop
19089 PROP_loops for now.
19090 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19091 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19092 * ipa-split.c: Include cfgloop.h.
19093 (split_function): Add the new return block to the loop tree root.
19094 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19095 whether we have removed the forwarder block.
19096 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19097 * cfgloop.h (place_new_loop): Declare.
19098 * cfgloopmanip.c (place_new_loop): Export.
19099 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19100 (tree-switch-conversion.o): Likewise.
19101 (tree-complex.o): Likewise.
19102 (tree-inline.o): Likewise.
19103 (tree-ssa-tailmerge.o): Likewise.
19104 (ipa-split.o): Likewise.
19105 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19106 (tree-ssa-copy.o): Likewise.
19107 * tree-switch-conversion.c: Include cfgloop.h
19108 (process_switch): If we emit a bit-test cascade, schedule loops
19110 * tree-complex.c: Include cfgloop.h.
19111 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19112 * asan.c: Include cfgloop.h.
19113 (create_cond_insert_point): Properly add new basic-blocks to
19114 loops, schedule loop fixup.
19115 * cfgloop.c (verify_loop_structure): Check that looks are not
19117 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19119 (expand_omp_for_generic): Likewise.
19120 (expand_omp_sections): Likewise.
19121 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19122 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19123 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19124 is initialized, not when loops are present.
19125 * tree-parloops.c (parallelize_loops): Remove checking here.
19126 * passes.c (init_optimization_passes): Schedule a copy-propagation
19127 pass before complete unrolling of inner loops.
19129 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19131 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19132 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19133 (colorize_init): Add argument to _WIN32 version.
19134 * toplev.c: Include diagnostic-color.h.
19135 (process_options): Default to -fdiagnostics-color=auto if
19136 GCC_COLORS env var is in the environment.
19137 * common.opt (fdiagnostics-color=): Add Var and Init.
19138 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19139 env var is in the environment, the default is auto rather than never.
19141 * diagnostic.h (file_name_as_prefix): Add context argument.
19142 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19143 the string as locus.
19144 * langhooks.c (lhd_print_error_function): Adjust caller.
19146 2013-04-25 Lawrence Crowl <crowl@google.com>
19148 * var-tracking.c (shared_hash_def::htab):
19149 Change type to hash_table. Update dependent calls and types.
19151 2013-04-25 Lawrence Crowl <crowl@google.com>
19153 * Makefile.in: Update as needed below.
19155 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19156 Move declaration to after the type's method definitons.
19158 * attribs.c (htab_t scoped_attributes::attribute_hash):
19159 Change type to hash_table. Update dependent calls and types.
19161 * bitmap.c (htab_t bitmap_desc_hash):
19162 Change type to hash_table. Update dependent calls and types.
19164 * cselib.c (htab_t cselib_hash_table):
19165 Change type to hash_table. Update dependent calls and types.
19167 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19168 (hash_string_slot_node): Move implementation into lto-streamer.h
19169 struct string_slot_hasher.
19170 (eq_string_slot_node): Likewise.
19172 * data-streamer-out.c: Update output_block::string_hash_table
19173 dependent calls and types.
19175 * dwarf2cfi.c (htab_t trace_index):
19176 Change type to hash_table. Update dependent calls and types.
19178 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19179 Change type to hash_table. Update dependent calls and types.
19180 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19181 (htab_t optimize_external_refs::map): Likewise.
19182 (htab_t output_comp_unit::extern_map): Likewise.
19183 (htab_t output_comdat_type_unit::extern_map): Likewise.
19184 (htab_t output_macinfo::macinfo_htab): Likewise.
19185 (htab_t optimize_location_lists::htab): Likewise.
19186 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19188 * except.c (htab_t ehspec_hash_type):
19189 Change type to hash_table. Update dependent calls and types.
19190 (assign_filter_values::ttypes): Likewise.
19191 (assign_filter_values::ehspec): Likewise.
19192 (sjlj_assign_call_site_values::ar_hash): Likewise.
19193 (convert_to_eh_region_ranges::ar_hash): Likewise.
19195 * gcse.c (htab_t pre_ldst_table):
19196 Change type to hash_table. Update dependent calls and types.
19198 * ggc-common.c (htab_t saving_htab):
19199 Change type to hash_table. Update dependent calls and types.
19200 (htab_t loc_hash): Likewise.
19201 (htab_t ptr_hash): Likewise.
19202 (call_count): Rename ggc_call_count.
19203 (call_alloc): Rename ggc_call_alloc.
19204 (loc_descriptor): Rename make_loc_descriptor.
19205 (add_statistics): Rename ggc_add_statistics.
19207 * ggc-common.c (saving_htab):
19208 Change type to hash_table. Update dependent calls and types.
19210 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19211 (push_gimplify_context): Likewise.
19212 (pop_gimplify_context): Likewise.
19213 (struct gimple_temp_hash_elt): Added.
19214 (struct gimplify_hasher): Likewise.
19215 (struct gimplify_ctx.temp_htab):
19216 Change type to hash_table. Update dependent calls and types.
19218 * gimple-fold.c: Include gimplify-ctx.h.
19220 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19221 Change type to hash_table. Update dependent calls and types.
19222 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19223 avoid potential global name collision.
19225 * gimplify.c: Include gimplify-ctx.h.
19226 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19227 (htab_t gimplify_ctx::temp_htab):
19228 Update dependent calls and types for new type hash_table.
19229 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19230 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19232 * gimplify-ctx.h: New.
19233 (struct gimple_temp_hash_elt): Move from gimplify.c.
19234 (class gimplify_hasher): New.
19235 (struct gimplify_ctx): Move from gimple.h.
19236 (htab_t gimplify_ctx::temp_htab):
19237 Change type to hash_table. Update dependent calls and types.
19239 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19240 (htab_t ivs_params::newivs_index):
19241 Change type to hash_table. Update dependent calls and types.
19242 (htab_t ivs_params::params_index): Likewise.
19243 (htab_t print_generated_program::params_index): Likewise.
19244 (htab_t gloog::newivs_index): Likewise.
19245 (htab_t gloog::params_index): Likewise.
19247 * graphite.c: Include graphite-htab.h.
19248 4htab_t graphite_transform_loops::bb_pbb_mapping):
19249 Change type to hash_table. Update dependent calls and types.
19251 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19252 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19253 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19255 * graphite-dependences.c: Include graphite-htab.h.
19256 (loop_is_parallel_p): Change hash table type of parameter.
19258 * graphite-htab.h: New.
19259 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19260 (extern find_pbb_via_hash): Move from graphite-poly.h.
19261 (extern loop_is_parallel_p): Move from graphite-poly.h.
19262 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19264 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19265 (extern loop_is_parallel_p): Move to graphite-htab.h.
19266 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19268 * haifa-sched.c (htab_t delay_htab):
19269 Change type to hash_table. Update dependent calls and types.
19270 (htab_t delay_htab_i2): Likewise.
19272 * ira-color.c (htab_t allocno_hard_regs_htab):
19273 Change type to hash_table. Update dependent calls and types.
19275 * ira-costs.c (htab_t cost_classes_htab):
19276 Change type to hash_table. Update dependent calls and types.
19278 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19279 Change type to hash_table. Update dependent calls and types.
19281 * loop-iv.c (htab_t bivs):
19282 Change type to hash_table. Update dependent calls and types.
19284 * loop-unroll.c (htab_t opt_info::insns_to_split):
19285 Change type to hash_table. Update dependent calls and types.
19286 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19288 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19289 (struct string_slot_hasher): New.
19290 (htab_t output_block::string_hash_table):
19291 Change type to hash_table. Update dependent calls and types.
19293 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19294 (htab_t file_name_hash_table):
19295 Change type to hash_table. Update dependent calls and types.
19297 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19300 * lto-streamer.c (htab_t tree_htab):
19301 Change type to hash_table. Update dependent calls and types.
19303 * omp-low.c: Include gimplify-ctx.h.
19305 * passes.c (htab_t name_to_pass_map):
19306 Change type to hash_table. Update dependent calls and types.
19307 (pass_traverse): Rename to passes_pass_traverse.
19309 * plugin.c (htab_t event_tab):
19310 Change type to hash_table. Update dependent calls and types.
19312 * postreload-gcse.c (htab_t expr_table):
19313 Change type to hash_table. Update dependent calls and types.
19314 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19316 * sese.c (debug_rename_map_1): Make extern.
19317 (htab_t copy_bb_and_scalar_dependences::rename_map):
19318 Change type to hash_table. Update dependent calls and types.
19320 * sese.h (extern debug_rename_map): Move to .c file.
19322 * store-motion.c (htab_t store_motion_mems_table):
19323 Change type to hash_table. Update dependent calls and types.
19325 * trans-mem.c (htab_t tm_new_mem_hash):
19326 Change type to hash_table. Update dependent calls and types.
19328 * tree-browser.c (htab_t TB_up_ht):
19329 Change type to hash_table. Update dependent calls and types.
19331 * tree-cfg.c (htab_t discriminator_per_locus):
19332 Change type to hash_table. Update dependent calls and types.
19334 * tree-complex.c: Include tree-hasher.h
19335 (htab_t complex_variable_components):
19336 Change type to hash_table. Update dependent calls and types.
19338 * tree-eh.c (htab_t finally_tree):
19339 Change type to hash_table. Update dependent calls and types.
19341 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19342 struct int_tree_hasher.
19343 (extern int_tree_map_eq): Likewise.
19344 (uid_decl_map_hash): Removed.
19345 (extern decl_tree_map_eq): Likewise.
19347 * tree-hasher.h: New.
19348 (struct int_tree_hasher): New.
19349 (typedef int_tree_htab_type): New.
19351 * tree-inline.c: Include gimplify-ctx.h.
19353 * tree-mudflap.c: Include gimplify-ctx.h.
19355 * tree-parloops.c: Include tree-hasher.h.
19356 (htab_t eliminate_local_variables_stmt::decl_address):
19357 Change type to hash_table. Update dependent calls and types.
19358 (htab_t separate_decls_in_region::decl_copies): Likewise.
19360 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19361 Change type to hash_table. Update dependent calls and types.
19363 * tree-sra.c (candidates):
19364 Change type to hash_table. Update dependent calls and types.
19366 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19368 (int_tree_map_hash): Likewise.
19370 * tree-ssa-dom.c (htab_t avail_exprs):
19371 Change type to hash_table. Update dependent calls and types.
19373 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19374 Change type to hash_table. Update dependent calls and types.
19376 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19377 Change type to hash_table. Update dependent calls and types.
19379 * tree-ssa-phiopt.c (seen_ssa_names):
19380 Change type to hash_table. Update dependent calls and types.
19382 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19383 Change type to hash_table. Update dependent calls and types.
19385 * tree-ssa-uncprop.c (equiv):
19386 Change type to hash_table. Update dependent calls and types.
19388 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19390 PR rtl-optimization/57003
19391 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19392 call note_stores with kill_clobbered_value callback again after
19393 killing regs_invalidated_by_call.
19395 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19397 * config/aarch64/aarch64-simd.md
19398 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19399 (aarch64_simd_bsl<mode>): Likewise.
19400 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19402 2013-04-25 Marek Polacek <polacek@redhat.com>
19404 PR tree-optimization/57066
19405 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19407 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19409 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19411 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19413 * config/aarch64/aarch64-builtins.c
19414 (aarch64_fold_builtin): New.
19415 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19416 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19417 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19418 * config/aarch64/arm_neon.h
19419 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19421 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19422 Tejas Belagod <tejas.belagod@arm.com>
19424 * config/aarch64/aarch64-builtins.c
19425 (aarch64_gimple_fold_builtin): New.
19426 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19427 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19428 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19429 (addvv4sf): Update.
19430 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19432 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19434 * config/aarch64/aarch64.md
19435 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19437 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19439 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19440 (*ngcsi_uxtw): New pattern.
19442 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19443 Julian Brown <julian@codesourcery.com>
19445 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19446 (TB_DREG): Add T_V4HF.
19447 (v4hf_UP): New macro.
19448 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19449 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19450 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19452 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19453 (arm_vector_mode_supported_p): Handle V4HF.
19454 (arm_mangle_map): Handle V4HFmode.
19455 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19456 * config/arm/arm_neon_builtins.def: Add entries for
19457 vcvtv4hfv4sf, vcvtv4sfv4hf.
19458 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19459 (neon_vcvtv4hfv4sf): Likewise.
19460 * config/arm/neon-gen.ml: Handle half-precision floating point
19462 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19463 * config/arm/arm_neon.h: Regenerate.
19464 * config/arm/neon.ml (type elts): Add F16.
19465 (type vectype): Add T_float16x4, T_floatHF.
19466 (type vecmode): Add V4HF.
19467 (type features): Add Requires_FP_bit feature.
19468 (elt_width): Handle F16.
19469 (elt_class): Likewise.
19470 (elt_of_class_width): Likewise.
19471 (mode_of_elt): Refactor.
19472 (type_for_elt): Handle F16, fix error messages.
19473 (vectype_size): Handle T_float16x4.
19474 (vcvt_sh): New function.
19475 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19476 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19477 (string_of_mode): Handle V4HF.
19478 * doc/arm-neon-intrinsics.texi: Regenerate.
19480 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19482 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19483 format specifier in 'X' case.
19485 2013-04-25 Alan Modra <amodra@gmail.com>
19488 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19489 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19490 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19491 Repeat for many other rotate/shift and mask patterns using subregs.
19492 Name lshiftrt insns.
19493 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19494 on WORDS_BIG_ENDIAN.
19496 2013-04-25 Alan Modra <amodra@gmail.com>
19498 * config.gcc: Support little-endian powerpc-linux targets.
19499 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19500 (LINK_OS_LINUX_SPEC): Define.
19501 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19502 Preserve MASK_LITTLE_ENDIAN.
19503 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19504 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19505 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19506 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19507 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19508 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19509 Correct fp word order for little-endian. Don't shift toc entries
19510 smaller than a word for little-endian.
19511 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19512 (bswapdi2 splits): Correct low-part subreg for little-endian.
19513 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19514 low/high where such is correct only for be.
19515 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19516 little-endian for -mcall-aixdesc.
19518 2013-04-25 Alan Modra <amodra@gmail.com>
19520 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19521 replace_equiv_address_nv.
19523 2013-04-25 Alan Modra <amodra@gmail.com>
19525 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19527 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19530 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19531 * rtl.h (struct rtx_def): ...
19533 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19535 PR rtl-optimizations/57046
19536 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19537 for multi-reg splits.
19539 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19541 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19543 2013-04-24 Sterling Augustine <saugustine@google.com>
19545 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19546 (comp_dir_string, debug_str_dwo_section): New.
19547 (DEBUG_STR_DWO_SECTION): Rename to ...
19548 (DEBUG_DWO_STR_SECTION): ... this.
19549 (DEBUG_NORM_STR_SECTION): Delete.
19550 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19551 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19552 (find_AT_string): Move most logic to ...
19553 (find_AT_string_in_table): ... here. New.
19554 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19555 add_skeleton_AT_string. Delete logic.
19556 (output_skeleton_debug_sections): Remove call to
19557 add_top_level_skeleton_die_attrs.
19558 (add_comp_dir_attribute): Move logic to comp_dir_string.
19559 (dwarf2out_init): Initialize debug_str_dwo_section.
19560 (output_indirect_string): Call find_string_form.
19561 (output_indirect_strings): Rewrite.
19562 (prune_unused_types): Empty skeleton_debug_str_hash.
19563 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19564 (dwarf2out_finish): Call output_indirect_strings.
19566 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19568 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19570 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19572 * rtl.h (struct rtx_def): Add comment for field jump.
19573 (LRA_SUBREG_P): New macro.
19574 * recog.c (register_operand): Check LRA_SUBREG_P.
19575 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19577 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19578 (lra_final_code_change): Skip subreg change for operators.
19579 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19580 if there are no operand changes.
19581 * lra-constraints.c (curr_insn_set): New.
19582 (match_reload): Set LRA_SUBREG_P.
19583 (emit_spill_move): Ditto.
19584 (check_and_process_move): Use curr_insn_set. Process only single
19585 set insns. Don't initialize sec_mem_p and change_p.
19586 (simplify_operand_subreg): Use LRA_SUBREG_P.
19587 (reg_in_class_p): New function.
19588 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19589 of #ifdef. Add code to remove cycling.
19590 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19591 non-null disp. Reload inner instead of disp when base and index
19592 are null. Try to put lo_sum into register.
19593 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19594 (check_and_process_move): Move code for move cost check to
19595 simple_move_p. Remove equiv_substitution.
19596 (simple_move_p): New function.
19597 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19598 curr_insn_set. Call check_and_process_move only for single set
19599 insns. Use the new function. Move call of check_and_process_move
19600 after operand equiv substitution and address process.
19602 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19604 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19605 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19606 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19608 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19610 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19612 2013-04-24 Marek Polacek <polacek@redhat.com>
19614 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19615 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19616 (select_loops_exit_conditions): Likewise.
19617 (number_of_iterations_for_all_loops): Likewise.
19618 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19619 (scev_analysis): Likewise.
19621 2013-04-02 Catherine Moore <clm@codesourcery.com>
19622 Chao-ying Fu <fu@mips.com>
19624 * config/mips/micromips.md (jraddiusp): New pattern.
19625 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19626 instruction if possible.
19628 2013-04-24 Alan Modra <amodra@gmail.com>
19630 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19632 2013-04-24 Julian Brown <julian@codesourcery.com>
19633 Chung-Lin Tang <cltang@codesourcery.com>
19635 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19636 dependency behavior in enumeration type DIE generation. Add TODO note
19637 to comments about future DW_FORM_sdata/udata re-work of related code.
19639 2013-04-23 Lawrence Crowl <crowl@google.com>
19641 * Makefile.in: Update as needed below.
19643 * hash-table.h (class hash_table):
19644 Correct many methods with parameter types compare_type to the correct
19645 value_type. (Correct code was unlikely to notice the change.)
19646 (hash_table::elements_with_deleted) New.
19647 (class hashtable::iterator): New.
19648 (hashtable::begin()): New.
19649 (hashtable::end()): New.
19650 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19652 * statistics.c (statistics_hashes):
19653 Change type to hash_table. Update dependent calls and types.
19655 * tree-into-ssa.c (var_infos):
19656 Change type to hash_table. Update dependent calls and types.
19658 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19659 Change type to hash_table. Update dependent calls and types.
19661 * tree-ssa-loop-im.c (struct mem_ref.refs):
19662 Change type to hash_table. Update dependent calls and types.
19664 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19665 Change type to hash_table. Update dependent calls and types.
19667 * tree-ssa-sccvn.c (vn_tables_s::nary):
19668 Change type to hash_table. Update dependent calls and types.
19669 (vn_tables_s::phis): Likewise.
19670 (vn_tables_s::references): Likewise.
19672 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19673 (vn_reference_eq): Update parameter and return types.
19675 * tree-ssa-structalias.c (pointer_equiv_class_table):
19676 Change type to hash_table. Update dependent calls and types.
19677 (location_equiv_class_table): Likewise.
19679 * tree-vect-data-refs.c: Consequential changes for making
19680 peeling a hash_table.
19682 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19683 (destroy_loop_vec_info): Dependent hash_table update.
19685 * tree-vectorizer.h (peeling_htab):
19686 Change type to hash_table. Update dependent calls and types.
19688 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19690 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19691 to check the register content is equal or not.
19692 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19693 to assign register content record.
19694 * lra-eliminations.c (update_reg_eliminate): Use
19695 lra_update_reg_val_offset to update register content offset.
19696 * lra-int.h (struct lra_reg): Add offset member.
19697 (lra_reg_val_equal_p): New static inline function.
19698 (lra_update_reg_val_offset): New static inline function.
19699 (lra_assign_reg_val): New static inline function.
19700 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19701 to assign register content record.
19702 (initialize_lra_reg_info_element): Initial offset to zero.
19704 2013-04-23 Catherine Moore <clm@codesourcery.com>
19706 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19707 operands. Record compression.
19709 2013-04-23 Xinliang David Li <davidxl@google.com>
19711 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19713 2013-04-23 Richard Biener <rguenther@suse.de>
19715 PR middle-end/57036
19716 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19717 parameter, only add abnormal goto edges from the copied body
19718 if the call could perform abnormal gotos.
19719 (copy_cfg_body): Adjust.
19721 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19723 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19725 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19727 * coretypes.h (gimple_stmt_iterator): Add struct to make
19730 2013-04-23 Richard Biener <rguenther@suse.de>
19732 PR tree-optimization/57026
19733 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19734 from SSA names occuring in abnormal PHI nodes.
19736 2013-04-22 Andi Kleen <ak@linux.intel.com>
19738 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19740 2013-04-22 Andi Kleen <ak@linux.intel.com>
19742 * lto/lto.c (print_lto_report_1): Declare early.
19743 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19745 2013-04-22 Andi Kleen <ak@linux.intel.com>
19747 * common.opt (-flto-report-wpa): Add.
19748 * doc/invoke.texi (-flto-report-wpa): Add.
19749 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19752 2013-04-22 Xinliang David Li <davidxl@google.com>
19754 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19755 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19756 * Makefile.in: New dependency
19758 David Daney <ddaney.cavm@gmail.com>
19760 * configure.ac (gcc_cv_as_micromips_support): Use the
19761 --fatal-warnings option.
19762 * configure: Regenerate.
19764 2013-04-22 Marek Polacek <polacek@redhat.com>
19767 * tsan.c (instrument_expr): Don't instrument expression
19768 in case its size is zero.
19770 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19774 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19776 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19778 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19780 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19781 (gimple_stmt_iterator): New typedef.
19782 * gimple.h (gimple_stmt_iterator): Rename to...
19783 (gimple_stmt_iterator_d): ... This.
19784 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19785 trees be valid for GIMPLE and GENERIC.
19786 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19787 * gimple-fold.c (gimple_fold_call): Call target hook
19788 gimple_fold_builtin.
19789 * hooks.c (hook_bool_gsiptr_false): New.
19790 * hooks.h (hook_bool_gsiptr_false): New.
19791 * target.def (fold_stmt): New.
19792 * doc/tm.texi: Regenerate.
19794 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19797 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19798 a set sp if no stack realignment.
19800 2013-04-22 Nick Clifton <nickc@redhat.com>
19802 * config.gcc (tilegx-linux): Extend extra_objs rather than
19804 (tilepro-linux): Likewise.
19806 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19808 * config/aarch64/aarch64-builtins.c
19810 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19811 (VAR<1-12>): Add MAP parameter.
19812 (BUILTIN_*): Likewise.
19813 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19814 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19815 (aarch64_ushl_n<mode>): Likewise.
19816 (aarch64_sshr_n<mode>): Likewise.
19817 (aarch64_ushr_n<mode>): Likewise.
19818 (aarch64_<maxmin><mode>): Likewise.
19819 (aarch64_sqrt<mode>): Likewise.
19820 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19821 (vshr<q>_n_*): Likewise.
19823 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19825 * config/aarch64/aarch64-builtins.c
19826 (aarch64_simd_builtin_type_mode): Handle SF types.
19828 (BUILTIN_GPF): Define.
19829 (aarch64_init_simd_builtins): Handle SF types.
19830 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19831 (frecps): Likewise.
19832 (frecpx): Likewise.
19833 * config/aarch64/aarch64-simd.md
19834 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19835 (aarch64_frecpe<mode>): New.
19836 (aarch64_frecps<mode>): Likewise.
19837 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19838 (v8type): Add frecp<esx>.
19839 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19840 (aarch64_frecps<mode>): Likewise.
19841 * config/aarch64/iterators.md (FRECP): New.
19842 (frecp_suffix): Likewise.
19843 * config/aarch64/arm_neon.h
19844 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19846 2013-04-22 Christian Bruel <christian.bruel@st.com>
19849 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19850 (REG_CLASS_NAMES): Idem.
19851 (REG_CLASS_CONTENTS): Idem.
19852 (REGCLASS_HAS_FP_REG): Idem.
19853 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19854 (sh_conditional_register_usage): Idem.
19856 2013-04-21 Jeff Law <law@redhat.com>
19858 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19859 (ssa_forward_propagate_and_combine): Use it.
19861 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19863 * lra.c: Update the flow chart diagram.
19865 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19867 PR rtl-optimization/56847
19868 * lra-constraints.c (process_alt_operands): Discourage alternative
19869 with non-matche doffsettable memory constraint fro memory with
19872 2013-04-19 Richard Biener <rguenther@suse.de>
19874 PR tree-optimization/56982
19875 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19877 * gimplify.c (gimplify_call_expr): Notice special calls.
19878 (gimplify_modify_expr): Likewise.
19879 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19880 abnormal control flow receivers.
19881 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19882 in the same way as cfun->has_nonlocal_labels.
19883 (gimple_purge_dead_abnormal_call_edges): Likewise.
19884 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19885 receivers start a basic-block.
19887 2013-04-19 Richard Biener <rguenther@suse.de>
19889 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19891 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19892 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19893 (SLP_TREE_LOAD_PERMUTATION): Add.
19894 (vect_transform_slp_perm_load): Adjust prototype.
19895 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19896 (vect_free_slp_instance): Likewise.
19897 (vect_create_new_slp_node): Likewise.
19898 (vect_supported_slp_permutation_p): Remove.
19899 (vect_slp_rearrange_stmts): Adjust.
19900 (vect_supported_load_permutation_p): Likewise. Inline
19901 vect_supported_slp_permutation_p here.
19902 (vect_analyze_slp_instance): Compute load permutations per
19903 slp node instead of per instance.
19904 (vect_get_slp_defs): Adjust.
19905 (vect_transform_slp_perm_load): Likewise.
19906 (vect_schedule_slp_instance): Remove redundant code.
19907 (vect_schedule_slp): Remove hack for PR56270, add it ...
19908 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19909 CSE loads for SLP. Adjust.
19911 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19913 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19914 spelling in two comments.
19916 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19919 * config/arm/arm.c (load_multiple_sequence): Require SP
19920 as base register for loads if SP is in the register list.
19922 2013-04-19 Martin Jambor <mjambor@suse.cz>
19924 PR tree-optimization/56718
19925 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19926 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19927 and made public. Adjusted all callers.
19928 (ipa_intraprocedural_devirtualization): New function.
19929 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19930 (ipa_intraprocedural_devirtualization): Likewise.
19931 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19933 2013-04-19 Richard Biener <rguenther@suse.de>
19935 PR tree-optimization/57000
19936 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19938 2013-04-19 Terry Guo <terry.guo@arm.com>
19940 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19942 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19943 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19944 (cortex_m4_fmacs): Use new reservations.
19945 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19947 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19949 PR rtl-optimization/56999
19950 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19952 (lra_coalesce): Remove split_origin_bitmap and related code.
19953 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19954 ranges if necessary.
19956 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19958 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19960 (ix86_expand_call): Remove clobbered_registers array and use
19961 x86_64_ms_sysv_extra_clobbered_registers instead.
19962 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19964 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19966 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19967 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19968 (*call_value_rex64_ms_sysv): Ditto.
19970 2013-04-18 Cary Coutant <ccoutant@google.com>
19972 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19973 parent instead of die_mark.
19975 2013-04-18 Diego Novillo <dnovillo@google.com>
19977 * gimple.c (create_gimple_tmp): New.
19978 (get_expr_type): New.
19979 (build_assign): New.
19980 (build_type_cast): New.
19981 * gimple.h (enum ssa_mode): Define.
19982 (gimple_seq_set_location): New.
19983 * asan.c (build_check_stmt): Change some gimple_build_* calls
19984 to use build_assign and build_type_cast.
19986 2013-04-18 Richard Biener <rguenther@suse.de>
19988 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19989 handle negative step. Remove redundant checks.
19990 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19991 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19992 for negative step and grouped loads fail to vectorize.
19994 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19996 * emit-rtl.c (reset_insn_used_flags): New function.
19997 (reset_all_used_flags): Use it.
19998 (verify_insn_sharing): New function.
19999 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20001 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20003 PR tree-optimization/56984
20004 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20005 and (x >> M) >= N don't register any assertion if N << M is the
20008 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20010 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20011 do not delete_insn insns that have not yet been emitted, only
20012 unlink them with remove_insn.
20013 * df-scan.c (df_insn_delete): Revert r197492.
20015 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20017 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20018 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20020 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20022 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20023 define_insn_and_split.
20024 (and_scc,ior_scc,negscc): Likewise.
20025 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20027 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20029 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20030 can benefit from using a sequence of LDRD instructions in epilogue
20031 instead of a single LDM instruction.
20033 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20036 * doc/extend.texi: Fix typo.
20038 2013-04-17 Richard Biener <rguenther@suse.de>
20040 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20041 (vect_build_slp_tree): ... here.
20042 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20043 match. Remove special-casing of mismatched complex loads.
20044 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20045 re-try the match with swapped commutative operands.
20046 (vect_supported_load_permutation_p): Remove special-casing of
20047 mismatched complex loads.
20048 (vect_analyze_slp_instance): Adjust.
20050 2013-04-17 Richard Biener <rguenther@suse.de>
20052 PR rtl-optimization/56921
20053 * cfgloop.h (struct loop): Add simple_loop_desc member.
20054 (struct niter_desc): Mark with GTY(()).
20055 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20056 * loop-iv.c (get_simple_loop_desc): Likewise.
20057 (free_simple_loop_desc): Likewise.
20060 2013-04-16 Richard Biener <rguenther@suse.de>
20062 PR rtl-optimization/56921
20063 * loop-init.c (pass_rtl_move_loop_invariants): Add
20064 TODO_do_not_ggc_collect to todo_flags_finish.
20065 (pass_rtl_unswitch): Same.
20066 (pass_rtl_unroll_and_peel_loops): Same.
20067 (pass_rtl_doloop): Same.
20069 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20071 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20072 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20073 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20075 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20076 * tree-streamer.c (record_common_node): Adjust reference in comment.
20078 2013-04-17 Terry Guo <terry.guo@arm.com>
20080 * config/arm/cortex-m4.md: Add a new bypass.
20082 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20084 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20086 (*subs_<optab><mode>_multp2): New pattern.
20087 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20088 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20090 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20092 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20093 (*subs_mul_imm_<mode>): New pattern.
20095 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20098 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20099 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20100 (vsx_movti_32bit): Change j->wa to O->wa.
20102 2013-04-16 Richard Biener <rguenther@suse.de>
20104 PR rtl-optimization/56921
20105 * loop-init.c (pass_rtl_move_loop_invariants): Add
20106 TODO_do_not_ggc_collect to todo_flags_finish.
20107 (pass_rtl_unswitch): Same.
20108 (pass_rtl_unroll_and_peel_loops): Same.
20109 (pass_rtl_doloop): Same.
20111 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20113 * config/arm/arm.c (emit_multi_reg_push): New declaration
20114 for an existing function.
20115 (arm_emit_strd_push): New function.
20116 (arm_expand_prologue): Used here.
20117 (arm_emit_ldrd_pop): New function.
20118 (arm_expand_epilogue): Used here.
20119 (arm_get_frame_offsets): Update condition.
20120 (arm_emit_multi_reg_pop): Add a special case for load of a single
20121 register with writeback.
20123 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20125 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20128 2013-04-16 Richard Biener <rguenther@suse.de>
20130 PR tree-optimization/56756
20131 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20132 (first_mem_ref_loc): New.
20133 (execute_sm): Place the load temporarily before a previous
20134 access instead of in the latch edge to ensure its SSA dependencies
20135 are defined at points dominating the load.
20137 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20139 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20140 correct fix by moving header and footer insn to the footer of
20141 the merged basic block. Clear BB_END of the merged-away block.
20143 PR middle-end/43631
20144 * emit-rtl.c (make_note_raw): New function.
20145 (link_insn_into_chain): New static inline function.
20146 (add_insn): Use it.
20147 (add_insn_before, add_insn_after): Factor insn chain linking code...
20148 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20149 using link_insn_into_chain.
20150 (note_outside_basic_block_p): New helper function for emit_note_after
20151 and emit_note_before.
20152 (emit_note_after): Use nobb variant of add_insn_after if the note
20153 should not be contained in a basic block.
20154 (emit_note_before): Use nobb variant of add_insn_before if the note
20155 should not be contained in a basic block.
20156 (emit_note_copy): Use make_note_raw.
20157 (emit_note): Likewise.
20158 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20159 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20160 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20161 the moved barrier the tail of the basic block it follows.
20162 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20164 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20166 PR tree-optimization/56962
20167 * gimple-ssa-strength-reduction.c (record_increment): Only set
20168 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20169 either rhs1 or rhs2 is equal to c->base_expr.
20171 2013-04-15 Richard Biener <rguenther@suse.de>
20173 PR tree-optimization/56933
20174 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20176 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20177 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20178 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20179 dependence check ...
20180 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20182 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20183 GROUP_READ_WRITE_DEPENDENCE.
20185 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20187 * emit-rtl.c (reset_all_used_flags): New function.
20188 (verify_rtl_sharing): Call reset_all_used_flags before and after
20189 performing the checks.
20191 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20193 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20194 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20195 * config/arm/constraints.md (De): New constraint.
20196 * config/arm/neon.md (anddi3_neon): Delete.
20197 (neon_vand<mode>): Expand to standard anddi3 pattern.
20198 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20199 Move earlier in the file.
20200 (neon_inv_logic_op2): Likewise.
20201 (arm_anddi_operand_neon): New predicate.
20203 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20205 * configure.ac (gcc_cv_ld_as_needed): Set
20206 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20207 Use -z ignore, -z record on *-*-solaris2*.
20208 (HAVE_LD_AS_NEEDED): Update comment.
20209 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20210 * configure: Regenerate.
20211 * config.in: Regenerate.
20212 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20213 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20214 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20215 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20216 equivalents. Fix markup.
20217 * doc/tm.texi: Regenerate.
20219 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20221 * config/i386/i386.opt: New option mstack-protector-guard=.
20222 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20223 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20224 TARGET_SSP_TLS_GUARD.
20225 * config/i386/i386.c (ix86_option_override_internal): Set
20226 ix86_stack_protector_guard.
20227 * config/i386/i386.md (stack_protect_set): Enable for
20228 TARGET_SSP_TLS_GUARD only.
20229 (stack_protect_set_<mode>): Ditto.
20230 (stack_protect_test): Ditto.
20231 (stack_protect_test_<mode>): Ditto.
20232 * doc/invoke.texi (i386 Option): Document.
20234 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20237 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20238 (S_MODES): Set H_MODE bit.
20239 (SF_MODES): Set only S_MODE and SF_MODE bits.
20240 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20241 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20242 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20243 <MODE_FLOAT>: Likewise.
20245 2013-04-15 Joey Ye <joey.ye@arm.com>
20247 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20249 2013-04-15 Joey Ye <joey.ye@arm.com>
20251 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20253 (thumb_far_jump_used_p): Count instruction size and set
20256 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20258 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20259 * resource.c (mark_target_live_regs): Reformat conditional block.
20261 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20263 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20264 notes, they are emitted only just before final.
20265 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20267 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20269 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20270 * cfgrtl.c (delete_insn): Call it here instead.
20271 * lra-spills.c (lra_final_code_change): Use delete_insn.
20272 * haifa-sched.c (sched_remove_insn): Likewise.
20273 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20274 returning to the nop pool.
20275 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20276 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20278 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20280 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20281 * doc/tm.texi: Regenerated.
20283 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20285 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20288 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20290 * df-core.c (df_find_def): Compare register numbers.
20291 (df_find_use): Likewise.
20293 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20296 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20297 lra_in_progress for return.
20299 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20301 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20302 define_insn into define_insn_and_split and emit movsicc patterns.
20304 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20306 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20308 2013-04-12 Richard Biener <rguenther@suse.de>
20310 * tree-pass.h (TODO_do_not_ggc_collect): New.
20311 * passes.c (execute_one_ipa_transform_pass): Honor
20312 TODO_do_not_ggc_collect.
20313 (execute_one_pass): Likewise.
20316 2013-04-10 Richard Biener <rguenther@suse.de>
20318 * passes.c (init_optimization_passes): Remove reload pass.
20319 * ira.c (do_reload): Merge into ...
20321 (rest_of_handle_reload): Remove.
20322 (pass_reload): Likewise.
20323 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20324 of reload for vzeroupper pass placement.
20326 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20328 PR tree-optimization/56918
20329 PR tree-optimization/56920
20330 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20331 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20332 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20333 use wide_mul_with_sign method.
20335 2013-04-12 Richard Biener <rguenther@suse.de>
20337 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20338 not be considered a gimple constant.
20340 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20342 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20343 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20346 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20347 Jakub Jelinek <jakub@redhat.com>
20349 * opts.c: Include diagnostic-color.h.
20350 (common_handle_option): Handle OPT_fdiagnostics_color_.
20351 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20352 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20353 (diagnostic-color.o): New.
20354 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20355 (diagnostic_color_rule): New enum.
20356 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20357 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20358 the location string.
20359 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20360 either NULL, or color kind.
20361 * diagnostic-color.c: New file.
20362 * diagnostic-color.h: New file.
20363 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20365 * doc/invoke.texi (-fdiagnostics-color): Document.
20366 * pretty-print.h (pp_show_color): Define.
20367 (struct pretty_print_info): Add show_color field.
20368 * diagnostic.c: Include diagnostic-color.h.
20369 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20370 macros. Colorize error:, warning: etc. strings and also the location
20372 (diagnostic_show_locus): Colorize the caret line.
20373 * pretty-print.c: Include diagnostic-color.h.
20374 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20375 inside of %< %> quotes or quoted through q format modifier.
20377 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20379 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20381 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20383 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20384 code in CC_NZ mode.
20385 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20388 2013-04-11 Marek Polacek <polacek@redhat.com>
20390 PR tree-optimization/48184
20391 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20393 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20395 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20396 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20397 (skip_simple_arithmetic): Tidy up.
20398 * tree.h (skip_simple_constant_arithmetic): Declare.
20400 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20402 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20404 2013-04-11 Richard Biener <rguenther@suse.de>
20406 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20407 generate vector constants.
20409 2013-04-11 Richard Biener <rguenther@suse.de>
20411 PR tree-optimization/56878
20412 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20413 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20415 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20416 Prefer to align the DR with the most invariant base address.
20418 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20420 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20423 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20425 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20426 floating-point vector comparisons against 0.
20428 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20430 PR tree-optimization/56899
20431 * fold-const.c (extract_muldiv_1): Apply distributive law
20432 only if TYPE_OVERFLOW_WRAPS (ctype).
20434 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20437 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20438 loading has side effect.
20440 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20442 * config/sparc/sparc.c: Include tree-pass.h.
20443 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20444 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20445 head of file. Change return type. Split off gate function.
20446 (sparc_gate_work_around_errata): New function.
20447 (pass_work_around_errata): New pass definition.
20448 (insert_pass_work_around_errata) New pass insert definition to
20449 insert pass_work_around_errata just after delayed-branch scheduling.
20450 (sparc_option_override): Insert the pass.
20451 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20453 2013-04-10 David S. Miller <davem@davemloft.net>
20455 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20456 or -mcpu=hypersparc.
20458 * target.def (cstore_mode): New hook.
20459 * target.h: Include insn-codes.h
20460 * targhooks.c: Likewise.
20461 (default_cstore_mode): New function.
20462 * targhooks.h: Declare it.
20463 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20464 * doc/tm.texi: Rebuild.
20465 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20466 target hook, rather than inspecting the insn_data.
20467 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20468 (TARGET_CSTORE_MODE): Redefine.
20469 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20471 * config/sparc/predicates.md (cstore_result_operand): New special
20473 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20474 Use it for operand 0.
20475 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20476 (*snesi_special): Likewise.
20477 (*snesi_zero): Likewise.
20478 (*seqsi_zero): Likewise.
20479 (*sltu_insn): Likewise.
20480 (*sgeu_insn): Likewise.
20481 (*seqdi_special): Make operand 0 and comparison operation be of
20483 (*snedi_special): Likewise.
20484 (*snedi_special_vis3): Likewise.
20485 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20486 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20487 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20488 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20489 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20490 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20491 (*sltu_extend_sp64): Likewise.
20492 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20493 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20494 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20495 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20496 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20498 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20500 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20501 (aarch64_start_file): Use the new function.
20503 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20504 Jason Merrill <jason@redhat.com>
20506 * common.opt: Add -gdwarf.
20507 * opts.c (common_handle_option): Handle it.
20508 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20510 2013-04-10 Richard Biener <rguenther@suse.de>
20512 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20513 (execute_one_ipa_transform_pass): But unconditionally here.
20514 (execute_one_pass): And here.
20515 (init_optimization_passes): Remove reload pass.
20516 * tree-pass.h (TODO_ggc_collect): Remove.
20517 (pass_reload): Likewise.
20518 * ira.c (do_reload): Merge into ...
20520 (rest_of_handle_reload): Remove.
20521 (pass_reload): Likewise.
20522 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20523 of reload for vzeroupper pass placement.
20524 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20525 and todo_flags_finish of all passes.
20527 2013-04-10 Richard Biener <rguenther@suse.de>
20529 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20530 first_const_oprnd field, rename first_def_type to first_op_type.
20531 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20532 (vect_get_and_check_slp_defs): Always use the type of the
20533 operand. Allow mixed vect_external_def, vect_constant_def types.
20534 (vect_get_constant_vectors): Handle mixed vect_external_def,
20535 vect_constant_def types.
20537 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20539 PR tree-optimization/55524
20540 * tree-ssa-math-opts.c
20541 (convert_mult_to_fma): Don't use an fms construct
20542 when we don't have an fms operation, but fnma, and it looks
20543 likely that we'll be able to use the latter.
20545 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20547 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20549 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20550 inline fail caused by overwritable functions.
20552 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20554 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20555 unnecessary bits in the constant power of two case.
20557 2013-04-10 Richard Biener <rguenther@suse.de>
20559 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20560 broken code swapping operands.
20561 (vect_build_slp_tree): Do not compute load permutations here.
20562 (vect_analyze_slp_instance): Compute load permutations here,
20563 after building the SLP tree.
20565 2013-04-09 Christian Bruel <christian.bruel@st.com>
20567 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20568 of next/prev_real_insn.
20570 2013-04-09 Jan Hubicka <jh@suse.cz>
20572 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20573 Drop aliased parameter.
20574 (function_and_variable_visibility): Do not handle alias pairs.
20575 * cgraph.c (varpool_externally_visible_p): Update prototype.
20576 * varpool.c (varpool_add_new_variable): Update.
20578 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20580 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20582 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20584 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20586 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20588 2013-04-09 Marek Polacek <polacek@redhat.com>
20590 PR tree-optimization/48762
20591 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20593 2013-04-09 Richard Biener <rguenther@suse.de>
20595 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20597 (vect_build_slp_tree): Likewise.
20598 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20599 calculating the cost of a SLP instance.
20600 (vect_analyze_slp_instance): Use it from here, after building
20603 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20605 PR middle-end/56883
20606 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20607 expand_omp_for_static_chunk): Use simple_p = true in
20608 force_gimple_operand_gsi calls when assigning to addressable decls.
20610 2013-04-09 Jeff Law <law@redhat.com>
20612 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20613 when the boolean was created by converting a wider object which
20614 had a boolean range.
20616 2013-04-09 Richard Biener <rguenther@suse.de>
20618 * tree-vectorizer.h (slp_void_p): Remove.
20619 (slp_tree): Typedef before _slp_tree declaration.
20620 (struct _slp_tree): Use a vector of slp_tree as children.
20621 (vect_get_place_in_interleaving_chain): Remove.
20622 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20624 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20626 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20627 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20628 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20629 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20630 Use slp_node instead of slp_void_p and adjust.
20632 2013-04-09 Richard Biener <rguenther@suse.de>
20634 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20635 work that is not necessary.
20637 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20639 PR tree-optimization/56854
20640 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20641 forward into clobber stmts if it would change MEM_REF lhs into
20644 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20646 * tree.c (type_hash_lookup, type_hash_add): Make static.
20647 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20649 2013-04-09 Richard Biener <rguenther@suse.de>
20651 * tree.h (unsave_expr_now): Remove.
20652 * tree-inline.c (mark_local_for_remap_r): Remove.
20653 (unsave_expr_1): Likewise.
20654 (unsave_r): Likewise.
20655 (unsave_expr_now): Likewise.
20656 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20657 (propagate_tree_value): Likewise.
20659 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20661 * doc/rtl.texi (sequence): Rewrite documentation to match the
20662 current use of SEQUENCE rtl objects.
20663 * rtl.def (SEQUENCE): Likewise.
20665 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20666 Update documentation.
20667 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20668 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20670 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20672 2013-04-08 Teresa Johnson <tejohnson@google.com>
20674 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20675 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20677 (estimate_edge_size_and_time): Add comment to suggest using rounding
20679 (estimate_node_size_and_time): Ditto.
20680 (remap_edge_change_prob): Use helper rounding divide methods.
20681 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20682 (gimple_mod_pow2_value_transform): Ditto.
20683 (gimple_mod_subtract_transform): Ditto.
20684 (gimple_ic_transform): Ditto.
20685 (gimple_stringops_transform): Ditto.
20686 * stmt.c (conditional_probability): Ditto.
20687 (emit_case_dispatch_table): Ditto.
20688 * lto-cgraph.c (merge_profile_summaries): Ditto.
20689 * tree-optimize.c (execute_fixup_cfg): Ditto.
20690 * cfgcleanup.c (try_forward_edges): Ditto.
20691 * cfgloopmanip.c (scale_loop_profile): Ditto.
20693 (duplicate_loop_to_header_edge): Ditto.
20694 (lv_adjust_loop_entry_edge): Ditto.
20695 * tree-vect-loop.c (vect_transform_loop): Ditto.
20696 * profile.c (compute_branch_probabilities): Ditto.
20697 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20698 * lto-streamer-in.c (input_cfg): Ditto.
20699 * gimple-streamer-in.c (input_bb): Ditto.
20700 * ipa-cp.c (update_profiling_info): Ditto.
20701 (update_specialized_profile): Ditto.
20702 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20703 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20705 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20706 (compute_trg_info): Ditto.
20707 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20708 (purge_dead_edges): Ditto.
20709 * loop-unswitch.c (unswitch_loop): Ditto.
20710 * cgraphclones.c (cgraph_clone_edge): Ditto.
20711 (cgraph_clone_node): Ditto.
20712 * tree-inline.c (copy_bb): Ditto.
20713 (copy_edges_for_bb): Ditto.
20714 (initialize_cfun): Ditto.
20715 (copy_cfg_body): Ditto.
20716 (expand_call_inline): Ditto.
20718 2013-04-08 Kai Tietz <ktietz@redhat.com>
20720 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20721 TARGET_CYGWIN64 by TARGET_64BIT.
20723 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20725 * config/epiphany/epiphany.md (GPR_1): New constant.
20726 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20727 * config/epiphany/epiphany.c (gen_compare_reg):
20728 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20729 is already in place.
20730 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20731 Don't require being called during rtl expansion; If y operlaps r0,
20733 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20734 (epiphany_expand_epilogue): Likewise.
20736 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20737 Don't use CC_FPmode for ORDERED / UNORDERED.
20738 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20740 * config/epiphany/constraints.md (CnL): New constraint.
20741 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20742 * config/epiphany/predicates.md (add_operand): Allow 1024.
20744 * config/epiphany/epiphany.md (logical_op): New code iterator.
20745 (op_mnc): New code attribute.
20746 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20747 (mov_f+1, mov_f+2): New peephole2 patterns.
20749 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20750 (cstoresi4): Also allow re-use of zero result when doing a NE
20751 comparison to a non-zero operand.
20752 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20754 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20755 Use gen_addsi3_i / gen_subsi3_i.
20757 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20761 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20762 contain anything but clobbers, at most one __builtin_stack_restore,
20763 optionally debug stmts and final resx, and if it has at least one
20764 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20765 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20766 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20767 which isn't defaut definition, remove them.
20768 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20769 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20770 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20771 with MEM_REF LHS with SSA_NAME address.
20773 2013-04-08 Jeff Law <law@redhat.com>
20775 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20777 2013-04-08 Richard Biener <rguenther@suse.de>
20779 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20781 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20782 determined vector type.
20783 (vect_analyze_data_refs): Likewise.
20784 (vect_get_new_vect_var): Adjust.
20785 (vect_create_destination_var): Preserve SSA name versions.
20786 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20787 not dump anything here.
20789 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20791 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20792 Add member lr_slot_known.
20793 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20795 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20796 Remove code that sets lr_slot_offset according to what a previous
20797 version of epiphany_emit_save_restore used to do.
20798 (epiphany_emit_save_restore): When doing an lr save or restore,
20799 set/verify lr_slot_known and lr_slot_offset.
20801 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20804 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20807 2013-04-08 Richard Biener <rguenther@suse.de>
20809 * alias.c (find_base_term): Fix thinko in previous change.
20811 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20813 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20814 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20815 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20816 if possible to compute val.
20817 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20818 For QImode integers don't require anything about precision. Use
20819 const_with_all_bytes_same to find out if the constant doesn't have
20820 repeated bytes in it.
20822 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20824 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20827 2013-04-08 Marek Polacek <polacek@redhat.com>
20829 PR rtl-optimization/48182
20830 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20833 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20836 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20837 nuses, make sure we have a label.
20839 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20842 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20843 (rs6000_emit_swdiv_low_precision): Remove.
20844 (rs6000_emit_swdiv): Rewrite to handle between one and four
20845 iterations of Newton-Raphson generally; modify required number of
20846 iterations for some cases.
20847 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20849 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20851 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20852 set-but-unused variable.
20854 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20855 basic blocks of released function bodies garbage-collectable.
20857 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20858 (struct rtl_opt_pass): Add TODO_df_finish.
20860 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20862 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20864 * config/arm/constraints.md (q): New constraint.
20865 * config/arm/ldrdstrd.md: New file.
20866 * config/arm/arm.md (ldrdstrd.md) New include.
20867 (arm_movdi): Use "q" instead of "r" constraint
20868 for double-word memory access.
20869 (movdf_soft_insn): Likewise.
20870 * config/arm/vfp.md (movdi_vfp): Likewise.
20871 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20872 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20873 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20874 (mem_ok_for_ldrd_strd): Likewise.
20875 (output_move_double): Update assertion.
20877 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20879 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20881 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20883 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20884 define_insn_and_split.
20885 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20887 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20889 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20890 define_insn_and_split.
20891 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20892 (shiftsi3_compare): New pattern.
20893 (rrx): New pattern.
20894 * config/arm/unspecs.md (UNSPEC_RRX): New.
20896 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20898 * config/arm/arm.md (negdi_extendsidi): New pattern.
20899 (negdi_zero_extendsidi): Likewise.
20901 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20903 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20904 define_insn_and_split.
20905 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20906 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20908 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20910 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20911 define_insn_and_split.
20912 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20913 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20915 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20917 * config/arm/arm.md (subsi3_carryin): New pattern.
20918 (subsi3_carryin_const): Likewise.
20919 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20920 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20922 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20924 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20926 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20928 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20929 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20931 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20933 * config/arm/arm.c (arm_expand_builtin): Change fcode
20934 type to unsigned int.
20936 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20938 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20940 2013-04-04 Ian Lance Taylor <iant@google.com>
20942 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20945 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20947 PR middle-end/56729
20948 * df-scan.c (df_insn_delete): Disable failing assert.
20950 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20952 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20953 New function prototype.
20954 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20955 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20956 (arm_builtin_vectorized_function): New function.
20958 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20960 * config/arm/arm_neon_builtins.def: New file.
20961 * config/arm/arm.c (neon_builtin_data): Move contents to
20962 arm_neon_builtins.def.
20963 (enum arm_builtins): Include neon builtin definitions.
20964 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20965 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20967 2013-04-04 Marek Polacek <polacek@redhat.com>
20969 PR tree-optimization/48186
20970 * predict.c (maybe_hot_frequency_p): Return false if
20971 HOT_BB_FREQUENCY_FRACTION is 0.
20972 (cgraph_maybe_hot_edge_p): Likewise.
20974 2013-04-04 Richard Biener <rguenther@suse.de>
20976 PR tree-optimization/56826
20977 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20980 2013-04-04 Richard Biener <rguenther@suse.de>
20982 PR tree-optimization/56213
20983 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20984 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20986 2013-04-04 Richard Biener <rguenther@suse.de>
20988 PR tree-optimization/56837
20989 * tree-loop-distribution.c (classify_partition): For non-zero
20990 values require that the value has the same precision as its
20991 mode to be useful as memset value.
20993 2013-04-03 Nick Clifton <nickc@redhat.com>
20995 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20996 (fmssf4): Use fmsf.s on E3V5 architectures.
20997 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20998 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21000 2013-04-03 Jeff Law <law@redhat.com>
21002 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21003 (lra-eliminations.o): Likewise.
21005 2013-04-03 Teresa Johnson <tejohnson@google.com>
21007 * gcov-io.c (compute_working_sets): Moved most of body of old
21008 compute_working_sets here from profile.c.
21009 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21010 (gcov_working_set_t): Moved typedef here from basic-block.h
21011 (compute_working_set): Declare.
21012 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21013 (get_working_sets): Renamed from compute_working_set,
21014 replace most of body with call to new compute_working_sets.
21015 (get_exec_counts): Replace call to compute_working_sets
21016 to get_working_sets.
21017 * profile.h (get_working_sets): Renamed from compute_working_set.
21018 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21019 to get_working_sets.
21020 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21021 * gcov-dump.c (dump_working_sets): New function.
21023 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21025 * hwint.c (sext_hwi, zext_hwi): New functions.
21026 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21027 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21028 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21029 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21030 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21031 (sext_hwi, zext_hwi): New functions.
21033 2013-04-03 Jeff Law <law@redhat.com>
21035 PR tree-optimization/56799
21036 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21037 back test for widening conversion erroneously dropped in prior change.
21039 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21042 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21043 instead of next_real_insn.
21045 2013-04-03 Marek Polacek <polacek@redhat.com>
21048 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21050 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21053 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21055 (thumb1_output_casesi): Likewise.
21056 (thumb2_output_casesi): Likewise.
21058 2013-04-03 Richard Biener <rguenther@suse.de>
21060 PR tree-optimization/56817
21061 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21063 (tree_unroll_loops_completely_1): ... new function to manually
21064 walk the loop tree, properly defering outer loops of unrolled
21065 loops to later iterations.
21067 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21069 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21070 (vectorizable_load): Likewise.
21071 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21072 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21074 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21076 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21079 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21081 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21083 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21085 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21087 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21089 PR tree-optimization/56790
21090 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21093 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21095 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21097 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21098 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21101 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21104 * tree.h (force_folding_builtin_constant_p): New decl.
21105 * builtins.c (force_folding_builtin_constant_p): New variable.
21106 (fold_builtin_constant_p): Fold immediately also if
21107 force_folding_builtin_constant_p.
21109 2013-04-03 Richard Biener <rguenther@suse.de>
21111 PR tree-optimization/56812
21112 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21113 DRs of the same interleaving chain are independent.
21115 2013-04-02 Jason Merrill <jason@redhat.com>
21117 * gdbinit.in (pbb): Use debug fn.
21119 2013-04-02 Lawrence Crowl <crowl@google.com>
21121 * sese.h (struct ivtype_map_elt_s): Remove unused.
21122 (extern debug_ivtype_map): Remove unused.
21123 (extern eq_ivtype_map_elts): Remove unused.
21124 * sese.c (debug_ivtype_map): Removed unused.
21125 (debug_ivtype_map_1): Removed unused.
21126 (debug_ivtype_elt): Remove unused.
21127 (eq_ivtype_map_elts): Remove unused.
21130 2013-04-02 Kai Tietz <ktietz@redhat.com>
21133 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21134 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21135 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21137 (legitimize_pe_coff_symbol): Likewise.
21138 (is_imported_p): New helper-function.
21139 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21140 for Windows x64 targets.
21141 (ix86_expand_prologue): Optimize for pe-coff targets.
21142 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21143 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21144 medium/large code-model.
21145 (legitimize_pic_address): Likewise.
21146 (legitimize_tls_address): Likewise.
21147 (ix86_expand_call): Likewise.
21148 (x86_output_mi_thunk): Likewise.
21149 (get_dllimport_decl): Add new beimport argument.
21150 (construct_plt_address): Don't assert for x64 pe-coff targets.
21151 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21153 (SYMBOL_FLAG_STUBVAR): New macro.
21154 (SYMBOL_REF_STUBVAR_P): Likewise.
21155 * config/i386/winnt.c (stub_list): New structure.
21156 (stub_head): New local variable.
21157 (i386_pe_record_stub): New function.
21158 (i386_pe_file_end): Emit refptr-stubs.
21160 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21162 PR rtl-optimization/56745
21163 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21164 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21167 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21168 and both of them are MEM_REFs, just compare first argument for
21169 equality and attempt to deal even with differing offsets.
21172 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21173 of gimple_clobber_p to be MEM_REF.
21174 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21175 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21176 after gimplification.
21177 * asan.c (get_mem_ref_of_assignment): Don't instrument
21178 gimple_clobber_p stmts.
21179 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21180 gimple_clobber_p stmt if they have MEM_REF lhs and
21181 are dead because of another gimple_clobber_p stmt.
21182 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21183 gimple_clobber_p stmts like debug stmts.
21184 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21185 that refer to unused VAR_DECLs or uninitialized values.
21186 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21187 gimple_clobber_p stmts if they refer to removed parameters.
21188 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21191 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21193 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21194 using SWI48 mode attribute.
21196 2013-04-02 Wei Mi <wmi@google.com>
21198 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21199 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21200 *<rotate_insn><mode>3_mask in i386.md.
21202 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21204 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21206 2013-04-02 Richard Biener <rguenther@suse.de>
21208 PR tree-optimization/56778
21209 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21210 Runtime alias tests are not supported for gather loads.
21211 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21212 stmts referenced from SSA operands before updating SSA form.
21214 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21215 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21217 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21218 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21219 * config/arm/cortex-a53.md: New file.
21220 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21221 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21222 * config/arm/arm.c (arm_issue_rate): Likewise.
21223 * config/arm/arm-tune.md: Regenerate
21224 * config/arm/arm-tables.opt: Regenerate.
21225 * config/arm/arm-cores.def: Add cortex-a53.
21227 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21229 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21232 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21234 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21235 scalar load/store operations using B/H registers.
21236 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21238 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21240 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21242 * config/aarch64/aarch64.c
21243 (aarch64_simd_scalar_immediate_valid_for_move): New.
21244 * config/aarch64/aarch64-protos.h
21245 (aarch64_simd_scalar_immediate_valid_for_move): New.
21246 * config/aarch64/constraints.md (Dh, Dq): New.
21247 * config/aarch64/iterators.md (hq): New.
21249 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21251 * reorg.c (get_branch_condition): Deal with conditional returns.
21252 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21254 2013-04-01 Wei Mi <wmi@google.com>
21256 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21257 Truncate operand 2 using %b asm operand modifier.
21258 (*<shift_insn><mode>3_mask): Ditto.
21259 (*<rotate_insn><mode>3_mask): Ditto.
21261 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21263 PR middle-end/56798
21264 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21266 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21268 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21270 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21272 2013-03-30 Lawrence Crowl <crowl@google.com>
21274 * dse.c (clear_alias_sets): Remove never set.
21275 (disqualified_clear_alias_sets): Remove never set.
21276 (clear_alias_mode_pool): Remove never set.
21277 (dse_step0): Remove condition that is never true.
21278 (canon_address): Remove condition that is never true.
21279 (dse_step7): Remove condition that is never true.
21280 (rest_of_handle_dse): Remove condition that is never true.
21281 (rest_of_handle_dse::did_global): Remove never read from above.
21282 (dse_step2_spill): Remove never called from above.
21283 (dse_step5_spill): Remove never called from above.
21285 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21287 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21288 JUMP_TABLE_DATA changes.
21289 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21290 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21291 (Insns) <jump_table_data>: New entry.
21292 * doc/tm.texi: Regenerate.
21294 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21296 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21297 for table jump at the end of a basic block using tablejump_p.
21298 * targhooks.c (default_invalid_within_doloop): Likewise.
21299 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21300 target hook implementation that is identical to the default hook.
21301 (rs6000_invalid_within_doloop): Remove.
21303 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21304 unused variable from tablejump_p call.
21306 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21307 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21308 (INSN_DELETED_P): Likewise.
21309 (emit_jump_table_data): New prototype.
21310 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21311 after 4th as unused.
21312 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21313 * sched-vis.c (print_insn): Likewise.
21314 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21315 insn for compatibility with back ends that use next_active_insn to
21316 identify jump table data.
21317 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21318 (remove_insn): Likewise.
21319 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21321 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21322 (emit_jump_table_data): New function.
21324 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21325 basic block, a JUMP_TABLE_DATA never is.
21326 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21327 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21328 off from code handling real insns.
21329 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21330 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21331 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21332 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21333 is not a NONDEBUG_INSN_P.
21334 * ira-costs.c (scan_one_insn): Likewise.
21335 * jump.c (mark_all_labels): Likewise.
21336 (mark_jump_label_1): Likewise.
21337 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21338 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21340 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21341 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21342 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21343 JUMP_TABLE_DATA_P insns.
21344 (calculate_elim_costs_all_insns): Likewise.
21345 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21346 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21347 (delete_output_reload): Code style fixups.
21348 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21349 insn flags on this non-insn.
21350 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21351 as scheduling barriers, for pre-change compatibility.
21352 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21353 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21355 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21356 redundant JUMP_TABLE_DATA_P test.
21357 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21358 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21359 (frv_for_each_packet): Likewise.
21360 * config/i386/i386.c (min_insn_size): Likewise.
21361 (ix86_avoid_jump_mispredicts): Likewise.
21362 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21363 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21364 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21365 (mips16_insn_length): Robustify.
21366 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21367 (mips16_split_long_branches): Likewise.
21368 * config/pa/pa.c (pa_combine_instructions): Likewise.
21369 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21370 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21371 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21372 as contributing to pool range lengths.
21373 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21374 Remove redundant JUMP_TABLE_DATA_P test.
21375 (sh_loop_align): Likewise.
21376 (split_branches): Likewise.
21377 (sh_insn_length_adjustment): Likewise.
21378 * config/spu/spu.c (get_branch_target): Likewise.
21380 2013-03-29 Jan Hubicka <jh@suse.cz>
21382 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21383 gcov streaming; stream hot bb threshold to ltrans.
21384 * predict.c (get_hot_bb_threshold): Break out from ....
21385 (maybe_hot_count_p): ... here.
21386 (set_hot_bb_threshold): New function.
21387 * lto-section-in.c (lto_section_name): Add profile.
21388 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21389 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21390 and data-streamer.h
21391 (histogram_entry): New structure.
21392 (histogram, histogram_pool): New global vars.
21393 (histogram_hash): New structure.
21394 (histogram_hash::hash): New method.
21395 (histogram_hash::equal): Likewise.
21396 (account_time_size): New function.
21397 (cmp_counts): New function.
21398 (dump_histogram): New function.
21399 (ipa_profile_generate_summary): New function.
21400 (ipa_profile_write_summary): New function.
21401 (ipa_profile_read_summary): New function.
21402 (ipa_profile): Decide on threshold.
21403 (pass_ipa_profile): Add ipa_profile_write_summary and
21404 ipa_profile_read_summary.
21405 * Makefile.in (ipa.o): Update dependencies.
21406 * lto-streamer.h (LTO_section_ipa_profile): New section.
21408 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21410 * tree.h (VAR_P): New.
21412 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21415 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21417 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21419 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21420 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21421 (control_flow_insn_p): Likewise.
21422 * cfgrtl.c (duplicate_insn_chain): Likewise.
21423 * final.c (get_attr_length_1): Likewise.
21424 (shorten_branches): Likewise.
21425 (final_scan_insn): Likewise.
21426 * function.c (instantiate_virtual_regs): Likewise.
21427 * gcse.c (insert_insn_end_basic_block): Likewise.
21428 * ira-costs.c (scan_one_insn): Likewise.
21429 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21430 * lra.c (check_rtl): Likewise.
21431 * reload1.c (elimination_costs_in_insn): Likewise.
21432 * reorg.c (follow_jumps): Likewise.
21434 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21435 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21436 (thumb_far_jump_used_p): Likewise.
21437 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21438 (workaround_speculation): Likewise.
21439 (add_sched_insns_for_speculation): Likewise.
21440 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21441 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21442 (frv_for_each_packet): Likewise.
21443 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21444 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21445 (final_emit_insn_group_barriers): Likewise.
21446 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21447 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21448 (mips16_insn_length): Likewise.
21449 * config/pa/pa.c (pa_reorg): Likewise.
21450 (pa_combine_instructions): Likewise.
21451 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21452 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21453 (sh_reorg): Likewise.
21454 (split_branches): Likewise.
21455 * config/spu/spu.c (get_branch_target): Likewise.
21457 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21460 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21462 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21463 Fix declaration name.
21465 2013-03-28 Lawrence Crowl <crowl@google.com>
21467 * graphds.h (struct graph.indicies): Remove unused.
21468 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21469 (SCOP_ORIGINAL_PDDRS): Remove unused.
21470 * sese.h (extern insert_loop_close_phis): Removed unused.
21471 (extern insert_guard_phis): Removed unused.
21472 (extern ivtype_map_elt_info): Removed unused.
21473 (new_ivtype_map_elt): Removed unused.
21474 * sese.c (ivtype_map_elt_info): Removed unused.
21476 2013-03-28 Lawrence Crowl <crowl@google.com>
21478 * Makefile.in: Add several missing include dependences.
21480 (test-dump.o): New. This object is not added to any executable,
21481 but is present for ad-hoc testing.
21483 (debug (const bitmap_head_def &)): New.
21484 (debug (const bitmap_head_def *)): New.
21486 (extern debug (const bitmap_head_def &)): New.
21487 (extern debug (const bitmap_head_def *)): New.
21489 (debug (edge_def &)): New.
21490 (debug (edge_def *)): New.
21492 (debug (basic_block_def &)): New.
21493 (debug (basic_block_def *)): New.
21495 (dump_node (const_tree, int, FILE *)): Correct source file.
21497 (debug (die_struct &)): New.
21498 (debug (die_struct *)): New.
21500 (extern debug (die_struct &)): New.
21501 (extern debug (die_struct *)): New.
21502 * gimple-pretty-print.c
21503 (debug (gimple_statement_d &)): New.
21504 (debug (gimple_statement_d *)): New.
21505 * gimple-pretty-print.h
21506 (extern debug (gimple_statement_d &)): New.
21507 (extern debug (gimple_statement_d *)): New.
21509 (debug (ira_allocno_copy &)): New.
21510 (debug (ira_allocno_copy *)): New.
21511 (debug (ira_allocno &)): New.
21512 (debug (ira_allocno *)): New.
21514 (extern debug (ira_allocno_copy &)): New.
21515 (extern debug (ira_allocno_copy *)): New.
21516 (extern debug (ira_allocno &)): New.
21517 (extern debug (ira_allocno *)): New.
21519 (debug (live_range &)): New.
21520 (debug (live_range *)): New.
21522 (debug (lra_live_range &)): New.
21523 (debug (lra_live_range *)): New.
21525 (debug (lra_live_range &)): New.
21526 (debug (lra_live_range *)): New.
21528 (debug (omega_pb_d &)): New.
21529 (debug (omega_pb_d *)): New.
21531 (extern debug (omega_pb_d &)): New.
21532 (extern debug (omega_pb_d *)): New.
21534 (debug (const rtx_def &)): New.
21535 (debug (const rtx_def *)): New.
21537 (debug_tree (tree): Move within file.
21538 (debug_raw (const tree_node &)): New.
21539 (debug_raw (const tree_node *)): New.
21540 (dump_tree_via_hooks (const tree_node *, int)): New.
21541 (debug (const tree_node &)): New.
21542 (debug (const tree_node *)): New.
21543 (debug_verbose (const tree_node &)): New.
21544 (debug_verbose (const tree_node *)): New.
21545 (debug_head (const tree_node &)): New.
21546 (debug_head (const tree_node *)): New.
21547 (debug_body (const tree_node &)): New.
21548 (debug_body (const tree_node *)): New.
21549 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21550 (debug (vec<tree, va_gc> &)): New.
21551 (debug (vec<tree, va_gc> *)): New.
21553 (extern debug (const rtx_def &)): New.
21554 (extern debug (const rtx_def *)): New.
21556 (debug_raw (simple_bitmap_def &)): New.
21557 (debug_raw (simple_bitmap_def *)): New.
21558 (debug (simple_bitmap_def &)): New.
21559 (debug (simple_bitmap_def *)): New.
21561 (extern debug (simple_bitmap_def &)): New.
21562 (extern debug (simple_bitmap_def *)): New.
21563 (extern debug_raw (simple_bitmap_def &)): New.
21564 (extern debug_raw (simple_bitmap_def *)): New.
21566 (debug (vinsn_def &)): New.
21567 (debug (vinsn_def *)): New.
21568 (debug_verbose (vinsn_def &)): New.
21569 (debug_verbose (vinsn_def *)): New.
21570 (debug (expr_def &)): New.
21571 (debug (expr_def *)): New.
21572 (debug_verbose (expr_def &)): New.
21573 (debug_verbose (expr_def *)): New.
21574 (debug (vec<rtx> &)): New.
21575 (debug (vec<rtx> *)): New.
21577 (extern debug (vinsn_def &)): New.
21578 (extern debug (vinsn_def *)): New.
21579 (extern debug_verbose (vinsn_def &)): New.
21580 (extern debug_verbose (vinsn_def *)): New.
21581 (extern debug (expr_def &)): New.
21582 (extern debug (expr_def *)): New.
21583 (extern debug_verbose (expr_def &)): New.
21584 (extern debug_verbose (expr_def *)): New.
21585 (extern debug (vec<rtx> &)): New.
21586 (extern debug (vec<rtx> *)): New.
21588 (_list_iter_cond_expr): Make inline instead of static.
21590 (debug (sreal &)): New.
21591 (debug (sreal *)): New.
21593 (extern debug (sreal &)): New.
21594 (extern debug (sreal *)): New.
21596 (extern debug_raw (const tree_node &)): New.
21597 (extern debug_raw (const tree_node *)): New.
21598 (extern debug (const tree_node &)): New.
21599 (extern debug (const tree_node *)): New.
21600 (extern debug_verbose (const tree_node &)): New.
21601 (extern debug_verbose (const tree_node *)): New.
21602 (extern debug_head (const tree_node &)): New.
21603 (extern debug_head (const tree_node *)): New.
21604 (extern debug_body (const tree_node &)): New.
21605 (extern debug_body (const tree_node *)): New.
21606 (extern debug (vec<tree, va_gc> &)): New.
21607 (extern debug (vec<tree, va_gc> *)): New.
21609 (debug (struct loop &)): New.
21610 (debug (struct loop *)): New.
21611 (debug_verbose (struct loop &)): New.
21612 (debug_verbose (struct loop *)): New.
21613 * tree-dump.c: Add header dependence.
21615 (extern debug (struct loop &)): New.
21616 (extern debug (struct loop *)): New.
21617 (extern debug_verbose (struct loop &)): New.
21618 (extern debug_verbose (struct loop *)): New.
21620 (debug (data_reference &)): New.
21621 (debug (data_reference *)): New.
21622 (debug (vec<data_reference_p> &)): New.
21623 (debug (vec<data_reference_p> *)): New.
21624 (debug (vec<ddr_p> &)): New.
21625 (debug (vec<ddr_p> *)): New.
21627 (extern debug (data_reference &)): New.
21628 (extern debug (data_reference *)): New.
21629 (extern debug (vec<data_reference_p> &)): New.
21630 (extern debug (vec<data_reference_p> *)): New.
21631 (extern debug (vec<ddr_p> &)): New.
21632 (extern debug (vec<ddr_p> *)): New.
21634 (debug (pt_solution &)): New.
21635 (debug (pt_solution *)): New.
21637 (extern debug (pt_solution &)): New.
21638 (extern debug (pt_solution *)): New.
21640 (debug (_var_map &)): New.
21641 (debug (_var_map *)): New.
21642 (debug (tree_live_info_d &)): New.
21643 (debug (tree_live_info_d *)): New.
21645 (extern debug (_var_map &)): New.
21646 (extern debug (_var_map *)): New.
21647 (extern debug (tree_live_info_d &)): New.
21648 (extern debug (tree_live_info_d *)): New.
21650 2013-03-28 Jan Hubicka <jh@suse.cz>
21652 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21654 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21656 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21657 record only when desired or required.
21659 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21661 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21662 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21664 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21666 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21667 (*andsi3_compare0_uxtw): New pattern.
21668 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21669 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21671 2013-03-28 Jan Hubicka <jh@suse.cz>
21673 * data-streamer-in.c (streamer_read_gcov_count): New function.
21674 * gimple-streamer-out.c: Include value-prof.h.
21675 (output_gimple_stmt): Output histogram.
21676 (output_bb): Use streamer_write_gcov_count.
21677 * value-prof.c: Include data-streamer.h
21678 (dump_histogram_value): Add HIST_TYPE_MAX.
21679 (stream_out_histogram_value): New function.
21680 (stream_in_histogram_value): New function.
21681 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21682 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21683 * data-streamer-out.c (streamer_write_gcov_count): New function.
21684 (streamer_write_gcov_count_stream): New function.
21685 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21686 (lto_output_node): Likewise.
21687 (input_node, input_edge): Likewise.
21688 * lto-streamer-out.c (output_cfg): Update streaming.
21689 * lto-streamer-in.c (input_cfg): Likewise.
21690 * data-streamer.h (streamer_write_gcov_count,
21691 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21692 * gimple-streamer-in.c: Include value-prof.h
21693 (input_gimple_stmt): Input histograms.
21694 (input_bb): Update profile streaming.
21696 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21698 * genmodes.c (emit_max_int): New function.
21699 (emit_insn_modes_h): Added call to emit_max_function.
21700 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21702 * machmode.def: Fixed comment.
21704 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21706 * combine.c (try_combine): Removed useless assert.
21707 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21709 2013-03-28 Marek Polacek <polacek@redhat.com>
21710 Richard Biener <rguenther@suse.de>
21712 PR tree-optimization/56695
21713 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21714 build signed result of a vector comparison.
21715 * tree-cfg.c (verify_gimple_comparison): Check that a result
21716 of a vector comparison has signed type.
21718 2013-03-28 Richard Biener <rguenther@suse.de>
21720 PR tree-optimization/37021
21721 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21722 do not restrict gaps between groups.
21723 * tree-vect-stmts.c (vectorizable_load): Properly account for
21724 a gap between groups.
21726 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21728 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21729 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21732 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21734 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21735 * doc/extend.texi (Named Address Spaces): Ditto.
21736 (Variable Attributes): Ditto.
21738 2013-03-27 Kai Tietz <ktietz@redhat.com>
21740 * config.build: Add support for cygwin x64 target.
21741 * config.gcc: Likewise.
21742 * config.host: Likewise.
21743 * configure.ac: Likewise
21744 * configure: Regenerated.
21746 2013-03-27 Kai Tietz <ktietz@redhat.com>
21748 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21749 * config/i386/t-cygwin-w64: New file.
21750 * config/i386/cygwin-w64.h: New file.
21751 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21752 and add support for x64-cygwin target.
21753 (CPP_SPEC): Likewise.
21754 (CXX_WRAP_SPEC_LIST): Undefine before define.
21755 (LIBGCJ_SONAME): Use 15 as version.
21757 2013-03-27 Richard Biener <rguenther@suse.de>
21759 PR tree-optimization/56716
21760 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21761 dumping for ref nodes.
21763 2013-03-27 Martin Jambor <mjambor@suse.cz>
21765 PR tree-optimization/55334
21766 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21767 restricted pointers to arrays.
21769 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21771 * Makefile.in (.SUFFIXES): Add .cc.
21772 (.c.o): Apply same recipe for implicit rule .cc.o.
21774 2013-03-27 Richard Biener <rguenther@suse.de>
21776 PR tree-optimization/37021
21777 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21778 REALPART/IMAGPART_EXPRs around the supported refs.
21779 * tree-ssa-structalias.c (find_func_aliases): Assume that
21780 floating-point values are not used to transfer pointers.
21782 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21784 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21785 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21786 * doc/tm.texi: Regenerate.
21787 * targhooks.h (default_has_ifunc_p): New.
21788 * targhooks.c (default_has_ifunc_p): Ditto.
21789 * config/linux-protos.h: New file.
21790 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21791 hook for linux which disables support of indirect functions in android.
21792 * config/linux-android.c: New file.
21793 * config/t-linux-android.c: Ditto.
21794 * config.gcc: Added new object file linux-android.o.
21795 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21796 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21797 * varasm.c (do_assemble_alias): Likewise.
21798 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21799 doesn't support indirect functions.
21800 * configure: Regenerate.
21802 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21805 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21806 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21809 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21811 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21813 2013-03-27 Terry Guo <terry.guo@arm.com>
21815 * config/arm/arm-cores.def: Added core cortex-r7.
21816 * config/arm/arm-tune.md: Regenerated.
21817 * config/arm/arm-tables.opt: Regenerated.
21818 * doc/invoke.texi: Added entry for core cortex-r7.
21820 2013-03-27 Walter Lee <walt@tilera.com>
21822 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21823 double-decrement of next_scratch_regno.
21825 2013-03-27 Walter Lee <walt@tilera.com>
21827 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21829 (insn_v1mulus): Ditto.
21830 (insn_v2muls): Ditto.
21832 2013-03-27 Walter Lee <walt@tilera.com>
21834 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21835 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21837 2013-03-27 Walter Lee <walt@tilera.com>
21839 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21840 (*sibcall_value): Ditto.
21842 2013-03-27 Walter Lee <walt@tilera.com>
21844 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21845 (insn_mnz_v8qi): ... this ...
21846 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21848 (insn_v<n>mnz): Replaced by ...
21849 (insn_v1mnz): ... this ...
21850 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21852 (insn_mz_<mode>): Replaced by ...
21853 (insn_mz_v8qi): ... this ...
21854 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21856 (insn_v<n>mz): Replaced by ...
21857 (insn_v1mz): ... this ...
21858 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21861 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21863 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21865 2013-03-26 Roland McGrath <mcgrathr@google.com>
21867 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21868 than fprintf with a non-constant, non-format string.
21870 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21872 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21873 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21874 operand 0 predicate.
21875 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21876 attribute. Use general_x64nomem_operand as operand 1 predicate.
21877 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21878 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21879 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21880 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21881 (mov<mode>_insv_1): Remove expander. Merge insn with
21882 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21883 Use general_x64nomem_operand as operand 1 predicate.
21884 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21885 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21886 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21887 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21888 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21889 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21890 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21891 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21892 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21893 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21894 (general_x64nomem_operand): Ditto.
21896 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21898 * config/rtems.opt: Add -pthread option.
21900 2013-03-26 Richard Biener <rguenther@suse.de>
21902 * alias.c (find_base_term): Avoid redundant and not used recursion.
21903 (base_alias_check): Get the initial base term from the caller.
21904 (true_dependence_1): Compute and pass base terms to base_alias_check.
21905 (write_dependence_p): Likewise.
21906 (may_alias_p): Likewise.
21908 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21910 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21911 PC-relative load in SI modes and above only.
21913 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21915 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21916 * config/arm/iwmmxt.md (WCGR0): Update.
21917 (WCGR1, WCGR2, WCGR3): Likewise.
21919 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21921 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21922 Use x64 and nox64 isa attributes.
21924 2013-03-26 Richard Biener <rguenther@suse.de>
21926 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21927 alignment computations and rely on get_object_alignment_1
21928 for the !TYPE_P case.
21929 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21931 2013-03-26 Walter Lee <walt@tilera.com>
21933 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21934 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21936 2013-03-25 Jeff Law <law@redhat.com>
21938 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21939 check for INTEGRAL_TYPE_P that was missing due to checking in
21940 wrong version of prior patch.
21942 2013-03-25 Walter Lee <walt@tilera.com>
21944 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21945 TILEGX_INSN_SHUFFLEBYTES1.
21946 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21948 (tilegx_builtins): Ditto.
21949 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21951 2013-03-25 Walter Lee <walt@tilera.com>
21953 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21954 (floatunssisf2): New pattern.
21955 (floatsidf2): New pattern.
21956 (floatunssidf2): New pattern.
21958 2013-03-25 Walter Lee <walt@tilera.com>
21960 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21961 tests for constraint J, K, N, P.
21963 2013-03-25 Walter Lee <walt@tilera.com>
21965 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21966 Use indirect/pcrel encoding.
21967 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21970 2013-03-25 Steve Ellcey <sellcey@mips.com>
21972 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21973 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21974 * config/mips/mips.c (mips_option_override): Set IMADD default.
21975 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21976 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21977 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21978 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21979 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21981 2013-03-25 Jeff Law <law@redhat.com>
21983 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21984 slightly to avoid creating and folding useless trees. Simplify
21985 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21987 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21989 * config/i386/i386.md (*zero_extendsidi2): Merge with
21990 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21991 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21992 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21993 targets. Clarify comment.
21995 2013-03-25 Martin Jambor <mjambor@suse.cz>
21997 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21998 pass-through jump functions differently.
21999 (ipa_read_jump_function): Likewise. Also use setter functions to set
22002 2013-03-25 Martin Jambor <mjambor@suse.cz>
22004 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22005 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22007 (ipa_get_indirect_edge_target): New function.
22008 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22009 ipa_get_indirect_edge_target. Update all callers.
22010 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22011 ipa_get_indirect_edge_target_1 instead of calling
22012 ipa_get_indirect_edge_target.
22013 (create_specialized_node): Pass aggvlas to
22014 ipcp_discover_new_direct_edges.
22016 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22018 * config/arm/arm.md (f_sels, f_seld): New types.
22019 (*cmov<mode>): New pattern.
22020 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22023 2013-03-25 Kai Tietz <ktietz@redhat.com>
22025 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22026 POSIX-printf for mingw-hosted builds.
22028 2013-03-25 Richard Biener <rguenther@suse.de>
22030 PR middle-end/56694
22031 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22032 must-not-throw stmt location.
22034 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22036 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22037 Emit load-acquire versions when acq is true.
22038 (arm_emit_store_exclusive): Add rel parameter.
22039 Emit store-release versions when rel is true.
22040 (arm_split_compare_and_swap): Use acquire-release instructions
22042 of barriers when appropriate.
22043 (arm_split_atomic_op): Likewise.
22044 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22045 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22046 (VUNSPEC_SLX): Likewise.
22047 (VUNSPEC_LDA): Likewise.
22048 (VUNSPEC_STL): Likewise.
22049 * config/arm/sync.md (atomic_load<mode>): New pattern.
22050 (atomic_store<mode>): Likewise.
22051 (arm_load_acquire_exclusive<mode>): Likewise.
22052 (arm_load_acquire_exclusivesi): Likewise.
22053 (arm_load_acquire_exclusivedi): Likewise.
22054 (arm_store_release_exclusive<mode>): Likewise.
22056 2013-03-25 Catherine Moore <clm@codesourcery.com>
22058 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22059 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22060 * config/mip/predicates.md (lwsp_swsp_operand,
22061 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22062 sb16_operand, db4_operand, db7_operand, ib3_operand,
22063 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22064 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22065 andi16_operand): New predicates.
22066 * config/mips/mips.md (compression): New attribute.
22067 (enabled): New attribute.
22068 (length): Consider compression in computing length.
22069 (shift_compression): New code attribute.
22070 (*add<mode>3): New operands. Record compression.
22071 (sub<mode>3): Likewise.
22072 (one_cmpl<mode>2): Likewise.
22073 (*and<mode>3): Likewise.
22074 (*ior<mode>3): Likewise.
22075 (unnamed pattern for xor): Likewise.
22076 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22077 (*<optab><mode>3): Likewise.
22078 (*mov<mode>_internal: Likewise.
22079 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22080 (mips_unsigned_immediate_p): New.
22081 (umips_lwsp_swsp_address_p): New.
22082 (m16_based_address_p): New.
22083 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22084 (mips_unsigned_immediate_p): New prototype.
22085 (lwsp_swsp_address_p): New prototype.
22086 (m16_based_address_p): New prototype.
22087 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22088 (mips_signed_immediate_p): New function.
22089 (m16_based_address_p): New function.
22090 (lwsp_swsp_address_p): New function.
22091 (mips_print_operand_punctuation): Recognize short delay slot insns
22092 for microMIPS.add<mode>3"
22094 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22097 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22098 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22100 2013-03-25 Richard Biener <rguenther@suse.de>
22102 PR tree-optimization/56689
22103 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22106 2013-03-25 Richard Biener <rguenther@suse.de>
22108 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22110 (memory_references): Likewise.
22111 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22112 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22113 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22114 (gather_mem_refs_in_loops): Fold into ...
22115 (analyze_memory_references): ... this. Move initialization
22116 to tree_ssa_lim_initialize.
22117 (fill_always_executed_in): Rename to ...
22118 (fill_always_executed_in_1): ... this.
22119 (fill_always_executed_in): Move contains_call computation to
22120 this new function from ...
22121 (tree_ssa_lim_initialize): ... here.
22122 (tree_ssa_lim): Call fill_always_executed_in.
22124 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22126 * postreload.c (reload_combine): Fix code detecting returns.
22128 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22130 * function.c (emit_use_return_register_into_block): On cc0 targets,
22131 do not emit the sequence between cc0 setter and user.
22133 2013-03-25 Kai Tietz <ktietz@redhat.com>
22135 * config/i386/predicates.md (local_symbolic_operand): Interpret
22136 dll-imported symbols as none-local.
22138 2013-03-25 Richard Biener <rguenther@suse.de>
22140 * tree-ssa-loop-im.c (struct depend): Remove.
22141 (struct lim_aux_data): Make depends a vec of gimples.
22142 (free_lim_aux_data): Adjust.
22143 (add_dependency): Likewise.
22144 (set_level): Likewise.
22146 2013-03-25 Richard Biener <rguenther@suse.de>
22148 PR middle-end/56434
22149 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22150 the pointer returned by calls with ECF_MALLOC set.
22152 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22154 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22156 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22158 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22159 using MMXMODE mode iterator.
22160 (*move<mode>_internal): Merge with *movv2sf_internal and
22161 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22163 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22165 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22166 (record_last_mem_set_info): Likewise.
22168 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22169 of XNEWVEC followed by memset.
22170 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22172 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22174 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22175 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22176 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22177 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22178 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22179 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22180 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22181 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22182 BARRIER_P instead of GET_CODE.
22184 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22186 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22187 inaccuracy in the probing code.
22189 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22190 (ctrapdi4): Likewise.
22192 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22194 * calls.c (expand_call): Add missing guard to code handling return
22195 of non-BLKmode structures in MSB.
22196 * function.c (expand_function_end): Likewise.
22198 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22200 * combine.c (try_combine): Adjust comment. Do not add the set of
22201 insn #0 if the destination indirectly is set or dies in insn #2.
22202 Tidy up code to distribute a new note.
22204 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22206 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22207 also for alternatives 16 and 17.
22209 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22211 * config/i386/sse.md (*mov<mode>_internal): Merge with
22212 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22213 Emit insn template depending on type attribute. Use
22214 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22215 movd instead of movq mnemonic for interunit moves. Rewrite mode
22216 attribute calculation. Remove unit attribute calculation.
22217 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22218 Set prefix_data16 attribute for DImode ssemov types.
22219 Use Ym instead of y for SSE-MMX conversion alternatives.
22220 Reorder operand constraints.
22222 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22224 * df.h (df_insn_delete): Adjust prototype.
22225 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22226 and let it decide whether mark the basic block dirty.
22227 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22228 * df-scan.c (df_insn_info_delete): New helper function, split
22229 off from df_insn_delete.
22230 (df_scan_free_bb_info): Use it.
22231 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22233 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22234 that the insn is actually an insn and it has a non-NULL basic block.
22235 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22237 2013-03-22 Richard Biener <rguenther@suse.de>
22239 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22241 (mem_ref_alloc): Do not allocate them.
22242 (refs_independent_p): Do not query or maintain a cache.
22244 2013-03-22 Richard Biener <rguenther@suse.de>
22246 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22247 (gather_mem_refs_in_loops): Do not compute it.
22248 (analyze_memory_references): Do not allocate it.
22249 (tree_ssa_lim_finalize): Do not free it.
22250 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22252 2013-03-22 Richard Biener <rguenther@suse.de>
22254 * is-a.h (as_a): Use gcc_checking_assert.
22256 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22258 * config/aarch64/aarch64.c (aarch64_print_operand): New
22259 format specifier for printing a constant in hex.
22260 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22261 format specifier for printing second operand.
22263 2013-03-22 Richard Biener <rguenther@suse.de>
22265 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22267 (gather_mem_refs_in_loops): Perform store accumulation here.
22268 (create_vop_ref_mapping_loop): Remove.
22269 (create_vop_ref_mapping): Likewise.
22270 (analyze_memory_references): Initialize refs_stored_in_loop.
22271 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22272 (record_indep_loop): Remove.
22273 (record_dep_loop): New function.
22274 (ref_indep_loop_p_1): Adjust to only walk over references
22275 in the loop, not its subloops.
22276 (ref_indep_loop_p): Rename to ...
22277 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22278 maintaining a more fine-grained cache.
22279 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22280 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22282 2013-03-22 Richard Biener <rguenther@suse.de>
22284 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22285 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22286 aggregate mem_ref_loc.
22287 (free_mem_ref_locs): Inline into ...
22288 (memref_free): ... this and adjust.
22289 (mem_ref_alloc): Adjust.
22290 (mem_ref_locs_alloc): Remove.
22291 (record_mem_ref_loc): Adjust.
22292 (get_all_locs_in_loop): Rewrite into ...
22293 (for_all_locs_in_loop): ... this iterator.
22294 (rewrite_mem_ref_loc): New functor.
22295 (rewrite_mem_refs): Use for_all_locs_in_loop.
22296 (sm_set_flag_if_changed): New functor.
22297 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22298 (ref_always_accessed): New functor.
22299 (ref_always_accessed_p): Use for_all_locs_in_loop.
22301 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22303 * tree-pass.h (PROP_gimple_lvec): New.
22304 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22305 (init_optimization_passes): Move pass_lower_vector.
22306 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22308 (pass_lower_vector): Provide PROP_gimple_lvec.
22309 (pass_lower_vector_ssa): Likewise.
22310 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22312 2013-03-21 Mark Wielaard <mjw@redhat.com>
22314 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22316 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22318 * config/i386/i386.md (*movdi_internal): Disparage slightly
22319 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22320 conversion alternatives.
22322 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22324 PR middle-end/48087
22325 * diagnostic.def (DK_WERROR): New kind.
22326 * diagnostic.h (werrorcount): Define.
22327 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22328 promoted to DK_ERROR, increment DK_WERROR counter instead of
22330 * toplev.c (toplev_main): Call print_ignored_options even if
22331 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22332 even if just werrorcount is non-zero.
22335 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22337 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22338 (string_cst_pool_decl): New function.
22339 (optimize_one_addr_into_implicit_ptr): New function.
22340 (resolve_addr_in_expr): Optimize DWARF location expression
22341 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22342 which doesn't live in memory, but has DW_AT_location or
22343 DW_AT_const_value, or refers to a string literal, into
22344 DW_OP_GNU_implicit_pointer.
22345 (optimize_location_into_implicit_ptr): New function.
22346 (resolve_addr): If removing DW_AT_location of a variable because
22347 it was DW_OP_addr of address of the variable, but the variable doesn't
22348 live in memory, try to emit const value attribute for the initializer.
22350 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22352 * tree.h (VECTOR_TYPE_P): New macro.
22353 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22354 TYPE_MODE): Use it.
22355 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22356 VEC_COND_EXPR cannot be lvalues.
22357 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22359 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22361 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22362 Restrict the transformation to equal modes.
22364 2013-03-21 Richard Biener <rguenther@suse.de>
22366 PR tree-optimization/39326
22367 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22368 (MEM_ANALYZABLE): Adjust.
22369 (record_mem_ref_loc): Move bitmap ops ...
22370 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22371 unanalyzable refs, do not record locations for it.
22372 (analyze_memory_references): Allocate ref zero as shared
22374 (refs_independent_p): Do not test for unanalyzed mems here.
22375 (ref_indep_loop_p_1): Special-case disambiguation against
22376 the unanalyzed ref.
22377 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22379 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22381 * config/arm/arm-protos.h (tune_params): Add
22382 prefer_neon_for_64bits field.
22383 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22384 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22385 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22386 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22387 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22388 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22389 (arm_option_override): Handle -mneon-for-64bits new option.
22390 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22391 (prefer_neon_for_64bits): Declare new variable.
22392 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22393 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22394 (arch_enabled): Handle new arch types. Remove support for onlya8
22396 (one_cmpldi2): Use new arch names.
22397 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22398 * config/arm/arm.opt (mneon-for-64bits): Add option.
22399 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22400 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22401 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22403 * doc/invoke.texi (-mneon-for-64bits): Document.
22405 2013-03-21 Richard Biener <rguenther@suse.de>
22407 PR tree-optimization/39326
22408 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22409 (sort_bbs_in_loop_postorder_cmp): New function.
22410 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22412 2013-03-21 Richard Biener <rguenther@suse.de>
22414 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22415 (vect_insert_into_interleaving_chain): Likewise.
22416 (vect_drs_dependent_in_basic_block): Inline ...
22417 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22419 (vect_analyze_data_ref_dependence): ... here. Simplify.
22420 (vect_check_interleaving): Simplify.
22421 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22422 (vect_slp_analyze_data_ref_dependences): ... this new function.
22423 (dr_group_sort_cmp): New function.
22424 (vect_analyze_data_ref_accesses): Compute data-reference groups
22425 here instead of in vect_analyze_data_ref_dependence. Use
22426 a more efficient algorithm.
22427 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22428 vect_slp_analyze_data_ref_dependences. Call
22429 vect_analyze_data_ref_accesses earlier.
22430 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22431 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22432 (vect_slp_analyze_data_ref_dependences): New prototype.
22434 2013-03-21 Richard Biener <rguenther@suse.de>
22436 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22437 ref is stored in the loop.
22438 (find_refs_for_sm): Walk only over all stores.
22439 (store_motion_loop): Allocate from lim_bitmap_obstack.
22440 (store_motion): Likewise.
22442 2013-03-21 Richard Biener <rguenther@suse.de>
22444 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22445 Update virtual SSA form.
22447 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22449 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22450 * configure: Regenerate.
22451 * config.in: Regenerate.
22452 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22453 if !HAVE_LD_EH_FRAME_CIEV3.
22455 2013-03-21 Richard Biener <rguenther@suse.de>
22457 * tree-cfg.c (verify_expr_no_block): New function.
22458 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22459 nor DECL_VALUE_EXPR have locations with associated blocks.
22460 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22461 (clear_unused_block_pointer): Remove code dealing with
22462 blocks in DECL_DEBUG_EXPR locations.
22464 2013-03-21 Richard Biener <rguenther@suse.de>
22466 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22467 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22468 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22469 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22470 instead of DECL_DEBUG_EXPR_IS_FROM.
22471 * gimplify.c (gimplify_modify_expr): Likewise.
22472 * tree-cfg.c (verify_expr_location_1): Likewise.
22473 * tree-complex.c (create_one_component_var): Likewise.
22474 * tree-sra.c (create_access_replacement): Likewise.
22475 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22476 (clear_unused_block_pointer): Likewise.
22477 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22478 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22479 * var-tracking.c (var_debug_decl): Likewise.
22480 (track_expr_p): Likewise.
22481 * tree-inline.c (add_local_variables): Likewise. Set
22482 DECL_HAS_DEBUG_EXPR_P after copying it.
22483 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22484 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22486 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22489 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22490 * configure: Regenerate.
22491 * config.in: Regenerate.
22492 * config/i386/i386.md (*movdf_internal): Use
22493 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22494 movd instead of movq mnemonic for interunit moves.
22495 (*movdi_internal): Ditto.
22497 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22499 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22500 (abd<mode>_3): New pattern.
22501 (aba<mode>_3): New pattern.
22502 (fabd<mode>_3): New pattern.
22504 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22506 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22507 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22508 occurrence of REGISTER_PREFIX as its empty string.
22510 2013-03-20 Jeff Law <law@redhat.com>
22512 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22513 addititional equivalences for equality comparisons between an SSA_NAME
22514 and a constant where the SSA_NAME was set from a widening conversion.
22516 2013-03-20 Walter Lee <walt@tilera.com>
22518 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22520 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22522 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22523 depending on type attribute.
22524 (*movti_internal): Ditto.
22525 (*movtf_internal): Ditto.
22526 (*movxf_internal): Ditto.
22527 (*movdf_internal): Ditto.
22528 (*movsf_internal): Ditto.
22530 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22532 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22533 maybe_vex for sselog1 and ssemov types.
22534 (*movdi_internal): Reorder operand constraints.
22535 (*movsi_internal): Ditto. Set prefix attribute to
22536 maybe_vex for sselog1 and ssemov types.
22537 (*movtf_internal): Set prefix attribute to maybe_vex
22538 for sselog1 and ssemov types.
22539 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22540 DImode ssemov types. Reorder operand constraints.
22541 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22542 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22543 attribute for SImode ssemov types. Reorder operand constraints.
22545 2013-03-20 Martin Jambor <mjambor@suse.cz>
22547 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22548 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22550 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22552 * config/rs6000/predicates.md (indexed_address, update_address_mem
22553 update_indexed_address_mem): New predicates.
22554 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22555 attribute for load/store instructions.
22556 * config/rs6000/dfp.md (movsd_store): Likewise.
22557 (movsd_load): Likewise.
22558 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22559 (unnamed HI->DI extend define_insn): Likewise.
22560 (unnamed SI->DI extend define_insn): Likewise.
22561 (unnamed QI->SI extend define_insn): Likewise.
22562 (unnamed QI->HI extend define_insn): Likewise.
22563 (unnamed HI->SI extend define_insn): Likewise.
22564 (unnamed HI->SI extend define_insn): Likewise.
22565 (extendsfdf2_fpr): Likewise.
22566 (movsi_internal1): Likewise.
22567 (movsi_internal1_single): Likewise.
22568 (movhi_internal): Likewise.
22569 (movqi_internal): Likewise.
22570 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22571 attribute for load/store instructions.
22572 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22574 (mov<mode>_softfloat): Likewise.
22575 (mov<mode>_hardfloat32): Likewise.
22576 (mov<mode>_hardfloat64): Likewise.
22577 (mov<mode>_softfloat64): Likewise.
22578 (movdi_internal32): Likewise.
22579 (movdi_internal64): Likewise.
22580 (probe_stack_<mode>): Likewise.
22582 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22584 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22585 floating point, and decimal floating point to reload iterator.
22587 * config/rs6000/constraints.md (wl constraint): New constraints to
22588 return FLOAT_REGS if certain options are used to reduce the number
22589 of separate patterns that exist in the file.
22590 (wx constraint): Likewise.
22591 (wz constraint): Likewise.
22593 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22594 -mdebug=reg, print wg, wl, wx, and wz constraints.
22595 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22596 Initialize the reload functions for 64-bit binary/decimal floating
22598 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22599 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22600 create the buffer on the stack to overcome not having a 32-bit
22602 (rs6000_emit_move): Likewise.
22603 (rs6000_secondary_memory_needed_rtx): Likewise.
22604 (rs6000_alloc_sdmode_stack_slot): Likewise.
22605 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22606 via xxlxor, just like DFmode 0.0.
22608 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22609 define as 1 if we are running on a power7 or newer.
22610 (enum r6000_reg_class_enum): Add new constraints.
22612 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22613 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22614 with other moves by using conditional constraits (wg). Use LFIWZX
22615 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22616 (movsd splitter): Likewise.
22617 (movsd_hardfloat): Likewise.
22618 (movsd_softfloat): Likewise.
22620 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22621 binary and decimal floating point moves.
22622 (fmove_ok): New attributes to combine binary and decimal floating
22623 point moves, and to combine power6x (mfpgpr) moves along normal
22625 (real_value_to_target): Likewise.
22626 (f32_lr): Likewise.
22627 (f32_lm): Likewise.
22628 (f32_li): Likewise.
22629 (f32_sr): Likewise.
22630 (f32_sm): Likewise.
22631 (f32_si): Likewise.
22632 (movsf): Combine binary and decimal floating point moves. Combine
22633 power6x (mfpgpr) moves with other moves by using conditional
22634 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22635 (mov<mode> for SFmode/SDmode); Likewise.
22636 (SFmode/SDmode splitters): Likewise.
22637 (movsf_hardfloat): Likewise.
22638 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22639 (movsf_softfloat): Likewise.
22640 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22642 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22643 wx and wz constraints.
22645 * config/rs6000/constraints.md (wg constraint): New constraint to
22646 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22648 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22651 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22652 -mdebug=reg, print wg, wl, wx, and wz constraints.
22653 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22654 Initialize the reload functions for 64-bit binary/decimal floating
22656 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22657 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22658 create the buffer on the stack to overcome not having a 32-bit
22660 (rs6000_emit_move): Likewise.
22661 (rs6000_secondary_memory_needed_rtx): Likewise.
22662 (rs6000_alloc_sdmode_stack_slot): Likewise.
22663 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22664 via xxlxor, just like DFmode 0.0.
22666 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22667 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22668 with other moves by using conditional constraits (wg). Use LFIWZX
22669 and STFIWX for loading SDmode on power7.
22670 (movdd splitters): Likewise.
22671 (movdd_hardfloat32): Likewise.
22672 (movdd_softfloat32): Likewise.
22673 (movdd_hardfloat64_mfpgpr): Likewise.
22674 (movdd_hardfloat64): Likewise.
22675 (movdd_softfloat64): Likewise.
22677 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22678 64-bit binary and decimal floating point moves.
22679 (FMOVE64X): Likewise.
22680 (movdf): Combine 64-bit binary and decimal floating point moves.
22681 Combine power6x (mfpgpr) moves with other moves by using
22682 conditional constraits (wg).
22683 (mov<mode> for DFmode/DDmode): Likewise.
22684 (DFmode/DDmode splitters): Likewise.
22685 (movdf_hardfloat32): Likewise.
22686 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22687 (movdf_softfloat32): Likewise.
22688 (movdf_hardfloat64_mfpgpr): Likewise.
22689 (movdf_hardfloat64): Likewise.
22690 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22691 (movdf_softfloat64): Likewise.
22692 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22693 (reload_<mode>_load): Move to later in the file so they aren't in
22694 the middle of the floating point move insns.
22695 (reload_<mode>_store): Likewise.
22697 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22700 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22701 constraint if -mdebug=reg.
22702 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22703 Enable using dd reload support if needed.
22705 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22706 binary and decimal floating point moves in rs6000.md.
22707 (movtd_internal): Likewise.
22709 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22710 decimal floating point moves.
22712 (movtf_internal): Likewise.
22713 (mov<mode>_internal, TDmode/TFmode): Likewise.
22714 (movtf_softfloat): Likewise.
22715 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22717 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22718 movdi_internal64, using wg constraint for move direct operations.
22719 (movdi_internal64): Likewise.
22721 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22722 MODES_TIEABLE_P for selected modes. Print the numerical value of
22723 the various virtual registers. Use GPR/FPR first/last values,
22724 instead of hard coding the register numbers. Print which modes
22725 have reload functions registered.
22726 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22727 settings before/after setting cpu, target and subtarget settings.
22728 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22729 and for secondary reload failures in rs6000_secondary_reload_inner.
22730 (rs6000_secondary_reload_fail): Likewise.
22731 (rs6000_secondary_reload_inner): Likewise.
22733 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22734 macros for first/last GPR and FPR registers.
22735 (LAST_GPR_REGNO): Likewise.
22736 (FIRST_FPR_REGNO): Likewise.
22737 (LAST_FPR_REGNO): Likewise.
22739 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22740 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22741 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22742 (vcond<mode><mode>): Likewise.
22743 (vcondu<mode><mode>): Likewise.
22744 (vector_gtu<mode>): Likewise.
22745 (vector_gte<mode>): Likewise.
22746 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22747 to prevent the compiler from converting DImode operations to TImode.
22748 (ior<mode>3): Likewise.
22749 (and<mode>3): Likewise.
22750 (one_cmpl<mode>2): Likewise.
22751 (nor<mode>3): Likewise.
22752 (andc<mode>3): Likewise.
22754 * config/rs6000/constraints.md (wt constraint): New constraint
22755 that returns VSX_REGS if TImode is allowed in VSX registers.
22757 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22758 constant under VSX.
22760 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22761 similar to TImode, but it is restricted to being in the GPRs.
22763 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22764 TImode to occupy a single VSX register.
22766 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22767 -mvsx-timode for power7/power8.
22768 (power7 cpu): Likewise.
22769 (power8 cpu): Likewise.
22771 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22772 sure that TFmode/TDmode take up two registers if they are ever
22773 allowed in the upper VSX registers.
22774 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22776 (rs6000_init_hard_regno_mode_ok): Likewise.
22777 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22778 constraint. Print if LRA is turned on.
22779 (rs6000_option_override_internal): Give an error if -mvsx-timode
22780 and VSX is not enabled.
22781 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22782 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22783 to reg+offset addressing. Use PTImode when checking offset
22784 addresses for validity.
22785 (reg_offset_addressing_ok_p): Likewise.
22786 (rs6000_legitimate_offset_address_p): Likewise.
22787 (rs6000_legitimize_address): Likewise.
22788 (rs6000_legitimize_reload_address): Likewise.
22789 (rs6000_legitimate_address_p): Likewise.
22790 (rs6000_eliminate_indexed_memrefs): Likewise.
22791 (rs6000_emit_move): Likewise.
22792 (rs6000_secondary_reload): Likewise.
22793 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22794 reloads to fpr registers to continue to use reg+offset addressing,
22795 but 64-bit reloads to altivec registers need reg+reg addressing.
22796 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22797 it. Treat LO_SUM like a PLUS operation.
22798 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22799 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22800 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22801 registers to share a register with a smaller sized type, since VSX
22802 puts scalars in the upper 64-bits.
22803 (print_operand): Add support for PTImode.
22804 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22805 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22806 registers, but don't have arithmetic support.
22807 (rs6000_memory_move_cost): Add test for VSX.
22808 (rs6000_opt_masks): Add -mvsx-timode.
22810 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22813 (VSr): Use wt constraint for TImode.
22814 (VSv): Drop TImode support.
22815 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22816 (vsx_movti_64bit): Likewise.
22817 (vsx_movti_32bit): Likewise.
22818 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22819 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22820 one '?' on the appropriate output constraint. Do not allow TImode
22821 logical operations on 32-bit systems.
22822 (vsx_ior<mode>3): Likewise.
22823 (vsx_xor<mode>3): Likewise.
22824 (vsx_one_cmpl<mode>2): Likewise.
22825 (vsx_nor<mode>3): Likewise.
22826 (vsx_andc<mode>3): Likewise.
22827 (vsx_concat_<mode>): Likewise.
22828 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22830 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22831 OPTION_MASK_VSX_TIMODE.
22832 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22833 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22835 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22836 (TI2 iterator): New iterator for TImode, PTImode.
22837 (wd mode attribute): Add values for vector types.
22838 (movti_string): Replace TI move operations with operations for TImode
22839 and PTImode. Add support for TImode being allowed in VSX registers.
22840 (mov<mode>_string, TImode/PTImode): Likewise.
22841 (movti_ppc64): Likewise.
22842 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22843 (TI mode splitters): Likewise.
22845 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22848 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22850 PR tree-optimization/56355
22851 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22852 Also handle integers with undefined overflow.
22854 2013-03-20 Catherine Moore <clm@codesourcery.com>
22855 Maciej W. Rozycki <macro@codesourcery.com>
22856 Tom de Vries <tom@codesourcery.com>
22857 Nathan Sidwell <nathan@codesourcery.com>
22858 Iain Sandoe <iain@codesourcery.com>
22859 Nathan Froyd <froydnj@codesourcery.com>
22860 Chao-ying Fu <fu@mips.com>
22862 * doc/extend.texi: (micromips, nomicromips, nocompression):
22863 Document new function attributes.
22864 * doc/invoke.texi (minterlink-compressed, mmicromips,
22865 m14k, m14ke, m14kec): Document new options.
22866 (minterlink-mips16): Update documentation.
22867 * doc/md.texi (ZC, ZD): Document new constraints.
22868 * configure.ac (gcc_cv_as_micromips): Check if linker
22869 supports the .set micromips directive.
22870 * configure: Regenerate.
22871 * config.in: Regenerate.
22872 * config/mips/mips-tables.opt: Regenerate.
22873 * config/mips/micromips.md: New file.
22874 * constraints.md (ZC, ZD): New constraints.
22875 * config/mips/predicates.md (movep_src_register): New predicate.
22876 (movep_src_operand): New predicate.
22877 (non_volatile_mem_operand): New predicate.
22878 * config/mips/mips.md (multimem): New type.
22879 (length): Differentiate between 17-bit and 18-bit branch offsets.
22880 (MOVEP1, MOVEP2): New mode iterator.
22881 (mov_<load>l): Use ZC constraint.
22882 (mov_<load>r): Likewise.
22883 (mov_<store>l): Likewise.
22884 (mov_<store>r): Likewise.
22885 (*branch_equality<mode>_inverted): Add microMIPS support.
22886 (*branch_equality<mode>): Likewise.
22887 (*jump_absolute): Likewise.
22888 (indirect_jump_<mode>): Likewise.
22889 (tablejump_<mode>): Likewise.
22890 (<optab>_internal): Likewise.
22891 (sibcall_internal): Likewise.
22892 (sibcall_value_internal): Likewise.
22893 (prefetch): Use constraint ZD.
22894 * config/mips/mips.opt (minterlink-compressed): New option.
22895 (minterlink-mips16): Now an alias for minterlink-compressed.
22896 (mmicromips): New option.
22897 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22898 (compare_and_swap_12): Likewise.
22899 (sync_add<mode>): Likewise.
22900 (sync_<optab>_12): Likewise.
22901 (sync_old_<optab>_12): Likewise.
22902 (sync_new_<optab>_12): Likewise.
22903 (sync_nand_12): Likewise.
22904 (sync_old_nand_12): Likewise.
22905 (sync_new_nand_12): Likewise.
22906 (sync_sub<mode>): Likewise.
22907 (sync_old_add<mode>): Likewise.
22908 (sync_old_sub<mode>): Likewise.
22909 (sync_new_add<mode>): Likewise.
22910 (sync_new_sub<mode>): Likewise.
22911 (sync_<optab><mode>): Likewise.
22912 (sync_old_<optab><mode>): Likewise.
22913 (sync_new_<optab><mode>): Likewise.
22914 (sync_nand<mode>): Likewise.
22915 (sync_old_nand<mode>): Likewise.
22916 (sync_new_nand<mode>): Likewise.
22917 (sync_lock_test_and_set<mode>): Likewise.
22918 (test_and_set_12): Likewise.
22919 (atomic_compare_and_swap<mode>): Likewise.
22920 (atomic_exchange<mode>_llsc): Likewise.
22921 (atomic_fetch_add<mode>_llsc): Likewise.
22922 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22923 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22924 (umips_save_restore_pattern_p): Likewise.
22925 (umips_load_store_pair_p): Likewise.
22926 (umips_output_load_store_pair): Likewise.
22927 (umips_movep_target_p): Likewise.
22928 (umips_12bit_offset_address_p): Likewise.
22929 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22930 (mips_base_mips16): Rename this...
22931 (mips_base_compression_flags): ...to this. Update all uses.
22932 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22933 (mips_mips16_decl_p): Delete.
22934 (mips_nomips16_decl_p): Delete.
22935 (mips_get_compress_on_flags): New function.
22936 (mips_get_compress_off_flags): New function.
22937 (mips_get_compress_mode): New function.
22938 (mips_get_compress_on_name): New function.
22939 (mips_get_compress_off_name): New function.
22940 (mips_insert_attributes): Support multiple compression types.
22941 (mips_merge_decl_attributes): Likewise.
22942 (umips_12bit_offset_address_p): New function.
22943 (mips_start_function_definition): Emit .set micromips directive.
22944 (mips_call_may_need_jalx_p): New function.
22945 (mips_function_ok_for_sibcall): Add microMIPS support.
22946 (mips_print_operand_punctuation): Support short delay slots and
22948 (umips_swm_mask, umips_swm_encoding): New.
22949 (umips_build_save_restore): New function.
22950 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22951 (was_mips16_p): Remove.
22952 (old_compression_mode): New.
22953 (mips_set_compression_mode): New function.
22954 (mips_set_current_function): Add microMIPS support.
22955 (mips_option_override): Likewise.
22956 (umips_save_restore_pattern_p): New function.
22957 (umips_output_save_restore): New function.
22958 (umips_load_store_pair_p_1): New function.
22959 (umips_load_store_pair_p): New function.
22960 (umips_output_load_store_pair_1): New function.
22961 (umips_output_load_store_pair): New function.
22962 (umips_movep_target_p) New function.
22963 (mips_prepare_pch_save): Add microMIPS support.
22964 * config/mips/mips.h (TARGET_COMPRESSION): New.
22965 (TARGET_CPU_CPP_BUILTINS): Update macro
22966 to use new compression flags and to support microMIPS.
22967 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22968 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22969 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22970 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22971 (ASM_SPEC): Support mmicromips and mno-micromips.
22972 (M16STORE_REG_P): New macro.
22973 (MIPS_CALL): Support TARGET_MICROMIPS.
22974 (MICROMIPS_J): New macro.
22975 (mips_base_mips16): Rename this...
22976 (mips_base_compression_flags): ...to this.
22977 (UMIPS_12BIT_OFFSET_P): New macro.
22978 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22979 (MULTILIB_DIRNAMES): Likewise.
22980 2013-03-20 Richard Biener <rguenther@suse.de>
22982 PR tree-optimization/56661
22983 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22984 the result does not have to be distinct.
22986 2013-03-20 Richard Biener <rguenther@suse.de>
22988 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22991 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22992 Steven Bosscher <steven@gcc.gnu.org>
22994 PR rtl-optimization/56605
22995 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22997 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23000 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23001 that require movd instead of movq.
23003 2013-03-20 Richard Biener <rguenther@suse.de>
23005 * tree-ssa-structalias.c (struct variable_info): Add pointer
23006 to the first field of an aggregate with sub-vars. Make
23007 this and the pointer to the next subfield its ID.
23008 (vi_next): New function.
23009 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23010 storedanything_id, integer_id): Increment by one.
23011 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23012 get_call_clobber_vi): Adjust.
23013 (solution_set_expand): Simplify and speedup.
23014 (solution_set_add): Inline into ...
23015 (set_union_with_increment): ... this. Adjust accordingly.
23016 (do_sd_constraint): Likewise.
23017 (do_ds_constraint): Likewise.
23018 (do_complex_constraint): Simplify.
23019 (build_pred_graph): Adjust.
23020 (solve_graph): Likewise. Simplify and speedup.
23021 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23022 get_constraint_for_component_ref, get_constraint_for_1,
23023 first_vi_for_offset, first_or_preceding_vi_for_offset,
23024 create_function_info_for, create_variable_info_for_1,
23025 create_variable_info_for, intra_create_variable_infos): Adjust.
23026 (init_base_vars): Push NULL for ID zero.
23027 (compute_points_to_sets): Adjust.
23029 2013-03-20 Richard Biener <rguenther@suse.de>
23031 * cfgloop.c (verify_loop_structure): Streamline and avoid
23032 ICEing on corrupt loop tree.
23033 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23036 2013-03-20 Richard Biener <rguenther@suse.de>
23038 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23039 check whether an SSA update is needed.
23041 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23043 * config/mips/constraints.md (T): Rename to...
23047 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23048 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23050 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23052 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23053 (*subsi3_carryin_uxtw): Likewise.
23055 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23057 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23058 (*rorsi3_insn_uxtw): Likewise.
23060 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23062 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23063 (*extrsi5_insn_uxtw): Likewise.
23065 2013-03-19 Richard Biener <rguenther@suse.de>
23067 PR tree-optimization/56273
23068 * passes.c (init_optimization_passes): Move second VRP after DOM.
23070 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23072 * config/i386/i386.md (*movti_internal): Merge from
23073 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23074 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23075 nox64 isa attributes.
23077 2013-03-18 Richard Biener <rguenther@suse.de>
23079 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23081 (merge_node_constraints): Likewise.
23082 (build_succ_graph): Likewise.
23083 (valid_graph_edge): Inline into single caller.
23084 (unify_nodes): Likewise. Use bitmap_set_bit return value
23086 (scc_visit): Fix formatting and variable use.
23087 (do_sd_constraint): Use gcc_checking_assert.
23088 (do_ds_constraint): Likewise.
23089 (do_complex_constraint): Likewise.
23090 (condense_visit): Likewise. Cleanup.
23091 (dump_pred_graph): New function.
23092 (perform_var_substitution): Dump the pred-graph before
23093 variable substitution.
23094 (find_equivalent_node): Use gcc_checking_assert.
23095 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23097 2013-03-18 Richard Biener <rguenther@suse.de>
23099 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23100 Remove cond_expr_stmt_list argument and do not gimplify the
23102 (vect_loop_versioning): Adjust.
23103 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23104 Cleanup to use less temporaries.
23105 (vect_create_data_ref_ptr): Cleanup.
23107 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23109 PR tree-optimization/56635
23110 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23111 require types_compatible_p types.
23113 2013-03-18 Nick Clifton <nickc@redhat.com>
23115 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23116 spurious backslash.
23118 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23119 Add missing line to comment describing function.
23121 2013-03-18 Richard Biener <rguenther@suse.de>
23123 PR tree-optimization/56210
23124 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23125 Handle string / character search functions.
23126 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23128 2013-03-18 Richard Biener <rguenther@suse.de>
23130 PR middle-end/56483
23131 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23132 and implement properly.
23133 * gimple.h (gimple_cond_single_var_p): Remove.
23135 2013-03-18 Richard Biener <rguenther@suse.de>
23137 * tree-data-ref.h (find_data_references_in_loop): Declare.
23138 * tree-data-ref.c (get_references_in_stmt): Use a stack
23139 vector pre-allocated in the callers.
23140 (find_data_references_in_stmt): Adjust.
23141 (graphite_find_data_references_in_stmt): Likewise.
23142 (create_rdg_vertices): Likewise.
23143 (find_data_references_in_loop): Export.
23144 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23145 Compute dependences here...
23146 (vect_analyze_data_refs): ...not here. When we encounter
23147 a non-vectorizable data reference in basic-block vectorization
23148 truncate the data reference vector. Do not bother to
23149 fixup data-dependence information for gather loads.
23150 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23151 of data references, as reported.
23153 2013-03-18 Richard Biener <rguenther@suse.de>
23155 PR tree-optimization/3713
23156 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23157 has_constants and expr.
23158 (stmt_has_constants): Properly valueize SSA names when deciding
23159 whether the stmt has constants.
23161 2013-03-18 Richard Biener <rguenther@suse.de>
23163 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23164 whole function when there is nothing to do.
23165 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23166 * tree-vectorizer.c (vectorize_loops): Update virtual and
23167 loop-closed SSA once.
23168 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23170 2013-03-18 Richard Biener <rguenther@suse.de>
23172 PR middle-end/56113
23173 * domwalk.c (bb_postorder): New global static.
23174 (cmp_bb_postorder): New function.
23175 (walk_dominator_tree): Replace scheme imposing an order for
23176 visiting dominator sons by one sorting them at the time they
23177 are pushed on the stack.
23179 2013-03-18 Richard Biener <rguenther@suse.de>
23181 PR tree-optimization/39326
23182 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23183 (struct mem_ref): Replace mem member with ao_ref typed member.
23184 (MEM_ANALYZABLE): Adjust.
23185 (memref_eq): Likewise.
23186 (mem_ref_alloc): Likewise.
23187 (gather_mem_refs_stmt): Likewise.
23188 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23189 (execute_sm_if_changed_flag_set): Adjust.
23190 (execute_sm): Likewise.
23191 (ref_always_accessed_p): Likewise.
23192 (refs_independent_p): Likewise.
23193 (can_sm_ref_p): Likewise.
23195 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23198 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23199 return 1 even for !unsignedp.
23201 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23203 * config/i386/i386.md (isa): Add x64 and nox64.
23204 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23205 (*pushtf): Enable *roF alternative for x64 isa only.
23206 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23207 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23208 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23209 (*movtf_internal): Merge from *movtf_internal_rex64 and
23210 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23211 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23212 nox64 isa attributes.
23213 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23214 nox64 isa attributes.
23215 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23217 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23219 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23221 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23224 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23225 class names. Remove trailing comma after "ALL_REGS".
23227 2013-03-16 Jan Hubicka <jh@suse.cz>
23229 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23230 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23231 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23232 of cgraph_get_create_node.
23233 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23235 2013-03-16 Jason Merrill <jason@redhat.com>
23238 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23239 with DW_AT_default_value.
23241 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23243 * BASE-VER: Set to 4.9.0.
23245 2013-03-14 Andi Kleen <ak@linux.intel.com>
23248 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23249 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23250 Document _x* TSX intrinsics.
23252 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23253 David Holsgrove <david.holsgrove@xilinx.com>
23255 * configure.ac: Add MicroBlaze TLS support detection.
23256 * configure: Regenerate.
23257 * config/microblaze/microblaze-protos.h
23258 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23259 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23260 * config/microblaze/microblaze.c (microblaze_address_type): Add
23261 ADDRESS_TLS and tls_reloc address types.
23262 (microblaze_address_info): Add tls_reloc.
23263 (TARGET_HAVE_TLS): Define.
23264 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23265 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23266 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23267 load_tls_operand, microblaze_call_tls_get_addr,
23268 microblaze_legitimize_tls_address): New functions.
23269 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23270 (get_base_reg): Use microblaze_tls_symbol_p.
23271 (microblaze_classify_address): Handle TLS.
23272 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23273 label_mentioned_p and microblaze_tls_referenced_p.
23274 (microblaze_legitimize_address): Handle TLS.
23275 (microblaze_address_insns): Handle ADDRESS_TLS.
23276 (pic_address_needs_scratch): Handle TLS.
23277 (print_operand_address): Handle TLS.
23278 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23279 (microblaze_expand_move): Handle TLS.
23280 (microblaze_legitimate_constant_p): Check
23281 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23282 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23283 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23284 (PIC_OFFSET_TABLE_REGNUM): Set.
23285 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23286 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23287 (addsi3, movsi_internal2, movdf_internal): Update constraints
23288 * config/microblaze/predicates.md (arith_plus_operand): Define
23289 (move_operand): Redefine as move_src_operand,
23290 check microblaze_tls_referenced_p.
23292 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23294 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23295 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23297 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23299 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23302 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23304 PR tree-optimization/53265
23305 * common.opt (Waggressive-loop-optimizations): New option.
23306 * tree-ssa-loop-niter.c: Include tree-pass.h.
23307 (do_warn_aggressive_loop_optimizations): New function.
23308 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23309 if number_of_latch_executions returned constant.
23310 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23311 early. If number_of_latch_executions returned constant, set
23312 nb_iterations_upper_bound back to it.
23313 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23315 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23316 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23318 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23319 (MULTILIB_OSDIRNAMES): Set.
23320 * genmultilib: If defaultosdirname doesn't start with :: , set
23321 defaultosdirname2 instead, clear it and emit two . multilib_raw
23322 entries instead of just one.
23324 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23326 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23327 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23328 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23329 (SUBTARGET_OVERRIDE_OPTIONS): New.
23331 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23334 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23335 (musermode): Convert to Var(TARGET_USERMODE).
23336 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23337 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23338 * config/sh/sh.c (sh_option_override): Use
23339 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23340 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23342 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23344 (udivsi3_i4_single, divsi3_i4_single): Use
23345 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23347 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23349 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23352 2013-03-13 Richard Biener <rguenther@suse.de>
23354 PR tree-optimization/56608
23355 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23356 calls when vectorizing basic-blocks.
23358 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23361 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23364 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23366 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23368 2013-03-11 Jan Hubicka <jh@suse.cz>
23371 * lto-streamer-out.c (output_symbol_p): Skip references from
23372 constructors of external variables.
23374 2013-03-11 Jan Hubicka <jh@suse.cz>
23376 PR middle-end/56571
23377 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23379 * emit-rtl.c (verify_rtx_sharing): Likewise.
23380 (copy_insn_1): Likewise.
23381 * rtl.c (copy_rtx): Likewise.
23383 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23386 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23387 output_operand_lossage message.
23389 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23392 * arm.c (shift_op): Validate RTL pattern on the fly.
23393 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23396 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23399 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23400 new call patterns that clobber %fr12.
23401 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23402 split and postreload patterns.
23403 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23404 registers %fr12 and %fr12R as call used.
23406 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23408 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23409 (canon_address, record_store, replace_read, check_mem_read_rtx,
23410 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23411 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23412 rest_of_handle_dse): Likewise.
23414 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23416 PR middle-end/56524
23417 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23419 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23420 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23421 (save_optabs_if_changed): Replace with...
23422 (init_tree_optimization_optabs): ...this.
23423 * optabs.c (save_optabs_if_changed): Rename to...
23424 (init_tree_optimization_optabs): ...this. Take the optimization node
23425 as argument. Do nothing if the base optabs are already correct.
23426 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23427 to recompute optabs.
23428 * function.h (function): Remove optabs field.
23429 * function.c (invoke_set_current_function_hook): Call
23430 init_tree_optimization_optabs. Use the result to initialize
23433 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23435 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23436 if GTMA_HAS_NO_INSTRUMENTATION.
23437 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23438 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23439 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23440 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23441 GTMA_HAS_NO_INSTRUMENTATION.
23443 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23445 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23448 2013-03-08 Marek Polacek <polacek@redhat.com>
23449 Jakub Jelinek <jakub@redhat.com>
23451 PR tree-optimization/56478
23452 * predict.c (is_comparison_with_loop_invariant_p): Change the
23453 type of loop_step to tree.
23454 (predict_loops): Adjust.
23455 (predict_iv_comparison): Perform the computations on double_ints.
23457 2013-03-08 Richard Biener <rguenther@suse.de>
23459 PR tree-optimization/56570
23460 * tree-cfg.c (verify_expr_location_1): Verify locations for
23462 * tree-sra.c (create_access_replacement): Strip locations
23463 from DECL_DEBUG_EXPRs.
23465 2013-03-08 Richard Biener <rguenther@suse.de>
23467 * tree-inline.c (expand_call_inline): Do not associate
23468 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23469 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23471 2013-03-08 Richard Biener <rguenther@suse.de>
23473 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23474 or block changes with -Og. Fix for location / block encoding
23475 changes and PHI arguments with locations.
23477 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23479 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23481 (struct output_info): Likewise.
23482 (register_overhead): Remove bad gcc_assert.
23483 (bitmap_find_bit): If there is only a single bitmap element, do not
23484 count a miss as a search.
23485 (print_statistics): Update for counter type changes.
23486 (dump_bitmap_statistics): Likewise. Print headers such that they
23487 are properly lined up with the printed counters.
23489 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23491 PR tree-optimization/56559
23492 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23493 check that it has only a single use.
23495 2013-03-07 Richard Biener <rguenther@suse.de>
23497 * doc/invoke.texi (fwhole-program): Discourage use in combination
23500 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23502 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23504 PR tree-optimization/56539
23505 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23506 instead of GSI_CONTINUE_LINKING as last argument to
23507 force_gimple_operand_gsi. Adjust function comment.
23509 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23512 PR middle-end/56548
23513 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23514 promoted mode, convert the result back to the original mode.
23516 2013-03-06 Richard Biener <rguenther@suse.de>
23518 PR middle-end/56294
23519 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23520 (insert_updated_phi_nodes_compare_uids): New function.
23521 (update_ssa): Sort symbols_to_rename after UID before
23522 traversing it to insert PHI nodes.
23524 2013-03-06 Richard Biener <rguenther@suse.de>
23526 PR middle-end/50494
23527 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23528 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23531 2013-02-13 Richard Biener <rguenther@suse.de>
23534 * varasm.c (output_constant_def_1): Get the decl representing
23535 the constant as argument.
23536 (output_constant_def): Wrap output_constant_def_1.
23537 (make_decl_rtl): Use output_constant_def_1 with the decl
23538 representing the constant.
23539 (build_constant_desc): Optionally re-use a decl already
23540 representing the constant.
23541 (tree_output_constant_def): Adjust.
23543 2013-03-06 Joey Ye <joey.ye@arm.com>
23546 * gcc.c (convert_white_space): New function.
23547 (main): Handles white space in function name.
23549 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23552 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23553 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23554 to SH_DIV_CALL_TABLE for TARGET_SH2.
23555 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23557 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23558 call-table options.
23560 2013-03-05 Sterling Augustine <saugustine@google.com>
23561 Cary Coutant <ccoutant@google.com>
23564 * dwarf2out.c (resolve_addr): Don't call
23565 remove_loc_list_addr_table_entries a second time for the same
23568 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23571 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23572 (avoid_complex_debug_insns): New function.
23573 (expand_debug_locations): Call it.
23575 PR rtl-optimization/56484
23576 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23577 lifetimes of hard registers on small register class machines.
23579 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23581 * config/microblaze/microblaze-protos.h: Rename
23582 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23583 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23585 (microblaze_fast_interrupt_function_p): New function.
23586 (microblaze_is_interrupt_handler): Rename to
23587 microblaze_is_interrupt_variant and add fast_interrupt check.
23588 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23589 (save_restore_insns): Likewise.
23590 (compute_frame_size): Likewise.
23591 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23592 (microblaze_globalize_label): Likewise.
23593 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23594 * config/microblaze/microblaze.md: Use wrapper
23595 microblaze_is_interrupt_variant.
23597 2013-03-05 Kai Tietz <ktietz@redhat.com>
23599 * sdbout.c (sdbout_one_type): Switch to current function's section
23600 supporting cold/hot.
23602 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23604 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23607 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23609 PR middle-end/56461
23610 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23611 if VALGRIND_GET_VBITS is defined, temporarily make object
23612 memory all defined, and restore previous valgrind addressability
23613 and definability afterwards. Free this_object at the end.
23615 PR middle-end/56461
23616 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23617 right before calling lra_create_live_ranges, also call it
23618 when clearing live_p. Only call lra_clear_live_ranges
23619 at the end if live_p.
23621 PR middle-end/56461
23622 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23624 2013-03-05 Richard Biener <rguenther@suse.de>
23626 PR tree-optimization/56521
23627 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23630 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23633 * except.h (remove_unreachable_eh_regions): New prototype.
23634 * except.c (remove_eh_handler_splicer): New function, split out
23635 of remove_eh_handler.
23636 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23637 warning about running it on many EH regions one at a time.
23638 (remove_unreachable_eh_regions_worker): New function, walk the
23639 EH tree in depth-first order and remove non-marked regions.
23640 (remove_unreachable_eh_regions): New function.
23641 * tree-eh.c (mark_reachable_handlers): New function, split out
23642 from remove_unreachable_handlers.
23643 (remove_unreachable_handlers): Use mark_reachable_handlers and
23644 remove_unreachable_eh_regions.
23645 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23646 and remove_unreachable_eh_regions.
23648 2013-03-05 Richard Biener <rguenther@suse.de>
23650 PR middle-end/56525
23651 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23652 not freeing them until the end.
23654 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23656 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23658 2013-03-05 Richard Biener <rguenther@suse.de>
23660 PR tree-optimization/56270
23661 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23662 of loads after scheduling an SLP instance.
23664 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23666 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23668 (check_gcc_parallelize): Run guality.exp as a separate job from
23669 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23670 struct-layout-1.exp with stackalign.exp.
23672 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23674 PR middle-end/56461
23675 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23676 load_index sbitmap even if some bit in it isn't set.
23678 PR middle-end/56461
23679 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23680 (discover_iteration_bound_by_body_walk): Change queues to
23681 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23682 spelling in comment. Call safe_push on queues[bound_index] directly.
23683 Release queues[queue_index] in every iteration unconditionally.
23684 Release bounds vector.
23686 PR middle-end/56461
23687 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23688 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23689 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23692 2013-03-05 Richard Biener <rguenther@suse.de>
23695 * tree-inline.c (remap_blocks_to_null): New function.
23696 (expand_call_inline): When expanding a call stmt without
23697 an associated block inline remap all callee blocks to NULL.
23699 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23701 PR rtl-optimization/56494
23702 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23703 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23704 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23706 PR middle-end/56461
23707 * sel-sched-ir.c (free_sched_pools): Release
23708 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23709 if succs_info_pool.max_top isn't -1.
23712 * opts.c (opts_obstack, opts_concat): Moved to...
23713 * opts-common.c (opts_obstack, opts_concat): ... here.
23715 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23717 PR middle-end/56461
23718 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23720 2013-03-04 Martin Jambor <mjambor@suse.cz>
23722 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23723 all appropriate places.
23725 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23727 PR tree-optimization/56424
23728 * ipa-split.c (split_function): Do not set the RSO flag if result is
23729 not by reference and its type is a register type.
23731 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23733 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23734 (microblaze_legitimate_pic_operand): Likewise
23735 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23736 new function microblaze_legitimate_pic_operand
23737 * config/microblaze/microblaze-protos.h
23738 (microblaze_legitimate_pic_operand): Declare.
23740 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23742 * config/microblaze/predicates.md (call_insn_simple_operand):
23743 New predicate for supported rtx code types.
23744 * config/microblaze/microblaze.md (call_internal1): Use
23745 call_insn_simple_operand predicate.
23747 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23749 PR middle-end/56461
23750 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23751 partitions.ordered_remove.
23753 PR middle-end/56461
23754 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23755 vec_oprnds0.create (1) for modifier == NONE.
23757 PR middle-end/56461
23758 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23759 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23760 vec_oprnds1 right before pushing anything to it for
23763 PR middle-end/56461
23764 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23765 set nbbs to 0 instead of having separate code path.
23766 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23767 instead of false as last argument if returning NULL.
23769 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23771 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23772 the attribute is now called "target" instead of "option".
23773 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23774 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23775 attribute/pragma name for TARGET_OPTION_VALID_P and
23776 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23777 * doc/tm.texi: Regenerated.
23779 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23781 * config/microblaze/microblaze.c:
23782 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23783 * config/microblaze/microblaze.h: Add -mxl-reorder to
23785 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23786 instructions emitted if TARGET_REORDER.
23787 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23788 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23791 2013-03-01 Xinliang David Li <davidxl@google.com>
23793 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23796 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23798 PR middle-end/56461
23799 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23800 vector even when returning true. Fix up function comment formatting.
23802 PR middle-end/56461
23803 * ira-build.c (ira_loop_nodes_count): New variable.
23804 (create_loop_tree_nodes): Initialize it.
23805 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23807 PR middle-end/56461
23808 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23809 method on dr_chain and result_chain.
23810 * tree-vect-stmts.c (vectorizable_store): Only call
23811 result_chain.create if j == 0.
23813 PR middle-end/56461
23814 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23815 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23816 before overwriting it.
23818 2013-03-01 Tobias Burnus <burnus@net-b.de>
23820 * doc/extended.texi (C Extensions): Change order in @menu
23822 (Other MIPS Built-in Functions): Move last MIPS entry before
23823 "picoChip Built-in Functions".
23824 (SH Built-in Functions): Move after RX Built-in Functions.
23825 * doc/gcc.texi (Introduction): Change order in @menu
23827 * doc/md.texi (Constraints): Ditto.
23828 * gty.texi (Type Information): Ditto.
23829 (User-provided marking routines for template types): Make
23831 * doc/invoke.texi (AArch64 Options): Move before
23832 "Adapteva Epiphany Options".
23834 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23835 Jakub Jelinek <jakub@redhat.com>
23838 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23839 no_address_safety_analysis attribute.
23840 * doc/extend.texi (no_address_safety_attribute): Rename to
23841 no_sanitize_address attribute, mention no_address_safety_analysis
23842 attribute as deprecated alias.
23844 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23846 PR middle-end/56461
23847 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23848 type to vec<vec<tree> > *.
23849 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23850 to be vec<tree> instead of vec<tree> *, set vec_defs
23851 to vNULL and call vec_defs.create (number_of_vects), adjust other
23853 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23854 vectorizable_condition): Adjust vect_get_slp_defs callers.
23856 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23858 * config/aarch64/aarch64.c
23859 (aarch64_float_const_representable): Remove unused variable.
23861 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23863 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23865 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23867 * config/aarch64/aarch64-builtins.c
23868 (aarch64_init_simd_builtins): Make static.
23870 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23872 * config/aarch64/aarch64.c
23873 (aarch64_simd_make_constant): Make static.
23875 2013-02-28 Martin Jambor <mjambor@suse.cz>
23877 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23878 with no initialization to the RHS of debug statements.
23880 2013-02-28 Martin Jambor <mjambor@suse.cz>
23882 PR tree-optimization/56294
23883 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23885 (get_access_replacement): Do not call create_access_replacement.
23886 Assert a replacement exists.
23887 (get_repl_default_def_ssa_name): Create the replacement declaration
23890 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23892 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23893 final_end_function.
23895 2013-02-28 Marek Polacek <polacek@redhat.com>
23897 PR rtl-optimization/56466
23898 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23899 if we're changing a loop.
23900 (peel_loops_completely): Likewise.
23902 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23905 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23907 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23910 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23911 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23912 INTX_FTYPE_FX, FX_FTYPE_INTX.
23913 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23915 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23917 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23918 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23919 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23920 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23921 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23922 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23923 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23924 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23925 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23926 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23927 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23928 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23929 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23930 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23931 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23932 (avrxmega6): Increase max flash segments from 5 to 6.
23933 * config/avr/t-multilib: Regenerate.
23934 * config/avr/avr-tables.opt: Regenerate.
23935 * doc/avr-mmcu.texi: Regenerate.
23937 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23939 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23940 (avr_device_to_arch): Rename to avr_device_to_ld.
23941 (avr_device_to_as): New prototype.
23942 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23943 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23944 * config/avr/driver-avr.c (avr_device_to_as): New.
23945 (avr_device_to_arch): Rename to avr_device_to_ld.
23947 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23949 PR middle-end/56461
23950 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23951 method on dr_chain and result_chain.
23953 PR middle-end/56461
23954 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23955 pointer_set_destroy on not_executed_last_iteration.
23957 PR middle-end/56461
23958 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23960 PR middle-end/56461
23961 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23962 FOR_EACH_DEFINED_FUNCTION when freeing state.
23964 PR middle-end/56461
23965 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23967 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23970 PR middle-end/56461
23971 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23972 known_aggs[i].items and release known_aggs vector.
23974 PR middle-end/56461
23975 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23977 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23979 * config/microblaze/microblaze.c (microblaze_emit_compare):
23980 Use xor for EQ/NE comparisions.
23981 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23982 (cbranchsf4): Adjust operator to comparison_operator.
23984 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23986 PR middle-end/56461
23987 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23989 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23990 vec_safe_push, always update *slot.
23991 (redirect_edge_var_map_clear): Use vec_free.
23992 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23993 (free_var_map_entry): Use vec_free.
23994 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23995 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23997 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23999 PR middle-end/45472
24000 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24001 when the may_trap_p bit of the exprs being merged differs.
24002 Reorder tests for speculativeness in the logical and operator.
24004 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24006 * incpath.c (add_standard_paths): Use reconcat instead of concat
24007 where appropriate and avoid leaking memory.
24009 * opts.h: Include obstack.h.
24010 (opts_concat): New prototype.
24011 (opts_obstack): New declaration.
24012 * opts.c (opts_concat): New function.
24013 (opts_obstack): New variable.
24014 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24015 (finish_options): Use opts_concat instead of concat
24016 and XOBNEWVEC instead of XNEWVEC.
24017 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24018 generate_option): Likewise.
24019 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24020 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24023 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24024 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24026 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24028 PR middle-end/56461
24029 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24031 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24033 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24034 (arm_block_move_unaligned_straight): Likewise.
24035 (arm_adjust_block_mem): Likewise.
24037 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24040 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24041 temp, cond and label.
24042 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24045 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24046 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24049 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24050 of prologue/epilogue functions.
24053 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24054 Remove unused variable cfa_offset.
24055 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24058 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24059 type promotion to unsigned.
24062 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24063 for HOST_WIDE_INT of 32 bit / same size as int.
24064 (arm_block_move_unaligned_straight): Likewise.
24065 (arm_adjust_block_mem): Likewise.
24068 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24071 2013-02-26 Marek Polacek <polacek@redhat.com>
24073 PR tree-optimization/56426
24074 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24076 2013-02-26 Richard Biener <rguenther@suse.de>
24079 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24080 unused variable loops.
24082 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24084 PR tree-optimization/56448
24085 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24086 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24087 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24088 later operands of the references, or even first operand for
24089 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24091 PR tree-optimization/56443
24092 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24093 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24094 to type_for_mode langhook.
24096 2013-02-25 Matt Turner <mattst88@gmail.com>
24098 * doc/invoke.texi: Document r4700.
24100 2013-02-25 Richard Biener <rguenther@suse.de>
24102 PR tree-optimization/56175
24103 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24105 (simplify_bitwise_binary): ... here. Also guard the conversion
24106 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24108 2013-02-25 Catherine Moore <clm@codesourcery.com>
24111 2013-02-24 Catherine Moore <clm@codesourcery.com>
24112 Maciej W. Rozycki <macro@codesourcery.com>
24113 Tom de Vries <tom@codesourcery.com>
24114 Nathan Sidwell <nathan@codesourcery.com>
24115 Iain Sandoe <iain@codesourcery.com>
24116 Nathan Froyd <froydnj@codesourcery.com>
24117 Chao-ying Fu <fu@mips.com>
24119 * doc/extend.texi: (micromips, nomicromips, nocompression):
24120 Document new function attributes.
24121 * doc/invoke.texi (minterlink-compressed, mmicromips,
24122 m14k, m14ke, m14kec): Document new options.
24123 (minterlink-mips16): Update documentation.
24124 * doc/md.texi (ZC, ZD): Document new constraints.
24125 * configure.ac (gcc_cv_as_micromips): Check if linker
24126 supports the .set micromips directive.
24127 * configure: Regenerate.
24128 * config.in: Regenerate.
24129 * config/mips/mips-tables.opt: Regenerate.
24130 * config/mips/micromips.md: New file.
24131 * constraints.md (ZC, AD): New constraints.
24132 * config/mips/predicates.md (movep_src_register): New predicate.
24133 (movep_src_operand): New predicate.
24134 (non_volatile_mem_operand): New predicate.
24135 * config/mips/mips.md (multimem): New type.
24136 (length): Differentiate between 17-bit and 18-bit branch offsets.
24137 (MOVEP1, MOVEP2): New mode iterator.
24138 (mov_<load>l): Use ZC constraint.
24139 (mov_<load>r): Likewise.
24140 (mov_<store>l): Likewise.
24141 (mov_<store>r): Likewise.
24142 (*branch_equality<mode>_inverted): Add microMIPS support.
24143 (*branch_equality<mode>): Likewise.
24144 (*jump_absolute): Likewise.
24145 (indirect_jump_<mode>): Likewise.
24146 (tablejump_<mode>): Likewise.
24147 (<optab>_internal): Likewise.
24148 (sibcall_internal): Likewise.
24149 (sibcall_value_internal): Likewise.
24150 (prefetch): Use constraint ZD.
24151 * config/mips/mips.opt (minterlink-compressed): New option.
24152 (minterlink-mips16): Now an alias for minterlink-compressed.
24153 (mmicromips): New option.
24154 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24155 (compare_and_swap_12): Likewise.
24156 (sync_add<mode>): Likewise.
24157 (sync_<optab>_12): Likewise.
24158 (sync_old_<optab>_12): Likewise.
24159 (sync_new_<optab>_12): Likewise.
24160 (sync_nand_12): Likewise.
24161 (sync_old_nand_12): Likewise.
24162 (sync_new_nand_12): Likewise.
24163 (sync_sub<mode>): Likewise.
24164 (sync_old_add<mode>): Likewise.
24165 (sync_old_sub<mode>): Likewise.
24166 (sync_new_add<mode>): Likewise.
24167 (sync_new_sub<mode>): Likewise.
24168 (sync_<optab><mode>): Likewise.
24169 (sync_old_<optab><mode>): Likewise.
24170 (sync_new_<optab><mode>): Likewise.
24171 (sync_nand<mode>): Likewise.
24172 (sync_old_nand<mode>): Likewise.
24173 (sync_new_nand<mode>): Likewise.
24174 (sync_lock_test_and_set<mode>): Likewise.
24175 (test_and_set_12): Likewise.
24176 (atomic_compare_and_swap<mode>): Likewise.
24177 (atomic_exchange<mode>_llsc): Likewise.
24178 (atomic_fetch_add<mode>_llsc): Likewise.
24179 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24180 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24181 (umips_save_restore_pattern_p): Likewise.
24182 (umips_load_store_pair_p): Likewise.
24183 (umips_output_load_store_pair): Likewise.
24184 (umips_movep_target_p): Likewise.
24185 (umips_12bit_offset_address_p): Likewise.
24186 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24187 (mips_base_mips16): Rename this...
24188 (mips_base_compression_flags): ...to this. Update all uses.
24189 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24190 (mips_mips16_decl_p): Delete.
24191 (mips_nomips16_decl_p): Delete.
24192 (mips_get_compress_on_flags): New function.
24193 (mips_get_compress_off_flags): New function.
24194 (mips_get_compress_mode): New function.
24195 (mips_get_compress_on_name): New function.
24196 (mips_get_compress_off_name): New function.
24197 (mips_insert_attributes): Support multiple compression types.
24198 (mips_merge_decl_attributes): Likewise.
24199 (umips_12bit_offset_address_p): New function.
24200 (mips_start_function_definition): Emit .set micromips directive.
24201 (mips_call_may_need_jalx_p): New function.
24202 (mips_function_ok_for_sibcall): Add microMIPS support.
24203 (mips_print_operand_punctuation): Support short delay slots and
24205 (umips_swm_mask, umips_swm_encoding): New.
24206 (umips_build_save_restore): New function.
24207 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24208 (was_mips16_p): Remove.
24209 (old_compression_mode): New.
24210 (mips_set_compression_mode): New function.
24211 (mips_set_current_function): Add microMIPS support.
24212 (mips_option_override): Likewise.
24213 (umips_save_restore_pattern_p): New function.
24214 (umips_output_save_restore): New function.
24215 (umips_load_store_pair_p_1): New function.
24216 (umips_load_store_pair_p): New function.
24217 (umips_output_load_store_pair_1): New function.
24218 (umips_output_load_store_pair): New function.
24219 (umips_movep_target_p) New function.
24220 (mips_prepare_pch_save): Add microMIPS support.
24221 * config/mips/mips.h (TARGET_COMPRESSION): New.
24222 (TARGET_CPU_CPP_BUILTINS): Update macro
24223 to use new compression flags and to support microMIPS.
24224 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24225 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24226 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24227 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24228 (ASM_SPEC): Support mmicromips and mno-micromips.
24229 (M16STORE_REG_P): New macro.
24230 (MIPS_CALL): Support TARGET_MICROMIPS.
24231 (MICROMIPS_J): New macro.
24232 (mips_base_mips16): Rename this...
24233 (mips_base_compression_flags): ...to this.
24234 (UMIPS_12BIT_OFFSET_P): New macro.
24235 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24236 (MULTILIB_DIRNAMES): Likewise.
24238 2013-02-25 Tom de Vries <tom@codesourcery.com>
24240 PR rtl-optimization/56131
24241 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24242 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24243 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24245 2013-02-25 Tobias Burnus <burnus@net-b.de>
24247 * doc/invoke.texi (-fsanitize=): Move from optimization
24248 to debugging options.
24250 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24252 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24254 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24255 Alexander Monakov <amonakov@ispras.ru>
24257 PR middle-end/56077
24258 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24259 flush pending lists also on non-jumps. Adjust comment.
24261 2013-02-24 Catherine Moore <clm@codesourcery.com>
24262 Maciej W. Rozycki <macro@codesourcery.com>
24263 Tom de Vries <tom@codesourcery.com>
24264 Nathan Sidwell <nathan@codesourcery.com>
24265 Iain Sandoe <iain@codesourcery.com>
24266 Nathan Froyd <froydnj@codesourcery.com>
24267 Chao-ying Fu <fu@mips.com>
24269 * doc/extend.texi: (micromips, nomicromips, nocompression):
24270 Document new function attributes.
24271 * doc/invoke.texi (minterlink-compressed, mmicromips,
24272 m14k, m14ke, m14kec): Document new options.
24273 (minterlink-mips16): Update documentation.
24274 * doc/md.texi (ZC, ZD): Document new constraints.
24275 * configure.ac (gcc_cv_as_micromips): Check if linker
24276 supports the .set micromips directive.
24277 * configure: Regenerate.
24278 * config.in: Regenerate.
24279 * config/mips/mips-tables.opt: Regenerate.
24280 * config/mips/micromips.md: New file.
24281 * constraints.md (ZC, AD): New constraints.
24282 * config/mips/predicates.md (movep_src_register): New predicate.
24283 (movep_src_operand): New predicate.
24284 (non_volatile_mem_operand): New predicate.
24285 * config/mips/mips.md (multimem): New type.
24286 (length): Differentiate between 17-bit and 18-bit branch offsets.
24287 (MOVEP1, MOVEP2): New mode iterator.
24288 (mov_<load>l): Use ZC constraint.
24289 (mov_<load>r): Likewise.
24290 (mov_<store>l): Likewise.
24291 (mov_<store>r): Likewise.
24292 (*branch_equality<mode>_inverted): Add microMIPS support.
24293 (*branch_equality<mode>): Likewise.
24294 (*jump_absolute): Likewise.
24295 (indirect_jump_<mode>): Likewise.
24296 (tablejump_<mode>): Likewise.
24297 (<optab>_internal): Likewise.
24298 (sibcall_internal): Likewise.
24299 (sibcall_value_internal): Likewise.
24300 (prefetch): Use constraint ZD.
24301 * config/mips/mips.opt (minterlink-compressed): New option.
24302 (minterlink-mips16): Now an alias for minterlink-compressed.
24303 (mmicromips): New option.
24304 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24305 (compare_and_swap_12): Likewise.
24306 (sync_add<mode>): Likewise.
24307 (sync_<optab>_12): Likewise.
24308 (sync_old_<optab>_12): Likewise.
24309 (sync_new_<optab>_12): Likewise.
24310 (sync_nand_12): Likewise.
24311 (sync_old_nand_12): Likewise.
24312 (sync_new_nand_12): Likewise.
24313 (sync_sub<mode>): Likewise.
24314 (sync_old_add<mode>): Likewise.
24315 (sync_old_sub<mode>): Likewise.
24316 (sync_new_add<mode>): Likewise.
24317 (sync_new_sub<mode>): Likewise.
24318 (sync_<optab><mode>): Likewise.
24319 (sync_old_<optab><mode>): Likewise.
24320 (sync_new_<optab><mode>): Likewise.
24321 (sync_nand<mode>): Likewise.
24322 (sync_old_nand<mode>): Likewise.
24323 (sync_new_nand<mode>): Likewise.
24324 (sync_lock_test_and_set<mode>): Likewise.
24325 (test_and_set_12): Likewise.
24326 (atomic_compare_and_swap<mode>): Likewise.
24327 (atomic_exchange<mode>_llsc): Likewise.
24328 (atomic_fetch_add<mode>_llsc): Likewise.
24329 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24330 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24331 (umips_save_restore_pattern_p): Likewise.
24332 (umips_load_store_pair_p): Likewise.
24333 (umips_output_load_store_pair): Likewise.
24334 (umips_movep_target_p): Likewise.
24335 (umips_12bit_offset_address_p): Likewise.
24336 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24337 (mips_base_mips16): Rename this...
24338 (mips_base_compression_flags): ...to this. Update all uses.
24339 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24340 (mips_mips16_decl_p): Delete.
24341 (mips_nomips16_decl_p): Delete.
24342 (mips_get_compress_on_flags): New function.
24343 (mips_get_compress_off_flags): New function.
24344 (mips_get_compress_mode): New function.
24345 (mips_get_compress_on_name): New function.
24346 (mips_get_compress_off_name): New function.
24347 (mips_insert_attributes): Support multiple compression types.
24348 (mips_merge_decl_attributes): Likewise.
24349 (umips_12bit_offset_address_p): New function.
24350 (mips_start_function_definition): Emit .set micromips directive.
24351 (mips_call_may_need_jalx_p): New function.
24352 (mips_function_ok_for_sibcall): Add microMIPS support.
24353 (mips_print_operand_punctuation): Support short delay slots and
24355 (umips_swm_mask, umips_swm_encoding): New.
24356 (umips_build_save_restore): New function.
24357 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24358 (was_mips16_p): Remove.
24359 (old_compression_mode): New.
24360 (mips_set_compression_mode): New function.
24361 (mips_set_current_function): Add microMIPS support.
24362 (mips_option_override): Likewise.
24363 (umips_save_restore_pattern_p): New function.
24364 (umips_output_save_restore): New function.
24365 (umips_load_store_pair_p_1): New function.
24366 (umips_load_store_pair_p): New function.
24367 (umips_output_load_store_pair_1): New function.
24368 (umips_output_load_store_pair): New function.
24369 (umips_movep_target_p) New function.
24370 (mips_prepare_pch_save): Add microMIPS support.
24371 * config/mips/mips.h (TARGET_COMPRESSION): New.
24372 (TARGET_CPU_CPP_BUILTINS): Update macro
24373 to use new compression flags and to support microMIPS.
24374 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24375 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24376 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24377 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24378 (ASM_SPEC): Support mmicromips and mno-micromips.
24379 (M16STORE_REG_P): New macro.
24380 (MIPS_CALL): Support TARGET_MICROMIPS.
24381 (MICROMIPS_J): New macro.
24382 (mips_base_mips16): Rename this...
24383 (mips_base_compression_flags): ...to this.
24384 (UMIPS_12BIT_OFFSET_P): New macro.
24385 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24386 (MULTILIB_DIRNAMES): Likewise.
24388 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24391 * target-globals.c (save_target_globals): For init_reg_sets and
24392 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24394 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24396 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24397 * config/aarch64/t-aarch64
24398 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24400 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24402 PR inline-asm/56148
24403 * lra-constraints.c (process_alt_operands): Reload operand
24404 conflicting with earlier clobber only if no more other conflicting
24407 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24410 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24411 if not linking a shared library.
24413 2013-02-22 Seth LaForge <sethml@google.com>
24415 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24417 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24419 * config/arm/arm.md (split for extendsidi): Update condition.
24420 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24421 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24422 (qhs_zextenddi_cstr): Likewise.
24424 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24426 PR middle-end/56420
24427 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24428 avoid signed wrapping.
24429 (expand_mult): Handle properly multiplication by
24430 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24431 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24432 in the compiler if coeff is HOST_WIDE_INT_MIN.
24433 (expand_divmod): Don't make ext_op1 static, change it's type to
24434 uhwi. Avoid undefined behavior in -INTVAL (op1).
24436 PR rtl-optimization/50339
24437 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24439 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24440 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24441 into splitting_ashiftrt field.
24442 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24444 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24447 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24449 PR middle-end/56108
24450 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24451 are sure to go irrevocable.
24453 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24455 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24456 scalars are valid operands.
24458 2013-02-21 Martin Jambor <mjambor@suse.cz>
24460 PR tree-optimization/56310
24461 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24462 only matching indices and non-negative final offsets.
24463 (intersect_aggregates_with_edge): Pass src_idx to
24464 agg_replacements_to_vector. Pass src_idx insstead of index to
24465 intersect_with_agg_replacements.
24467 2013-02-21 Martin Jambor <mjambor@suse.cz>
24469 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24470 instead of hard-wired defaults.
24472 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24474 * doc/invoke.texi (MIPS Options): Update documentation of the
24475 floating-point multiply-accumulate instruction restrictions.
24477 2013-02-21 Kostya Serebryany <kcc@google.com>
24479 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24480 asan_shadow_offset on x86_64 linux.
24482 2013-02-21 Richard Biener <rguenther@suse.de>
24484 PR tree-optimization/56415
24486 2013-02-11 Richard Biener <rguenther@suse.de>
24488 PR tree-optimization/56273
24489 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24492 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24495 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24498 PR inline-asm/56405
24499 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24500 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24502 2013-02-20 Jan Hubicka <jh@suse.cz>
24504 PR tree-optimization/56265
24505 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24506 when target is referenced for first time.
24508 2013-02-20 Richard Biener <rguenther@suse.de>
24510 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24511 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24512 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24513 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24514 not return anything.
24515 (rename_ssa_copies): Do not remove unused locals.
24516 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24517 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24518 * passes.c (execute_function_todo): Do not schedule unused locals
24519 removal if cleanup_tree_cfg did something.
24520 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24521 about the number of removed locals.
24523 2013-02-20 Richard Biener <rguenther@suse.de>
24525 PR tree-optimization/56398
24526 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24528 2013-02-20 Martin Jambor <mjambor@suse.cz>
24530 PR tree-optimization/55334
24531 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24532 restricted pointers to arrays.
24534 2013-02-20 Richard Biener <rguenther@suse.de>
24535 Jakub Jelinek <jakub@redhat.com>
24537 PR tree-optimization/56396
24538 * tree-ssa-ccp.c (n_const_val): New static variable.
24539 (get_value): Return NULL for SSA names we don't have a lattice
24541 (ccp_initialize): Initialize n_const_val.
24542 * tree-ssa-copy.c (n_copy_of): New static variable.
24543 (init_copy_prop): Initialize n_copy_of.
24544 (get_value): Return NULL_TREE for SSA names we don't have a
24547 2013-02-20 Martin Jambor <mjambor@suse.cz>
24549 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24551 2013-02-20 Richard Biener <rguenther@suse.de>
24553 * genpreds.c (write_lookup_constraint): Do not compare first
24554 letter of the constraint again.
24556 2013-02-20 Richard Biener <rguenther@suse.de>
24558 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24560 (get_use_iv_cost): Terminate hashtable walk when coming across
24563 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24565 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24566 reassociation for avx2 targets.
24568 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24570 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24571 Add version check for v8.10.a to enable microblaze_has_clz
24572 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24573 version and TARGET_PATTERN_COMPARE check
24574 * config/microblaze/microblaze.md: New clzsi2 instruction
24576 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24578 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24579 function before branching.
24581 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24583 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24585 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24587 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24589 PR middle-end/55889
24590 * sel-sched.c: Include ira.h.
24591 (implicit_clobber_conflict_p): New function.
24592 (moveup_expr): Use it.
24593 * Makefile.in (sel-sched.o): Depend on ira.h.
24595 2013-02-19 Richard Biener <rguenther@suse.de>
24597 PR tree-optimization/56384
24598 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24599 (vn_hash_type): Split out from ...
24600 (vn_hash_constant_with_type): ... here.
24601 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24602 (vn_phi_eq): Compare types from vn_phi_s structure.
24603 (vn_phi_lookup): Populate vn_phi_s type.
24604 (vn_phi_insert): Likewise.
24606 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24608 PR tree-optimization/56350
24609 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24610 if haven't found reduction or nested cycle operand, rather than
24611 asserting we must find it.
24613 PR tree-optimization/56381
24614 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24617 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24618 Jakub Jelinek <jakub@redhat.com>
24621 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24622 (swap_optab_enable): Same.
24623 (init_all_optabs): Use argument instead of global.
24624 * tree.h (struct tree_optimization_option): New field target_optabs.
24625 * expr.h (init_all_optabs): Add argument to prototype.
24626 (TREE_OPTIMIZATION_OPTABS): New.
24627 (save_optabs_if_changed): Protoize.
24628 * optabs.h: Declare this_fn_optabs.
24629 * optabs.c (save_optabs_if_changed): New.
24630 Declare this_fn_optabs.
24631 (init_optabs): Add argument to init_all_optabs() call.
24632 * function.c (invoke_set_current_function_hook): Handle per
24634 * function.h (struct function): New field optabs.
24635 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24636 optimization_current_node has changed.
24637 * target-globals.h (save_target_globals_default_opts): Protoize.
24638 * target-globals.c (save_target_globals_default_opts): New.
24640 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24643 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24644 registers %fr12 and %fr12R as call used.
24647 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24648 and HImode, require all displacements to be an integer multiple of
24650 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24651 only allow QImode and HImode when reload is in progress and strict is
24652 true. Likewise for symbolic addresses. Use base14_operand to check
24653 displacements in REG+BASE addresses.
24655 2013-02-18 Richard Biener <rguenther@suse.de>
24657 PR tree-optimization/56366
24658 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24659 handle sign-conversion of outer-loop initial induction value.
24661 2013-02-18 Richard Biener <rguenther@suse.de>
24663 PR middle-end/56349
24664 * cfghooks.c (merge_blocks): If we merge a latch into another
24665 block adjust references to it.
24666 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24667 (verify_loop_structure): Verify that a recorded latch is in fact
24670 2013-02-18 Richard Biener <rguenther@suse.de>
24672 PR tree-optimization/56321
24673 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24674 order SSA name release and virtual operand unlinking.
24676 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24678 * config/microblaze/microblaze.md (save_stack_block): Define.
24679 (restore_stack_block): Likewise.
24681 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24683 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24684 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24685 * config/microblaze/microblaze.c (microblaze_option_override):
24686 Bail out early for PIC modes when target does not support PIC.
24688 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24690 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24691 Replace with a microblaze version.
24692 (microblaze_trampoline_init): Adapt for microblaze.
24693 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24696 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24697 Dodji Seketeli <dodji@redhat.com>
24700 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24701 (instrument_mem_region_access): Do not forget to always put
24702 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24703 0) statement, even for cases where either 'base' or 'base + len'
24704 are not instrumented -- because they have been previously
24705 instrumented. Simplify the logic by putting all the statements
24706 instrument 'base + len' inside a sequence, and then insert that
24707 sequence right before the current insertion point. Then, to
24708 instrument 'base + len', just get an iterator on that statement.
24709 And do not forget to update the pointer to iterator the function
24710 received as argument.
24712 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24714 PR rtl-optimization/56348
24715 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24717 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24719 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24720 (clean_graph_dump_file): Pass base to start_graph_dump.
24722 2013-02-14 Richard Henderson <rth@redhat.com>
24725 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24727 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24729 * collect2-aix.h: Define F_LOADONLY.
24731 2013-02-14 Richard Biener <rguenther@suse.de>
24734 * varasm.c (output_constant_def_1): Get the decl representing
24735 the constant as argument.
24736 (output_constant_def): Wrap output_constant_def_1.
24737 (make_decl_rtl): Use output_constant_def_1 with the decl
24738 representing the constant.
24739 (build_constant_desc): Optionally re-use a decl already
24740 representing the constant.
24741 (tree_output_constant_def): Adjust.
24743 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24746 * asan.c (instrument_builtin_call): Really put the length of the
24747 second source argument into src1_len.
24749 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24751 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24752 argument. If it is false, don't create edge from then_bb to
24754 (insert_if_then_before_iter): Pass true to it.
24755 (build_check_stmt): Pass false to it.
24756 (transform_statements): Flush hash table only on extended basic
24757 block boundaries, rather than at the beginning of every bb.
24758 Don't flush hash table on nonfreeing_call_p calls.
24759 * tree-flow.h (nonfreeing_call_p): New prototype.
24760 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24762 2013-02-13 David S. Miller <davem@davemloft.net>
24764 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24766 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24769 * ira.c (max_regno_before_ira): Move from ...
24771 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24774 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24776 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24778 2013-02-13 Richard Biener <rguenther@suse.de>
24781 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24782 globals in MEM_REFs.
24784 2013-02-13 Richard Biener <rguenther@suse.de>
24786 * loop-init.c (loop_optimizer_init): Clear loop state when
24787 re-initializing preserved loops.
24788 * loop-unswitch.c (unswitch_single_loop): Return whether
24789 we unswitched the loop. Do not verify loop state here.
24790 (unswitch_loops): When we unswitched a loop discover new loops.
24792 2013-02-13 Kostya Serebryany <kcc@google.com>
24794 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24796 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24798 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24800 Avoid instrumenting duplicated memory access in the same basic block
24801 * Makefile.in (asan.o): Add new dependency on hash-table.h
24802 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24803 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24804 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24805 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24806 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24807 (get_mem_ref_of_assignment): New functions.
24808 (get_mem_refs_of_builtin_call): Extract from
24809 instrument_builtin_call and tweak a little bit to make it fit with
24811 (instrument_builtin_call): Use the new
24812 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24813 of is_gimple_builtin_call.
24814 (instrument_derefs, instrument_mem_region_access): Insert the
24815 instrumented memory reference into the hash table.
24816 (maybe_instrument_assignment): Renamed instrument_assignment into
24817 this, and change it to advance the iterator when instrumentation
24818 actually happened and return true in that case. This makes it
24819 homogeneous with maybe_instrument_assignment, and thus give a
24820 chance to callers to be more 'regular'.
24821 (transform_statements): Clear the memory reference hash table
24822 whenever we enter a new BB, when we cross a function call, or when
24823 we are done transforming statements. Use
24824 maybe_instrument_assignment instead of instrumentation. No more
24825 need to special case maybe_instrument_assignment and advance the
24826 iterator after calling it; it's now handled just like
24827 maybe_instrument_call. Update comment.
24829 2013-02-13 Richard Biener <rguenther@suse.de>
24831 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24832 Fix loop discovery code.
24834 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24836 PR inline-asm/56148
24837 * lra-constraints.c (process_alt_operands): Match early clobber
24838 operand with itself. Check conflicts with earlyclobber only if
24839 the operand is not reloaded. Prefer to reload conflicting operand
24840 if earlyclobber and matching operands are the same.
24842 2013-02-12 Richard Biener <rguenther@suse.de>
24845 * lto-streamer-out.c (write_symbol): Do not output symbols
24846 for hard register variables.
24848 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24851 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24852 (umulsidi3_insn, mulsidi3_insn): New insns.
24854 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24856 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24857 (struct tune_params): Add vec_costs field.
24858 * config/arm/arm.c (arm_builtin_vectorization_cost)
24859 (arm_add_stmt_cost): New functions.
24860 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24861 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24862 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24863 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24864 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24865 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24866 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24868 2013-02-12 Richard Biener <rguenther@suse.de>
24871 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24872 decls again if possible.
24874 2013-02-12 Richard Biener <rguenther@suse.de>
24876 PR middle-end/56288
24877 * tree-ssa.c (verify_ssa_name): Fix check, move
24878 SSA_NAME_IN_FREE_LIST check up.
24880 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24881 Steven Bosscher <steven@gcc.gnu.org>
24883 PR rtl-optimization/56151
24884 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24885 equal to op0 or op1, and last_insn pattern is CODE operation
24886 with MEM dest and one of the operands matches that MEM.
24888 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24890 * doc/extend.texi: Document Function Multiversioning and "default"
24891 parameter string to target attribute.
24892 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24893 target attribute parameter is "default".
24894 (ix86_compare_version_priority): Remove checks for target attribute.
24895 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24896 Remove check for target attribute equal to NULL. Add assert.
24897 (ix86_generate_version_dispatcher_body): Change error to sorry.
24899 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24900 Jack Howarth <howarth@bromo.med.uc.edu>
24901 Patrick Marlier <patrick.marlier@gmail.com>
24904 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24905 define ENDFILE_SPEC as TM_DESTRUCTOR.
24906 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24908 2013-02-11 Alexander Potapenko <glider@google.com>
24909 Jack Howarth <howarth@bromo.med.uc.edu>
24910 Jakub Jelinek <jakub@redhat.com>
24913 * config/darwin.c (cdtor_record): Rename ctor_record.
24914 (sort_cdtor_records): Rename sort_ctor_records.
24915 (finalize_dtors): New routine to sort destructors by
24916 priority before use in assemble_integer.
24917 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24919 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24921 PR rtl-optimization/56275
24922 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24923 offset is non-negative and less than cmode size before
24924 calling simplify_subreg.
24926 2013-02-11 Richard Biener <rguenther@suse.de>
24928 PR tree-optimization/56264
24929 * cfgloop.h (fix_loop_structure): Adjust prototype.
24930 * loop-init.c (fix_loop_structure): Return the number of
24931 newly discovered loops.
24932 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24933 are discovered, do a full loop-closed SSA rewrite.
24935 2013-02-11 Richard Biener <rguenther@suse.de>
24937 PR tree-optimization/56273
24938 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24940 (check_array_ref): Fix missing newline in dumps.
24941 (search_for_addr_array): Likewise.
24943 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24945 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24947 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24950 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24952 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24954 PR rtl-optimization/56246
24955 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24957 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24958 constraints are satisfied.
24960 2013-02-08 Jeff Law <law@redhat.com>
24963 * emit-rtl.c (reg_is_parm_p): New function.
24964 * regs.h (reg_is_parm_p): New prototype.
24965 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24966 callee-clobbered registers.
24968 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24971 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24972 If there is no implicit builtin declaration, just return NULL.
24974 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24976 * config/i386/sse.md (FMAMODEM): New mode iterator.
24977 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24978 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24980 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24982 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24983 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24984 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24986 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24988 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24989 (microblaze*-*-elf): Likewise.
24990 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24992 * config/microblaze/microblaze-c.c: Add builtin defines for
24993 _LITTLE_ENDIAN and _BIG_ENDIAN.
24994 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24995 add to TARGET_DEFAULT flags.
24996 Expand ASM_SPEC and LINK_SPEC.
24997 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24998 * config/microblaze/microblaze.md: Update extendsidi2 and
24999 movdi_internal instructions to use low-order / high-order reg
25001 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25002 options and inversemask / mask of LITTLE_ENDIAN.
25003 * config/microblaze/t-microblaze: Expand multilib options to
25004 include mlittle-endian (le) and update exceptions patterns.
25006 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25008 PR rtl-optimization/56195
25009 * lra-constraints.c (get_reload_reg): Don't reuse regs
25010 if they have smaller mode than requested, if they have
25011 wider mode than requested, try to return a SUBREG.
25013 PR tree-optimization/56250
25014 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25015 if type is unsigned and code isn't MULT_EXPR.
25017 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25019 PR tree-optimization/56064
25020 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25021 bits according to mode.
25022 * fixed-value.h (fixed_from_double_int)
25023 (const_fixed_from_double_int): Adjust comments.
25025 2013-02-08 Richard Biener <rguenther@suse.de>
25028 * lto-streamer.h (struct data_in): Remove current_file, current_line
25029 and current_col members.
25030 * lto-streamer-out.c (lto_output_location): Stream changed bits
25031 en-block for efficiency.
25032 * lto-streamer-in.c (clear_line_info): Remove.
25033 (lto_input_location): Cache current file, line and column
25034 globally via local statics. Read changed bits en-block.
25035 (input_function): Do not call clear_line_info.
25036 (lto_read_body): Likewise.
25037 (lto_input_toplevel_asms): Likewise.
25039 2013-02-08 Michael Matz <matz@suse.de>
25041 PR tree-optimization/52448
25042 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25043 (nt_call_phase): New static.
25044 (add_or_mark_expr): Only mark accesses with newer phase than any
25046 (nonfreeing_call_p): New.
25047 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25048 (nt_fini_block): Keep blocks marked as visited.
25049 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25051 2013-02-08 Richard Biener <rguenther@suse.de>
25053 * ira.c (ira): Free broken dominator information.
25055 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25057 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25059 2013-02-08 Marek Polacek <polacek@redhat.com>
25061 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25063 2013-02-08 Richard Biener <rguenther@suse.de>
25065 PR middle-end/56181
25066 * cfgloop.h (flow_loops_find): Adjust.
25067 (bb_loop_header_p): Declare.
25068 * cfgloop.c (bb_loop_header_p): New function split out from ...
25069 (flow_loops_find): ... here. Adjust function signature,
25070 support incremental loop structure update.
25071 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25072 * cfgloopmanip.c (fix_loop_structure): Move ...
25073 * loop-init.c (fix_loop_structure): ... here.
25074 (apply_loop_flags): Split out from ...
25075 (loop_optimizer_init): ... here.
25076 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25077 in incremental mode, only remove dead loops here.
25079 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25082 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25083 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25084 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25085 (*round<mode>3.libgcc): New insns for fixed-modes.
25086 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25087 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25088 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25089 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25090 implementations. Define to __builtin_avr_absFX,
25091 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25092 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25093 __builtin_avr_countlsFX, respectively.
25094 * config/avr/avr-c.c (target.h): Include it.
25095 (enum avr_builtin_id): New enum.
25096 (avr_resolve_overloaded_builtin): New static function.
25097 (avr_register_target_pragmas): Use it to set
25098 targetm.resolve_overloaded_builtin.
25099 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25100 tree nodes used by DEF_BUILTIN.
25101 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25102 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25103 <AVR_BUILTIN_xxBITS>: Same.
25105 2013-02-08 Richard Biener <rguenther@suse.de>
25107 * cfgloop.c (verify_loop_structure): Properly handle
25108 a loop exiting to another loop header.
25109 * ira-int.h (ira_loops): Remove.
25110 * ira.c (ira_loops): Remove.
25111 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25112 (do_reload): Use loop_optimizer_finalize.
25113 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25114 number_of_loops to access the loop tree.
25115 (more_one_region_p): Likewise.
25116 (finish_loop_tree_nodes): Likewise.
25117 (rebuild_regno_allocno_maps): Likewise.
25118 (mark_loops_for_removal): Likewise.
25119 (mark_all_loops_for_removal): Likewise.
25120 (remove_unnecessary_regions): Likewise.
25121 (ira_build): Likewise.
25122 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25124 2013-02-08 Richard Biener <rguenther@suse.de>
25126 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25127 * ipa-pure-const.c (analyze_function): Avoid calling
25128 mark_irreducible_loops twice.
25129 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25131 2013-02-07 David S. Miller <davem@davemloft.net>
25133 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25135 * var-tracking.c (vt_add_function_parameter): Test the presence of
25136 HAVE_window_save properly and do not remap argument registers when
25137 we have a leaf function.
25139 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25142 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25144 * config/i386/i386.c (ix86_print_operand): Ditto.
25146 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25148 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25150 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25152 PR rtl-optimization/56225
25153 * lra-constraints.c (process_alt_operands): Check that reload hard
25154 reg can hold value for strict_low_part.
25156 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25159 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25160 dwarf2out_end_function.
25161 (in_first_function_p, maybe_at_text_label_p,
25162 first_loclabel_num_not_at_text_label): New variables.
25163 (dwarf2out_var_location): In the first function find out
25164 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25165 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25168 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25170 PR rtl-optimization/56178
25171 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25172 SUBREG of a register. Tidy up related block of code.
25173 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25174 note if the source is a register or a SUBREG of a register.
25176 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25179 * config/rs6000/rs6000.md (ptrm): New mode attr.
25180 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25181 call_value_indirect_aix<pttrsize>,
25182 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25185 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25187 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25188 if -bnortl. Convert to strcmp and strncmp.
25190 2013-02-07 Alan Modra <amodra@gmail.com>
25193 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25194 addresses won't wrap when offsetting.
25195 (rs6000_secondary_reload): Provide secondary reloads needed for
25196 wrapping LO_SUM addresses.
25198 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25200 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25201 MACH, just __MACH__.
25203 2013-02-06 Richard Biener <rguenther@suse.de>
25205 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25206 instead of calling fix_loop_structure.
25208 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25210 PR middle-end/56217
25211 * omp-low.c (use_pointer_for_field): Return false if
25212 lower_send_shared_vars doesn't generate any copy-out code.
25214 2013-02-06 Tom de Vries <tom@codesourcery.com>
25216 PR rtl-optimization/56131
25217 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25218 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25219 of the label is NULL. Add comment.
25221 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25223 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25226 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25227 (STATIC_LIBTSAN_LIBS): Likewise.
25228 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25229 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25230 is defined, don't add anything else beyond that.
25231 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25232 (LINK_COMMAND_SPEC): Use them.
25234 PR tree-optimization/56205
25235 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25236 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25237 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25239 2013-02-05 Richard Biener <rguenther@suse.de>
25241 PR tree-optimization/53342
25242 PR tree-optimization/53185
25243 * tree-vectorizer.h (vect_check_strided_load): Remove.
25244 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25245 not disallow peeling for vectorized strided loads.
25246 (vect_check_strided_load): Make static and simplify.
25247 (vect_analyze_data_refs): Adjust.
25248 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25249 correctly when vectorizing strided loads.
25251 2013-02-05 Richard Biener <rguenther@suse.de>
25253 * doc/install.texi: Refer to ISL, not PPL.
25255 2013-02-05 Jan Hubicka <jh@suse.cz>
25257 PR tree-optimization/55789
25258 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25260 2013-02-05 Jan Hubicka <jh@suse.cz>
25262 PR tree-optimization/55789
25263 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25264 the dead call anyway.
25266 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25269 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25271 2013-02-04 Alexander Potapenko <glider@google.com>
25272 Jack Howarth <howarth@bromo.med.uc.edu>
25273 Jakub Jelinek <jakub@redhat.com>
25276 * config/darwin.c (sort_ctor_records): Stabilized qsort
25277 on constructor priority by using original position.
25278 (finalize_ctors): New routine to sort constructors by
25279 priority before use in assemble_integer.
25280 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25282 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25285 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25286 about visibility on artificial decls.
25287 * config/sol2.c (solaris_assemble_visibility): Likewise.
25289 2013-02-04 Kai Tietz <ktietz@redhat.com>
25292 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25293 argument and improve checking of return-argument types for 16-byte
25295 (ix86_function_value_1): Add additional valtype argument on call
25296 of function_value_64.
25297 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25298 handling infunction_value_64 function.
25300 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25302 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25304 2013-02-04 Richard Biener <rguenther@suse.de>
25306 PR tree-optimization/56188
25307 * tree-ssa-structalias.c (label_visit): Consider case with
25308 initially non-empty points-to set.
25309 (perform_var_substitution): Dump node mapping and clean up.
25311 2013-02-04 Richard Guenther <rguenther@suse.de>
25314 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25315 node prevail as last resort.
25316 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25317 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25319 2013-02-04 Richard Biener <rguenther@suse.de>
25321 PR tree-optimization/56113
25322 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25324 (equiv_class_lookup_or_add): ... this.
25325 (label_visit): Adjust and fix error in previous patch.
25326 (perform_var_substitution): Adjust.
25328 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25330 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25331 * config/sh/sh4-300.md: Likewise.
25332 * config/sh/sh4a.md: Likewise.
25333 * config/sh/constraints.md: Likewise.
25334 * config/sh/sh.md: Likewise.
25335 * config/sh/netbsd-elf.h: Likewise.
25336 * config/sh/predicates.md: Likewise.
25337 * config/sh/sh-protos.h: Likewise.
25338 * config/sh/ushmedia.h: Likewise.
25339 * config/sh/linux.h: Likewise.
25340 * config/sh/sh.c: Likewise.
25341 * config/sh/superh.h: Likewise.
25342 * config/sh/elf.h: Likewise.
25343 * config/sh/sh4.md: Likewise.
25344 * config/sh/sh.h: Likewise.
25346 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25348 * config/pa/constraints.md: Adjust unused letters. Change "T"
25349 constraint to match_test floating_point_store_memory_operand().
25350 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25351 (base14_operand): New.
25352 (floating_point_store_memory_operand): New.
25353 (integer_store_memory_operand): Revise to use base14_operand and
25354 reg_plus_base_memory_operand.
25355 (move_dest_operand): Allow symbolic_memory_operands.
25356 (symbolic_memory_operand): Check for LO_SOM.
25357 (symbolic_operand): Change default case to break.
25358 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25359 CONST_DOUBLE values to be reloaded by putting them into memory when
25360 the destination is a floating point register.
25361 (movdf): Remove code to handle CONST_DOUBLE.
25363 (reload_indf_r1): New.
25364 (reload_insf_r1): New.
25365 Consistently use "Q" and "T" constraints with integer and floating
25366 point move instructions, respectively.
25367 (movdi): Remove FAIL.
25368 Change predicate for source operand unamed DImode move from
25369 general_operand to move_src_operand.
25370 (umulsidi3): Change predicate for destination operand to
25372 Likewise for similar unamed patterns.
25373 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25374 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25375 (hppa_legitimize_address): Simplify mask calculation.
25376 (pa_emit_move_sequence): Revised handling of secondary reloads from
25377 REG+D addresses for floating point loads and stores. Directly handle
25378 loading CONST0_RTX (mode) to a floating point register.
25379 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25380 to floating point registers. Don't restrict secondary reloads to
25381 floating point registers to integer modes. Revise some comments and
25383 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25384 (pa_legitimate_address_p): New.
25385 (pa_legitimize_reload_address): New.
25386 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25387 (STRICT_REG_OK_FOR_BASE_P): New.
25388 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25389 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25391 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25392 Andrew Dixie <andrewd@gentrack.com>
25394 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25397 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25399 * expmed.c (extract_bit_field_1): Pass the full width of the
25400 structure to get_best_reg_extraction_insn.
25402 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25405 * configure.ac (use_cxa_atexit): Add AIX.
25406 * configure: Regenerate.
25408 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25410 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25413 * final.c (need_profile_function): New variable.
25414 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25415 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25416 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25417 notes, targetm.asm_out.function_prologue doesn't emit anything,
25418 HAVE_prologue and profiler should be emitted before prologue,
25419 set need_profile_function instead of emitting it.
25420 (final_scan_insn): If need_profile_function, emit
25421 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25422 NOTE_INSN_FUNCTION_BEG note.
25424 2013-02-01 Richard Henderson <rth@redhat.com>
25426 * config/rs6000/rs6000.md (smulditi3): New.
25429 * config/alpha/alpha.md (umulditi3): New.
25431 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25433 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25434 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25436 2013-02-01 Richard Biener <rguenther@suse.de>
25438 PR tree-optimization/56113
25439 * tree-ssa-structalias.c (label_visit): Reduce work for
25440 single-predecessor nodes.
25442 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25444 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25445 range isn't testing for zero.
25447 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25449 PR middle-end/56113
25450 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25452 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25453 Nick Clifton <nickc@redhat.com>
25455 * config/v850/constraints.md (Q): Define as a memory constraint.
25456 * config/v850/predicates.md (label_ref_operand): New predicate.
25457 (e3v5_shift_operand): New predicate.
25458 (ior_operator): New predicate.
25459 * config/v850/t-v850: Add e3v5 multilib.
25460 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25461 (v850_gen_movdi): Prototype.
25462 * config/v850/v850.c: Add support for e3v5 architecture.
25463 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25465 (construct_save_jarl): Add e3v5 long JARL support.
25466 (v850_adjust_insn_length): New function. Adjust length of call
25467 insns when using e3v5 instructions.
25468 (v850_gen_movdi): New function: Generate instructions to move a
25470 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25471 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25472 (TARGET_USE_FPU): Enable for e3v5.
25473 (CONST_OK_FOR_W): New macro.
25474 (ADJUST_INSN_LENGTH): Define.
25475 * config/v850/v850.md (UNSPEC_LOOP): Define.
25476 (attr cpu): Add v850e3v5.
25477 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25478 (movdi): New pattern.
25479 (movdi_internal): New pattern.
25480 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25481 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25482 (cstoresf4): Likewise.
25483 (cstoredf4): Likewise.
25484 (insv): New pattern.
25485 (rotlso3_a): New pattern.
25486 (rotlsi3_b): New pattern
25487 (rotlsi3_v850e3v5): New pattern.
25488 (doloop_begin): New pattern.
25489 (fix_loop_counter): New pattern.
25490 (doloop_end): New pattern.
25491 (branch_normal): Add e3v5 long branch support.
25492 (branch_invert): Likewise.
25493 (branch_z_normal): Likewise.
25494 (branch_z_invert): Likewise.
25495 (branch_nz_normal): Likewise.
25496 (branch_nz_invert): Likewise.
25497 (call_internal_short): Add e3v5 register-indirect JARL support.
25498 (call_internal_long): Likewise.
25499 (call_value_internal_short): Likewise.
25500 (call_value_internal_long): Likewise.
25501 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25502 (mloop): New option.
25503 * config.gcc: Add support for configuring v840e3v5 target.
25504 * doc/invoke.texi: Document new v850 specific command line options.
25506 2013-01-31 Paul Koning <ni1d@arrl.net>
25510 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25511 children if parent is a class.
25512 (prune_unused_types_prune): Don't add DW_AT_declaration.
25514 2013-01-31 Richard Biener <rguenther@suse.de>
25516 PR tree-optimization/56157
25517 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25518 match up operand with SLP child.
25520 2013-01-31 Jason Merrill <jason@redhat.com>
25523 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25524 parameters the first time.
25525 (gen_scheduled_generic_parms_dies): Check completeness here.
25527 2013-01-31 Richard Biener <rguenther@suse.de>
25529 PR middle-end/53073
25530 * common.opt (faggressive-loop-optimizations): New flag,
25531 enabled by default.
25532 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25533 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25534 infer_loop_bounds_from_undefined by it.
25536 2013-01-31 Richard Biener <rguenther@suse.de>
25538 PR tree-optimization/56150
25539 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25540 visit virtual operands.
25541 (find_uses_to_rename_bb): Likewise.
25543 2013-01-31 Richard Biener <rguenther@suse.de>
25545 PR tree-optimization/56150
25546 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25547 mixed store non-store stmts.
25549 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25552 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25553 LIBASAN_EARLY_SPEC is defined.
25554 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25555 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25557 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25560 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25561 invalid args instead of ICEing on it.
25562 (ix86_valid_target_attribute_tree): Return error_mark_node if
25563 ix86_valid_target_attribute_inner_p failed.
25564 (ix86_valid_target_attribute_p): Return false only if
25565 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25566 target("default") attribute.
25567 (sorted_attr_string): Change argument from const char * to tree,
25568 merge in all target attribute arguments rather than just one.
25569 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25570 instead of free. Avoid using strcat.
25571 (ix86_mangle_function_version_assembler_name): Mangle
25572 target("default") as if no target attribute is present. Adjust
25573 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25574 instead of xmalloc and XDELETEVEC instead of free.
25575 (ix86_function_versions): Don't return true if one of the decls
25576 doesn't have target attribute. If they don't and one of the decls
25577 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25578 sorted_attr_string caller. Use XDELETEVEC instead of free.
25579 (ix86_supports_function_versions): Remove.
25580 (make_name): Fix up formatting.
25581 (make_dispatcher_decl): Remove resolver_name and its initialization.
25582 Avoid leaking memory.
25583 (is_function_default_version): Return true if there is
25584 target("default") attribute rather than no target attribute at all.
25585 (make_resolver_func): Avoid leaking memory.
25586 (ix86_generate_version_dispatcher_body): Likewise.
25587 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25588 * target.def (supports_function_versions): Remove.
25589 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25590 * doc/tm.texi: Regenerated.
25592 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25594 PR rtl-optimization/56144
25595 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25596 for values with side effects.
25598 2013-01-30 Richard Biener <rguenther@suse.de>
25600 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25601 (sparseset_pop): Likewise.
25602 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25603 to be able to use quick_push in the worker loop.
25605 2013-01-30 Marek Polacek <polacek@redhat.com>
25607 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25609 2013-01-30 Richard Biener <rguenther@suse.de>
25612 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25614 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25616 PR tree-optimization/56064
25617 * fixed-value.c (fixed_from_double_int): New function.
25618 * fixed-value.h (fixed_from_double_int): New prototype.
25619 (const_fixed_from_double_int): New static inline function.
25620 * fold-const.c (native_interpret_fixed): New static function.
25621 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25622 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25623 (native_encode_fixed): New static function.
25624 (native_encode_expr) <FIXED_CST>: Use it.
25625 (native_interpret_int): Move double_int worker code to...
25626 * double-int.c (double_int::from_buffer): ...this new static method.
25627 * double-int.h (double_int::from_buffer): Prototype it.
25629 2013-01-30 Richard Biener <rguenther@suse.de>
25631 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25632 New pointer-map and obstack.
25633 (init_alias_vars): Allocate pointer-map and obstack.
25634 (delete_points_to_sets): Free them.
25635 (find_what_var_points_to): Cache result.
25636 (find_what_p_points_to): Adjust for changed interface of
25637 find_what_var_points_to.
25638 (compute_points_to_sets): Likewise.
25639 (ipa_pta_execute): Likewise.
25641 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25643 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25644 * configure: Regenerate.
25645 * config.in: Regenerate.
25646 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25647 #nobits/#progbits if supported.
25649 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25652 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25653 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25654 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25656 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25658 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25659 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25661 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25663 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25664 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25666 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25668 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25670 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25671 * config/arm/cortex-a7.md: New bypasses using
25672 arm_mac_accumulator_is_result.
25674 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25676 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25677 (cortex_a7_neon_mla): Likewise.
25678 (cortex_a7_fpfmad): New reservation.
25679 (cortex_a7_fpmacs): Use ffmas and update required units.
25680 (cortex_a7_fpmuld): Update required units and latency.
25681 (cortex_a7_fpmacd): Likewise.
25682 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25683 (cortex_a7_neon). Likewise.
25684 (bypass) Update participating units.
25686 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25688 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25689 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25691 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25692 (vfp_fmul): Use ffmad.
25693 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25694 (cortex_r4_fmacd): Use ffmad.
25695 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25696 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25697 (cortex_a9_fmacd): Use ffmad.
25698 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25699 (cortex_a8_vfp_macd): Use ffmad.
25700 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25701 (cortex_a5_fpmacd): Use ffmad.
25702 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25703 (cortex_a15_vfp_macd): Use ffmad.
25704 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25706 2013-01-29 Jason Merrill <jason@redhat.com>
25709 * varasm.c (default_assemble_visibility): Don't warn about
25710 visibility on artificial decls.
25712 2013-01-29 Richard Biener <rguenther@suse.de>
25714 PR tree-optimization/56113
25715 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25717 (label_visit): Free duplicate bitmaps and record the leader instead.
25718 (perform_var_substitution): Adjust.
25720 2013-01-29 Richard Biener <rguenther@suse.de>
25722 PR tree-optimization/55270
25723 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25724 the CFG, schedule loops for fixup.
25726 2013-01-29 Nick Clifton <nickc@redhat.com>
25728 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25731 2013-01-28 Leif Ekblad <leif@rdos.net>
25733 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25734 * config/i386/i386.h (TARGET_RDOS): New macro.
25735 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25736 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25737 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25738 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25739 DEFAULT_LARGE_SECTION_THRESHOLD.
25740 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25741 * config/i386/rdos.h: New file.
25742 * config/i386/rdos64.h: New file.
25744 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25747 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25750 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25752 PR rtl-optimization/56117
25753 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25754 call cselib_lookup_from_insn on the MEM before calling
25755 add_insn_mem_dependence.
25757 2013-01-28 Richard Biener <rguenther@suse.de>
25759 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25760 to a stmt that didn't have one.
25761 (copy_phis_for_bb): Likewise for PHI arguments.
25762 (copy_debug_stmt): Likewise for debug stmts.
25764 2013-01-28 Richard Biener <rguenther@suse.de>
25766 PR tree-optimization/56034
25767 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25768 (partition_builtin_p): Adjust.
25769 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25770 it is the last partition.
25771 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25772 up the vertex for the definition.
25773 (classify_partition): Classify whether a partition is a
25774 PKIND_REDUCTION, thus has uses outside of the loop.
25775 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25776 Merge all PKIND_REDUCTION partitions into the last partition.
25777 (tree_loop_distribution): Seed partitions from reductions as well.
25779 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25781 PR tree-optimization/56125
25782 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25783 pow(x,c) into sqrt(x) * powi(x, n/2) or
25784 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25785 optimizing for size.
25786 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25787 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25790 PR tree-optimization/56094
25791 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25792 to UNKNOWN_LOCATION while gimplifying expr.
25794 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25797 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25798 operand 0 in movabs insn template for -masm=intel asm alternative.
25799 (*movabs<mode>_2): Ditto for operand 1.
25801 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25804 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25807 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25809 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25812 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25814 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25815 Undef to avoid warning.
25817 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25819 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25820 * configure: Regenerate.
25822 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25824 PR tree-optimization/56098
25825 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25826 for stmts with volatile ops.
25827 (cond_store_replacement): Don't optimize if assign has volatile ops.
25828 (cond_if_else_store_replacement_1): Don't optimize if either
25829 then_assign or else_assign have volatile ops.
25830 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25833 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25835 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25837 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25839 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25840 missing ':' in asm example.
25842 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25844 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25845 entries into lane and laneq entries.
25846 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25847 Remove AdvSIMD scalar modes.
25848 (aarch64_sq<r>dmulh_laneq<mode>): New.
25849 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25851 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25852 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25853 * config/aarch64/iterators.md (VCOND): New.
25856 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25859 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25860 Add NULL LIBNAME argument to existing definitions.
25861 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25862 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25863 * config/avr/avr.c (DEF_BUILTIN): Same.
25864 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25865 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25866 implementation is available (DECL_ASSEMBLER_NAME is set).
25867 (avr_fold_absfx): New static function.
25868 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25869 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25870 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25871 AVR_BUILTIN_ABSLLK.
25872 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25873 (abshk, absk, abslk, absllk): Provide as static inline functions.
25875 2013-01-25 Marek Polacek <polacek@redhat.com>
25877 PR tree-optimization/56035
25878 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25880 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25882 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25883 (*movtf_internal_rex64): Add (!o,C) alternative
25884 (*movxf_internal_rex64): Ditto.
25885 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25887 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25889 * doc/invoke.texi: fix typo.
25890 * doc/objc.texi: fix typo.
25892 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25894 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25895 for the first two alternatives.
25897 2013-01-24 Diego Novillo <dnovillo@google.com>
25899 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25900 (ggc-zone.o): Remove.
25901 * configure.ac: Remove option --with-gc.
25902 * configure: Re-generate.
25903 * doc/install.texi: Remove documentation for --with-gc.
25904 * gengtype.c (write_enum_defn): Remove. Update all users.
25905 (write_Types_process_field): Remove generation of gt_e_* argument.
25906 (output_type_enum): Remove. Update all users.
25907 (write_enum_defn): Remove. Update all users.
25908 (enum alloc_zone): Remove. Update all users.
25909 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25910 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25911 Update all callers.
25912 (struct ptr_data): Remove field TYPE. Update all users.
25913 (gt_pch_note_object): Remove argument TYPE. Update all users.
25914 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25916 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25917 (struct alloc_zone): Remove.
25918 (ggc_internal_alloc_zone_stat): Remove.
25919 (ggc_internal_cleared_alloc_zone_stat): Remove.
25920 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25921 (ggc_pch_count_object): Remove last argument. Update all users.
25922 (ggc_pch_alloc_object): Remove last argument. Update all users.
25923 (struct alloc_zone): Remove.
25924 * ggc-zone.c: Remove.
25925 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25926 (struct alloc_zone): Remove.
25927 (ggc_alloc_typed_stat): Remove.
25928 (ggc_alloc_typed): Remove.
25929 (ggc_splay_alloc): Remove first argument.
25930 (rtl_zone): Remove. Update all users.
25931 (tree_zone): Remove. Update all users.
25932 (tree_id_zone): Remove. Update all users.
25933 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25934 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25935 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25936 * tree-ssanames.c: Remove references to zone allocator in comments.
25938 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25940 * config/avr/avr.c (avr_out_fract): Make register numbers that
25941 might be outside of source operand signed.
25943 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25945 * config/i386/constraints.md (Yf): New constraint.
25946 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25947 of f constraint to conditionaly disable x87 register preferences.
25948 (*movdf_internal): Ditto.
25949 (*movsf_internal): Ditto.
25951 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25953 PR inline-asm/55934
25954 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25955 that have operands with impossible constraints.
25956 Add a FIXME for a speed-up opportunity.
25957 * lra-constraints.c (process_alt_operands): Verify that a class
25958 selected from constraints on asms is valid for the operand mode.
25959 (curr_insn_transform): Remove incorrect comment.
25961 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25963 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25964 TOC operand is a valid symbol ref in the constant pool.
25966 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25968 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25970 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25973 * config/avr/stdfix.h: New file.
25974 * t-avr (stdfix-gcc.h): New rule to build it.
25975 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25977 2013-01-23 Kostya Serebryany <kcc@google.com>
25979 * config/darwin.h: remove dependency on
25980 CoreFoundation (asan on Mac OS).
25982 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25985 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25986 instead of cmpdi_operand for first comparison operand.
25987 Don't assert that comparison operands aren't both constants.
25989 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25991 * doc/install.texi (Downloading the Source): Update references to
25992 downloading separate components.
25994 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25996 * doc/extend.texi (__int128): Improve grammar.
25998 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26001 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26002 alternative to (o,r).
26003 (*movdi_internal_rex64): Remove (!o,n) alternative.
26004 (DImode immediate->memory splitter): Remove.
26005 (DImode immediate->memory peephole2): Remove.
26006 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26007 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26008 alternative to (!o,*r).
26009 (*movtf_internal_sse): New pattern.
26010 (*movxf_internal_rex64): New pattern.
26011 (*movxf_internal): Disable for TARGET_64BIT.
26012 (*movdf_internal_rex64): Remove (!o,F) alternative.
26014 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26016 PR middle-end/56074
26017 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26018 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26019 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26020 stmt locations where LOCATION_LOCUS of the stmt location is
26021 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26024 * config/i386/i386.md (UNSPEC_STOS): New.
26025 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26026 *strsetqi_1): Add UNSPEC_STOS.
26028 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26031 * doc/invoke.texi: Remove left over -Wsynth example.
26033 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26035 PR tree-optimization/56051
26036 * fold-const.c (fold_binary_loc): Don't fold
26037 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26038 a narrowing conversion, or widening conversion from signed
26041 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26043 PR rtl-optimization/56023
26044 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26045 dependent on debug instruction.
26047 2013-01-21 Martin Jambor <mjambor@suse.cz>
26049 PR middle-end/56022
26050 * function.c (allocate_struct_function): Call
26051 invoke_set_current_function_hook earlier.
26053 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26055 * reload1.c (init_reload): Only initialize reload_obstack
26056 during the first call.
26058 2013-01-21 Marek Polacek <polacek@redhat.com>
26060 * cfgloop.c (verify_loop_structure): Fix up grammar.
26062 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26064 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26065 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26067 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26070 * config/arm/marvell-pj4.md: Update copyright year.
26071 Fix up use of alu to alu_reg and simple_alu_imm.
26073 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26075 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26077 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26080 * lra-constraints.c (curr_insn_transform): Don't reuse original
26081 insn for secondary memory move when memory mode should be different.
26083 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26085 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26086 atomic_storedi_1): New patterns.
26088 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26090 btver2 pipeline descriptions.
26091 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26093 * config/i386/i386.md (btver2_decode): New type attributes.
26094 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26096 * config/i386/btver2.md: New file describing btver2 pipelines.
26098 2013-01-19 Andrew Pinski <apinski@cavium.com>
26100 PR tree-optimization/52631
26101 * tree-ssa-sccvn (visit_use): Before looking up the original
26102 statement, try looking up the simplified expression.
26104 2013-01-19 Anthony Green <green@moxielogic.com>
26106 * config/moxie/moxie.c (moxie_expand_prologue): Set
26107 current_function_static_stack_size.
26109 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26111 PR tree-optimization/56029
26112 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26113 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26115 2013-01-18 Sharad Singhai <singhai@google.com>
26117 PR tree-optimization/55995
26118 * dumpfile.c (dump_loc): Print location only if available.
26119 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26121 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26124 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26125 secondary memory move.
26126 (inherit_reload_reg): Use rclass instead of cl for
26127 check_secondary_memory_needed_p.
26129 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26131 PR middle-end/56015
26132 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26133 the case where writing real complex part of target modifies op1.
26135 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26137 * config/aarch64/aarch64-simd.md
26138 (aarch64_vcond_internal<mode>): Handle unordered cases.
26139 * config/aarch64/iterators.md (v_cmp_result): New.
26141 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26142 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26144 * config/arm/marvell-pj4.md: New file.
26145 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26146 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26147 (generic_vfp): Likewise.
26148 * config/arm/arm-cores.def: Add marvell-pj4.
26149 * config/arm/arm-tune.md: Regenerate.
26150 * config/arm/arm-tables.opt: Regenerate.
26151 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26152 * doc/invoke.texi: Document marvell-pj4.
26154 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26156 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26158 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26163 * var-tracking.c (negative_power_of_two_p): New.
26164 (global_get_addr_cache, local_get_addr_cache): New.
26165 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26166 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26168 (vt_stack_offset_p): Remove.
26169 (vt_canon_true_dep): Always canonicalize loc's address.
26170 (clobber_overlapping_mems): Make sure we have a MEM.
26171 (local_get_addr_clear_given_value): New.
26172 (val_reset): Clear local cached entries.
26173 (compute_bb_dataflow): Create and release the local cache.
26174 Disable duplicate MEMs clobbering.
26175 (emit_notes_in_bb): Clobber MEMs likewise.
26176 (vt_emit_notes): Create and release the local cache.
26177 (vt_initialize, vt_finalize): Create and release the global
26178 cache, respectively.
26179 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26181 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26183 PR libmudflap/53359
26184 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26185 not found in the symtab.
26187 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26190 PR rtl-optimization/55547
26191 PR rtl-optimization/53827
26194 * alias.c (offset_overlap_p): New, factored out of...
26195 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26196 the conservative special case for symbolic constants. Don't
26197 adjust zero sizes on alignment.
26199 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26201 PR rtl-optimization/52573
26202 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26203 REG_UNUSED for the same register.
26205 2013-01-17 Richard Biener <rguenther@suse.de>
26206 Marek Polacek <polacek@redhat.com>
26208 PR rtl-optimization/55833
26209 * loop-unswitch.c (unswitch_loops): Move loop verification...
26210 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26211 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26212 Set it to true when we're removing a loop from hierarchy tree in
26213 an irreducible region.
26214 (fix_bb_placements): Adjust caller.
26215 (fix_loop_placements): Likewise.
26217 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26219 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26220 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26221 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26222 Remove ID. Adjust comments.
26223 * config/avr/avr-c.c (avr_builtin_name): Remove.
26224 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26225 * config/avr/avr.c (avr_tolower): New static function.
26226 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26227 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26228 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26231 2013-01-17 Jan Hubicka <jh@suse.cz>
26233 PR tree-optimization/55273
26234 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26236 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26239 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26240 store through atomic_store<mode>_1.
26241 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26243 2013-01-17 Martin Jambor <mjambor@suse.cz>
26245 PR tree-optimizations/55264
26246 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26247 for virtual methods.
26248 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26249 virtual methods before inlining is over.
26250 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26252 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26255 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26257 PR rtl-optimization/56005
26258 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26259 pending reads for prefetch.
26261 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26263 * config/aarch64/aarch64.md
26264 (*cstoresi_neg_uxtw): New pattern.
26265 (*cmovsi_insn_uxtw): New pattern.
26266 (*<optab>si3_uxtw): New pattern.
26267 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26268 (*<optab>si3_insn_uxtw): New pattern.
26269 (*bswapsi2_uxtw): New pattern.
26271 2013-01-16 Richard Biener <rguenther@suse.de>
26273 * tree-inline.c (tree_function_versioning): Remove set but
26274 never used variable.
26276 2013-01-16 Richard Biener <rguenther@suse.de>
26278 PR tree-optimization/55964
26279 * tree-flow.h (rename_variables_in_loop): Remove.
26280 (rename_variables_in_bb): Likewise.
26281 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26282 (copy_loop_before): Adjust and delete update-ssa status.
26283 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26284 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26285 (rename_variables_in_loop): Remove.
26286 (slpeel_update_phis_for_duplicate_loop): Likewise.
26287 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26288 use available cfg machinery instead of duplicating it.
26289 Update PHI nodes and perform poor-mans SSA update here.
26290 (slpeel_tree_peel_loop_to_edge): Adjust.
26292 2013-01-16 Richard Biener <rguenther@suse.de>
26294 PR tree-optimization/54767
26295 PR tree-optimization/53465
26296 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26297 (vrp_visit_phi_node): For PHI arguments coming via backedges
26298 drop all symbolical range information.
26299 (execute_vrp): Compute backedges.
26301 2013-01-16 Richard Biener <rguenther@suse.de>
26303 * doc/install.texi: Update CLooG and ISL requirements to
26306 2013-01-16 Christian Bruel <christian.bruel@st.com>
26309 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26310 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26311 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26313 2013-01-16 DJ Delorie <dj@redhat.com>
26315 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26316 (UNSPECV_SP_SWITCH_E): New.
26317 (sp_switch_1): Change to an unspec.
26318 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26321 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26323 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26324 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26325 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26326 (expand_mem_thread_fence): Ditto.
26327 (expand_mem_signal_fence): Ditto.
26328 (expand_atomic_load): Ditto.
26329 (expand_atomic_store): Ditto.
26331 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26333 PR rtl-optimization/55547
26334 PR rtl-optimization/53827
26337 * alias.c (memrefs_conflict_p): Set sizes to negative after
26340 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26343 * function.c (thread_prologue_and_epilogue_insns): Always
26344 add crtl->drap_reg to set_up_by_prologue.set, even if
26345 stack_realign_drap is false.
26347 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26349 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26350 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26351 *call): Fix indention.
26353 2013-01-15 Tom de Vries <tom@codesourcery.com>
26356 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26359 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26361 PR rtl-optimization/55153
26362 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26364 2013-01-15 Martin Jambor <mjambor@suse.cz>
26366 PR tree-optimization/55920
26367 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26368 accesses as grp_to_be_debug_replaced.
26370 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26372 PR tree-optimization/55920
26373 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26374 there is non-useless type conversion needed from debug rhs to lhs,
26375 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26377 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26378 Mikael Pettersson <mikpe@it.uu.se>
26381 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26383 (ASM_OUTPUT_CASE_LABEL): Remove.
26384 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26385 * final.c (shorten_branches): Update alignment of labels before
26386 jump tables if CASE_VECTOR_SHORTEN_MODE.
26388 2013-01-15 Richard Biener <rguenther@suse.de>
26391 * system.h: Do not include gmp.h for building host tools.
26393 2013-01-15 Richard Biener <rguenther@suse.de>
26395 PR middle-end/55882
26396 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26397 account for bitpos when computing alignment.
26399 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26401 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26402 (ix86_target_macros_internal): Likewise.
26404 * config/i386/i386.c (m_CORE2I7): Removed.
26405 (m_CORE_HASWELL): New macro.
26406 (m_CORE_ALL): Likewise.
26407 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26408 (initial_ix86_arch_features): Likewise.
26409 (processor_target_table): Initializations for Core avx2.
26410 (cpu_names): New names "core-avx2".
26411 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26412 PROCESSOR_CORE_HASWELL.
26413 (ix86_issue_rate): New case.
26414 (ia32_multipass_dfa_lookahead): Likewise.
26415 (ix86_sched_init_global): Likewise.
26417 * config/i386/i386.h (TARGET_HASWELL): New macro.
26418 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26419 (processor_type): New PROCESSOR_HASWELL.
26421 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26423 PR tree-optimization/55955
26424 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26425 *SHIFT_EXPR and *ROTATE_EXPR codes.
26427 PR tree-optimization/48766
26428 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26429 -ftrapv disable -fwrapv.
26431 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26434 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26435 etc. to 1 and not to __flash.
26436 Use LL suffix for __INT24_MAX__ with -mint8.
26437 Use ULL suffix for __UINT24_MAX__ with -mint8.
26439 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26441 * config/avr/avr-arch.h
26442 (struct base_arch_s): Use typedef avr_arch_t instead.
26443 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26444 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26445 * config/avr/avr.c: Same.
26446 * config/avr/avr-devices.c: Same.
26447 * config/avr/driver-avr.c: Same.
26448 * config/avr/gen-avr-mmcu-texi.c: Same.
26449 * config/avr/avr-mcus.def: Adjust comment.
26451 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26453 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26454 * config/aarch64/iterators.md (VALLDI): New.
26456 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26457 Andi Kleen <ak@linux.intel.com>
26460 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26461 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26464 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26466 * config/avr/avr-stdint.h: Remove trailing blanks.
26467 * config/avr/avr-log.h: Same.
26468 * config/avr/avr-arch.h: Same.
26469 * config/avr/avr-devices.c: Same.
26470 * config/avr/avr-dimode.md: Same.
26471 * config/avr/predicates.md: Same.
26472 * config/avr/avr-c.c: Same. And fix typo.
26474 * config/avr/avr-protos.h: Same. And:
26475 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26476 (init_cumulative_args): Rename to avr_init_cumulative_args.
26477 (expand_prologue): Rename to avr_expand_prologue.
26478 (expand_epilogue): Rename to avr_expand_epilogue.
26479 (adjust_insn_length): Rename to avr_adjust_insn_length.
26480 (notice_update_cc): Rename to avr_notice_update_cc.
26481 (final_prescan_insn): Rename to avr_final_prescan_insn.
26482 * config/avr/avr.c: Same.
26483 * config/avr/avr.h: Same.
26484 * config/avr/avr.md: Remove trailing blanks.
26485 (prologue): Use avr_expand_prologue.
26486 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26488 2013-01-14 Richard Biener <rguenther@suse.de>
26490 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26491 verify_location, collect_subblocks): New functions.
26492 (verify_gimple_in_cfg): Verify that locations only reference
26493 BLOCKs in the functions BLOCK tree.
26495 2013-01-14 Richard Biener <rguenther@suse.de>
26497 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26499 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26501 (insert_out_of_ssa_copy_on_edge): Likewise.
26502 (rewrite_close_phi_out_of_ssa): Likewise.
26503 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26505 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26506 propagated constants.
26507 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26510 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26512 * config/avr/avr-modes.def: Add GPL copyright notice.
26514 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26516 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26517 MEMMODEL_MASK to determine memory model.
26518 (atomic_store<mode>): Ditto from operands[2].
26519 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26521 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26524 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26525 (fold_gimple_assign): Don't call unshare_expr here.
26526 (fold_ctor_reference): Call unshare_expr.
26528 2013-01-13 Terry Guo <terry.guo@arm.com>
26530 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26531 * doc/fragments.texi: Document MULTILIB_REUSE.
26532 * gcc.c (multilib_reuse): New internal spec.
26533 (set_multilib_dir): Also search multilib from multilib_reuse.
26534 * genmultilib (tmpmultilib3): Refactor code.
26535 (tmpmultilib4): Ditto.
26536 (multilib_reuse): New multilib argument.
26538 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26540 * Makefile.in: Update copyright.
26542 2013-01-12 Tom de Vries <tom@codesourcery.com>
26544 PR middle-end/55890
26545 * calls.c (expand_call): Check if arg_nr is valid.
26547 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26549 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26550 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26551 documentation. Add missing '__' in front of
26552 __builtin_ia32_packssdw256.
26554 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26557 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26558 NO_REGS for larl operands.
26559 (s390_reload_larl_operand): Use s390_load_address instead of
26562 2013-01-11 Richard Biener <rguenther@suse.de>
26564 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26565 (verify_node_sharing): ... here.
26566 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26568 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26570 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26571 Substitute TREECHECKING.
26572 * configure: Regenerate.
26573 * Makefile.in (TREECHECKING): New.
26575 2013-01-11 Richard Guenther <rguenther@suse.de>
26577 PR tree-optimization/44061
26578 * tree-vrp.c (extract_range_basic): Compute zero as
26579 value-range for __builtin_constant_p of function parameters.
26581 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26583 Update copyright years.
26585 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26587 PR rtl-optimization/55672
26588 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26589 const to be eliminable.
26591 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26593 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26594 * configure: Regenerate.
26596 2013-01-10 Richard Biener <rguenther@suse.de>
26598 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26600 2013-01-10 Richard Biener <rguenther@suse.de>
26603 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26604 locations for virtual PHI arguments.
26605 (rewrite_update_phi_arguments): Likewise.
26607 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26609 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26612 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26614 PR tree-optimization/55921
26615 * tree-complex.c (expand_complex_asm): New function.
26616 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26618 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26621 * config/s390/s390.c (s390_symref_operand_p)
26622 (s390_loadrelative_operand_p): Merge the two functions.
26623 (s390_check_qrst_address, print_operand_address): Add parameters
26624 to s390_loadrelative_operand_p invokation.
26625 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26626 (s390_reload_larl_operand, s390_secondary_reload): Use
26627 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26628 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26630 2013-01-09 Mike Stump <mikestump@comcast.net>
26632 * dse.c (record_store): Remove unnecessary assert.
26634 2013-01-09 Jan Hubicka <jh@suse.cz>
26636 PR tree-optimization/55569
26637 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26638 * cfgloop.h (scale_loop_profile): Likewise.
26640 2013-01-09 Jan Hubicka <jh@suse.cz>
26643 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26645 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26647 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26649 PR middle-end/55114
26650 * expr.h (maybe_emit_group_store): Declare.
26651 * expr.c (maybe_emit_group_store): New function.
26652 * builtins.c (expand_builtin_int_roundingfn): Call it.
26653 (expand_builtin_int_roundingfn_2): Likewise.
26655 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26657 PR rtl-optimization/55829
26658 * lra-constraints.c (match_reload): Add code for absent output.
26659 (curr_insn_transform): Add code for reloads of matched inputs
26662 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26664 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26665 attribute of movddup insn to DF.
26666 (*vec_interleave_lowv2df): Ditto.
26667 (vec_dupv2df): Ditto.
26669 2013-01-09 Jan Hubicka <jh@suse.cz>
26671 PR tree-optimiation/55875
26672 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26673 EVERY_ITERATION parameter.
26674 (number_of_iterations_exit): Check if exit is executed every iteration.
26675 (idx_infer_loop_bounds): Similarly here.
26676 (n_of_executions_at_most): Simplify
26677 to only test for cases where statement is dominated by the
26678 particular bound; handle correctly the "postdominance" test.
26679 (scev_probably_wraps_p): Use max loop iterations info
26680 as a global bound first.
26682 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26683 Nick Clifton <nickc@redhat.com>
26685 * config/v850/v850.md (cbranchsf4): New pattern.
26686 (cstoresf4): New pattern.
26687 (cbranchdf4): New pattern.
26688 (cstoredf4): New pattern.
26689 (movsicc): Disallow floating point comparisons.
26690 (cmpsf_le_insn): Fix order of operators.
26691 (cmpsf_lt_insn): Likewise.
26692 (cmpsf_eq_insn): Likewise.
26693 (cmpdf_le_insn): Likewise.
26694 (cmpdf_lt_insn): Likewise.
26695 (cmpdf_eq_insn): Likewise.
26696 (cmpsf_ge_insn): Use LE comparison.
26697 (cmpdf_ge_insn): Likewise.
26698 (cmpsf_gt_insn): Use LT comparison.
26699 (cmpdf_gt_insn): Likewise.
26700 (cmpsf_ne_insn): Delete pattern.
26701 (cmpdf_ne_insn): Delete pattern.
26702 * config/v850/v850.c (v850_gen_float_compare): Use
26703 gen_cmpdf_eq_insn for NE comparison.
26704 (v850_float_z_comparison_operator)
26705 (v850_float_nz_comparison_operator): Move from here ...
26706 * config/v850/predicates.md: ... to here. Move GT and GE
26707 comparisons into v850_float_z_comparison_operator.
26708 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26710 (v850_float_nz_comparison_operator): Likewise.
26712 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26714 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26715 with calls to gen_insvsi/gen_insvdi.
26717 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26719 * config/i386/i386.c (initial_ix86_tune_features): Set up
26720 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26722 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26723 Jakub Jelinek <jakub@redhat.com>
26725 PR tree-optimization/48189
26726 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26727 If nitercst is 0, don't predict the exit edge.
26729 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26731 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26732 in asm_fprintf with reg_names.
26733 (aarch64_print_operand_address): Likewise.
26734 (aarch64_return_addr): Likewise.
26735 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26737 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26739 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26740 (INT_U6_BITS): Likewise.
26741 * config/pa/predicates.md (uint6_operand): New predicate.
26742 (shift5_operand, shift6_operand): Likewise.
26743 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26745 (lshrdi3): Use shift6_operand.
26746 (shrpsi4, shrpdi4): New insn patterns.
26747 (extzv): Delete expander.
26748 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26749 predicates in unamed zero extract patterns. Tighten common constraint.
26750 (extv): Delete expander.
26751 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26752 predicates in unamed sign extract patterns. Tighten common constraint.
26753 (insv): Delete expander.
26754 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26755 predicates in unamed insert patterns. Tighten common constraint.
26756 Change uint32_operand predicate to uint6_operand predicate in unamed
26757 DImode pattern to insert constant values of type 1...1xxxx.
26759 2013-01-04 Jan Hubicka <jh@suse.cz>
26761 PR tree-optimization/55823
26762 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26765 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26766 Uros Bizjak <ubizjak@gmail.com>
26768 PR rtl-optimization/55845
26769 * df-problems.c (can_move_insns_across): Stop scanning at
26770 volatile_insn_p source instruction or give up if
26771 across_from .. across_to range contains any volatile_insn_p
26774 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26776 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26777 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26779 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26780 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26782 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26785 * asan.c (asan_clear_shadow): New function.
26786 (asan_emit_stack_protection): Use it.
26788 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26790 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26791 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26792 with tab instead of space.
26794 2013-01-08 Nick Clifton <nickc@redhat.com>
26796 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26797 register bank 0 at the start of an interrupt handler.
26798 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26801 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26803 * config/aarch64/aarch64-simd.md
26804 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26805 (aarch64_simd_bsl): Likewise.
26806 (aarch64_vcond_internal<mode>): Likewise.
26807 (vcond<mode><mode>): Likewise.
26808 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26809 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26811 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26813 * config/aarch64/aarch64-builtins.c
26814 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26816 2013-01-08 Martin Jambor <mjambor@suse.cz>
26819 * tree-sra.c (analyze_access_subtree): Return true also after
26820 potentially creating a debug-only replacement.
26822 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26824 PR middle-end/55890
26825 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26827 PR tree-optimization/54120
26828 * tree-vrp.c (range_fits_type_p): Don't allow
26829 src_precision < precision from signed vr to unsigned_p
26830 if vr->min or vr->max is negative.
26831 (simplify_float_conversion_using_ranges): Test can_float_p
26832 against CODE_FOR_nothing.
26834 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26835 Richard Biener <rguenther@suse.de>
26837 PR middle-end/55851
26838 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26839 types instead of just INTEGER_TYPE types.
26841 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26843 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26846 2013-01-07 Steve Ellcey <sellcey@mips.com>
26849 * config/mips/mips.opt: Change mad to mmad to match documentation.
26851 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26854 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26855 .progmemx.data now.
26857 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26860 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26861 (avr_addrspace_t): Add .section_name field.
26862 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26864 (avr_addrspace): Same. Initialize .section_name. Remove last
26865 NULL entry. Put __memx into .progmemx.data.
26866 (progmem_section_prefix): Remove.
26867 (avr_asm_init_sections): No need to initialize progmem_section.
26868 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26869 section name prefix.
26870 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26871 retrieve the progmem section.
26872 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26873 boundary to run over avr_addrspace[].
26874 (avr_register_target_pragmas): Ditto.
26876 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26878 * varasm.c (output_constant_def_contents): For asan_protect_global
26879 protected strings, adjust DECL_ALIGN if needed, before testing for
26881 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26882 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26884 (output_object_block): For asan protected decls, emit asan padding
26885 after their contents.
26886 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26887 (asan_finish_file): Test it here instead.
26889 2013-01-07 Nick Clifton <nickc@redhat.com>
26890 Matthias Klose <doko@debian.org>
26891 Doug Kwan <dougkwan@google.com>
26892 H.J. Lu <hongjiu.lu@intel.com>
26895 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26897 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26899 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26901 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26903 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26905 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26908 * doc/install.texi (Cross-Compiler-Specific Options): Document
26911 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26913 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26914 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26915 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26916 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26917 vqmovun_high_s64): Fix source operand number and update copyright.
26919 2013-01-07 Richard Biener <rguenther@suse.de>
26921 PR middle-end/55890
26922 * gimple.h (gimple_call_builtin_p): New overload.
26923 * gimple.c (validate_call): New function.
26924 (gimple_call_builtin_p): Likewise.
26925 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26926 Use gimple_call_builtin_p.
26927 (find_func_clobbers): Likewise.
26928 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26929 (strlen_optimize_stmt): Likewise.
26931 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26933 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26934 (vld1q_dup_*): Likewise.
26935 (vld1_*): Likewise.
26936 (vld1q_*): Likewise.
26937 (vld1_lane_*): Likewise.
26938 (vld1q_lane_*): Likewise.
26940 2013-01-07 Richard Biener <rguenther@suse.de>
26942 * lto-streamer.h (LTO_minor_version): Bump to 2.
26944 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26946 * config/aarch64/aarch64-protos.h
26947 (aarch64_const_double_zero_rtx_p): Rename to...
26948 (aarch64_float_const_zero_rtx_p): ...this.
26949 (aarch64_float_const_representable_p): New.
26950 (aarch64_output_simd_mov_immediate): Likewise.
26951 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26952 move immediate case.
26953 * config/aarch64/aarch64.c
26954 (aarch64_const_double_zero_rtx_p): Rename to...
26955 (aarch64_float_const_zero_rtx_p): ...this.
26956 (aarch64_print_operand): Allow printing of new constants.
26957 (aarch64_valid_floating_const): New.
26958 (aarch64_legitimate_constant_p): Check for valid floating-point
26960 (aarch64_simd_valid_immediate): Likewise.
26961 (aarch64_vect_float_const_representable_p): New.
26962 (aarch64_float_const_representable_p): Likewise.
26963 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26964 (aarch64_output_simd_mov_immediate): New.
26965 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26966 (*movdf_aarch64): Likewise.
26967 * config/aarch64/constraints.md (Ufc): New.
26968 (Y): call aarch64_float_const_zero_rtx.
26969 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26971 2013-01-07 Richard Biener <rguenther@suse.de>
26973 PR tree-optimization/55888
26974 PR tree-optimization/55862
26975 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26976 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26977 not if it is contained therein.
26979 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26981 * config/avr/t-avr: Typo.
26983 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26986 * config/avr/t-avr: Don't automatically rebuild
26987 $(srcdir)/config/avr/t-multilib
26988 $(srcdir)/config/avr/avr-tables.opt
26989 $(srcdir)/doc/avr-mmcu.texi
26990 (avr-mcus): New phony target to build them on request.
26991 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26992 * avr/avr-mcus.def: Adjust comments.
26994 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26996 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26998 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27000 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27002 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27004 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27006 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27008 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27009 to generate profiling.
27010 * config/rs6000/aix64.h (LIB_SPEC): Same.
27012 2013-01-04 Andrew Pinski <apinski@cavium.com>
27014 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27016 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27018 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27020 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27022 (ix86_expand_move): Ditto.
27023 (ix86_zero_extend_to_Pmode): Ditto.
27024 (ix86_expand_call): Ditto.
27025 (ix86_expand_special_args_builtin): Ditto.
27026 (ix86_expand_builtin): Ditto.
27028 2013-01-04 Richard Biener <rguenther@suse.de>
27030 PR tree-optimization/55862
27031 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27032 translating them through PHI nodes.
27034 2013-01-04 Martin Jambor <mjambor@suse.cz>
27036 PR tree-optimization/55755
27037 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27038 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27040 2013-01-04 Marek Polacek <polacek@redhat.com>
27042 PR middle-end/55859
27043 * opts.c (default_options_optimization): Clarify error message.
27045 2013-01-04 Richard Biener <rguenther@suse.de>
27047 PR middle-end/55863
27048 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27051 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27054 * config/pa/pa.md (movsi): Revert previous change.
27055 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27058 2013-01-03 Richard Henderson <rth@redhat.com>
27060 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27061 after eliminating TLS symbols.
27063 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27066 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27067 * graphite-poly.c (debug_gmp_value): Likewise.
27069 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27072 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27073 selected code model, define __code_mode_small__, __code_model_medium__,
27074 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27075 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27076 xchg temporary register with %k. Declare temporary register as
27078 [__x86_64__]: For medium and large code models, preserve %rbx register.
27080 2013-01-03 Richard Biener <rguenther@suse.de>
27082 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27083 (dump_subscript): Adjust.
27084 (finalize_ddr_dependent): Do not dump redundant info.
27085 (analyze_siv_subscript): Adjust.
27086 (subscript_dependence_tester): Likewise.
27087 (compute_affine_dependence): Likewise.
27089 2013-01-03 Richard Biener <rguenther@suse.de>
27092 2013-01-03 Richard Biener <rguenther@suse.de>
27094 PR tree-optimization/55857
27095 * tree-vect-stmts.c (vectorizable_load): Do not setup
27096 re-alignment for invariant loads.
27098 2013-01-02 Richard Biener <rguenther@suse.de>
27100 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27101 invariant load do not generate a vector load from the scalar location.
27103 2013-01-03 Richard Biener <rguenther@suse.de>
27105 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27106 for not vectorizing.
27107 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27108 not build INDIRECT_REFs, call get_name once only.
27109 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27110 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27112 2013-01-03 Richard Biener <rguenther@suse.de>
27114 PR tree-optimization/55857
27115 * tree-vect-stmts.c (vectorizable_load): Do not setup
27116 re-alignment for invariant loads.
27118 2013-01-03 Richard Biener <rguenther@suse.de>
27121 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27122 prefer a built-in decl.
27124 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27126 * gcc.c (process_command): Update copyright notice dates.
27127 * gcov.c (print_version): Likewise.
27128 * gcov-dump.c (print_version): Likewise.
27130 PR rtl-optimization/55838
27131 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27132 iv0.step, iv1.step and step.
27134 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27135 Marc Glisse <marc.glisse@inria.fr>
27137 PR tree-optimization/55832
27138 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27139 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27140 integer_{one,zero}_node.
27142 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27145 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27146 * var-tracking.c (reverse_op): Don't add reverse ops to
27147 VALUEs that have already
27148 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27150 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27152 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27154 2013-01-02 Teresa Johnson <tejohnson@google.com>
27156 * dumpfile.c (dump_loc): Print filename with location.
27157 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27158 new location_t parameter to emit complete unroll message with
27159 new dump framework.
27160 (canonicalize_loop_induction_variables): Compute loops location
27161 and pass to try_unroll_loop_completely.
27162 * loop-unroll.c (report_unroll_peel): New function.
27163 (peel_loops_completely): Use new dump format with location
27164 for main dumpfile message, and invoke report_unroll_peel on success.
27165 (decide_unrolling_and_peeling): Ditto.
27166 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27167 by report_unroll_peel.
27168 (decide_peel_completely): Ditto.
27169 (decide_unroll_constant_iterations): Ditto.
27170 (decide_unroll_runtime_iterations): Ditto.
27171 (decide_peel_simple): Ditto.
27172 (decide_unroll_stupid): Ditto.
27173 * cfgloop.c (get_loop_location): New function.
27174 * cfgloop.h (get_loop_location): Declare.
27176 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27178 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27181 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27183 PR middle-end/55198
27184 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27185 BLKmode objects when EXPAND_MEMORY is specified.
27187 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27189 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27191 (fold_builtin_cpu): Do not share cpu model decls across statements.
27193 2013-01-02 Jason Merrill <jason@redhat.com>
27196 * tree.c (build_array_type_1): Revert earlier change.
27198 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27200 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27202 * config/aarch64/aarch64-tune.md: Re-generate.
27204 2013-01-02 Richard Biener <rguenther@suse.de>
27206 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27207 invariant load do not generate a vector load from the scalar location.
27209 2013-01-02 Richard Biener <rguenther@suse.de>
27212 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27213 * configure: Regenerate.
27215 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27217 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27218 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27219 (expand_builtin_int_roundingfn_2): Keep the original target around
27220 for the fallback case.
27222 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27224 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27225 to be clear for sign changes.
27227 2013-01-01 Jan Hubicka <jh@suse.cz>
27229 * ipa-inline-analysis.c: Fix formatting.
27231 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27233 PR tree-optimization/55831
27234 * tree-vect-loop.c (get_initial_def_for_induction): Use
27235 gsi_after_labels instead of gsi_start_bb.
27237 Copyright (C) 2013 Free Software Foundation, Inc.
27239 Copying and distribution of this file, with or without modification,
27240 are permitted in any medium without royalty provided the copyright
27241 notice and this notice are preserved.