1 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
6 2013-10-21 Richard Biener <rguenther@suse.de>
8 PR tree-optimization/58794
9 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
10 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
12 2013-10-21 Richard Biener <rguenther@suse.de>
15 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
16 to (T) X for sign-changing conversions (or no conversion).
18 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
20 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
22 2013-10-20 Jan Hubicka <jh@suse.cz>
24 * config/i386/i386-tune.def: Add comment; organize into categories
26 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
28 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
29 argument. Update function comment.
30 (expand_set_or_movmem_via_rep): New function combining
31 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
32 (expand_movmem_via_rep_mov): Remove.
33 expand_setmem_via_rep_stos): Remove.
34 (expand_movmem_epilogue): Update calls correspondingly.
35 (expand_setmem_epilogue_via_loop): Likewise.
37 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
38 (expand_set_or_movmem_prologue): New function combining
39 expand_movmem_prologue and expand_setmem_prologue.
40 (expand_movmem_prologue): Remove.
41 (expand_setmem_prologue): Remove.
42 (expand_set_or_movmem_constant_prologue): New function combining
43 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
44 (expand_constant_movmem_prologue): Remove.
45 (expand_constant_setmem_prologue): Remove.
46 (promote_duplicated_reg): Allow vector-const0 value.
47 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
48 and ix86_expand_setmem.
49 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
50 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
52 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
54 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
56 (vec_unpacku_hi_v8hi): Likewise.
57 (vec_unpacku_lo_v16qi): Likewise.
58 (vec_unpacku_lo_v8hi): Likewise.
60 2013-10-20 Jan Hubicka <jh@suse.cz>
62 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
63 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
64 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
66 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
68 * config/mips/mips.h (ISA_HAS_WSBH): Define.
69 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
71 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
73 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
76 * system.h: Undef m_slot.
78 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
80 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
81 all elements for both endian flavors.
83 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
86 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
87 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
88 and SI_REG for 64bit SYSV ABI targets.
90 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
92 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
93 to multi_reg_return. Clarify that we are skipping USEs of multiple
94 return registers. Use bool type where appropriate.
96 2013-10-18 Jan Hubicka <jh@suse.cz>
98 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
100 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
101 (X86_TUNE_PUSH_MEMORY): Likewise.
102 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
103 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
104 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
106 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
107 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
109 (ix86_option_override_internal): Update to use tune features instead
112 2013-10-18 Cong Hou <congh@google.com>
114 PR tree-optimization/58508
115 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
116 statement that contains data refs with zero-step.
118 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
120 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
121 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
122 * gimple-low.c (gimple_check_call_arg,
123 gimple_check_call_matching_types): Move to cgraph.c.
124 * gimple-low.h: Remove prototype.
125 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
126 Relocate from gimple-low.c.
127 * cgraph.h: Add prototype. Don't include basic-block.h.
128 * gimplify.c: Add gimple-low to include list.
129 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
130 * tree-eh.c: Add gimple-low to include list.
131 * tree-nested.c: Likewise.
132 * cfgexpand.c: Add tree-ssa-address.h to include list.
134 * gimple-fold.c: Likewise.
135 * gimple-ssa-strength-reduction.c: Likewise.
136 * trans-mem.c: Likewise.
137 * tree-mudflap.c: Likewise.
138 * tree-ssa-loop-ivopts.c: Likewise.
139 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
140 (degenerate_phi_result): Move to tree-phinodes.c.
141 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
142 * tree-ssa-threadedge.c: Likewise.
143 * tree-vrp.c: Likewise.
144 * tree-phinodes.c (degenerate_phi_result): Relocate here.
145 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
146 * tree-phinodes.h (degenerate_phi_result): Add prototype.
147 * tree-ssa-copy.c: Include tree-ssa-dom.h.
148 * tree-ssa-forwprop.c: Likewise.
149 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
150 pass_data_cleanup_cfg_post_optimizing,
151 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
152 * tree-optimize.c: Delete File.
153 * graphite.c: Include tree-cfgcleanup.h.
154 * passes.c: Likewise.
155 * tree-cfg.c: Likewise.
156 * tree-profile.c: Likewise.
157 * tree-ssa-dse.c: Likewise.
158 * tree-ssa-loop-ivcanon.c: Likewise.
159 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
160 * tree-outof-ssa.c: Include sbitmap.h.
161 * tree-ssa-live.c: Likewise.
162 * tree-ssa-propagate.c: Likewise.
163 * tree-ssa-structalias.c: Likewise.
164 * tree-stdarg.c: Likewise.
165 * Makefile.in (OBJS): Delete tree-optimize.o.
166 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
167 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
168 * varasm.c: Include basic-block.h.
169 * cfgloop.h: Include function.h instead of basic-block.h
170 (bb_loop_depth): Move to cfgloop.c.
171 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
173 2013-10-18 Teresa Johnson <tejohnson@google.com>
175 * predict.c (probably_never_executed): Compare frequency-based
176 count to number of training runs.
177 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
179 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
181 * config/arm/arm.c (cortexa9_extra_costs): New table.
182 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
184 2013-10-18 Jeff Law <law@redhat.com>
186 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
188 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
189 (register_jump_thread): Add "registered_jump_thread" debug
191 * dbgcnt.def (registered_jump_thread): New debug counter.
193 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
195 * config/rs6000/rs6000.c: Include cgraph.h.
197 2013-10-18 Teresa Johnson <tejohnson@google.com>
199 * tree-ssa-tail-merge.c (replace_block_by): Update edge
200 weights during merging.
202 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
204 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
205 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
206 * tree-ssa-operands.h: Remove prototype.
207 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
208 * gimple.c (virtual_operand_p): Relocate from gimple.c.
209 * gimple.h: Add prototype.
210 * gimple-ssa.h: Include tree-ssa-operands.h.
211 * tree-dump.c: Add tree-cfg.h to include list.
212 * tree-ssa-alias.c: Add ipa-reference.h to include list.
213 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
214 * config/i386/i386.c: Don't include tree-flow.h.
215 * config/rs6000/rs6000.c: Likewise.
217 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
219 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
221 2013-10-18 Richard Biener <rguenther@suse.de>
223 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
224 or TYPE_UNSIGNED of integral types.
225 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
226 NULL_TREE for zero-precision integral types.
228 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
230 * config/aarch64/arm_neon.h
231 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
233 2013-10-17 Sriraman Tallam <tmsriram@google.com>
235 * opth-gen.awk: Define target_flags_explicit.
237 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
239 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
240 fields to the reg_addr array that describes the valid addressing
241 mode for any register, general purpose registers, floating point
242 registers, and Altivec registers.
243 (FIRST_RELOAD_REG_CLASS): Likewise.
244 (LAST_RELOAD_REG_CLASS): Likewise.
245 (struct reload_reg_map_type): Likewise.
246 (reload_reg_map_type): Likewise.
247 (RELOAD_REG_VALID): Likewise.
248 (RELOAD_REG_MULTIPLE): Likewise.
249 (RELOAD_REG_INDEXED): Likewise.
250 (RELOAD_REG_OFFSET): Likewise.
251 (RELOAD_REG_PRE_INCDEC): Likewise.
252 (RELOAD_REG_PRE_MODIFY): Likewise.
253 (reg_addr): Likewise.
254 (mode_supports_pre_incdec_p): New helper functions to say whether
255 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
256 (mode_supports_pre_modify_p): Likewise.
257 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
258 print the valid address mode bits for each mode.
259 (rs6000_debug_print_mode): Likewise.
260 (rs6000_debug_reg_global): Likewise.
261 (rs6000_setup_reg_addr_masks): New function to set up the address
262 mask bits for each type.
263 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
264 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
265 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
266 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
267 PRE_MODIFY are supported.
268 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
269 registers, instead of {src,dest}_av_p.
270 (rs6000_print_options_internal): Tweak the debug output slightly.
272 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
274 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
277 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
279 * tree-flow.h (struct omp_region): Move to omp-low.c.
280 Remove omp_ prototypes and variables.
281 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
282 (copy_var_decl): Relocate prototype from tree-flow.h.
283 * gimple.c (copy_var_decl): Relocate from omp-low.c.
284 * tree.h: Move prototype to omp-low.h.
285 * omp-low.h: New File. Relocate prototypes here.
286 * omp-low.c (struct omp_region): Make local here.
287 (root_omp_region): Make static.
288 (copy_var_decl) Move to gimple.c.
289 (new_omp_region): Make static.
290 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
291 * tree-cfg.c: Include omp-low.h.
292 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
293 * gimplify.c: Include omp-low.h.
294 * tree-parloops.c: Likewise.
296 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
298 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
299 operands are in memory, prefer to force non-matched operand 1 to
302 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
305 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
306 restrict TImode addresses to single indirect registers if both
307 -mquad-memory and -mvsx-timode are used.
308 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
309 we should emit load/store quad. Remove using %y for quad memory
312 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
313 constraints to allow load/store quad on machines where TImode is
314 not allowed in VSX registers. Use 'n' instead of 'F' constraint
315 for TImode to load integer constants.
317 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
319 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
321 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
323 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
324 handling of STACK_REG.
326 2013-10-17 Richard Biener <rguenther@suse.de>
328 PR tree-optimization/58143
329 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
331 (rewrite_to_defined_overflow): Likewise.
332 (move_computations_dom_walker::before_dom): Rewrite stmts
333 with undefined signed overflow that are not always executed
334 into unsigned arithmetic.
336 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
339 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
340 explicit isa flag to be an options variable, instead of using
341 global_options_set. Remove define from rs6000.h.
342 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
344 * config/rs6000/rs6000.c (rs6000_option_override_internal):
345 Initialize rs6000_isa_flags_explicit.
346 (rs6000_function_specific_save): Add gcc_options* parameter, so
347 that the powerpc builds after the 2013-10-15 changes.
348 (rs6000_function_specific_restore): Likewise.
350 2013-10-16 DJ Delorie <dj@redhat.com>
352 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
353 op is a REG before checking REGNO.
354 (rl78_alloc_physical_registers): Verify pattern is a SET before
357 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
359 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
361 (vec_unpacks_lo_v4sf): Likewise.
362 (vec_unpacks_float_hi_v4si): Likewise.
363 (vec_unpacks_float_lo_v4si): Likewise.
364 (vec_unpacku_float_hi_v4si): Likewise.
365 (vec_unpacku_float_lo_v4si): Likewise.
367 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
369 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
370 (vsx_concat_v2sf): Likewise.
372 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
374 * config/aarch64/aarch64.md
375 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
377 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
379 PR tree-optimization/58697
380 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
381 estimated_loop_iterations_int.
382 (max_stmt_executions_int): Call get_max_loop_iterations_int.
383 (get_max_loop_iterations_int): New. HWINT version of
384 get_max_loop_iterations.
385 * cfgloop.h: Add prototypes.
386 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
387 * loop-unroll.c (decide_peel_once_rolling): Call
388 get_estimated_loop_iterations_int.
389 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
390 * tree-ssa-loop-niter.h: Tweak prototypes.
392 2013-10-16 David Malcolm <dmalcolm@redhat.com>
394 * gengtype-parse.c (struct_field_seq): Ignore access-control
395 keywords ("public:" etc).
397 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
399 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
400 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
402 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
404 * config/arm/arm.opt (mlra): New option.
405 * config/arm/arm.c (arm_lra_p): New function.
406 (TARGET_LRA_P): Define.
408 2013-10-16 Paulo Matos <pmatos@broadcom.com>
410 * tree-core.h (tree_code_name): Remove.
411 * tree.h (get_tree_code_name): New prototype.
412 * tree.c (tree_code_name): Make static.
413 (get_tree_code_name): New function.
414 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
415 tree_class_check_failed, tree_range_check_failed,
416 tree_not_class_check_failed, omp_clause_check_failed,
417 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
418 wrapper get_tree_code_name instead of calling tree_code_name directly.
419 * tree-vrp.c (dump_asserts_for): Likewise.
420 * tree-dump.c (dequeue_and_dump): Likewise.
421 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
422 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
423 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
424 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
425 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
426 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
427 dump_gimple_omp_for): Likewise.
428 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
429 * tree-ssa-pre.c (print_pre_expr): Likewise.
430 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
431 * print-tree.c (print_node_brief, print_node): Likewise.
432 * gimple.c (gimple_check_failed): Likewise.
433 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
434 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
435 * config/mep/mep.c (mep_validate_vliw): Likewise.
436 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
437 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
439 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
441 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
444 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
446 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
447 (MULTILIB_MATCHES): Add multilib for -march=v8.
449 2013-10-15 Sriraman Tallam <tmsriram@google.com>
452 * optc-save-gen.awk: Add extra parameter to the save and restore
454 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
455 * tree.c (build_optimization_node): New parameter. Add extra parameter
456 to call to cl_optimization_save.
457 (build_target_option_node): New parameter. Add extra parameter
458 to call to cl_target_option_save.
459 * tree.h (build_optimization_node): New parameter.
460 (build_target_option_node): New parameter.
461 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
462 build_optimization_node and build_target_option_node.
463 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
464 (handle_pragma_push_options): Ditto.
465 * toplev.c (process_options): Ditto.
466 * opts.c (init_options_struct): Check for opts_set non-null.
467 * target.def (target_option.save): New parameter.
468 (target_option.restore): New parameter.
470 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
471 (ix86_pragma_target_parse): Ditto.
472 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
474 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
475 to build_optimization_node and build_target_option_node.
476 (rs6000_valid_attribute_p): Ditto.
477 (rs6000_pragma_target_parse): Ditto.
478 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
480 * config/i386/i386.h:
481 TARGET_64BIT_P: New Macro
482 TARGET_MMX_P: New Macro.
483 TARGET_3DNOW_P: New Macro.
484 TARGET_3DNOW_A_P: New Macro.
485 TARGET_SSE_P: New Macro.
486 TARGET_SSE2_P: New Macro.
487 TARGET_SSE3_P: New Macro.
488 TARGET_SSSE3_P: New Macro.
489 TARGET_SSE4_1_P: New Macro.
490 TARGET_SSE4_2_P: New Macro.
491 TARGET_AVX_P: New Macro.
492 TARGET_AVX2_P: New Macro.
493 TARGET_AVX512F_P: New Macro.
494 TARGET_AVX512PF_P: New Macro.
495 TARGET_AVX512ER_P: New Macro.
496 TARGET_AVX512CD_P: New Macro.
497 TARGET_FMA_P: New Macro.
498 TARGET_SSE4A_P: New Macro.
499 TARGET_FMA4_P: New Macro.
500 TARGET_XOP_P: New Macro.
501 TARGET_LWP_P: New Macro.
502 TARGET_ABM_P: New Macro.
503 TARGET_BMI_P: New Macro.
504 TARGET_BMI2_P: New Macro.
505 TARGET_LZCNT_P: New Macro.
506 TARGET_TBM_P: New Macro.
507 TARGET_POPCNT_P: New Macro.
508 TARGET_SAHF_P: New Macro.
509 TARGET_MOVBE_P: New Macro.
510 TARGET_CRC32_P: New Macro.
511 TARGET_AES_P: New Macro.
512 TARGET_PCLMUL_P: New Macro.
513 TARGET_CMPXCHG16B_P: New Macro.
514 TARGET_FSGSBASE_P: New Macro.
515 TARGET_RDRND_P: New Macro.
516 TARGET_F16C_P: New Macro.
517 TARGET_RTM_P: New Macro.
518 TARGET_HLE_P: New Macro.
519 TARGET_RDSEED_P: New Macro.
520 TARGET_PRFCHW_P: New Macro.
521 TARGET_ADX_P: New Macro.
522 TARGET_FXSR_P: New Macro.
523 TARGET_XSAVE_P: New Macro.
524 TARGET_XSAVEOPT_P: New Macro.
525 TARGET_LP64_P: New Macro.
526 TARGET_X32_P: New Macro.
527 TARGET_FPMATH_DEFAULT_P: New Macro.
528 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
529 * config/i386/i386.c (ix86_option_override_internal): New parameters.
531 Change ix86_tune_string to access opts->x_ix86_tune_string.
532 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
533 Change ix86_arch_string to access opts->x_ix86_arch_string.
534 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
535 Change ix86_pmode to access opts->x_ix86_pmode.
536 Change ix86_abi to access opts->x_ix86_abi.
537 Change ix86_cmodel to access opts->x_ix86_cmodel.
538 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
539 Change ix86_isa_flags_explicit to access
540 opts->x_ix86_isa_flags_explicit.
541 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
542 Change ix86_regparm to access opts->x_ix86_regparm.
543 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
544 Change ix86_preferred_stack_boundary_arg to access
545 opts->x_ix86_preferred_stack_boundary_arg.
546 Change ix86_force_align_arg_pointer to access
547 opts->x_ix86_force_align_arg_pointer.
548 Change ix86_incoming_stack_boundar_arg to access
549 opts->x_ix86_incoming_stack_boundar_arg.
550 Change ix86_fpmath to access opts->x_ix86_fpmath.
551 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
552 Change ix86_recip_name to access opts->x_ix86_recip_name.
553 Change ix86_stack_protector_guard to access
554 opts->x_ix86_stack_protector_guard.
555 Change ix86_tune_memcpy_strategy to access
556 opts->x_ix86_tune_memcpy_strategy.
557 Change ix86_tune_memset_strategy to access
558 opts->x_ix86_tune_memset_strategy.
559 Change global_options to access opts.
560 Change global_options_set to access opts_set.
561 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
562 Change TARGET_MMX to TARGET_MMX_P (opts->...).
563 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
564 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
565 Change TARGET_SSE to TARGET_SSE_P (opts->...).
566 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
567 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
568 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
569 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
570 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
571 Change TARGET_AVX to TARGET_AVX_P (opts->...).
572 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
573 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
574 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
575 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
576 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
577 Change TARGET_FMA to TARGET_FMA_P (opts->...).
578 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
579 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
580 Change TARGET_XOP to TARGET_XOP_P (opts->...).
581 Change TARGET_LWP to TARGET_LWP_P (opts->...).
582 Change TARGET_ABM to TARGET_ABM_P (opts->...).
583 Change TARGET_BMI to TARGET_BMI_P (opts->...).
584 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
585 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
586 Change TARGET_TBM to TARGET_TBM_P (opts->...).
587 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
588 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
589 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
590 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
591 Change TARGET_AES to TARGET_AES_P (opts->...).
592 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
593 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
594 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
595 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
596 Change TARGET_F16C to TARGET_F16C_P (opts->...).
597 Change TARGET_RTM to TARGET_RTM_P (opts->...).
598 Change TARGET_HLE to TARGET_HLE_P (opts->...).
599 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
600 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
601 Change TARGET_ADX to TARGET_ADX_P (opts->...).
602 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
603 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
604 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
605 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
606 Change TARGET_X32 to TARGET_X32_P (opts->...).
607 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
608 Change TARGET_FLOAT_RETURNS_IN_80387 to
609 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
610 (ix86_function_specific_save): New parameter. Use opts-> fields
611 to replace global fields.
612 (ix86_function_specific_restore): Ditto.
613 (ix86_valid_target_attribute_inner_p): New parameters.
615 Fix call to ix86_handle_option and set_option.
616 (ix86_valid_target_attribute_tree): New parameters.
617 Change global_options to access opts.
618 Change global_options_set to access opts_set.
619 Fix call to ix86_valid_target_attribute_inner_p.
620 Change ix86_tune_string to access opts->x_ix86_tune_string.
621 Change ix86_arch_string to access opts->x_ix86_arch_string.
622 Change ix86_fpmath to access opts->x_ix86_fpmath
623 Fix call to ix86_option_override_internal.
624 Fix call to ix86_add_new_builtins.
625 Fix calls to build_optimization_node and build_target_option_node.
626 (ix86_valid_target_attribute_p): Remove access to global_options.
627 Use new gcc_options structure func_options.
628 Fix call to ix86_valid_target_attribute_tree.
629 Fix call to build_optimization_node.
630 (get_builtin_code_for_version): Fix call to
631 ix86_valid_target_attribute_tree.
633 2013-10-15 David Malcolm <dmalcolm@redhat.com>
635 * Makefile.in (PICFLAG): New.
636 (enable_host_shared): New.
637 (INTERNAL_CFLAGS): Use PICFLAG.
638 (LIBIBERTY): Use pic build of libiberty.a if configured with
639 --enable-host-shared.
640 * configure.ac: Add --enable-host-shared, setting up new
642 * configure: Regenerate.
643 * doc/install.texi (--enable-shared): Add note contrasting it with ...
644 (--enable-host-shared): New option.
646 2013-10-15 Richard Biener <rguenther@suse.de>
648 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
649 for built-in functions.
651 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
653 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
654 (optimize_range_tests_1): New function,
655 extracted from optimize_range_tests.
656 (optimize_range_tests_xor): Similarly.
657 (optimize_range_tests_diff): New function.
658 (optimize_range_tests): Use optimize_range_tests_1.
660 2013-10-15 Cong Hou <congh@google.com>
662 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
663 requirement of the reduction pattern so that one operand of the
664 reduction operation can come from outside of the loop.
666 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
668 * config/arm/neon-schedgen.ml: Remove.
669 * config/arm/cortex-a9-neon.md: Remove comment regarding
672 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
674 * config/arm/types: Remove old neon types.
676 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
678 * config/arm/cortex-a7.md
679 (cortex_a7_neon_type): New.
680 (cortex_a7_neon_mul): Update for new types.
681 (cortex_a7_neon_mla): Likewise.
682 (cortex_a7_neon): Likewise.
684 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
686 * config/arm/cortex-a15-neon.md
687 (cortex_a15_neon_type): New,
689 (cortex_a15_neon_int_1): Remove.
690 (cortex_a15_neon_int_2): Likewise.
691 (cortex_a15_neon_int_3): Likewise.
692 (cortex_a15_neon_int_4): Likewise.
693 (cortex_a15_neon_int_5): Likewise.
694 (cortex_a15_neon_vqneg_vqabs): Likewise.
695 (cortex_a15_neon_vmov): Likewise.
696 (cortex_a15_neon_vaba): Likewise.
697 (cortex_a15_neon_vaba_qqq): Likewise.
698 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
699 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
700 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
702 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
703 (cortex_a15_neon_mla_qqq_8_16): Likewise.
704 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
705 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
706 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
707 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
708 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
709 (cortex_a15_neon_shift_1): Likewise.
710 (cortex_a15_neon_shift_2): Likewise.
711 (cortex_a15_neon_shift_3): Likewise.
712 (cortex_a15_neon_vshl_ddd): Likewise.
713 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
714 (cortex_a15_neon_vsra_vrsra): Likewise.
715 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
716 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
717 (cortex_a15_neon_bp_3cycle): Likewise.
718 (cortex_a15_neon_ldm_2): Likewise.
719 (cortex_a15_neon_stm_2): Likewise.
720 (cortex_a15_neon_mcr): Likewise.
721 (cortex_a15_neon_mrc): Likewise.
722 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
723 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
724 (cortex_a15_neon_fp_vmul_ddd): Likewise.
725 (cortex_a15_neon_fp_vmul_qqd): Likewise.
726 (cortex_a15_neon_fp_vmla_ddd): Likewise.
727 (cortex_a15_neon_fp_vmla_qqq): Likewise.
728 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
729 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
730 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
731 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
732 (cortex_a15_neon_bp_simple): Likewise.
733 (cortex_a15_neon_bp_2cycle): Likewise.
734 (cortex_a15_neon_bp_3cycle): Likewise.
735 (cortex_a15_neon_vld1_1_2_regs): Likewise.
736 (cortex_a15_neon_vld1_3_4_regs): Likewise.
737 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
738 (cortex_a15_neon_vld2_4_regs): Likewise.
739 (cortex_a15_neon_vld3_vld4): Likewise.
740 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
741 (cortex_a15_neon_vst1_3_4_regs): Likewise.
742 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
743 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
744 (cortex_a15_neon_vst3_vst4): Rename to...
745 (cortex_a15_neon_vst4): This, update for new attributes.
746 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
747 (cortex_a15_neon_vld3_vld4_lane): Likewise.
748 (cortex_a15_neon_vst1_vst2_lane): Likewise.
749 (cortex_a15_neon_vst3_vst4_lane): Likewise.
750 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
751 (cortex_a15_neon_ldm_2): Likewise.
752 (cortex_a15_neon_stm_2): Likewise.
753 (cortex_a15_neon_mcr): Likewise.
754 (cortex_a15_neon_mcr_2_mcrr): Likewise.
755 (cortex_a15_neon_mrc): Likewise.
756 (cortex_a15_neon_mrrc): Likewise.
758 (cortex_a15_neon_abd): New.
759 (cortex_a15_neon_abd_q): Likewise.
760 (cortex_a15_neon_aba): Likewise.
761 (cortex_a15_neon_aba_q): Likewise.
762 (cortex_a15_neon_acc): Likewise.
763 (cortex_a15_neon_acc_q): Likewise.
764 (cortex_a15_neon_arith_basic): Likewise.
765 (cortex_a15_neon_arith_complex): Likewise.
766 (cortex_a15_neon_multiply): Likewise.
767 (cortex_a15_neon_multiply_q): Likewise.
768 (cortex_a15_neon_mla): Likewise.
769 (cortex_a15_neon_mla_q): Likewise.
770 (cortex_a15_neon_sat_mla_long): Likewise.
771 (cortex_a15_neon_shift_acc): Likewise.
772 (cortex_a15_neon_shift_imm_basic): Likewise.
773 (cortex_a15_neon_shift_imm_complex): Likewise.
774 (cortex_a15_neon_shift_reg_basic): Likewise.
775 (cortex_a15_neon_shift_reg_basic_q): Likewise.
776 (cortex_a15_neon_shift_reg_complex): Likewise.
777 (cortex_a15_neon_shift_reg_complex_q): Likewise.
778 (cortex_a15_neon_fp_negabs): Likewise
779 (cortex_a15_neon_fp_arith): Likewise
780 (cortex_a15_neon_fp_arith_q): Likewise
781 (cortex_a15_neon_fp_cvt_int): Likewise
782 (cortex_a15_neon_fp_cvt_int_q): Likewise
783 (cortex_a15_neon_fp_cvt_16): Likewise
784 (cortex_a15_neon_fp_mul): Likewise
785 (cortex_a15_neon_fp_mul_q): Likewise
786 (cortex_a15_neon_fp_mla): Likewise
787 (cortex_a15_neon_fp_mla_q): Likewise
788 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
789 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
790 (cortex_a15_neon_bitops): Likewise.
791 (cortex_a15_neon_bitops_q): Likewise.
792 (cortex_a15_neon_from_gp): Likewise.
793 (cortex_a15_neon_from_gp_q): Likewise.
794 (cortex_a15_neon_tbl3_tbl4): Likewise.
795 (cortex_a15_neon_zip_q): Likewise.
796 (cortex_a15_neon_to_gp): Likewise.
797 (cortex_a15_neon_load_a): Likewise.
798 (cortex_a15_neon_load_b): Likewise.
799 (cortex_a15_neon_load_c): Likewise.
800 (cortex_a15_neon_load_d): Likewise.
801 (cortex_a15_neon_load_e): Likewise.
802 (cortex_a15_neon_load_f): Likewise.
803 (cortex_a15_neon_store_a): Likewise.
804 (cortex_a15_neon_store_b): Likewise.
805 (cortex_a15_neon_store_c): Likewise.
806 (cortex_a15_neon_store_d): Likewise.
807 (cortex_a15_neon_store_e): Likewise.
808 (cortex_a15_neon_store_f): Likewise.
809 (cortex_a15_neon_store_g): Likewise.
810 (cortex_a15_neon_store_h): Likewise.
811 (cortex_a15_vfp_to_from_gp): Likewise.
813 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
815 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
817 (cortex_a9_neon_vshl_ddd): Remove.
818 (cortex_a9_neon_vst3_vst4): Likewise.
819 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
821 (cortex_a9_neon_bit_ops_q): New.
823 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
824 (cortex_a9_neon_int_2): Likewise.
825 (cortex_a9_neon_int_3): Likewise.
826 (cortex_a9_neon_int_4): Likewise.
827 (cortex_a9_neon_int_5): Likewise.
828 (cortex_a9_neon_vqneg_vqabs): Likewise.
829 (cortex_a9_neon_vmov): Likewise.
830 (cortex_a9_neon_vaba): Likewise.
831 (cortex_a9_neon_vaba_qqq): Likewise.
832 (cortex_a9_neon_shift_1): Likewise.
833 (cortex_a9_neon_shift_2): Likewise.
834 (cortex_a9_neon_shift_3): Likewise.
835 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
836 (cortex_a9_neon_vsra_vrsra): Likewise.
837 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
838 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
839 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
841 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
842 (cortex_a9_neon_mla_qqq_8_16): Likewise.
843 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
845 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
846 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
847 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
848 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
849 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
850 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
851 (cortex_a9_neon_fp_vsum): Likewise.
852 (cortex_a9_neon_fp_vmul_ddd): Likewise.
853 (cortex_a9_neon_fp_vmul_qqd): Likewise.
854 (cortex_a9_neon_fp_vmla_ddd): Likewise.
855 (cortex_a9_neon_fp_vmla_qqq): Likewise.
856 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
857 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
858 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
859 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
860 (cortex_a9_neon_bp_simple): Likewise.
861 (cortex_a9_neon_bp_2cycle): Likewise.
862 (cortex_a9_neon_bp_3cycle): Likewise.
863 (cortex_a9_neon_ldr): Likewise.
864 (cortex_a9_neon_str): Likewise.
865 (cortex_a9_neon_vld1_1_2_regs): Likewise.
866 (cortex_a9_neon_vld1_3_4_regs): Likewise.
867 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
868 (cortex_a9_neon_vld2_4_regs): Likewise.
869 (cortex_a9_neon_vld3_vld4): Likewise.
870 (cortex_a9_neon_vld1_vld2_lane): Likewise.
871 (cortex_a9_neon_vld3_vld4_lane): Likewise.
872 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
873 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
874 (cortex_a9_neon_vst1_3_4_regs): Likewise.
875 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
876 (cortex_a9_neon_vst1_vst2_lane): Likewise.
877 (cortex_a9_neon_vst3_vst4_lane): Likewise.
878 (cortex_a9_neon_mcr): Likewise.
879 (cortex_a9_neon_mcr_2_mcrr): Likewise.
880 (cortex_a9_neon_mrc): Likewise.
881 (cortex_a9_neon_mrrc): Likewise.
883 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
885 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
887 (cortex_a8_neon_vshl_ddd): Remove.
888 (cortex_a8_neon_vst3_vst4): Likewise.
889 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
891 (cortex_a8_neon_bit_ops_q): New.
893 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
894 (cortex_a8_neon_int_2): Likewise..
895 (cortex_a8_neon_int_3): Likewise.
896 (cortex_a8_neon_int_5): Likewise.
897 (cortex_a8_neon_vqneg_vqabs): Likewise.
898 (cortex_a8_neon_int_4): Likewise.
899 (cortex_a8_neon_vaba): Likewise.
900 (cortex_a8_neon_vaba_qqq): Likewise.
901 (cortex_a8_neon_shift_1): Likewise.
902 (cortex_a8_neon_shift_2): Likewise.
903 (cortex_a8_neon_shift_3): Likewise.
904 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
905 (cortex_a8_neon_vsra_vrsra): Likewise.
906 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
907 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
908 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
910 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
911 (cortex_a8_neon_mla_qqq_8_16): Likewise.
912 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
914 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
915 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
916 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
917 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
918 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
919 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
920 (cortex_a8_neon_fp_vsum): Likewise.
921 (cortex_a8_neon_fp_vmul_ddd): Likewise.
922 (cortex_a8_neon_fp_vmul_qqd): Likewise.
923 (cortex_a8_neon_fp_vmla_ddd): Likewise.
924 (cortex_a8_neon_fp_vmla_qqq): Likewise.
925 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
926 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
927 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
928 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
929 (cortex_a8_neon_bp_simple): Likewise.
930 (cortex_a8_neon_bp_2cycle): Likewise.
931 (cortex_a8_neon_bp_3cycle): Likewise.
932 (cortex_a8_neon_ldr): Likewise.
933 (cortex_a8_neon_str): Likewise.
934 (cortex_a8_neon_vld1_1_2_regs): Likewise.
935 (cortex_a8_neon_vld1_3_4_regs): Likewise.
936 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
937 (cortex_a8_neon_vld2_4_regs): Likewise.
938 (cortex_a8_neon_vld3_vld4): Likewise.
939 (cortex_a8_neon_vld1_vld2_lane): Likewise.
940 (cortex_a8_neon_vld3_vld4_lane): Likewise.
941 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
942 (cortex_a8_neon_vst1_3_4_regs): Likewise.
943 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
944 (cortex_a8_neon_vst1_vst2_lane): Likewise.
945 (cortex_a8_neon_vst3_vst4_lane): Likewise.
946 (cortex_a8_neon_mcr): Likewise.
947 (cortex_a8_neon_mcr_2_mcrr): Likewise.
948 (cortex_a8_neon_mrc): Likewise.
949 (cortex_a8_neon_mrrc): Likewise.
951 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
953 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
955 * config/aarch64/aarch64-simd.md (neon_type): Remove.
956 (aarch64_simd_dup<mode>): Add "type" attribute.
957 (aarch64_dup_lane<mode>): Likewise.
958 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
959 (*aarch64_simd_mov<mode>): Likewise.
960 (aarch64_simd_mov_from_<mode>low): Likewise.
961 (aarch64_simd_mov_from_<mode>high): Likewise.
962 (orn<mode>3): Likewise.
963 (bic<mode>3): Likewise.
964 (add<mode>3): Likewise.
965 (sub<mode>3): Likewise.
966 (mul<mode>3): Likewise.
967 (*aarch64_mul3_elt<mode>): Likewise.
968 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
969 (*aarch64_mul3_elt_to_128df): Likewise.
970 (*aarch64_mul3_elt_to_64v2df): Likewise.
971 (neg<mode>2): Likewise.
972 (abs<mode>2): Likewise.
973 (abd<mode>_3): Likewise.
974 (aba<mode>_3): Likewise.
975 (fabd<mode>_3): Likewise.
976 (*fabd_scalar<mode>3): Likewise.
977 (and<mode>3): Likewise.
978 (ior<mode>3): Likewise.
979 (xor<mode>3): Likewise.
980 (one_cmpl<mode>2): Likewise.
981 (aarch64_simd_vec_set<mode>): Likewise.
982 (aarch64_simd_lshr<mode>): Likewise.
983 (aarch64_simd_ashr<mode>): Likewise.
984 (aarch64_simd_imm_shl<mode>): Likewise.
985 (aarch64_simd_reg_sshl<mode): Likewise.
986 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
987 (aarch64_simd_reg_shl<mode>_signed): Likewise.
988 (aarch64_simd_vec_setv2di): Likewise.
989 (aarch64_simd_vec_set<mode>): Likewise.
990 (aarch64_mla<mode>): Likewise.
991 (*aarch64_mla_elt<mode>): Likewise.
992 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
993 (aarch64_mls<mode>): Likewise.
994 (*aarch64_mls_elt<mode>): Likewise.
995 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
996 (<su><maxmin><mode>3): Likewise.
997 (move_lo_quad_<mode>): Likewise.
998 (aarch64_simd_move_hi_quad_<mode>): Likewise.
999 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1000 (vec_pack_trunc_<mode>): Likewise.
1001 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1002 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1003 (*aarch64_<su>mlal_lo<mode>): Likewise.
1004 (*aarch64_<su>mlal_hi<mode>): Likewise.
1005 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1006 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1007 (*aarch64_<su>mlal<mode>): Likewise.
1008 (*aarch64_<su>mlsl<mode>): Likewise.
1009 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1010 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1011 (add<mode>3): Likewise.
1012 (sub<mode>3): Likewise.
1013 (mul<mode>3): Likewise.
1014 (div<mode>3): Likewise.
1015 (neg<mode>2): Likewise.
1016 (abs<mode>2): Likewise.
1017 (fma<mode>4): Likewise.
1018 (*aarch64_fma4_elt<mode>): Likewise.
1019 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1020 (*aarch64_fma4_elt_to_128df): Likewise.
1021 (*aarch64_fma4_elt_to_64v2df): Likewise.
1022 (fnma<mode>4): Likewise.
1023 (*aarch64_fnma4_elt<mode>): Likewise.
1024 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1025 (*aarch64_fnma4_elt_to_128df): Likewise.
1026 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1027 (<frint_pattern><mode>2): Likewise.
1028 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1029 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1030 (vec_unpacks_lo_v4sf): Likewise.
1031 (aarch64_float_extend_lo_v2df): Likewise.
1032 (vec_unpacks_hi_v4sf): Likewise.
1033 (aarch64_float_truncate_lo_v2sf): Likewise.
1034 (aarch64_float_truncate_hi_v4sf): Likewise.
1035 (aarch64_vmls<mode>): Likewise.
1036 (<su><maxmin><mode>3): Likewise.
1037 (<maxmin_uns><mode>3): Likewise.
1038 (reduc_<sur>plus_<mode>): Likewise.
1039 (reduc_<sur>plus_v2di): Likewise.
1040 (reduc_<sur>plus_v2si): Likewise.
1041 (reduc_<sur>plus_<mode>): Likewise.
1042 (aarch64_addpv4sf): Likewise.
1043 (clz<mode>2): Likewise.
1044 (reduc_<maxmin_uns>_<mode>): Likewise.
1045 (reduc_<maxmin_uns>_v2di): Likewise.
1046 (reduc_<maxmin_uns>_v2si): Likewise.
1047 (reduc_<maxmin_uns>_<mode>): Likewise.
1048 (reduc_<maxmin_uns>_v4sf): Likewise.
1049 (aarch64_simd_bsl<mode>_internal): Likewise.
1050 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1051 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1052 (aarch64_get_lane<mode>): Likewise.
1053 (*aarch64_combinez<mode>): Likewise.
1054 (aarch64_combine<mode>): Likewise.
1055 (aarch64_simd_combine<mode>): Likewise.
1056 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1057 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1058 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1059 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1060 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1061 (aarch64_<sur>h<addsub><mode>): Likewise.
1062 (aarch64_<sur><addsub>hn<mode>): Likewise.
1063 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1064 (aarch64_pmul<mode>): Likewise.
1065 (aarch64_<su_optab><optab><mode>): Likewise.
1066 (aarch64_<sur>qadd<mode>): Likewise.
1067 (aarch64_sqmovun<mode>): Likewise.
1068 (aarch64_<sur>qmovn<mode>): Likewise.
1069 (aarch64_s<optab><mode>): Likewise.
1070 (aarch64_sq<r>dmulh<mode>): Likewise.
1071 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1072 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1073 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1074 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1075 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1076 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1077 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1078 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1079 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1080 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1081 (aarch64_sqdmull<mode>): Likewise.
1082 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1083 (aarch64_sqdmull_n<mode>): Likewise.
1084 (aarch64_sqdmull2<mode>_internal): Likewise.
1085 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1086 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1087 (aarch64_<sur>shl<mode>): Likewise.
1088 (aarch64_<sur>q<r>shl<mode>
1089 (aarch64_<sur>shll_n<mode>): Likewise.
1090 (aarch64_<sur>shll2_n<mode>): Likewise.
1091 (aarch64_<sur>shr_n<mode>): Likewise.
1092 (aarch64_<sur>sra_n<mode>): Likewise.
1093 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1094 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1095 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1096 (aarch64_cm<optab><mode>): Likewise.
1097 (aarch64_cm<optab>di): Likewise.
1098 (aarch64_cm<optab><mode>): Likewise.
1099 (aarch64_cm<optab>di): Likewise.
1100 (aarch64_cmtst<mode>): Likewise.
1101 (aarch64_cmtstdi): Likewise.
1102 (aarch64_cm<optab><mode>): Likewise.
1103 (*aarch64_fac<optab><mode>): Likewise.
1104 (aarch64_addp<mode>): Likewise.
1105 (aarch64_addpdi): Likewise.
1106 (sqrt<mode>2): Likewise.
1107 (vec_load_lanesoi<mode>): Likewise.
1108 (vec_store_lanesoi<mode>): Likewise.
1109 (vec_load_lanesci<mode>): Likewise.
1110 (vec_store_lanesci<mode>): Likewise.
1111 (vec_load_lanesxi<mode>): Likewise.
1112 (vec_store_lanesxi<mode>): Likewise.
1113 (*aarch64_mov<mode>): Likewise.
1114 (aarch64_ld2<mode>_dreg): Likewise.
1115 (aarch64_ld2<mode>_dreg): Likewise.
1116 (aarch64_ld3<mode>_dreg): Likewise.
1117 (aarch64_ld3<mode>_dreg): Likewise.
1118 (aarch64_ld4<mode>_dreg): Likewise.
1119 (aarch64_ld4<mode>_dreg): Likewise.
1120 (aarch64_tbl1<mode>): Likewise.
1121 (aarch64_tbl2v16qi): Likewise.
1122 (aarch64_combinev16qi): Likewise.
1123 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1124 (aarch64_st2<mode>_dreg): Likewise.
1125 (aarch64_st2<mode>_dreg): Likewise.
1126 (aarch64_st3<mode>_dreg): Likewise.
1127 (aarch64_st3<mode>_dreg): Likewise.
1128 (aarch64_st4<mode>_dreg): Likewise.
1129 (aarch64_st4<mode>_dreg): Likewise.
1130 (*aarch64_simd_ld1r<mode>): Likewise.
1131 (aarch64_frecpe<mode>): Likewise.
1132 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1133 (aarch64_frecps<mode>): Likewise.
1135 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1137 * config/arm/iterators.md (V_elem_ch): New.
1139 (VQH_type): Likewise.
1140 * config/arm/arm.md (is_neon_type): New.
1141 (conds): Use is_neon_type.
1142 (anddi3_insn): Update type attribute.
1143 (xordi3_insn): Likewise.
1144 (one_cmpldi2): Likewise.
1145 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1146 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1147 (*movmisalign<mode>_neon_store): Likewise.
1148 (*movmisalign<mode>_neon_load): Likewise.
1149 (vec_set<mode>_internal): Likewise.
1150 (vec_set<mode>_internal): Likewise.
1151 (vec_setv2di_internal): Likewise.
1152 (vec_extract<mode>): Likewise.
1153 (vec_extract<mode>): Likewise.
1154 (vec_extractv2di): Likewise.
1155 (*add<mode>3_neon): Likewise.
1156 (adddi3_neon): Likewise.
1157 (*sub<mode>3_neon): Likewise.
1158 (subdi3_neon): Likewise.
1159 (fma<VCVTF:mode>4): Likewise.
1160 (fma<VCVTF:mode>4_intrinsic): Likewise.
1161 (*fmsub<VCVTF:mode>4): Likewise.
1162 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1163 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1164 (ior<mode>3): Likewise.
1165 (and<mode>3): Likewise.
1166 (orn<mode>3_neon): Likewise.
1167 (orndi3_neon): Likewise.
1168 (bic<mode>3_neon): Likewise.
1169 (bicdi3_neon): Likewise.
1170 (xor<mode>3): Likewise.
1171 (one_cmpl<mode>2): Likewise.
1172 (abs<mode>2): Likewise.
1173 (neg<mode>2): Likewise.
1174 (negdi2_neon): Likewise.
1175 (*umin<mode>3_neon): Likewise.
1176 (*umax<mode>3_neon): Likewise.
1177 (*smin<mode>3_neon): Likewise.
1178 (*smax<mode>3_neon): Likewise.
1179 (vashl<mode>3): Likewise.
1180 (vashr<mode>3_imm): Likewise.
1181 (vlshr<mode>3_imm): Likewise.
1182 (ashl<mode>3_signed): Likewise.
1183 (ashl<mode>3_unsigned): Likewise.
1184 (neon_load_count): Likewise.
1185 (ashldi3_neon_noclobber): Likewise.
1186 (ashldi3_neon): Likewise.
1187 (signed_shift_di3_neon): Likewise.
1188 (unsigned_shift_di3_neon): Likewise.
1189 (ashrdi3_neon_imm_noclobber): Likewise.
1190 (lshrdi3_neon_imm_noclobber): Likewise.
1191 (<shift>di3_neon): Likewise.
1192 (widen_ssum<mode>3): Likewise.
1193 (widen_usum<mode>3): Likewise.
1194 (quad_halves_<code>v4si): Likewise.
1195 (quad_halves_<code>v4sf): Likewise.
1196 (quad_halves_<code>v8hi): Likewise.
1197 (quad_halves_<code>v16qi): Likewise.
1198 (reduc_splus_v2di): Likewise.
1199 (neon_vpadd_internal<mode>): Likewise.
1200 (neon_vpsmin<mode>): Likewise.
1201 (neon_vpsmax<mode>): Likewise.
1202 (neon_vpumin<mode>): Likewise.
1203 (neon_vpumax<mode>): Likewise.
1204 (*ss_add<mode>_neon): Likewise.
1205 (*us_add<mode>_neon): Likewise.
1206 (*ss_sub<mode>_neon): Likewise.
1207 (*us_sub<mode>_neon): Likewise.
1208 (neon_vadd<mode>_unspec): Likewise.
1209 (neon_vaddl<mode>): Likewise.
1210 (neon_vaddw<mode>): Likewise.
1211 (neon_vhadd<mode>): Likewise.
1212 (neon_vqadd<mode>): Likewise.
1213 (neon_vaddhn<mode>): Likewise.
1214 (neon_vmul<mode>): Likewise.
1215 (neon_vfms<VCVTF:mode>): Likewise.
1216 (neon_vmlal<mode>): Likewise.
1217 (neon_vmls<mode>): Likewise.
1218 (neon_vmlsl<mode>): Likewise.
1219 (neon_vqdmulh<mode>): Likewise.
1220 (neon_vqdmlal<mode>): Likewise.
1221 (neon_vqdmlsl<mode>): Likewise.
1222 (neon_vmull<mode>): Likewise.
1223 (neon_vqdmull<mode>): Likewise.
1224 (neon_vsub<mode>_unspec): Likewise.
1225 (neon_vsubl<mode>): Likewise.
1226 (neon_vsubw<mode>): Likewise.
1227 (neon_vqsub<mode>): Likewise.
1228 (neon_vhsub<mode>): Likewise.
1229 (neon_vsubhn<mode>): Likewise.
1230 (neon_vceq<mode>): Likewise.
1231 (neon_vcge<mode>): Likewise.
1232 (neon_vcgeu<mode>): Likewise.
1233 (neon_vcgt<mode>): Likewise.
1234 (neon_vcgtu<mode>): Likewise.
1235 (neon_vcle<mode>): Likewise.
1236 (neon_vclt<mode>): Likewise.
1237 (neon_vcage<mode>): Likewise.
1238 (neon_vcagt<mode>): Likewise.
1239 (neon_vtst<mode>): Likewise.
1240 (neon_vabd<mode>): Likewise.
1241 (neon_vabdl<mode>): Likewise.
1242 (neon_vaba<mode>): Likewise.
1243 (neon_vabal<mode>): Likewise.
1244 (neon_vmax<mode>): Likewise.
1245 (neon_vmin<mode>): Likewise.
1246 (neon_vpaddl<mode>): Likewise.
1247 (neon_vpadal<mode>): Likewise.
1248 (neon_vpmax<mode>): Likewise.
1249 (neon_vpmin<mode>): Likewise.
1250 (neon_vrecps<mode>): Likewise.
1251 (neon_vrsqrts<mode>): Likewise.
1252 (neon_vqabs<mode>): Likewise.
1253 (neon_vqneg<mode>): Likewise.
1254 (neon_vcls<mode>): Likewise.
1255 (clz<mode>2): Likewise.
1256 (popcount<mode>2): Likewise.
1257 (neon_vrecpe<mode>): Likewise.
1258 (neon_vrsqrte<mode>): Likewise.
1259 (neon_vget_lane<mode>_sext_internal): Likewise.
1260 (neon_vget_lane<mode>_zext_internal): Likewise.
1261 (neon_vdup_n<mode>): Likewise.
1262 (neon_vdup_n<mode>): Likewise.
1263 (neon_vdup_nv2di): Likewise.
1264 (neon_vdup_lane<mode>_interal): Likewise.
1265 (*neon_vswp<mode>): Likewise.
1266 (neon_vcombine<mode>): Likewise.
1267 (float<mode><V_cvtto>2): Likewise.
1268 (floatuns<mode><V_cvtto>2): Likewise.
1269 (fix_trunc<mode><V_cvtto>2): Likewise.
1270 (fixuns_trunc<mode><V_cvtto>2
1271 (neon_vcvt<mode>): Likewise.
1272 (neon_vcvt<mode>): Likewise.
1273 (neon_vcvtv4sfv4hf): Likewise.
1274 (neon_vcvtv4hfv4sf): Likewise.
1275 (neon_vcvt_n<mode>): Likewise.
1276 (neon_vcvt_n<mode>): Likewise.
1277 (neon_vmovn<mode>): Likewise.
1278 (neon_vqmovn<mode>): Likewise.
1279 (neon_vqmovun<mode>): Likewise.
1280 (neon_vmovl<mode>): Likewise.
1281 (neon_vmul_lane<mode>): Likewise.
1282 (neon_vmul_lane<mode>): Likewise.
1283 (neon_vmull_lane<mode>): Likewise.
1284 (neon_vqdmull_lane<mode>): Likewise.
1285 (neon_vqdmulh_lane<mode>): Likewise.
1286 (neon_vqdmulh_lane<mode>): Likewise.
1287 (neon_vmla_lane<mode>): Likewise.
1288 (neon_vmla_lane<mode>): Likewise.
1289 (neon_vmlal_lane<mode>): Likewise.
1290 (neon_vqdmlal_lane<mode>): Likewise.
1291 (neon_vmls_lane<mode>): Likewise.
1292 (neon_vmls_lane<mode>): Likewise.
1293 (neon_vmlsl_lane<mode>): Likewise.
1294 (neon_vqdmlsl_lane<mode>): Likewise.
1295 (neon_vext<mode>): Likewise.
1296 (neon_vrev64<mode>): Likewise.
1297 (neon_vrev32<mode>): Likewise.
1298 (neon_vrev16<mode>): Likewise.
1299 (neon_vbsl<mode>_internal): Likewise.
1300 (neon_vshl<mode>): Likewise.
1301 (neon_vqshl<mode>): Likewise.
1302 (neon_vshr_n<mode>): Likewise.
1303 (neon_vshrn_n<mode>): Likewise.
1304 (neon_vqshrn_n<mode>): Likewise.
1305 (neon_vqshrun_n<mode>): Likewise.
1306 (neon_vshl_n<mode>): Likewise.
1307 (neon_vqshl_n<mode>): Likewise.
1308 (neon_vqshlu_n<mode>): Likewise.
1309 (neon_vshll_n<mode>): Likewise.
1310 (neon_vsra_n<mode>): Likewise.
1311 (neon_vsri_n<mode>): Likewise.
1312 (neon_vsli_n<mode>): Likewise.
1313 (neon_vtbl1v8qi): Likewise.
1314 (neon_vtbl2v8qi): Likewise.
1315 (neon_vtbl3v8qi): Likewise.
1316 (neon_vtbl4v8qi): Likewise.
1317 (neon_vtbl1v16qi): Likewise.
1318 (neon_vtbl2v16qi): Likewise.
1319 (neon_vcombinev16qi): Likewise.
1320 (neon_vtbx1v8qi): Likewise.
1321 (neon_vtbx2v8qi): Likewise.
1322 (neon_vtbx3v8qi): Likewise.
1323 (neon_vtbx4v8qi): Likewise.
1324 (*neon_vtrn<mode>_insn): Likewise.
1325 (*neon_vzip<mode>_insn): Likewise.
1326 (*neon_vuzp<mode>_insn): Likewise.
1327 (neon_vld1<mode>): Likewise.
1328 (neon_vld1_lane<mode>): Likewise.
1329 (neon_vld1_lane<mode>): Likewise.
1330 (neon_vld1_dup<mode>): Likewise.
1331 (neon_vld1_dup<mode>): Likewise.
1332 (neon_vld1_dupv2di): Likewise.
1333 (neon_vst1<mode>): Likewise.
1334 (neon_vst1_lane<mode>): Likewise.
1335 (neon_vst1_lane<mode>): Likewise.
1336 (neon_vld2<mode>): Likewise.
1337 (neon_vld2<mode>): Likewise.
1338 (neon_vld2_lane<mode>): Likewise.
1339 (neon_vld2_lane<mode>): Likewise.
1340 (neon_vld2_dup<mode>): Likewise.
1341 (neon_vst2<mode>): Likewise.
1342 (neon_vst2<mode>): Likewise.
1343 (neon_vst2_lane<mode>): Likewise.
1344 (neon_vst2_lane<mode>): Likewise.
1345 (neon_vld3<mode>): Likewise.
1346 (neon_vld3qa<mode>): Likewise.
1347 (neon_vld3qb<mode>): Likewise.
1348 (neon_vld3_lane<mode>): Likewise.
1349 (neon_vld3_lane<mode>): Likewise.
1350 (neon_vld3_dup<mode>): Likewise.
1351 (neon_vst3<mode>): Likewise.
1352 (neon_vst3qa<mode>): Likewise.
1353 (neon_vst3qb<mode>): Likewise.
1354 (neon_vst3_lane<mode>): Likewise.
1355 (neon_vst3_lane<mode>): Likewise.
1356 (neon_vld4<mode>): Likewise.
1357 (neon_vld4qa<mode>): Likewise.
1358 (neon_vld4qb<mode>): Likewise.
1359 (neon_vld4_lane<mode>): Likewise.
1360 (neon_vld4_lane<mode>): Likewise.
1361 (neon_vld4_dup<mode>): Likewise.
1362 (neon_vst4<mode>): Likewise.
1363 (neon_vst4qa<mode>): Likewise.
1364 (neon_vst4qb<mode>): Likewise.
1365 (neon_vst4_lane<mode>): Likewise.
1366 (neon_vst4_lane<mode>): Likewise.
1367 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1368 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1369 (neon_vec_<US>mult_lo_<mode>): Likewise.
1370 (neon_vec_<US>mult_hi_<mode>): Likewise.
1371 (neon_vec_<US>shiftl_<mode>): Likewise.
1372 (neon_unpack<US>_<mode>): Likewise.
1373 (neon_vec_<US>mult_<mode>): Likewise.
1374 (vec_pack_trunc_<mode>): Likewise.
1375 (neon_vec_pack_trunc_<mode>): Likewise.
1376 (neon_vabd<mode>_2): Likewise.
1377 (neon_vabd<mode>_3): Likewise.
1379 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1381 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1382 (load_pair): Update type attribute.
1383 (store_pair): Update type attribute.
1384 * config/aarch64/iterators.md (q): New.
1386 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1388 * config/arm/types.md: Add new types for Neon insns.
1390 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1391 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1392 Sergey Lega <sergey.s.lega@intel.com>
1393 Anna Tikhonova <anna.tikhonova@intel.com>
1394 Ilya Tocar <ilya.tocar@intel.com>
1395 Andrey Turetskiy <andrey.turetskiy@intel.com>
1396 Ilya Verbin <ilya.verbin@intel.com>
1397 Kirill Yukhin <kirill.yukhin@intel.com>
1398 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1400 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1401 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1402 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1404 (srcp14<mode>): Ditto.
1405 (rsqrt14<mode>): Ditto.
1406 (rsqrt14<mode>): Ditto.
1407 (avx512f_vmscalef<mode>): Ditto.
1408 (avx512f_scalef<mode>): Ditto.
1409 (avx512f_getexp<mode>): Ditto.
1410 (avx512f_sgetexp<mode>): Ditto.
1411 (avx512f_fixupimm<mode>): Ditto.
1412 (avx512f_sfixupimm<mode>): Ditto.
1413 (avx512f_rndscale<mode>): Ditto.
1414 (*avx512er_exp2<mode>): Ditto.
1415 (*avx512er_rcp28<mode>): Ditto.
1416 (avx512er_rsqrt28<mode>): Ditto.
1417 (avx512f_getmant<mode>): Ditto.
1418 (avx512f_getmant<mode>): Ditto.
1419 (avx512f_rndscale<mode>): Fix formatting.
1421 2013-10-15 Martin Jambor <mjambor@suse.cz>
1423 * ipa-utils.h (ipa_edge_within_scc): Declare.
1424 * ipa-cp.c (edge_within_scc): Moved...
1425 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1427 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1428 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1429 Sergey Lega <sergey.s.lega@intel.com>
1430 Anna Tikhonova <anna.tikhonova@intel.com>
1431 Ilya Tocar <ilya.tocar@intel.com>
1432 Andrey Turetskiy <andrey.turetskiy@intel.com>
1433 Ilya Verbin <ilya.verbin@intel.com>
1434 Kirill Yukhin <kirill.yukhin@intel.com>
1435 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1437 * config/i386/predicates.md (const_8_to_15_operand): New.
1438 (const_16_to_31_operand): Ditto.
1439 * config/i386/sse.md (V8FI): New.
1441 (reduc_splus_v8df): Ditto.
1442 (reduc_splus_v16sf): Ditto.
1443 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1444 (vec_extract_hi_<mode>): Ditto.
1445 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1446 (vec_set_lo_<mode>): Ditto.
1447 (vec_set_hi_<mode>): Ditto.
1448 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1449 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1450 (avx512f_pshufd_1): Ditto.
1451 (avx512f_broadcast<mode>): Ditto.
1452 (avx512f_broadcast<mode>): Ditto.
1453 (define_split): Split vec_extract_lo into move.
1454 (ssequartermode): Ditto.
1455 (ssedoublemode): Extened with wider modes.
1456 (vec_extract_lo_<mode>): Ditto.
1458 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1459 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1460 Sergey Lega <sergey.s.lega@intel.com>
1461 Anna Tikhonova <anna.tikhonova@intel.com>
1462 Ilya Tocar <ilya.tocar@intel.com>
1463 Andrey Turetskiy <andrey.turetskiy@intel.com>
1464 Ilya Verbin <ilya.verbin@intel.com>
1465 Kirill Yukhin <kirill.yukhin@intel.com>
1466 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1468 * config/i386/predicates.md (register_or_constm1_operand): New.
1469 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1470 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1471 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1472 UNSPEC_SCATTER_PREFETCH
1474 (avx512f_ucmp<mode>3): Ditto.
1475 (avx512f_vternlog<mode>): Ditto.
1476 (avx512f_align<mode>): Ditto.
1477 (<shift_insn><mode>3): Ditto.
1478 (avx512f_<rotate>v<mode>): Ditto.
1479 (avx512f_<rotate><mode>): Ditto.
1480 (avx512f_eq<mode>3): Ditto.
1481 (avx512f_eq<mode>3_1): Ditto.
1482 (avx512f_gt<mode>3): Ditto.
1483 (avx512f_testm<mode>3): Ditto.
1484 (avx512f_testnm<mode>3): Ditto.
1485 (avx512pf_gatherpf<mode>): Ditto.
1486 (*avx512pf_gatherpf<mode>_mask): Ditto.
1487 (*avx512pf_gatherpf<mode>): Ditto.
1488 (avx512pf_scatterpf<mode>): Ditto.
1489 (*avx512pf_scatterpf<mode>_mask): Ditto.
1490 (*avx512pf_scatterpf<mode>): Ditto.
1491 (avx512f_vec_dup_gpr<mode>): Ditto.
1492 (clz<mode>2): Ditto.
1493 (conflict<mode>): Ditto.
1494 (REDUC_SMINMAX_MODE): Extened with wider modes.
1495 (reduc_<code>_<mode>): Ditto.
1496 (vlshr<mode>3): Ditto.
1497 (vashl<mode>3): Ditto.
1499 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1500 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1501 Sergey Lega <sergey.s.lega@intel.com>
1502 Anna Tikhonova <anna.tikhonova@intel.com>
1503 Ilya Tocar <ilya.tocar@intel.com>
1504 Andrey Turetskiy <andrey.turetskiy@intel.com>
1505 Ilya Verbin <ilya.verbin@intel.com>
1506 Kirill Yukhin <kirill.yukhin@intel.com>
1507 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1509 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1512 (avx512fmaskmode): Ditto.
1513 (bcstscalarsuff): Ditto.
1514 (avx512f_blendm<mode>): Ditto.
1515 (cmp_imm_predicate): Ditto.
1516 (avx512f_cmp<mode>3): Ditto.
1517 (avx512f_vec_dup<mode>): Ditto.
1518 (avx512f_vec_dup_mem<mode>): Ditto.
1519 (avx512f_vpermi2var<mode>3): Ditto.
1520 (avx512f_vpermt2var<mode>3): Ditto.
1521 (vec_init<mode>): Ditto.
1522 (avx512f_gathersi<mode>): Ditto.
1523 (*avx512f_gathersi<mode>): Ditto.
1524 (*avx512f_gathersi<mode>_2): Ditto.
1525 (avx512f_gatherdi<mode>): Ditto.
1526 (*avx512f_gatherdi<mode>): Ditto.
1527 (*avx512f_gatherdi<mode>_2): Ditto.
1528 (avx512f_scattersi<mode>): Ditto.
1529 (*avx512f_scattersi<mode>): Ditto.
1530 (avx512f_scatterdi<mode>): Ditto.
1531 (*avx512f_scatterdi<mode>): Ditto.
1532 (sseintprefix): Extened with wider modes.
1533 (VEC_GATHER_IDXSI): Ditto.
1534 (VEC_GATHER_IDXDI): Ditto.
1535 (VEC_GATHER_SRCDI): Ditto.
1537 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1538 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1540 * config/arm/t-aprofile: New file.
1541 * config.gcc: Handle --with-multilib-list option.
1543 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1545 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1548 2013-10-15 Richard Biener <rguenther@suse.de>
1550 * tree-loop-distribution.c (build_empty_rdg): Inline into
1552 (rdg_flag_vertex): Inline into single user.
1553 (rdg_flag_vertex_and_dependent): Likewise.
1554 (build_rdg_partition_for_vertex): Remove processed bitmap.
1555 (rdg_build_partitions): Simplify.
1557 2013-10-15 Richard Biener <rguenther@suse.de>
1559 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1560 Restructure forwarding through conversions and copies to
1561 avoid performing copy-propagation the wrong way. Adjust
1562 recursion invocations.
1563 (forward_propagate_addr_expr): Add argument stating if we
1564 are recursing from a single-use.
1565 (ssa_forward_propagate_and_combine): Adjust.
1567 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1569 * dumpfile.h (gcc::dump_manager): New class, to hold state
1570 relating to dumpfile management.
1571 (get_dump_file_name): Remove in favor of method of dump_manager.
1572 (dump_initialized_p): Likewise.
1573 (dump_start): Likewise.
1574 (dump_finish): Likewise.
1575 (dump_switch_p): Likewise.
1576 (dump_register): Likewise.
1577 (get_dump_file_info): Likewise.
1578 * context.c (gcc::context::context): Construct the dump_manager
1580 * context.h (gcc::context::get_dumps): New.
1581 (gcc::context::m_dumps): New.
1582 * coverage.c (coverage_init): Port to dump_manager API.
1583 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1584 (extra_dump_files_in_use): Likewise.
1585 (extra_dump_files_alloced): Likewise.
1586 (gcc::dump_manager::dump_manager): New.
1587 (dump_register): Convert to...
1588 (gcc::dump_manager::dump_register): ...method, replacing
1589 function-static next_dump with m_next_dump field.
1590 (get_dump_file_info): Convert to...
1591 (gcc::dump_manager::get_dump_file_info): ...method.
1592 (get_dump_file_name): Convert to...
1593 (gcc::dump_manager::get_dump_file_name): ...method.
1594 (dump_start): Convert to...
1595 (gcc::dump_manager::dump_start): ...method.
1596 (dump_finish): Convert to...
1597 (gcc::dump_manager::dump_finish): ...method.
1598 (dump_begin): Replace body with...
1599 (gcc::dump_manager::dump_begin): ...new method.
1600 (dump_phase_enabled_p): Convert to...
1601 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1602 (dump_phase_enabled_p): Convert to...
1603 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1604 (dump_initialized_p): Convert to...
1605 (gcc::dump_manager::dump_initialized_p): ...method.
1606 (dump_flag_name): Replace body with...
1607 (gcc::dump_manager::dump_flag_name): ...new method.
1608 (dump_enable_all): Convert to...
1609 (gcc::dump_manager::dump_enable_all): ...new method.
1610 (opt_info_enable_passes): Convert to...
1611 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1612 (dump_switch_p_1): Convert to...
1613 (gcc::dump_manager::dump_switch_p_1): ...new method.
1614 (dump_switch_p): Convert to...
1615 (gcc::dump_manager::dump_switch_p): ...new method.
1616 (opt_info_switch_p): Port to dump_manager API.
1617 (enable_rtl_dump_file): Likewise.
1618 * opts-global.c (handle_common_deferred_options): Port to new
1620 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1621 (pass_manager::register_one_dump_file): Likewise.
1622 (pass_manager::register_pass): Likewise.
1623 (pass_init_dump_file): Likewise.
1624 (pass_fini_dump_file): Likewise.
1625 * statistics.c (statistics_early_init): Likewise.
1627 2013-10-14 Richard Biener <rguenther@suse.de>
1629 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1630 iterative_hash_canonical_type, gimple_canonical_type_hash,
1631 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1632 gimple_register_canonical_type, print_gimple_types_stats,
1633 free_gimple_type_tables): Move to lto/lto.c
1634 (gt-gimple.h): Do not include.
1635 * gimple.h (gimple_register_canonical_type,
1636 print_gimple_types_stats, free_gimple_type_tables): Remove.
1637 * Makefile.in (GTFILES): Remove gimple.c.
1639 2013-10-14 Travis Snoozy <quandary@remstate.com>
1642 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1643 scanning for msp430x targets.
1645 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1648 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1649 (registers_ok_for_ldd_peep): Move around.
1650 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1651 * config/sparc/sparc.md (widening peepholes): Use it.
1653 2013-10-14 Richard Biener <rguenther@suse.de>
1657 * gimple.c (iterative_hash_canonical_type): Make sure to
1658 record the hash into the correct hashtable slot.
1660 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1662 PR rtl-optimization/58662
1663 * combine.c (try_combine): Take into account death nodes on I2 when
1664 splitting a PARALLEL of two independent SETs. Fix dump message.
1666 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1669 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1670 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1671 * config/sh/t-sh (sh_treg_combine.o): New entry.
1672 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1673 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1674 (register_sh_passes): New function. Register sh_treg_combine pass.
1675 (sh_option_override): Invoke it.
1676 (sh_canonicalize_comparison): Handle op0_preserve_value.
1677 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1678 opportunities. Canonicalize branch condition.
1679 (nott): Allow only if pseudos can be created for non-SH2A.
1681 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1684 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1685 (ix86_expand_movmem): Replace copy_addr_to_reg with
1686 ix86_copy_addr_to_reg.
1687 (ix86_expand_setmem): Likewise.
1689 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1691 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1692 provided by ix86_fp_compare_mode instead of CCFPUmode.
1694 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1696 * config/aarch64/arm_neon.h
1697 (vtbx<1,3>_<psu>8): Fix register constriants.
1699 2013-10-11 Jeff Law <law@redhat.com>
1701 PR tree-optimization/58640
1702 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1703 threading paths that cross over two loop entry points.
1705 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1707 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1708 handle vector float as well.
1709 (*vsx_le_perm_load_v4si): Likewise.
1710 (*vsx_le_perm_store_v2di): Likewise.
1711 (*vsx_le_perm_store_v4si): Likewise.
1713 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1715 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1716 directly to circumvent subtract from splat{31} workaround.
1717 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1719 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1720 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1721 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1722 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1723 separate big and little endian logic.
1724 (*altivec_vperm_<mode>_internal): New define_insn.
1725 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1726 separate big and little endian logic.
1727 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1728 (vec_permv16qi): Add little endian logic.
1730 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1732 * doc/extend.texi (returns_nonnull): Remove arguments.
1734 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1735 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1736 Sergey Lega <sergey.s.lega@intel.com>
1737 Anna Tikhonova <anna.tikhonova@intel.com>
1738 Ilya Tocar <ilya.tocar@intel.com>
1739 Andrey Turetskiy <andrey.turetskiy@intel.com>
1740 Ilya Verbin <ilya.verbin@intel.com>
1741 Kirill Yukhin <kirill.yukhin@intel.com>
1742 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1744 * config/i386/sse.md (VI48F_256_512): New.
1745 (avx2_permvar<mode>): Change to ...
1746 (<avx2_avx512f>_permvar<mode>): This.
1748 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1749 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1750 Sergey Lega <sergey.s.lega@intel.com>
1751 Anna Tikhonova <anna.tikhonova@intel.com>
1752 Ilya Tocar <ilya.tocar@intel.com>
1753 Andrey Turetskiy <andrey.turetskiy@intel.com>
1754 Ilya Verbin <ilya.verbin@intel.com>
1755 Kirill Yukhin <kirill.yukhin@intel.com>
1756 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1758 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1759 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1760 * config/i386/sse.md (VI4_AVX): New.
1761 (sf2simodelower): Ditto.
1762 (sse2_cvtps2dq): Change to ...
1763 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1765 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1766 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1767 Sergey Lega <sergey.s.lega@intel.com>
1768 Anna Tikhonova <anna.tikhonova@intel.com>
1769 Ilya Tocar <ilya.tocar@intel.com>
1770 Andrey Turetskiy <andrey.turetskiy@intel.com>
1771 Ilya Verbin <ilya.verbin@intel.com>
1772 Kirill Yukhin <kirill.yukhin@intel.com>
1773 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1775 * config/i386/sse.md (V_512): New.
1777 (vcond<V_512:mode><VF_512:mode>): Ditto.
1778 (vcond<V_512:mode><VI_512:mode>): Ditto.
1779 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1781 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1782 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1783 Sergey Lega <sergey.s.lega@intel.com>
1784 Anna Tikhonova <anna.tikhonova@intel.com>
1785 Ilya Tocar <ilya.tocar@intel.com>
1786 Andrey Turetskiy <andrey.turetskiy@intel.com>
1787 Ilya Verbin <ilya.verbin@intel.com>
1788 Kirill Yukhin <kirill.yukhin@intel.com>
1789 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1791 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1792 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1794 (fma<mode>4): Removed condition.
1795 (fms<mode>4): Ditto.
1796 (fnma<mode>4): Ditto.
1797 (fnms<mode>4): Ditto.
1798 (fma4i_fmadd_<mode>): Ditto.
1799 (*fma_fmadd_<mode>): Ditto.
1800 (*fma_fmsub_<mode>): Ditto.
1801 (*fma_fnmadd_<mode>): Ditto.
1802 (*fma_fnmsub_<mode>): Ditto.
1803 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1804 (*fma_fmaddsub_<mode>): Ditto.
1805 (*fma_fmsubadd_<mode>): Ditto.
1806 (*fmai_fmadd_<mode>): Ditto.
1807 (*fmai_fmsub_<mode>): Ditto.
1808 (*fmai_fnmadd_<mode>): Ditto.
1809 (*fmai_fnmsub_<mode>): Ditto.
1811 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1812 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1813 Sergey Lega <sergey.s.lega@intel.com>
1814 Anna Tikhonova <anna.tikhonova@intel.com>
1815 Ilya Tocar <ilya.tocar@intel.com>
1816 Andrey Turetskiy <andrey.turetskiy@intel.com>
1817 Ilya Verbin <ilya.verbin@intel.com>
1818 Kirill Yukhin <kirill.yukhin@intel.com>
1819 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1821 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1822 (VI124_256): Changed to ...
1823 (VI124_256_48_512): This.
1824 (ssepackmode): Extended with wider modes.
1825 (<code><mode>3): Changed iterator.
1826 (*avx2_<code><mode>3): Ditto.
1827 (vec_pack_trunc_<mode>): Ditto.
1829 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1830 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1831 Sergey Lega <sergey.s.lega@intel.com>
1832 Anna Tikhonova <anna.tikhonova@intel.com>
1833 Ilya Tocar <ilya.tocar@intel.com>
1834 Andrey Turetskiy <andrey.turetskiy@intel.com>
1835 Ilya Verbin <ilya.verbin@intel.com>
1836 Kirill Yukhin <kirill.yukhin@intel.com>
1837 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1839 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1840 (VI8F_256_512): Ditto.
1841 (abs<mode>2): Changed iterator.
1842 (avx2_perm<mode>): Changed to ...
1843 (<avx2_avx512f>_perm<mode>): This.
1844 (avx2_perm<mode>_1): Changed to ...
1845 (<avx2_avx512f>_perm<mode>_1): This.
1847 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1848 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1849 Sergey Lega <sergey.s.lega@intel.com>
1850 Anna Tikhonova <anna.tikhonova@intel.com>
1851 Ilya Tocar <ilya.tocar@intel.com>
1852 Andrey Turetskiy <andrey.turetskiy@intel.com>
1853 Ilya Verbin <ilya.verbin@intel.com>
1854 Kirill Yukhin <kirill.yukhin@intel.com>
1855 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1857 * config/i386/sse.md (VI48_AVX512F): New.
1858 (VI48_AVX2): Changed to ...
1859 (VI48_AVX2_48_AVX512F): This.
1860 (avx2_ashrv<mode>): Changed to ...
1861 (<avx2_avx512f>_ashrv<mode>): This.
1862 (avx2_<shift_insn>v<mode>): Changed to ...
1863 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1865 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1866 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1867 Sergey Lega <sergey.s.lega@intel.com>
1868 Anna Tikhonova <anna.tikhonova@intel.com>
1869 Ilya Tocar <ilya.tocar@intel.com>
1870 Andrey Turetskiy <andrey.turetskiy@intel.com>
1871 Ilya Verbin <ilya.verbin@intel.com>
1872 Kirill Yukhin <kirill.yukhin@intel.com>
1873 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1875 * config/i386/sse.md (VI4_AVX512F): New.
1876 (VI8_AVX2_AVX512F): Ditto.
1877 (mul<mode>3): Extended with wider modes.
1878 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1879 (mul<mode>3): Ditto.
1880 (vec_widen_<s>mult_odd_<mode>): Ditto.
1882 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1883 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1884 Sergey Lega <sergey.s.lega@intel.com>
1885 Anna Tikhonova <anna.tikhonova@intel.com>
1886 Ilya Tocar <ilya.tocar@intel.com>
1887 Andrey Turetskiy <andrey.turetskiy@intel.com>
1888 Ilya Verbin <ilya.verbin@intel.com>
1889 Kirill Yukhin <kirill.yukhin@intel.com>
1890 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1892 * config/i386/sse.md (VI2_AVX512F): New.
1893 (VI124_AVX512F): Ditto.
1894 (sseunpackmode): Extended with wider modes.
1895 (sseunpackfltmode): Ditto.
1896 (vec_unpacks_float_hi_<mode>): Ditto.
1897 (vec_unpacks_float_lo_<mode>): Ditto.
1898 (vec_unpacku_float_hi_<mode>): Ditto.
1899 (vec_unpacku_float_lo_<mode>): Ditto.
1900 (vec_unpacks_lo_<mode>): Ditto.
1901 (vec_unpacks_hi_<mode>): Ditto.
1902 (vec_unpacku_lo_<mode>): Ditto.
1903 (vec_unpacku_hi_<mode>): Ditto.
1905 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1906 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1907 Sergey Lega <sergey.s.lega@intel.com>
1908 Anna Tikhonova <anna.tikhonova@intel.com>
1909 Ilya Tocar <ilya.tocar@intel.com>
1910 Andrey Turetskiy <andrey.turetskiy@intel.com>
1911 Ilya Verbin <ilya.verbin@intel.com>
1912 Kirill Yukhin <kirill.yukhin@intel.com>
1913 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1915 * config/i386/i386.md (multdiv): New.
1916 (multdiv_mnemonic): Ditto.
1917 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1918 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1919 (<sse>_vmdiv<mode>3): Removed.
1921 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1922 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1923 Sergey Lega <sergey.s.lega@intel.com>
1924 Anna Tikhonova <anna.tikhonova@intel.com>
1925 Ilya Tocar <ilya.tocar@intel.com>
1926 Andrey Turetskiy <andrey.turetskiy@intel.com>
1927 Ilya Verbin <ilya.verbin@intel.com>
1928 Kirill Yukhin <kirill.yukhin@intel.com>
1929 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1931 * config/i386/sse.md (V): Extended with wider modes.
1933 (ssehalfvecmode): Ditto.
1935 (ssepackfltmode): Ditto.
1936 (avx_vec_concat<mode>): Ditto.
1937 (V_256_512): New iterator.
1938 (VF2_512_256): Ditto.
1939 (si2dfmode): New attribute.
1940 (si2dfmodelower): Ditto.
1942 (concat_tg_mode): Ditto.
1943 (floatv4siv4df2): Changed to ...
1944 (float<si2dfmodelower><mode>2): This.
1945 (avx_cvtps2pd256): Changed to ...
1946 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1947 (vec_pack_trunc_v4df): Changed to ...
1948 (vec_pack_trunc_<mode>): This.
1949 (avx_vpermil<mode>): Changed to ...
1950 (<sse2_avx_avx512f>_vpermil<mode>): This.
1951 (<fixsuffix>fix_truncv8dfv8si2): New.
1952 (vec_pack_sfix_trunc_v8df): Ditto.
1953 (avx512f_rndscale<mode>): Ditto.
1954 (avx512f_roundpd512): Ditto.
1955 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1957 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1958 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1959 Sergey Lega <sergey.s.lega@intel.com>
1960 Anna Tikhonova <anna.tikhonova@intel.com>
1961 Ilya Tocar <ilya.tocar@intel.com>
1962 Andrey Turetskiy <andrey.turetskiy@intel.com>
1963 Ilya Verbin <ilya.verbin@intel.com>
1964 Kirill Yukhin <kirill.yukhin@intel.com>
1965 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1967 * config/i386/i386.md (any_fix): New iterator.
1968 (fixsuffix): New attribute.
1969 * config/i386/sse.md (VF1): Extened with wider modes.
1973 (sseintvecmodelower): Ditto.
1974 (ssescalarmode): Ditto.
1975 (ssescalarnum): Ditto.
1977 (ssexmmmode): Ditto.
1978 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1979 (<sse>_rcp<mode>2): Change iterator.
1980 (rsqrt<mode>2): Ditto.
1981 (<sse>_rsqrt<mode>2): Ditto.
1982 (avx2_vec_dup<mode>): Ditto.
1983 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1984 (round<mode>2_sfix): Ditto.
1985 (avx2_pbroadcast<mode>): Ditto.
1986 (*andnot<mode>3): Handle XI mode.
1987 (*<code><mode>3): Ditto.
1988 (AVXTOSSEMODE): Removed.
1989 (avx_vpermil<mode>): Changed to ...
1990 (<sse2_avx_avx512f>_vpermil<mode>): This.
1992 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1993 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1994 Sergey Lega <sergey.s.lega@intel.com>
1995 Anna Tikhonova <anna.tikhonova@intel.com>
1996 Ilya Tocar <ilya.tocar@intel.com>
1997 Andrey Turetskiy <andrey.turetskiy@intel.com>
1998 Ilya Verbin <ilya.verbin@intel.com>
1999 Kirill Yukhin <kirill.yukhin@intel.com>
2000 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2002 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2003 (<sse>_comi): Ditto.
2004 (<sse>_ucomi): Ditto.
2005 (sse_cvtss2siq_2): Ditto.
2006 (sse2_cvtsd2si): Ditto.
2007 (sse2_cvtsd2siq): Ditto.
2008 (sse2_cvttsd2si): Ditto.
2009 (sse2_cvttsd2siq): Ditto.
2010 (<shift_insn><mode>3): Ditto.
2011 (sse2_cvtsi2sdq): Update constraint and prefix.
2012 (sse_cvtsi2ss): Update prefix.
2013 (sse_cvtsi2ssq): Ditto.
2015 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2017 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2018 ignore internal calls.
2020 2013-10-11 Richard Biener <rguenther@suse.de>
2022 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2023 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2024 TREE_OVERFLOW is set.
2026 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2028 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2030 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2031 * gimple.h: Likewise.
2033 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2035 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2038 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2040 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2041 * configure: Regenerate.
2043 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2045 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2046 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2047 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2048 similarly pp_right_paren.
2049 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2050 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2051 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2052 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2054 (omp_declare_simd_clauses_equal,
2055 omp_remove_redundant_declare_simd_attrs): New functions.
2056 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2057 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2058 * tree.h (OMP_LOOP_CHECK): Define.
2059 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2060 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2061 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2062 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2063 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2064 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2065 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2066 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2067 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2068 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2069 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2070 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2071 (omp_remove_redundant_declare_simd_attrs): New prototype.
2072 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2073 GIMPLE_OMP_TEAMS): New codes.
2074 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2075 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2077 (target_nesting_level): New variable.
2078 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2079 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2080 for collapse > 1 static schedule unless ordered.
2081 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2082 (determine_parallel_type): Adjust get_ws_args_for caller.
2083 (install_var_field): Handle mask & 4 for double indirection.
2084 (scan_sharing_clauses): Ignore shared clause on teams construct.
2085 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2086 (create_omp_child_function): If inside target or declare target
2087 constructs, set "omp declare target" attribute on the child function.
2088 (find_combined_for): New function.
2089 (scan_omp_parallel): Handle combined loops.
2090 (scan_omp_target, scan_omp_teams): New functions.
2091 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2092 restrictions and set ctx->cancellable for cancellable constructs.
2093 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2094 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2095 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2096 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2097 (omp_clause_aligned_alignment): New function.
2098 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2099 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2100 on teams constructs. Handle user defined reductions and new
2102 (lower_reduction_clauses): Don't set placeholder to address of ref
2103 if it has already the right type.
2104 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2105 (expand_parallel_call): Use the new non-_start suffixed builtins,
2106 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2107 and GOMP_parallel_end after the call.
2108 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2109 (expand_omp_for_init_counts): Handle combined loops.
2110 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2112 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2113 end of cancellable loops.
2114 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2115 Likewise. Handle collapse > 1 loops.
2116 (expand_omp_simd): Handle combined loops.
2117 (expand_omp_for): Add inner_stmt argument, adjust callers of
2118 expand_omp_for* functions, use expand_omp_for_static*chunk even
2119 for collapse > 1 unless ordered.
2120 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2121 of cancellable sections.
2122 (expand_omp_single): Remove need_barrier variable, just rely on
2123 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2124 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2125 (expand_omp_atomic_load, expand_omp_atomic_store,
2126 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2127 (expand_omp_target): New function.
2128 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2129 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2130 (build_omp_regions_1): Immediately close region for
2131 GF_OMP_TARGET_KIND_UPDATE.
2132 (maybe_add_implicit_barrier_cancel): New function.
2133 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2135 (lower_omp_single): Likewise. Add clobber after the barrier.
2136 (lower_omp_taskgroup): New function.
2137 (lower_omp_for): Handle combined loops. Adjust
2138 lower_rec_input_clauses caller. Handle cancellation.
2139 (lower_depend_clauses): New function.
2140 (lower_omp_taskreg): Lower depend clauses. Adjust
2141 lower_rec_input_clauses caller. Add clobber after the call. Handle
2143 (lower_omp_target, lower_omp_teams): New functions.
2144 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2145 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2146 and GOMP_cancellation_point calls.
2147 (lower_omp): Fold stmts inside of target region.
2148 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2149 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2150 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2151 (BT_FN_VOID_OMPFN_PTR_UINT,
2152 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2153 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2154 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2155 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2156 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2157 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2158 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2159 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2160 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2161 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2162 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2163 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2164 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2165 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2166 BUILT_IN_GOMP_PARALLEL_END.
2167 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2168 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2169 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2170 GF_OMP_FOR_KIND_DISTRIBUTE.
2171 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2172 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2173 (dump_gimple_omp_return): Print lhs if it has any.
2174 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2175 gimple_omp_atomic_seq_cst_p.
2176 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2177 and GIMPLE_OMP_TEAMS.
2178 * langhooks.c (lhd_omp_mappable_type): New function.
2179 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2180 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2182 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2183 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2184 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2185 (struct gimplify_omp_ctx): Add combined_loop field.
2186 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2187 on stmts inside of target region.
2188 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2189 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2190 ORT_TARGET and ORT_TARGET_DATA.
2191 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2193 (omp_notice_threadprivate_variable): Complain about threadprivate
2194 variables in target region.
2195 (omp_notice_variable): Complain about vars with non-mappable type
2196 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2197 (omp_check_private): Ignore ORT_TARGET* regions.
2198 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2199 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2200 (find_combined_omp_for): New function.
2201 (gimplify_omp_for): Handle gimplification of combined loops.
2202 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2204 (gimplify_omp_target_update): New function.
2205 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2206 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2207 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2208 (gimplify_body): If fndecl has "omp declare target" attribute, add
2209 implicit ORT_TARGET context around it.
2210 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2211 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2212 * tree-nested.c (convert_nonlocal_reference_stmt,
2213 convert_local_reference_stmt, convert_gimple_call): Handle
2214 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2215 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2216 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2217 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2218 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2219 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2220 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2221 BUILT_IN_GOMP_LOOP_END_CANCEL,
2222 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2223 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2224 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2225 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2226 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2227 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2228 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2229 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2230 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2231 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2232 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2233 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2234 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2235 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2236 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2237 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2238 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2239 gimple_build_omp_teams): New functions.
2240 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2241 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2242 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2243 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2244 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2245 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2246 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2247 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2248 GF_OMP_ATOMIC_SEQ_CST): New.
2249 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2250 gimple_build_omp_teams): New prototypes.
2251 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2252 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2253 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2254 GIMPLE_OMP_SINGLE as end of range.
2255 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2256 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2257 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2258 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2259 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2260 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2261 gimple_omp_target_kind, gimple_omp_target_set_kind,
2262 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2263 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2264 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2265 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2266 gimple_omp_teams_set_clauses): New inlines.
2267 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2268 and GIMPLE_OMP_TASKGROUP.
2269 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2270 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2271 enum omp_clause_proc_bind_kind): New.
2272 (union omp_clause_subcode): Add depend_kind, map_kind and
2273 proc_bind_kind fields.
2274 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2275 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2276 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2277 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2278 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2280 2013-10-10 Teresa Johnson <tejohnson@google.com>
2282 * predict.c (tree_estimate_probability): Add new parameter
2283 for estimate_bb_frequencies.
2284 (estimate_bb_frequencies): Add new parameter to force estimation.
2285 (rebuild_frequencies): When max frequency in function is small,
2286 recompute counts from frequencies.
2287 * predict.h (estimate_bb_frequencies): New parameter.
2289 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2291 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2292 optimizations are disabled.
2294 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2296 * coverage.c (coverage_finish): Fix leak of da_file_name.
2298 2013-10-10 Jan Hubicka <jh@suse.cz>
2300 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2301 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2304 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2307 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2308 if any labels are in FALLTHRU_BB, use a special label emitted
2309 immediately after the asm goto insn rather than label_rtx
2311 (expand_asm_stmt): Adjust caller.
2312 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2313 edge if the last insn in predecessor is a jump with single successor,
2314 but it isn't simplejump_p.
2316 2013-10-10 Richard Biener <rguenther@suse.de>
2318 PR tree-optimization/58656
2319 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2321 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2323 * gimplify.c: Include expr.h and tm_p.h for targets with special
2324 va-arg padding requirements.
2326 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2328 * tree-flow.h: Move some prototypes to gimple.h.
2329 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2330 * gimple.h: Relocate some prototypes from tree-flow.h
2331 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2333 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2334 (build_va_arg_indirect_ref): Relocate and make static.
2335 (std_gimplify_va_arg_expr): Relocate here.
2336 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2337 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2339 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2341 * doc/md.texi: Document the mnemonic attribute.
2343 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2346 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2347 (set_attr_alternative x ...) when searching for user defined
2350 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2352 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2354 2013-10-09 Easwaran Raman <eraman@google.com>
2356 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2357 * cfgexpand.c (defer_stack_allocation): ...use here
2358 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2360 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2362 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2363 (operand_equal_for_value_replacement): New function, extracted from
2364 value_replacement and enhanced to catch more cases.
2365 (value_replacement): Use operand_equal_for_value_replacement.
2367 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2369 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2370 get_max_loop_iterations.
2372 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2374 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2375 Place comment above function.
2377 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2379 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2380 are not related to tree-cfg.c.
2381 * tree-ssa-address.h: New file. Relocate prototypes.
2382 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2383 (addr_for_mem_ref): New. Combine call to get_address_description and
2384 return addr_for_mem_ref.
2385 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2386 * tree-ssa-live.h: Adjust prototypes.
2387 * passes.c: Include tree-ssa-live.h.
2388 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2389 * graphite.c (graphite_transform_loops): Make static.
2390 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2391 make_pass_graphite, pass_data_graphite_transforms,
2392 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2393 * ipa-pure-const.c (warn_function_noreturn): Make static.
2394 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2395 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2396 Relocate from tree-cfg.c
2397 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2399 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2400 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2401 Move to ipa-pure-const.c.
2402 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2403 Relocate from tree-optimize.c.
2404 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2405 make_pass_fixup_cfg): Move to tree-cfg.c.
2406 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2407 Relocate some prototypes.
2408 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2409 * tree-dump.c (dump_function_to_file): Remove prototype.
2410 Add tree-flow.h to the include file.
2411 * tree-dump.h: Remove prototype.
2412 * tree-parloops.h: New File. Add prototypes.
2413 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2414 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2415 from tree-ssa-loop.c.
2416 * tree-predcom.c (run_tree_predictive_commoning,
2417 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2418 Relocate here from tree-ssa-loop.c.
2419 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2420 ssa_name_values.release ().
2421 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2422 (ssa_name_values): Relocate from tree-flow.h.
2423 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2424 * tree-ssa-loop.c (run_tree_predictive_commoning,
2425 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2426 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2427 make_pass_graphite, pass_data_graphite_transforms,
2428 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2429 tree_parallelize_loops, pass_data_parallelize_loops,
2430 make_pass_parallelize_loops): Move to other files.
2431 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2433 * tree.h: Remove prototypes from tree-address.c.
2435 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2437 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2438 (struct int_tree_map): Move to tree-hasher.h
2439 (SCALE, LABEL, PERCENT): Move to gimple.h
2440 * tree-flow-inline.h: Delete. Move functions to other files.
2441 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2442 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2443 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2444 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2445 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2446 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2447 * tree-hasher.h: Don't include tree-flow.h.
2448 (struct int_tree_map): Relocate from tree-flow.h.
2449 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2450 tree-flow-inline.h and make static.
2451 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2453 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2454 tree-flow-inline.h and make static.
2455 * tree.h (is_global_var, may_be_aliased): Relocate from
2457 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2458 * value-prof.c: No longer include tree-flow-inline.h.
2459 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2461 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2463 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2464 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2465 (enum move_pos): Move to tree-ssa-loop-im.h
2466 * cfgloop.h: Move some prototypes.
2467 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2468 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2469 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2470 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2471 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2472 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2473 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2474 (enum move_pos): Relocate here.
2475 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2477 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2478 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2479 make_pass_lim): Relocate here from tree-ssa-loop.c.
2480 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2482 (loop_edge_to_cancel, unloop_loops): Make static.
2483 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2484 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2485 (tree_complete_unroll, gate_tree_complete_unroll,
2486 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2487 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2488 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2489 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2490 (stmt_invariant_in_loop_p): Remove unused function.
2491 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2492 * tree-ssa-loop-manip.h: New file. Add prototypes.
2493 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2494 (gcov_type_to_double_int): Move to cfgloop.h.
2495 (double_int_cmp, bound_index,
2496 estimate_numbers_of_iterations_loop): Make static.
2497 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2498 (max_loop_iterations): Factor out get_max_loop_iterations.
2499 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2501 * tree-ssa-loop-niter.h: New file. Add prototypes.
2502 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2503 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2504 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2505 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2506 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2507 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2508 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2509 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2510 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2511 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2512 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2513 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2514 pass_data_complete_unroll, make_pass_complete_unroll,
2515 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2516 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2517 tree-ssa-loop-ivcanon.c.
2518 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2519 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2520 tree-ssa-loop-prefetch.c.
2521 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2523 (get_lsm_tmp_name): Relocate and add suffix parameter.
2524 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2525 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2526 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2527 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2528 (get_estimated_loop_iterations): Factor out accessor from
2529 estimated_loop_iterations in tree-ssa-loop-niter.c.
2530 (get_max_loop_iterations): Factor out accessor from
2531 _max_loop_iterations in tree-ssa-niter.c.
2532 * loop-unroll.c (decide_unroll_constant_iterations,
2533 decide_unroll_runtime_iterations, decide_peel_simple,
2534 decide_unroll_stupid): Use new get_* accessors.
2536 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2538 PR tree-optimization/20318
2539 * doc/extend.texi (returns_nonnull): New function attribute.
2540 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2542 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2543 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2545 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2548 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2549 false if both components are bitfields.
2551 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2553 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2554 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2555 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2556 * config/aarch64/aarch64.h
2557 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2558 * config/aarch64/aarch64-simd-builtins.def
2559 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2561 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2563 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2565 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2567 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2569 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2571 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2572 (vneg_f64): New intrinsic.
2573 (vneg_s8): Asm replaced with C.
2574 (vneg_s16): Likewise.
2575 (vneg_s32): Likewise.
2576 (vneg_s64): New intrinsic.
2577 (vnegq_f32): Asm replaced with C.
2578 (vnegq_f64): Likewise.
2579 (vnegq_s8): Likewise.
2580 (vnegq_s16): Likewise.
2581 (vnegq_s32): Likewise.
2582 (vnegq_s64): Likewise.
2584 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2586 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2588 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2590 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2591 packed stack special handling.
2592 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2593 back to fixed stack slots for FPRs saved due to stdarg.
2595 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2597 * config/s390/s390.c (s390_frame_info): Restructure function.
2599 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2601 * config/s390/s390.c (struct s390_frame_layout): New field
2603 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2604 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2605 regs_ever_clobbered to char*.
2606 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2607 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2608 (s390_register_info): Enable FPR save slots. Move/Copy some
2609 functionality into ...
2610 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2611 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2613 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2614 not imply a stack frame.
2615 (s390_init_frame_layout): Remove variable clobbered_regs.
2616 (s390_update_register_info): Remove function.
2617 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2618 cannot be used for register renaming.
2619 (s390_hard_regno_scratch_ok): New function.
2620 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2621 (s390_initial_elimination_offset): Change offset calculation of
2622 the return address pointer.
2623 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2624 (restore_gprs): Set frame related flag.
2625 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2626 (s390_emit_prologue): Call s390_register_info instead of
2627 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2628 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2629 (s390_optimize_prologue): Call s390_optimize_register_info.
2630 Try to remove also FPR slot save/restore INSNs. Remove frame
2631 related flags from restore INSNs.
2633 2013-10-08 DJ Delorie <dj@redhat.com>
2635 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2638 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2639 avoid conflict with the MI use of %c.
2640 * config/rl78/rl78-real.md: change %c to %C throughout.
2641 * config/rl78/rl78-virt.md: Likewise.
2643 2013-10-08 Jan Hubicka <jh@suse.cz>
2645 * config/i386/i386.c (ix86_option_override_internal): Switch
2646 to SSE math for -ffast-math when target ISA supports SSE2.
2648 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2650 * tree-flow.h: Remove some prototypes.
2651 * tree.h: Remove some protypes, add a couple.
2652 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2653 using_eh_for_cleanups_p): Add interface routines for front ends.
2654 * tree-eh.h: New file. Add protoptyes.
2655 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2656 (add_stmt_to_eh_lp_fn): Make static.
2657 (lower_try_finally): Use new using_eh_for_cleanups_p.
2658 * emit-rtl.c: Include tree-eh.h.
2659 * gimple.h: Include tree-eh.h.
2661 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2663 PR tree-optimization/58480
2664 * tree-vrp.c (infer_nonnull_range): New function.
2665 (infer_value_range): Call infer_nonnull_range.
2667 2013-10-08 Dehao Chen <dehao@google.com>
2669 PR tree-optimization/58619
2670 * tree-inline.c (copy_phis_for_bb): Combine location data
2673 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2676 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2677 RTX_FRAME_RELATED_P on INSN.
2679 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2681 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2682 (altivec_expand_vec_perm_const): Call it.
2684 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2686 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2687 sequences for LE VSX loads and stores at expand time.
2688 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2690 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2691 (rs6000_gen_le_vsx_permute): New.
2692 (rs6000_gen_le_vsx_load): New.
2693 (rs6000_gen_le_vsx_store): New.
2694 (rs6000_gen_le_vsx_move): New.
2695 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2696 (*vsx_le_perm_load_v4si): New.
2697 (*vsx_le_perm_load_v8hi): New.
2698 (*vsx_le_perm_load_v16qi): New.
2699 (*vsx_le_perm_store_v2di): New.
2700 (*vsx_le_perm_store_v4si): New.
2701 (*vsx_le_perm_store_v8hi): New.
2702 (*vsx_le_perm_store_v16qi): New.
2703 (*vsx_xxpermdi2_le_<mode>): New.
2704 (*vsx_xxpermdi4_le_<mode>): New.
2705 (*vsx_xxpermdi8_le_V8HI): New.
2706 (*vsx_xxpermdi16_le_V16QI): New.
2707 (*vsx_lxvd2x2_le_<mode>): New.
2708 (*vsx_lxvd2x4_le_<mode>): New.
2709 (*vsx_lxvd2x8_le_V8HI): New.
2710 (*vsx_lxvd2x16_le_V16QI): New.
2711 (*vsx_stxvd2x2_le_<mode>): New.
2712 (*vsx_stxvd2x4_le_<mode>): New.
2713 (*vsx_stxvd2x8_le_V8HI): New.
2714 (*vsx_stxvd2x16_le_V16QI): New.
2716 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2718 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2720 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2722 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2723 loop to work also for 31bit ABI.
2724 Save the stack pointer for frame_size > 0.
2726 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2728 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2729 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2730 constraint letters from expanders.
2731 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2732 retry count to general_operand.
2733 ("tabort"): Give operand 0 a mode.
2734 ("tabort_1"): Add mode and constraint letter for operand 0.
2735 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2737 2013-10-04 Jeff Law <law@redhat.com>
2739 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2741 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2743 (thread_across_edge): Here. Call it.
2745 2013-10-04 Cary Coutant <ccoutant@google.com>
2747 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2748 discarding a location list expression (or a piece of one).
2750 2013-10-03 Jan Hubicka <jh@suse.cz>
2752 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2753 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2754 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2756 2013-10-03 Jan Hubicka <jh@suse.cz>
2758 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2759 accumulate-outgoing-args when producing unwind info.
2761 2013-10-03 Wei Mi <wmi@google.com>
2763 * lra-constraints.c (insert_move_for_subreg): New function
2764 extracted from simplify_operand_subreg.
2765 (simplify_operand_subreg): Add reload for paradoxical subreg.
2767 2013-10-03 Rong Xu <xur@google.com>
2769 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2770 the candidate of builtin_expect such that we should fix the
2771 size/time estimation.
2772 (estimate_function_body_sizes): Do the acutally size/time fix-up
2775 2013-10-03 Rong Xu <xur@google.com>
2777 * predict.c (tree_predict_by_opcode): Get the probability
2778 for builtin_expect from param builtin_expect_probability.
2779 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2780 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2781 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2783 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2786 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2787 * calls.c (alloca_call_p): Use get_callee_fndecl.
2788 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2789 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2791 (vrp_visit_stmt): Remove duplicated code.
2793 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2795 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2796 ceildf2, btruncdf2, instead of vsx_* name.
2798 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2799 iterators to only do V2DF and V4SF here. Move the DF code to
2800 rs6000.md where it is combined with SF mode. Replace <VSv> with
2801 just 'v' since only vector operations are handled with these insns
2802 after moving the DF support to rs6000.md.
2803 (vsx_sub<mode>3): Likewise.
2804 (vsx_mul<mode>3): Likewise.
2805 (vsx_div<mode>3): Likewise.
2806 (vsx_fre<mode>2): Likewise.
2807 (vsx_neg<mode>2): Likewise.
2808 (vsx_abs<mode>2): Likewise.
2809 (vsx_nabs<mode>2): Likewise.
2810 (vsx_smax<mode>3): Likewise.
2811 (vsx_smin<mode>3): Likewise.
2812 (vsx_sqrt<mode>2): Likewise.
2813 (vsx_rsqrte<mode>2): Likewise.
2814 (vsx_fms<mode>4): Likewise.
2815 (vsx_nfma<mode>4): Likewise.
2816 (vsx_copysign<mode>3): Likewise.
2817 (vsx_btrunc<mode>2): Likewise.
2818 (vsx_floor<mode>2): Likewise.
2819 (vsx_ceil<mode>2): Likewise.
2820 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2821 (vsx_sminsf3): Likewise.
2822 (vsx_fmadf4): Likewise.
2823 (vsx_fmsdf4): Likewise.
2824 (vsx_nfmadf4): Likewise.
2825 (vsx_nfmsdf4): Likewise.
2826 (vsx_cmpdf_internal1): Likewise.
2828 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2829 simpler to select whether a target has SPE or traditional floating
2830 point support in iterators.
2831 (TARGET_DF_SPE): Likewise.
2832 (TARGET_SF_FPR): Likewise.
2833 (TARGET_DF_FPR): Likewise.
2834 (TARGET_SF_INSN): Macros to say whether floating point support
2835 exists for a given operation for expanders.
2836 (TARGET_DF_INSN): Likewise.
2838 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2839 combining of SF/DF mode operations, using both traditional and VSX
2847 (abs<mode>2): Combine SF/DF modes using traditional floating point
2848 instructions. Add support for using the upper DF registers with
2849 VSX support, and SF registers with power8-vector support. Update
2850 expanders for operations supported by both the SPE and traditional
2851 floating point units.
2852 (abs<mode>2_fpr): Likewise.
2853 (nabs<mode>2): Likewise.
2854 (nabs<mode>2_fpr): Likewise.
2855 (neg<mode>2): Likewise.
2856 (neg<mode>2_fpr): Likewise.
2857 (add<mode>3): Likewise.
2858 (add<mode>3_fpr): Likewise.
2859 (sub<mode>3): Likewise.
2860 (sub<mode>3_fpr): Likewise.
2861 (mul<mode>3): Likewise.
2862 (mul<mode>3_fpr): Likewise.
2863 (div<mode>3): Likewise.
2864 (div<mode>3_fpr): Likewise.
2865 (sqrt<mode>3): Likewise.
2866 (sqrt<mode>3_fpr): Likewise.
2867 (fre<Fs>): Likewise.
2868 (rsqrt<mode>2): Likewise.
2869 (cmp<mode>_fpr): Likewise.
2870 (smax<mode>3): Likewise.
2871 (smin<mode>3): Likewise.
2872 (smax<mode>3_vsx): Likewise.
2873 (smin<mode>3_vsx): Likewise.
2874 (negsf2): Delete SF operations that are merged with DF.
2881 (fmasf4_fpr): Likewise.
2882 (fmssf4_fpr): Likewise.
2883 (nfmasf4_fpr): Likewise.
2884 (nfmssf4_fpr): Likewise.
2885 (sqrtsf2): Likewise.
2886 (rsqrtsf_internal1): Likewise.
2887 (smaxsf3): Likewise.
2888 (sminsf3): Likewise.
2889 (cmpsf_internal1): Likewise.
2890 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2891 (negdf2): Delete DF operations that are merged with SF.
2893 (nabsdf2): Likewise.
2899 (rsqrtdf_internal1): Likewise.
2900 (fmadf4_fpr): Likewise.
2901 (fmsdf4_fpr): Likewise.
2902 (nfmadf4_fpr): Likewise.
2903 (nfmsdf4_fpr): Likewise.
2904 (sqrtdf2): Likewise.
2905 (smaxdf3): Likewise.
2906 (smindf3): Likewise.
2907 (cmpdf_internal1): Likewise.
2908 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2909 (btrunc<mode>2): Delete separate expander, and combine with the
2910 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2911 (btrunc<mode>2_fpr): Likewise.
2912 (ceil<mode>2): Likewise.
2913 (ceil<mode>2_fpr): Likewise.
2914 (floor<mode>2): Likewise.
2915 (floor<mode>2_fpr): Likewise.
2916 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2917 Add support for using the upper registers with VSX and
2918 power8-vector. Move insns to be closer to the define_expands. On
2919 VSX systems, prefer the traditional form of FMA over the VSX
2920 version, since the traditional form allows the target not to
2921 overlap with the inputs.
2922 (fms<mode>4_fpr): Likewise.
2923 (nfma<mode>4_fpr): Likewise.
2924 (nfms<mode>4_fpr): Likewise.
2926 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2927 Richard Earnshaw <richard.earnshaw@arm.com>
2929 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2930 (struct mult_cost_table): Likewise.
2931 (struct mem_cost_table): Likewise.
2932 (struct fp_cost_table): Likewise.
2933 (struct vector_cost_table): Likewise.
2934 (cpu_cost_table): Likewise.
2935 * config/arm/arm.opt (mold-rts-costs): New option.
2936 (mnew-generic-costs): Likewise.
2937 * config/arm/arm.c (generic_extra_costs): New table.
2938 (cortexa15_extra_costs): Likewise.
2939 (arm_slowmul_tune): Use NULL as new costs.
2940 (arm_fastmul_tune): Likewise.
2941 (arm_strongarm_tune): Likewise.
2942 (arm_xscale_tune): Likewise.
2943 (arm_9e_tune): Likewise.
2944 (arm_v6t2_tune): Likewise.
2945 (arm_cortex_a5_tune): Likewise.
2946 (arm_cortex_a9_tune): Likewise.
2947 (arm_v6m_tune): Likewise.
2948 (arm_fa726te_tune): Likewise.
2949 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2950 (arm_cortex_tune): Use generict_extra_costs.
2951 (shifter_op_p): New function.
2952 (arm_unspec_cost): Likewise.
2953 (LIBCALL_COST): Define.
2954 (arm_new_rtx_costs): New function.
2955 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2956 table is available. Use old costs otherwise unless mnew-generic-costs
2958 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2959 (cpu_cost_table): Declare.
2961 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2964 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2965 (*subs_mul_imm_<mode>)
2966 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2967 (*sub_<shift>_<mode>)
2968 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2969 Remove k constraint.
2971 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2973 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2975 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2977 2013-10-02 Teresa Johnson <tejohnson@google.com>
2979 * predict.c (probably_never_executed): New function.
2980 (probably_never_executed_bb_p): Invoke probably_never_executed.
2981 (probably_never_executed_edge_p): Ditto.
2982 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2983 Treat profile insanities conservatively.
2985 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2987 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2989 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2991 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2992 use it. Use smaller increase for scratch. Don't increase reject
2993 for early clobber scratch.
2994 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2995 setting eliminated regs except setting fp from hfp.
2996 (lra_eliminate): Check lra_insn_recog_data on NULL.
2998 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3001 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3002 setting -mvsx-timode by default until the underlying problem is fixed.
3003 (RS6000_CPU, power7 defaults): Likewise.
3005 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3007 * config/x-linux (host-linux.o): Remove header dependencies.
3008 Use $(COMPILE) and $(POSTCOMPILE).
3009 * config/t-linux-android (linux-android.o): Ditto.
3011 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3013 * Makefile.in (expmed.o-warn): Remove.
3015 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3017 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3018 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3020 2013-10-02 Teresa Johnson <tejohnson@google.com>
3022 * dojump.c (do_jump_1): Divide probability between
3023 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3025 2013-10-02 Tom Tromey <tromey@redhat.com>
3027 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3029 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3031 * tree-flow.h: Remove some prototypes.
3032 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3033 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3034 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3035 mark_virtual_phi_result_for_renaming): Relocate here.
3036 * tree-into-ssa.h: Add prototypes.
3037 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3038 single_pred_before_succ_order.
3039 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3040 (nonfreeing_call_p) Move to gimple.c.
3041 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3043 * basic-block.h (single_pred_before_succ_order): Add prototype.
3044 * gimple.c (nonfreeing_call_p): Relocate here.
3045 * gimple.h: Add prototype.
3046 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3047 * tree-ssa-dom.h: New file. Relocate prototypes here.
3048 * tree-ssa.h: Include tree-ssa-dom.h.
3050 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3052 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3053 Use $(COMPILE) and $(POSTCOMPILE).
3055 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3057 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3059 * tree-flow.h: Remove some prototypes.
3060 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3061 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3062 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3063 propagate_tree_value*): Move from here to...
3064 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3065 propagate_tree_value*): Relocate here.
3066 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3067 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3068 * gimple-fold.c: Remove gimple-fold.h from include list.
3069 * tree-vrp.c: Remove gimple-fold.h from include list.
3070 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3071 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3072 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3073 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3074 * sese.c: Add tree-ssa-propagate.h to include list.
3076 2013-10-02 Richard Biener <rguenther@suse.de>
3078 * tree-loop-distribution.c: Include tree-vectorizer.h for
3080 (enum partition_kind): Remove PKIND_REDUCTION.
3081 (struct partition_s): Remove has_writes member, add reduction_p member.
3082 (partition_alloc): Adjust.
3083 (partition_builtin_p): Likewise.
3084 (partition_has_writes): Remove.
3085 (partition_reduction_p): New function.
3086 (partition_merge_into): Likewise.
3087 (generate_code_for_partition): Commonize builtin partition
3089 (rdg_cannot_recompute_vertex_p): Remove.
3090 (already_processed_vertex_p): Likewise.
3091 (rdg_flag_vertex): Do not set has_writes.
3092 (classify_partition): Adjust.
3093 (rdg_build_partitions): Do not set has_writes, treat all
3094 partitions as useful.
3095 (distribute_loop): Record number of library calls generated. Adjust.
3096 (tree_loop_distribution): Report number of loops and library
3097 calls generated as opt-info.
3099 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3101 * tree-flow.h: Include new .h files. Move prototypes.
3102 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3103 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3104 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3105 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3106 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3107 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3108 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3109 * tree.h (get_ref_base_and_extent): Move prototype out.
3110 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3112 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3113 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3114 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3115 * tree-scalar-evolution.c: Include tree.h.
3116 * sese.c: Include tree.h.
3117 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3118 * dwarf2out.c: Include tree-dfa.h.
3119 * tree-chrec.c: Include tree.h.
3120 * tree-data-ref.c: Include tree.h.
3122 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3124 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3127 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3129 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3130 Use $(COMPILE) and $(POSTCOMPILE).
3131 (sol2-cxx.o): Likewise.
3132 (sol2-stubs.o): Likewise.
3134 * config/x-solaris (host-solaris.o): Likewise.
3136 * config/sparc/t-sparc (sparc.o): Remove.
3137 (sparc-c.o): Remove header dependencies.
3138 Use $(COMPILE) and $(POSTCOMPILE).
3139 * config/sparc/x-sparc: Likewise.
3141 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3143 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3144 * config/arc/arc700.md: Likewise.
3145 * config/arc/arc-modes.def: Likewise.
3146 * config/arc/arc-simd.h: Likewise.
3147 * config/arc/t-arc-uClibc: Likewise.
3148 * config/arc/t-arc-newlib: Likewise.
3150 2013-10-02 Renlin Li <renlin.li@arm.com>
3152 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3154 (aarch64_expand_epilogue): Likewise.
3156 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3157 Yufeng Zhang <yufeng.zhang@arm.com>
3159 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3161 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3162 'base_in' represent a conversion and legal_cast_p_1 holds; set
3163 'base_in' with the returned value from get_unwidened.
3165 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3167 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3168 plus_constant is not used.
3170 2013-10-01 Wei Mi <wmi@google.com>
3172 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3173 * config/i386/i386.md: Add define_peephole2 to
3174 break partial reg stall for cvtss2sd/cvtsd2ss.
3176 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3178 * config/arc/arc.c (pass_arc_ifcvt::clone):
3179 Update for ctxt_ -> m_ctxt change.
3181 2013-10-01 Jeff Law <law@redhat.com>
3183 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3184 outgoing_edge and intermediate_edge fields. Instead store the path.
3185 (redirection_data::hash): Hash on the last edge's destination index.
3186 (redirection_data::equal): Check the entire thread path.
3187 (lookup_redirectio_data): Corresponding changes.
3188 (create_edge_and_update_destination_phis): Likewise.
3189 (thread_single_edge): Likewise.
3191 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3192 Diego Novillo <dnovillo@google.com>
3194 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3195 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3196 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3197 versions of these patterns.
3199 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3200 (__builtin_arc_aligned): Likewise.
3202 * config/arc/arc.md: Expand adc_0 comment stating the intended
3203 purpose and why it isn't ready.
3204 Replace commented out call_value_via_label_mixed with a
3205 plain comment about bl_s.
3207 * config/arc/arc.c (stdio.h): Don't include directly.
3208 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3209 Remove if (1) condition.
3210 (arc_encode_section_info): Fix comment.
3212 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3214 * config/arc/arc.c (arc_conditional_register_usage):
3215 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3216 Also set reg_alloc_order for DMA config regs.
3218 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3219 Jeremy Bennett <jeremy.bennett@embecosm.com>
3221 * doc/install.texi (--with-cpu): Mention ARC.
3222 (arc-*-elf32): New paragraph.
3223 (arc-linux-uclibc): Likewise.
3224 * doc/md.texi (Machine Constraints): Add ARC part.
3225 * doc/invoke.texi: (menu): Add ARC Options.
3226 (Machine Dependent Options) <ARC Options>: Add synopsis.
3227 (node ARC Options): Add.
3228 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3229 (ARC Built-in Functions): New section defining
3230 generic ARC built-in functions.
3231 (ARC SIMD Built-in Functions): New section defining SIMD specific
3233 (Declaring Attributes of Functions): Extended
3234 description of short_call and long_call attributes for ARC and
3235 added index entries.
3237 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3238 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3239 Joern Rennecke <joern.rennecke@embecosm.com>
3240 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3241 Brendan Kehoe <brendan@zen.org>
3242 Michael Eager <eager@eagercon.com>
3243 Simon Cook <simon.cook@embecosm.com>
3244 Jeremy Bennett <jeremy.bennett@embecosm.com>
3246 * config/arc, common/config/arc: New directories.
3248 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3249 Brendan Kehoe <brendan@zen.org>
3250 Simon Cook <simon.cook@embecosm.com>
3252 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3254 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3256 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3257 * tree-ssa-coalesce.h: New. Move prototype to here.
3258 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3259 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3260 (gimple_can_coalesce_p): Move to...
3261 * gimple.c (gimple_can_coalesce_p): Here.
3263 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3265 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3266 (dump_decl_set): Move to gimple.c.
3267 * gimple.h: Don't include tree-ssa-operands.h.
3268 (dump_decl_set): Add prototype.
3269 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3270 Move to gimple-ssa.h.
3271 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3272 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3273 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3274 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3275 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3277 (dump_decl_set): Relocate here.
3278 * gimple-ssa.h: New file.
3279 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3280 Relocate from gimple.h.
3281 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3282 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3283 swap_tree_operands and remove non-ssa path.
3284 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3285 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3287 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3288 vect_is_simple_reduction_1): Use swap_ssa_operands.
3289 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3290 (enum need_phi_state): Move to tree-into-ssa.c.
3291 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3292 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3293 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3294 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3295 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3296 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3297 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3298 num_imm_uses): Move to ssa-iterators.h.
3299 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3300 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3302 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3303 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3304 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3305 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3306 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3307 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3308 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3309 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3310 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3311 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3312 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3313 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3314 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3315 (set_phi_nodes): Move to tree-phinodes.h.
3316 * tree-ssa-operands.h (enum ssa_op_iter_type,
3317 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3318 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3319 (dump_decl_set): Remove prototype.
3320 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3321 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3322 (set_phi_nodes): Relocate from tree-flow-inline.h.
3323 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3325 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3326 include list. Temporarily add gimple.h to include list.
3327 * ssa-iterators.h: New file.
3328 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3329 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3330 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3331 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3332 Relocate from tree-ssa-operands.h.
3333 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3334 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3335 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3336 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3337 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3338 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3339 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3340 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3341 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3342 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3343 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3344 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3345 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3346 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3347 Relocate from tree-flow-inline.h.
3348 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3350 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3353 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3354 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3355 Insert '\t' to output template.
3356 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3357 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3358 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3359 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3361 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3363 * doc/install.texi (Host/target specific installation notes for GCC):
3364 Put @anchor before @heading.
3365 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3366 Use @email for email addresses.
3368 2013-10-01 Jeff Law <law@redhat.com>
3370 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3371 a vec. Only delete the path if we create one without successfully
3372 registering a jump thread.
3373 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3375 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3376 (paths): New vector of jump threading paths.
3377 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3378 (THREAD_PATH): New accessor macro for the entire thread path.
3379 (lookup_redirection_data): Get intermediate and final outgoing edge
3380 from the thread path.
3381 (create_edge_and_update_destination_phis): Copy the threading path.
3382 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3383 jump threading path.
3384 (ssa_redirect_edges): Get edges and block types from the jump threading
3385 path. Free the path vector.
3386 (thread_block): Get edges from the jump threading path. Look at the
3387 entire path to see if we thread to a loop exit. If we cancel a jump
3388 thread request, then free the path vector.
3389 (thread_single_edge): Get edges and block types from the jump threading
3390 path. Free the path vector.
3391 (thread_through_loop_header): Get edges and block types from the jump
3392 threading path. Free the path vector.
3393 (mark_threaded_blocks): Iterate over the vector of paths and store
3394 the path on the appropriate edge. Get edges and block types from the
3395 jump threading path.
3396 (mark_threaded_blocks): Get edges and block types from the jump
3397 threading path. Free the path vector.
3398 (thread_through_all_blocks): Use the vector of paths rather than
3399 a vector of 3-edge sets.
3400 (register_jump_thread): Accept pointer to a path vector rather
3401 than the path vector itself. Store the path vector for later use.
3404 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3405 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3408 * config/s390/s390.c (s390_split_branches): Modify check for table
3410 (s390_chunkify_start): Rearrange table jump insn check in order to
3411 deal with compare and branch insns correctly.
3413 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3417 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3418 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3419 define_insn_and_split.
3420 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3421 (shiftsi3_compare): New pattern.
3423 * config/arm/unspecs.md (UNSPEC_RRX): New.
3425 2013-10-01 Alan Modra <amodra@gmail.com>
3427 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3428 casing inout operands.
3430 2013-10-01 Richard Biener <rguenther@suse.de>
3432 PR tree-optimization/58553
3433 * tree-loop-distribution.c (struct partition_s): Add niter member.
3434 (classify_partition): Populate niter member for the partition
3435 and properly identify whether the relevant store happens before
3436 or after the loop exit.
3437 (generate_memset_builtin): Use niter member from the partition.
3438 (generate_memcpy_builtin): Likewise.
3440 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3442 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3443 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3445 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3447 * basic-block.h (edge_list): Prefix member names with "m_".
3448 * context.h (context): Likewise.
3449 * domwalk.h (dom_walker): Likewise.
3450 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3451 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3452 * hash-table.h (hash_table): Likewise.
3453 * machmode.h (bit_field_mode_iterator): Likewise.
3454 * pass_manager.h (pass_list): Likewise.
3455 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3456 * tree-pass.h (pass_data): Likewise.
3457 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3458 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3459 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3460 * asan.c (pass_data_asan): Update accordingly.
3461 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3462 (control_dependences::control_dependences): Likewise.
3463 (control_dependences::~control_dependences): Likewise.
3464 (control_dependences::~control_dependences): Likewise.
3465 (control_dependences::get_edges_dependent_on): Likewise.
3466 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3467 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3468 * context.c (gcc::context::context): Likewise.
3469 * cprop.c (pass_rtl_cprop::clone): Likewise.
3470 * domwalk.c (dom_walker::walk): Likewise.
3471 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3472 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3473 * mode-switching.c (pass_mode_switching::clone): Likewise.
3474 * passes.c (opt_pass::opt_pass): Likewise.
3475 (pass_manager::pass_manager): Likewise.
3476 * predict.c (pass_strip_predict_hints::clone): Likewise.
3477 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3478 (pass_split_all_insns::clone): Likewise.
3479 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3481 (bit_field_mode_iterator::next_mode): Likewise.
3482 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3483 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3484 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3485 * tree-complex.c (pass_lower_complex::clone): Likewise.
3486 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3487 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3488 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3489 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3490 (pass_fold_builtins::clone): Likewise.
3491 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3492 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3493 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3494 (pass_cd_dce::clone): Likewise.
3495 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3496 (pass_phi_only_cprop::clone): Likewise.
3497 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3498 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3499 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3500 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3501 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3502 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3503 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3504 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3505 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3506 * tree-vrp.c (pass_vrp::clone): Likewise.
3507 * tsan.c (pass_tsan::clone): Likewise.
3509 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3512 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3513 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3516 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3517 optimization, punt if sign_bit_p looked through any zero extension.
3519 2013-09-30 Teresa Johnson <tejohnson@google.com>
3521 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3522 Update redirected out edge count in joiner case.
3523 (ssa_redirect_edges): Common the joiner and non-joiner cases
3524 so that joiner case gets profile updates.
3526 2013-09-30 Richard Biener <rguenther@suse.de>
3528 PR tree-optimization/58554
3529 * tree-loop-distribution.c (classify_partition): Require
3530 unconditionally executed stores for memcpy and memset recognition.
3531 (tree_loop_distribution): Calculate dominance info.
3533 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3535 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3536 (NO_PROFILE_COUNTERS): Likewise.
3537 (PROFILE_HOOK): Likewise.
3538 (FUNCTION_PROFILER): Likewise.
3539 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3541 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3543 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3544 calls and defines in TARGET_MACHO conditional.
3545 (load_macho_picbase_di): Likewise.
3546 (reload_macho_picbase): Likewise.
3547 (reload_macho_picbase_si): Likewise.
3548 (reload_macho_picbase_di): Likewise.
3549 (nonlocal_goto_receiver): Likewise.
3551 2013-09-30 Nick Clifton <nickc@redhat.com>
3553 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3554 that use the MSP430X ISA.
3555 (msp430_option_override): Scan -mmcu command line option for any
3556 MCU name that supports the MSP430X ISA.
3557 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3558 -mmcu options which enable the MSP430X ISA.
3560 2013-09-30 Richard Biener <rguenther@suse.de>
3563 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3564 before looking for setjmp-like calls.
3566 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3569 * config/darwin-protos.h (machopic_get_function_picbase): New.
3570 * config/darwin.c (machopic_get_function_picbase): New.
3571 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3572 label for a new func. (load_macho_picbase_di): Likewise.
3573 (reload_macho_picbase): New expand.
3574 (reload_macho_picbase_si): New insn.
3575 (reload_macho_picbase_di): New insn.
3576 (nonlocal_goto_receiver): New define and split.
3577 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3578 (unspecv enum): Add UNSPECV_NLGR.
3580 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3582 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3583 that altivec registers are correctly sized on Darwin.
3585 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3587 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3588 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3589 * config/x-darwin (host-darwin.o): Likewise.
3590 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3591 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3592 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3594 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3596 * doc/invoke.texi: Fix usage of @tie{} command.
3598 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3600 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3602 2013-09-28 Jan Hubicka <jh@suse.cz>
3604 * config/i386/x86-tune.def: Add documentation for each of the options;
3607 2013-09-28 Jan Hubicka <jh@suse.cz>
3609 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3611 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3612 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3613 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3614 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3615 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3617 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3619 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3620 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3621 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3622 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3623 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3624 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3625 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3626 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3627 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3628 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3629 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3630 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3631 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3632 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3633 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3634 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3635 gimple.h, godump.c, graphite-clast-to-gimple.c,
3636 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3637 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3638 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3639 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3640 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3641 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3642 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3643 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3644 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3645 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3646 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3647 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3648 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3649 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3650 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3651 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3652 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3653 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3654 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3655 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3656 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3657 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3658 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3659 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3660 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3661 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3662 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3663 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3664 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3665 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3666 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3667 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3668 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3669 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3670 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3671 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3672 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3673 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3674 whitespace before "(".
3676 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3678 * expr.h (extract_bit_field): Remove packedp parameter.
3679 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3680 from forward declaration.
3681 (store_split_bit_field): Remove packedp arg from calls to
3682 extract_fixed_bit_field.
3683 (extract_bit_field_1): Remove packedp parameter and packedp
3684 argument from recursive calls and calls to extract_fixed_bit_field.
3685 (extract_bit_field): Remove packedp parameter and corresponding
3686 arg to extract_bit_field_1.
3687 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3689 (extract_split_bit_field): Remove packedp arg from call to
3690 extract_fixed_bit_field.
3691 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3692 (copy_blkmode_from_reg): Likewise.
3693 (copy_blkmode_to_reg): Likewise.
3694 (read_complex_part): Likewise.
3695 (store_field): Likewise.
3696 (expand_expr_real_1): Likewise.
3697 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3698 to extract_bit_field.
3699 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3700 call to extract_bit_field.
3701 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3702 call to extract_bit_field.
3703 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3704 and special packedp behavior from -fstrict-volatile-bitfields
3707 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3709 * lra-eliminations.c (init_elim_table): Guard value_p.
3711 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3713 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3714 DFmode, DImode, and SFmode in the upper VSX registers based on the
3715 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3716 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3717 of the VSX register handling.
3719 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3721 (zero_extendsidi2_lfiwzx): Add support for loading into the
3722 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3723 only do VSX memory options on Altivec registers.
3724 (extendsidi2_lfiwax): Likewise.
3725 (extendsfdf2_fpr): Likewise.
3726 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3727 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3728 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3729 (movdi_internal64): Likewise.
3731 2013-09-27 Xinliang David Li <davidxl@google.com>
3733 * opts.c (finish_options): Adjust parameters
3734 according to vect cost model.
3735 (common_handle_option): Set dynamic vect cost
3737 targhooks.c (default_add_stmt_cost): Compute stmt cost
3739 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3740 Use helper function.
3741 * tree-vectorizer.h (unlimited_cost_model): New function.
3742 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3743 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3745 (vect_enhance_data_refs_alignment): Ditto.
3746 * flag-types.h: New enum.
3747 * common/config/i386/i386-common.c (ix86_option_init_struct):
3748 No need to initialize vect_cost_model flag.
3749 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3752 2013-09-27 Diego Novillo <dnovillo@google.com>
3754 * gimple.h (enum ssa_mode): Remove.
3756 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3758 * cfgloop.h (number_of_loops): Fix typo in check for null.
3760 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3763 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3764 are children of outermost saved_cfun's loop, and set it up to
3765 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3766 if loop != loop0 and SESE region contains bbs that belong to loop0.
3768 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3770 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3771 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3772 (set_address_segment, set_address_base, set_address_index)
3773 (set_address_disp): Accept the argument unconditionally.
3774 (baseness): Remove must_be_base_p and must_be_index_p checks.
3775 (decompose_normal_address): Classify as much as possible in the
3778 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3780 * cse.c (count_reg_usage): Handle INT_LIST.
3781 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3782 * reginfo.c (reg_scan_mark_refs): Likewise.
3783 * reload1.c (eliminate_regs_1): Likewise.
3785 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3788 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3791 2013-09-27 Richard Biener <rguenther@suse.de>
3793 PR tree-optimization/58459
3794 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3795 restriction not propagating into loops.
3797 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3799 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3800 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3801 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3803 2013-09-26 Richard Biener <rguenther@suse.de>
3805 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3807 2013-09-26 Richard Biener <rguenther@suse.de>
3809 * alias.h (component_uses_parent_alias_set): Rename to ...
3810 (component_uses_parent_alias_set_from): ... this.
3811 * alias.c (component_uses_parent_alias_set): Rename to ...
3812 (component_uses_parent_alias_set_from): ... this and return
3814 (reference_alias_ptr_type_1): Use the result from
3815 component_uses_parent_alias_set_from instead of stripping
3816 components one at a time.
3817 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3819 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3821 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3822 Move prototypes to...
3823 * tree-ssa-ter.h: New File. Move prototypes here.
3824 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3825 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3826 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3827 from is_replaceable_p.
3828 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3829 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3830 (process_replaceable): Use ter_is_replaceable_p.
3831 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3832 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3833 newly refactored ssa_is_replaceable_p.
3834 * cfgexpand.c: Include tree-outof-ssa.h.
3835 * ssaexpand.h: Delete.
3837 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3839 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3840 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3842 (create_gimple_tmp): Delete.
3843 (get_expr_type, build_assign, build_type_cast): Move to...
3844 * gimple-builder.c: New File.
3845 (get_expr_type): Relocate from gimple.c.
3846 (build_assign, build_type_cast): Change to only create ssanames.
3847 * gimple.h: Move prototypes to...
3848 * gimple-builder.h: New File. Here.
3849 * tree-ssa.h: And here.
3850 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3851 count_uses_and_derefs): Relocate from gimple.c.
3852 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3853 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3854 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3855 gimple_replace_ssa_lhs.
3856 * asan.c: Include gimple-builder.h.
3857 * Makefile.in: Add gimple-builder.o.
3859 2013-09-26 Richard Biener <rguenther@suse.de>
3861 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3862 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3863 (loe_visit_block): Use gcc_checking_assert.
3864 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3865 gimple_assign_ssa_name_copy_p.
3866 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3869 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3871 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3872 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3874 2013-09-26 Richard Biener <rguenther@suse.de>
3876 PR tree-optimization/58539
3877 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3878 the fact that debug statements are not taking part in loop-closed
3881 2013-09-26 Nick Clifton <nickc@redhat.com>
3883 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3884 time warning message.
3885 (msp430_print_operand_raw): Delete unused letter parameter.
3886 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3887 (msp430_print_operand_address): New function.
3888 (msp430_print_operand): Move address printing code from here to
3890 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3892 (zero_extendpsisi2): Likewise.
3893 (extendpsisi2): New pattern.
3894 (andneghi3): New pattern.
3896 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3898 * config/aarch64/aarch64.opt (mlra): New option.
3899 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3900 (TARGET_LRA_P): Define.
3902 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3904 * expr.c (expand_assignment): Remove obsolete comment.
3906 2013-09-25 Jeff Law <law@redhat.com>
3908 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3909 tree-ssa-threadupdate.h.
3910 (register_jump_thread): Similarly.
3911 * tree-ssa-threadupdate.h: New header file.
3912 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3913 * tree-vrp.c: Likewise.
3914 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3915 (thread_around_empty_blocks): Change type of path vector argument to
3916 an edge,type pair from just an edge. Initialize both elements when
3917 appending to a jump threading path. Tweak references to elements
3919 (thread_across_edge): Similarly. Release memory for the elements
3921 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3922 (dump_jump_thread_path): New function broken out from
3923 register_jump_thread.
3924 (register_jump_thread): Use dump_jump_thread_path. Change type of
3925 path vector entries. Search the path for NULL edges and dump
3926 the path if one is found. Tweak the conversion of path to 3-edge
3927 form to use the block copy type information embedded in the path.
3929 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3931 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3933 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3934 Vladimir Makarov <vmakarov@redhat.com>
3936 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3937 from the least significant bit.
3938 (strip_address_mutations): Add bitfield operations handling.
3939 (must_be_index_p): Add shifting and rotate operations handling.
3940 (set_address_base): Use must_be_base_p predicate.
3941 (set_address_index): Use must_be_index_p predicate.
3943 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3944 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3945 Sergey Lega <sergey.s.lega@intel.com>
3946 Anna Tikhonova <anna.tikhonova@intel.com>
3947 Ilya Tocar <ilya.tocar@intel.com>
3948 Andrey Turetskiy <andrey.turetskiy@intel.com>
3949 Ilya Verbin <ilya.verbin@intel.com>
3950 Kirill Yukhin <kirill.yukhin@intel.com>
3951 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3953 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3955 (ix86_expand_vector_move_misalign): Support new unaligned load and
3956 stores and use new names.
3957 (CODE_FOR_sse2_storedqu): Rename to ...
3958 (CODE_FOR_sse2_storedquv16qi): ... this.
3959 (CODE_FOR_sse2_loaddqu): Rename to ...
3960 (CODE_FOR_sse2_loaddquv16qi): ... this.
3961 (CODE_FOR_avx_loaddqu256): Rename to ...
3962 (CODE_FOR_avx_loaddquv32qi): ... this.
3963 (CODE_FOR_avx_storedqu256): Rename to ...
3964 (CODE_FOR_avx_storedquv32qi): ... this.
3965 * config/i386/i386.md (fpint_logic): New.
3966 * config/i386/sse.md (VMOVE): Extend for AVX512.
3970 (VI_UNALIGNED_LOADSTORE): Ditto.
3971 (sse2_avx_avx512f): Ditto.
3972 (sse2_avx2): Extend for AVX512.
3973 (sse4_1_avx2): Ditto.
3974 (avx2_avx512f): New.
3975 (sse): Extend for AVX512.
3978 (avxsizesuffix): Ditto.
3979 (sseintvecmode): Ditto.
3981 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3982 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3983 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3984 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3985 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3986 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3987 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3988 (STORENT_MODE): Extend for AVX512.
3989 (*absneg<mode>2): Replace constraint "x" with "v".
3990 (*mul<mode>3): Ditto.
3991 (*ieee_smin<mode>3): Ditto.
3992 (*ieee_smax<mode>3): Ditto.
3993 (avx_cmp<mode>3): Replace VF with VF_128_256.
3994 (*<sse>_maskcmp<mode>3_comm): Ditto.
3995 (<sse>_maskcmp<mode>3): Ditto.
3996 (<sse>_andnot<mode>3): Extend for AVX512.
3997 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3998 (<code><mode>3, fpint_logic): New.
3999 (*<code><mode>3): Extend for AVX512.
4000 (avx512flogicsuff): New.
4001 (avx512f_<logic><mode>): Ditto.
4002 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4004 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4005 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4006 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4007 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4008 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4009 (xop_vpermil2<mode>3): Ditto.
4010 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4011 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4012 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4013 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4015 2013-09-25 Tom Tromey <tromey@redhat.com>
4017 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4018 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4019 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4020 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4021 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4022 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4023 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4024 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4025 (GRAPHITE_HTAB_H): Remove.
4027 2013-09-25 Tom Tromey <tromey@redhat.com>
4029 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4031 2013-09-25 Tom Tromey <tromey@redhat.com>
4033 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4035 2013-09-25 Tom Tromey <tromey@redhat.com>
4037 * config/i386/t-i386 (i386.o): Remove.
4038 (i386-c.o): Use COMPILE and POSTCOMPILE.
4040 2013-09-25 Tom Tromey <tromey@redhat.com>
4042 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4044 2013-09-25 Tom Tromey <tromey@redhat.com>
4046 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4047 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4048 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4049 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4050 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4051 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4052 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4053 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4054 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4055 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4056 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4057 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4058 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4059 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4060 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4061 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4062 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4063 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4064 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4065 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4066 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4067 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4068 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4069 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4070 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4071 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4072 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4073 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4074 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4075 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4076 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4077 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4078 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4079 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4080 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4081 (graphite-clast-to-gimple.o, graphite-dependences.o)
4082 (graphite-interchange.o, graphite-poly.o)
4083 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4084 (graphite-optimize-isl.o, tree-vect-loop.o)
4085 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4086 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4087 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4088 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4089 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4090 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4091 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4092 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4093 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4094 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4095 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4096 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4097 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4098 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4099 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4100 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4101 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4102 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4103 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4104 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4105 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4106 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4107 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4108 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4109 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4110 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4111 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4112 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4113 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4114 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4115 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4116 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4117 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4118 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4119 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4120 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4121 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4122 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4123 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4124 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4125 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4126 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4127 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4128 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4129 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4130 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4131 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4132 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4133 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4134 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4135 (gcov-dump.o): Remove.
4136 (default-c.o): Use COMPILE and POSTCOMPILE.
4137 (CFLAGS-gcc.o): New variable.
4138 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4140 2013-09-25 Tom Tromey <tromey@redhat.com>
4142 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4143 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4144 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4145 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4146 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4147 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4148 (c-family/array-notation-common.o, c-family/stub-objc.o)
4149 (c-family/c-ubsan.o): Remove.
4151 2013-09-25 Tom Tromey <tromey@redhat.com>
4153 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4155 2013-09-25 Tom Tromey <tromey@redhat.com>
4157 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4158 to add -DENABLE_SHARED_LIBGCC.
4159 (gcc.o): Don't use subshell.
4161 2013-09-25 Tom Tromey <tromey@redhat.com>
4163 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4164 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4165 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4166 * configure, config.in: Rebuild.
4168 2013-09-25 Tom Tromey <tromey@redhat.com>
4170 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4171 (COMPILE, POSTCOMPILE): New variables.
4172 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4173 (DEPFILES): New variable.
4174 Include ".Po" files.
4175 * configure.ac: Add checks for dependency checking.
4176 * configure, aclocal.m4: Regenerate.
4178 2013-09-25 Tom Tromey <tromey@redhat.com>
4180 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4181 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4183 2013-09-25 Tom Tromey <tromey@redhat.com>
4185 * Makefile.in (generated_files): Add options.h,
4186 target-hooks-def.h, insn-opinit.h,
4187 common/common-target-hooks-def.h, pass-instances.def,
4188 c-family/c-target-hooks-def.h.
4190 2013-09-25 Jeff Law <law@redhat.com>
4192 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4193 than foo[foo.length () - 1] to access last member in a vec.
4194 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4196 2013-09-25 Richard Biener <rguenther@suse.de>
4199 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4201 2013-09-25 Jan Hubicka <jh@suse.cz>
4203 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4206 2013-09-25 Marek Polacek <polacek@redhat.com>
4209 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4210 when determining the type name.
4212 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4214 * config/sh/sh.md: Fix formatting.
4216 2013-09-24 Xinliang David Li <davidxl@google.com>
4218 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4219 max peel iterations parameter.
4220 * param.def: New parameter.
4221 * doc/invoke.texi: Document New parameter.
4223 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4225 * gimple-pretty-print.c: Various whitespace tweaks.
4226 * tree-core.h: Likewise.
4227 * tree-pretty-print.c: Likewise.
4228 * tree-ssa-alias.c: Likewise.
4229 * tree-ssa-copy.c: Likewise.
4230 * tree-ssanames.c: Likewise.
4231 * tree-ssanames.h: Likewise.
4232 * tree-vrp.c: Likewise.
4234 2013-09-24 Alan Modra <amodra@gmail.com>
4238 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4239 for output operands that disallow regs. Don't use EXPAND_WRITE on
4242 2013-09-24 Richard Biener <rguenther@suse.de>
4245 * tree.c (reference_alias_ptr_type): Move ...
4246 * alias.c (reference_alias_ptr_type): ... here and implement
4247 in terms of the new reference_alias_ptr_type_1.
4248 (ref_all_alias_ptr_type_p): New helper.
4249 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4250 use ref_all_alias_ptr_type_p.
4251 (get_deref_alias_set): Add flag_strict_aliasing check here.
4252 (reference_alias_ptr_type_1): New function, split out from ...
4253 (get_alias_set): ... here.
4254 (alias_ptr_types_compatible_p): New function.
4255 * alias.h (reference_alias_ptr_type): Declare.
4256 (alias_ptr_types_compatible_p): Likewise.
4257 * tree.h (reference_alias_ptr_type): Remove.
4258 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4259 to compare MEM_REF alias types.
4261 2013-09-24 Richard Biener <rguenther@suse.de>
4263 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4265 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4267 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4268 reload helper function arrays into a single array reg_addr.
4269 (reload_fpr_gpr): Likewise.
4270 (reload_gpr_vsx): Likewise.
4271 (reload_vsx_gpr): Likewise.
4272 (struct rs6000_reg_addr): Likewise.
4273 (reg_addr): Likewise.
4274 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4275 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4276 (rs6000_init_hard_regno_mode_ok): Likewise.
4277 (rs6000_secondary_reload_direct_move): Likewise.
4278 (rs6000_secondary_reload): Likewise.
4280 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4281 constraints: wu, ww, and wy. Repurpose wv constraint added during
4282 power8 changes. Put wg constraint in alphabetical order.
4284 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4285 for future work to add ISA 2.07 VSX single precision support.
4286 (-mvsx-scalar-double): Change default from -1 to 1, update
4287 documentation comment.
4288 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4289 (-mupper-regs-df): New debug switch to control whether DF values
4290 can go in the traditional Altivec registers.
4291 (-mupper-regs-sf): New debug switch to control whether SF values
4292 can go in the traditional Altivec registers.
4294 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4296 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4297 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4298 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4299 (rs6000_debug_legitimate_address_p): Print if we are running
4300 before, during, or after reload.
4301 (rs6000_secondary_reload): Add a comment.
4302 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4304 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4305 constraints. Update documentation string.
4306 (wd constraint): Likewise.
4307 (wf constraint): Likewise.
4308 (wg constraint): Likewise.
4309 (wn constraint): Likewise.
4310 (ws constraint): Likewise.
4311 (wt constraint): Likewise.
4312 (wx constraint): Likewise.
4313 (wz constraint): Likewise.
4314 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4316 (ww constraint): Likewise.
4317 (wy constraint): Likewise.
4318 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4319 in the previous submissions.
4320 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4322 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4324 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4325 (REG_BR_PROB): Say that the probability is stored in an int_list.
4326 * reg-notes.def: Update commentary to mention INT_LIST.
4327 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4328 (INT_LIST): New rtx.
4329 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4330 * rtlanal.c (int_reg_note_p): New function.
4331 (alloc_reg_note): Assert that the note does not have an int argument.
4332 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4333 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4334 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4335 rather than an INSN_LIST. Handle INT_LIST.
4336 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4337 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4338 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4339 Manipulate them as ints rather than rtxes.
4340 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4341 * regmove.c (copy_src_to_dest): Likewise.
4342 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4344 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4345 into the cases that need it.
4346 * config/arm/arm.c (arm_unwind_emit): Likewise.
4348 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4349 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4350 * loop-doloop.c (add_test, doloop_modify): Likewise.
4351 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4352 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4353 * predict.c (combine_predictions_for_insn): Likewise.
4354 * print-rtl.c (print_rtx): Handle INT_LIST.
4355 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4356 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4357 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4358 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4359 (ix86_split_fp_branch, predict_jump): Likewise.
4360 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4361 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4362 * config/spu/spu.c (ea_load_store_inline): Likewise.
4364 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4365 value of a REG_BR_PROB note.
4366 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4367 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4368 * emit-rtl.c (try_split): Likewise.
4369 * predict.c (br_prob_note_reliable_p): Likewise.
4370 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4371 * reorg.c (mostly_true_jump): Likewise.
4372 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4373 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4374 * config/i386/i386.c (ix86_print_operand): Likewise.
4375 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4376 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4377 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4378 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4379 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4380 * config/sparc/sparc.c (output_cbranch): Likewise.
4381 * config/spu/spu.c (get_branch_target): Likewise.
4382 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4383 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4385 2013-09-23 Jan Hubicka <jh@suse.cz>
4387 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4389 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4390 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4392 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4393 (gimple_fold_call): Dump inconsistent devirtualizations; add
4394 sanity check for type based devirtualizations.
4395 * ipa-prop.c: Include ipa-utils.h
4396 (ipa_intraprocedural_devirtualization): Add sanity check.
4397 (try_make_edge_direct_virtual_call): Likewise.
4399 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4401 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4402 assignment statements.
4404 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4406 * gimple-pretty-print.c (dump_ssaname_info): New function.
4407 (dump_gimple_phi): Call it.
4408 (pp_gimple_stmt_1): Likewise.
4409 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4410 (range_info_def): Declare.
4411 * tree-pretty-print.c (pp_double_int): New function.
4412 (dump_generic_node): Call it.
4413 * tree-pretty-print.h (pp_double_int): Declare.
4414 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4415 * tree-ssanames.h (range_info_def): New structure.
4416 (value_range_type): Move definition here.
4417 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4419 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4421 (set_range_info): New function.
4422 (get_range_info): Likewise.
4423 (duplicate_ssa_name_range_info): Likewise.
4424 (duplicate_ssa_name_fn): Check pointer type and call
4425 duplicate_ssa_name_range_info.
4426 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4427 * tree-vrp.c (value_range_type): Remove definition, now in
4429 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4430 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4431 (SSA_NAME_RANGE_INFO): New macro.
4433 2013-09-23 Richard Biener <rguenther@suse.de>
4435 PR tree-optimization/58464
4436 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4437 (phi_trans_add): Change to add conditionally on being not present.
4438 (phi_translate_1): Remove recursion detection here.
4439 (phi_translate): Pre-seed the cache with NULL to catch
4440 recursion here in a more generic way.
4441 (bitmap_find_leader): Adjust comment.
4442 (get_representative_for): Dump value-numbers.
4443 (create_expression_by_pieces): Likewise.
4444 (insert_into_preds_of_block): Likewise.
4446 2013-09-23 Christian Bruel <christian.bruel@st.com>
4449 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4450 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4452 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4455 2013-09-20 Renlin Li <renlin.li@arm.com>
4457 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4459 (aarch64_expand_epilogue): Likewise.
4460 (aarch64_legitimize_reload_address): Likewise.
4462 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4464 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4465 NULL_TREE before pushing them onto the vector. Likewise for labels.
4467 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4469 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4470 comparison operators when -fno-trapping-math is in effect.
4471 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4472 comparison operators in TFmode and assert that unsupported operators
4474 (ia64_print_operand): Likewise.
4476 2013-09-21 Jan Hubicka <jh@suse.cz>
4478 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4479 (sse_typeless_stores): Enable for core
4480 (sse_load0_by_pxor): Likewise.
4481 (four_jump_limit): Disable for core.
4482 (pad_returns): Likewise.
4483 (avoid_vector_decode): Likewise.
4484 (fuse_cmp_and_branch): Enable for cores.
4485 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4487 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4490 * config/pa/pa.c (pa_option_override): Disable auto increment and
4491 decrement instructions until reload is completed.
4493 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4494 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4495 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4497 2013-09-20 DJ Delorie <dj@redhat.com>
4498 Nick Clifton <nickc@redhat.com>
4500 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4501 (need_to_save): Save bank 0 on interrupts.
4502 (characterize_address): Strip far address wrappers.
4503 (rl78_as_legitimate_address): Likewise.
4504 (transcode_memory_rtx): Likewise.
4505 (rl78_peep_movhi_p): Disable this peephole after devirt.
4506 (rl78_propogate_register_origins): Forget all origins when a
4508 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4509 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4511 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4512 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4513 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4515 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4517 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4519 * config/pa/pa.md: In "scc" insn patterns, change output template to
4520 handle const0_rtx in reg_or_0_operand operands.
4522 2013-09-20 Martin Husemann <martin@NetBSD.org>
4525 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4526 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4528 2013-09-20 Richard Biener <rguenther@suse.de>
4531 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4532 remembering SSA name version and block index.
4533 (new_scev_info_str): Adjust.
4534 (hash_scev_info): Likewise. Also hash the block index.
4535 (eq_scev_info): Adjust.
4536 (find_var_scev_info): Likewise.
4537 (struct instantiate_cache_entry): Remove.
4538 (struct instantiate_cache_type): Use a htab to map name, block
4540 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4541 (get_instantiated_value_entry): Likewise.
4542 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4543 (instantiate_scev_name): Adjust.
4545 2013-09-20 Jeff Law <law@redhat.com>
4547 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4549 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4551 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4552 Call aarch64_simd_expand_args to update op[argc].
4554 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4556 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4559 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4561 * gengtype.c (file_rules): Added rule for *.cc files.
4562 (get_output_file_with_visibility): Give fatal message when no
4565 2013-09-20 Renlin Li <renlin.li@arm.com>
4567 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4568 (aarch64_expand_epilogue): Likewise.
4569 (aarch64_legitimize_reload_address): Likewise.
4571 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4574 * expr.c (expand_assignment): Remove misalignp code path.
4576 2013-09-20 Marek Polacek <polacek@redhat.com>
4579 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4580 TYPE_PRECISION. Add asserts.
4582 2013-09-20 Richard Biener <rguenther@suse.de>
4584 PR tree-optimization/58453
4585 * tree-loop-distribution.c (distribute_loop): Apply the cost
4586 model for -ftree-loop-distribute-patterns, too.
4588 2013-09-20 Richard Biener <rguenther@suse.de>
4591 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4592 make type comparison less strict.
4594 2013-09-20 Alan Modra <amodra@gmail.com>
4596 * configure: Regenerate.
4597 * aclocal.m4: Regenerate.
4599 2013-09-20 Marek Polacek <polacek@redhat.com>
4602 * doc/extend.texi: Document that attribute used is meant to be used
4603 on variables with static storage duration.
4605 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4607 PR tree-optimization/58472
4608 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4609 simd_lane_access set inv_p = false.
4610 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4611 the simduid magic VAR_DECL.
4613 2013-09-19 Jan Hubicka <jh@suse.cz>
4615 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4617 2013-09-17 Jeff Law <law@redhat.com>
4619 * tree-ssa-dom.c (record_temporary_equivalences): New function
4620 split out of dom_opt_dom_walker::after_dom_children.
4621 (dom_opt_dom_walker::thread_across_edge): Move common code
4622 in here from dom_opt_dom_walker::after_dom_children.
4623 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4625 2013-09-19 Jan Hubicka <jh@suse.cz>
4627 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4628 (TARGET_GENERIC): Use PROCESOR_GENERIC
4629 (enum processor_type): Unify generic32 and 64.
4630 * i386.md (cpu): Likewise.
4631 * x86-tune.def (use_leave): Enable for generic32.
4632 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4633 * athlon.md: Change generic64 to generic in all occurences.
4634 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4635 (ix86_target_macros_internal): Likewise.
4636 * driver-i386.c (host_detect_local_cpu): Likewise.
4637 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4639 (generic_memcpy, generic_memset, generic_cost): This one.
4640 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4641 (m_GENERIC32, m_GENERIC64): Remove.
4642 (m_GENERIC): Turn into one flag.
4643 (processor_target): Unify generic tunnings.
4644 (ix86_option_override_internal): Replace generic32/64 by generic.
4645 (ix86_issue_rate): Likewise.
4646 (ix86_adjust_cost): Likewise.
4648 2013-09-19 Jan Hubicka <jh@suse.cz>
4650 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4651 of speculative flag.
4653 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4655 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4656 GOMP_sections_start, even if !exit_reachable.
4658 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4660 * lra-constraints.c (need_for_all_save_p): Use macro
4661 HARD_REGNO_CALL_PART_CLOBBERED.
4662 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4663 set up pseudo conflict hard regs.
4665 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4668 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4671 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4673 PR rtl-optimization/58438
4674 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4675 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4676 from previous subpasses.
4678 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4680 * arm.c (arm_get_frame_offsets): Validate architecture supports
4681 LDRD/STRD before accepting the tuning preference.
4682 (arm_expand_prologue): Likewise.
4683 (arm_expand_epilogue): Likewise.
4685 2013-09-18 Richard Biener <rguenther@suse.de>
4687 PR tree-optimization/58417
4688 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4689 have chrecs with symbols defined in the loop as operands.
4690 (chrec_fold_multiply): Likewise.
4691 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4692 parameters before folding binary operations.
4693 (struct instantiate_cache_entry_hasher): Remove.
4694 (struct instantiate_cache_type): Use a pointer-map.
4695 (instantiate_cache_type::instantiate_cache_type): New function.
4696 (instantiate_cache_type::get): Likewise.
4697 (instantiate_cache_type::set): Likewise.
4698 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4699 (get_instantiated_value_entry): Likewise.
4700 (global_cache): New global.
4701 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4702 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4703 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4704 (instantiate_scev_name): Adjust.
4705 (instantiate_scev): Construct global instead of local cache.
4706 (resolve_mixers): Likewise.
4708 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4709 Paolo Carlini <paolo.carlini@oracle.com>
4712 * doc/implement-cxx.texi: Fix references to the C++ standards.
4714 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4716 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4717 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4718 simd_lane_access drs, update also DR_ALIGNED_TO.
4720 2013-09-18 Marek Polacek <polacek@redhat.com>
4723 * doc/extend.texi: Document no_sanitize_undefined attribute.
4724 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4725 no_sanitize_undefined attribute.
4727 2013-09-18 Nick Clifton <nickc@redhat.com>
4729 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4730 (ASM_DECLARE_FUNCTION_NAME): Define.
4732 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4734 * compare-elim.c (find_comparison_dom_walker): New class
4735 (find_comparisons_in_bb): Rename to
4736 find_comparison_dom_walker::before_dom_children
4737 (find_comparisons): Adjust
4738 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4740 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4741 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4742 (init_walk_dominator_tree): Remove declaration.
4743 (fini_walk_dominator_tree): Remove declaration.
4744 * fwprop.c (single_def_use_dom_walker): New class
4745 (single_def_use_enter_block): Convert to
4746 single_def_use_dom_walker::before_dom_children.
4747 (single_def_use_leave_block): Convert to
4748 single_def_use_dom_walker::after_dom_children.
4749 (build_single_def_use_links): Adjust.
4750 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4752 (find_candidates_in_block): Convert to
4753 find_candidates_dom_walker::before_dom_children.
4754 (execute_strength_reduction): Adjust.
4755 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4756 (sese_dom_walker): New class.
4757 (sese_dom_walker::sese_dom_walker): New constructor.
4758 (sese_dom_walker::~sese_dom_walker): New destructor.
4759 (build_sese_conditions_before): Convert to
4760 sese_dom_walker::before_dom_children.
4761 (build_sese_conditions_after): Convert to
4762 sese_dom_walker::after_dom_children.
4763 (build_poly_scop): Adjust
4764 * tree-into-ssa.c (rewrite_dom_walker): New class
4765 (rewrite_enter_block): Convert to
4766 rewrite_dom_walker::before_dom_children.
4767 (rewrite_leave_block): Convert to
4768 rewrite_dom_walker::after_dom_children.
4769 (rewrite_update_dom_walker): New class.
4770 (rewrite_update_enter_block): Convert to
4771 rewrite_update_dom_walker::before_dom_children.
4772 (rewrite_update_leave_block): Convert to
4773 rewrite_update_dom_walker::after_dom_children.
4774 (rewrite_blocks, rewrite_into_ssa): Adjust.
4775 (mark_def_dom_walker): New class.
4776 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4777 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4778 (mark_def_sites_blocks): Convert to
4779 mark_def_dom_walker::before_dom_children.
4780 (mark_def_site_blocks): Remove.
4781 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4782 (tree_ssa_dominator_optimize): Adjust.
4783 (dom_thread_across_edge): Convert to method
4784 dom_opt_dom_walker::thread_across_edge.
4785 (dom_opt_enter_block): Convert to member function
4786 dom_opt_dom_walker::before_dom_children.
4787 (dom_opt_leave_block): Convert to member function
4788 dom_opt_dom_walker::after_dom_children.
4789 * tree-ssa-dse.c (dse_dom_walker): New class.
4790 (dse_enter_block): Convert to member function
4791 dse_dom_walker::before_dom_children.
4792 (tree_ssa_dse): Adjust.
4793 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4794 (determine_invariantness_stmt): Convert to method
4795 invariantness_dom_walker::before_dom_children.
4796 (determine_invariantness): Remove
4797 (move_computations_dom_walker): New class.
4798 (move_computations_stmt): Convert to method
4799 move_computations_dom_walker::before_dom_children.
4800 (move_computations, tree_ssa_lim): Adjust.
4801 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4802 (nt_init_block): Make method
4803 notrappping_dom_walker::before_dom_children.
4804 (nt_fini_block): Make
4805 method nontrapping_dom_walker::after_dom_children.
4806 (get_non_trapping): Adjust.
4807 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4808 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4809 (eliminate_leave_block): Make method.
4810 eliminate_dom_walker::after_dom_children.
4812 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4813 (strlen_enter_block): Make method
4814 strlen_dom_walker::before_dom_children.
4815 (strlen_leave_block): Make
4816 method strlen_dom_walker::after_dom_children.
4817 (tree_ssa_strlen): Adjust.
4818 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4819 (tree_ssa_uncprop): Adjust.
4820 (uncprop_leave_block): Make method
4821 uncprop_dom_walker::after_dom_children.
4822 (uncprop_leave_block): Make method
4823 uncprop_dom_walker::before_dom_children.
4825 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4827 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4828 branch in same basic block. Check both src and dest of the move insn.
4830 2013-09-17 Nick Clifton <nickc@redhat.com>
4832 * config/rl78/rl78-real.md (bf): New pattern.
4834 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4835 (rl78_print_operand): Do not put a # before a %B.
4836 * config/rl78/rl78.opt: Tweak doc strings.
4838 2013-09-17 DJ Delorie <dj@redhat.com>
4840 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4841 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4842 (rl78_option_override): Likewise, if -mallregs.
4843 (is_virtual_register): Likewise.
4844 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4845 (REGNO_OK_FOR_BASE_P): Likewise.
4846 * config/rl78/rl78.opt (-mallregs): New.
4848 2013-09-17 Nick Clifton <nickc@redhat.com>
4850 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4851 For interrupt functions: save all call clobbered registers if the
4852 interrupt handler is not a leaf function.
4853 (rl78_expand_prologue): Always recompute the frame information.
4854 For interrupt functions: only select bank 0 if one of the bank 0
4855 registers is going to be psuhed.
4857 2013-09-17 DJ Delorie <dj@redhat.com>
4859 * config/rl78/constraints.md: For each W* constraint, rename to C*
4860 and create a W* constraint that checks for an optional ES: prefix
4862 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4863 (es_addr): New. Used to wrap far addresses.
4864 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4865 (rl78_es_base): New.
4866 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4867 wrapped far addresses.
4868 (rl78_print_operand_1): Unwrap far addresses before processing.
4869 (rl78_lo16): Wrap far addresses in unspecs.
4870 (rl78_es_addr): New.
4871 (rl78_es_base): New.
4872 (insn_ok_now): Check for not-yet-wrapped far addresses.
4873 (transcode_memory_rtx): Properly re-wrap far addresses.
4875 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4877 * config/sparc/t-rtems: Add leon3 multilibs.
4879 2013-09-17 Cong Hou <congh@google.com>
4881 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4882 when checking the dot production pattern. The type of rhs operand
4883 of multiply is now checked correctly.
4885 2013-09-17 Jeff Law <law@redhat.com>
4887 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4888 edge implied equivalences into successor phis.
4889 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4890 here from tree-ssa-threadedge.c.
4891 (mark_threaded_blocks): When threading through a joiner, if both
4892 successors of the joiner's clone reach the same block, verify the
4893 PHI arguments are equal. If not, cancel the jump threading request.
4894 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4895 tree-ssa-threadupdate.c
4896 (thread_across_edge): Don't check PHI argument equality when
4897 threading through joiner block here.
4899 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4901 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4902 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4903 (warn_uninit, warn_uninitialized_vars,
4904 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4905 Move to tree-ssa-uninit.c.
4906 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4907 (has_undefined_value_p): New. Pass dependant parts of
4908 ssa_undefined_value_p.
4909 (uninit_undefined_value_p): Use has_undefined_value_p.
4910 (warn_uninit, warn_uninitialized_vars,
4911 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4912 Move from tree-ssa.c.
4913 * tree-ssa.h: Adjust prototypes.
4915 2013-09-17 Jan Hubicka <jh@suse.cz>
4918 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4919 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4920 FUNCTION_NOT_OPTIMIZED.
4921 * ipa-inline-analysis.c (compute_inline_parameters): Function
4922 not optimized is not inlinable unless it is alwaysinline.
4923 (inline_analyze_function): Force calls in not optimized
4924 function not inlinable.
4926 2013-09-17 Jan Hubicka <jh@suse.cz>
4929 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4931 * ipa.c (function_and_variable_visibility): Likewise.
4932 * ipa-profile.c (ipa_profile): Likewise.
4934 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4937 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4938 attribute, and don't inline the resolver in this case.
4940 2013-09-17 Teresa Johnson <tejohnson@google.com>
4942 * coverage.c (get_coverage_counts): Add missing newline.
4944 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4946 PR tree-optimization/58088
4947 * fold-const.c (mask_with_trailing_zeros): New function.
4948 (fold_binary_loc): Make sure we don't recurse infinitely
4949 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4950 Use mask_with_trailing_zeros where appropriate.
4952 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4954 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4955 of 'prev' var to get better distance estimation.
4957 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4959 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4960 * tree-inline.c (is_parameter_of): New predicate.
4961 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4962 a parameter has been remapped.
4963 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4964 (optimize_inline_calls): Initialize transform_parameter.
4965 (copy_gimple_seq_and_replace_locals): Likewise.
4966 (tree_function_versioning): Likewise.
4967 (maybe_inline_call_in_expr): Likewise.
4969 2013-09-17 Nick Clifton <nickc@redhat.com>
4971 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4972 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4974 (is_attr_func): New function.
4975 (msp430_is_interrupt_func): New function.
4976 (is_naked_func): New function.
4977 (is_reentrant_func): New function.
4978 (is_critical_func): New function.
4979 (msp430_start_function): Add annotations for function attributes.
4980 (msp430_attr): New function.
4981 (msp430_attribute_table): New.
4982 (msp430_function_section): New function.
4983 (TARGET_ASM_FUNCTION_SECTION): Define.
4984 (msp430_builtin): New enum.
4985 (msp430_init_builtins): New function.
4986 (msp430_builtin_devl): New function.
4987 (msp430_expand_builtin): New function.
4988 (TARGET_INIT_BUILTINS): Define.
4989 (TARGET_EXPAND_BUILTINS): Define.
4990 (TARGET_BUILTIN_DECL): Define.
4991 (msp430_expand_prologue): Add support for naked, interrupt,
4992 critical and reentrant functions.
4993 (msp430_expand_epilogue): Likewise.
4994 (msp430_print_operand): Handle 'O' character.
4995 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4997 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4998 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4999 (pushm): Use a 'n' rather than an 'i' constraint.
5000 (msp_return): Add generation of the interrupt return instruction.
5001 (disable_interrupts): New pattern.
5002 (enable_interrupts): New pattern.
5003 (push_intr_state): New pattern.
5004 (pop_intr_state): New pattern.
5005 (bic_SR): New pattern.
5006 (bis_SR): New pattern.
5007 * doc/extend.texi: Document MSP430 function attributes and builtin
5010 2013-09-17 Richard Biener <rguenther@suse.de>
5012 PR tree-optimization/58432
5013 * tree-loop-distribution.c (tree_loop_distribution): Also
5014 scan PHIs for outside loop uses and seed a partition from them.
5016 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5018 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5019 (restructure_reference): Call backtrace_base_for_ref.
5021 2013-09-17 Alan Modra <amodra@gmail.com>
5024 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5027 2013-09-16 DJ Delorie <dj@redhat.com>
5029 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5030 vregs location for RL78/G10.
5031 (rl78_expand_prologue): Avoid SEL on G10.
5032 (rl78_expand_epilogue): Likewise.
5033 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5034 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5035 __RL78_G10__ when appropriate.
5036 (ASM_SPEC): Pass -mg10 along to the assembler.
5037 * config/rl78/rl78.md (sel_rb): Disable for G10.
5038 * config/rl78/rl78.opt: Add -mg10 option.
5039 * config/rl78/t-rl78: Add -mg10 multilib.
5041 2013-09-16 Xinliang David Li <davidxl@google.com>
5043 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5044 * omp-low.c (omp_max_vf): Ditto.
5045 (expand_omp_simd): Ditto.
5046 * tree-vectorizer.c (vectorize_loops): Ditto.
5047 (gate_vect_slp): Ditto.
5048 (gate_increase_alignment): Ditto.
5049 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5050 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5051 (gate_tree_vectorize): Name change.
5052 (tree_vectorize): Ditto.
5053 (pass_vectorize::gate): Call new function.
5054 (pass_vectorize::execute): Ditto.
5055 * opts.c: O3 default setting change.
5056 (finish_options): Check new flag.
5057 * doc/invoke.texi: Document new flags.
5058 * common.opt: New flags.
5060 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5062 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5063 (Condition Code): Don't reference it.
5064 * doc/tm.texi: Regenerate.
5066 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5069 * lra-constraints.c (undo_optional_reloads): Consider all optional
5070 reload even if it did not get a hard reg.
5072 2013-09-16 Teresa Johnson <tejohnson@google.com>
5074 * dumpfile.c (dump_loc): Remove newline emission.
5075 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5076 emission to dump_printf_loc calls where missing.
5077 (vect_mark_for_runtime_alias_test): Ditto.
5078 (vect_analyze_data_ref_dependence): Ditto.
5079 (vect_analyze_data_ref_dependences): Ditto.
5080 (vect_slp_analyze_data_ref_dependence): Ditto.
5081 (vect_slp_analyze_data_ref_dependences): Ditto.
5082 (vect_compute_data_ref_alignment): Ditto.
5083 (vect_update_misalignment_for_peel): Ditto.
5084 (vect_verify_datarefs_alignment): Ditto.
5085 (vector_alignment_reachable_p): Ditto.
5086 (vect_get_data_access_cost): Ditto.
5087 (vect_enhance_data_refs_alignment): Ditto.
5088 (vect_find_same_alignment_drs): Ditto.
5089 (vect_analyze_data_refs_alignment): Ditto.
5090 (vect_analyze_group_access): Ditto.
5091 (vect_analyze_data_ref_access): Ditto.
5092 (vect_analyze_data_ref_accesses): Ditto.
5093 (vect_prune_runtime_alias_test_list): Ditto.
5094 (vect_analyze_data_refs): Ditto.
5095 (vect_create_addr_base_for_vector_ref): Ditto.
5096 (vect_create_data_ref_ptr): Ditto.
5097 (vect_grouped_store_supported): Ditto.
5098 (vect_grouped_load_supported): Ditto.
5099 * value-prof.c (check_counter): Ditto.
5100 (check_ic_target): Ditto.
5101 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5102 (vect_recog_widen_mult_pattern): Ditto.
5103 (vect_recog_widen_sum_pattern): Ditto.
5104 (vect_recog_over_widening_pattern): Ditto.
5105 (vect_recog_widen_shift_pattern): Ditto.
5106 (vect_recog_rotate_pattern): Ditto.
5107 (vect_recog_vector_vector_shift_pattern): Ditto.
5108 (vect_recog_divmod_pattern): Ditto.
5109 (vect_recog_mixed_size_cond_pattern): Ditto.
5110 (vect_recog_bool_pattern): Ditto.
5111 (vect_pattern_recog_1): Ditto.
5112 (vect_pattern_recog): Ditto.
5113 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5114 (vect_is_simple_iv_evolution): Ditto.
5115 (vect_analyze_scalar_cycles_1): Ditto.
5116 (vect_get_loop_niters): Ditto.
5117 (vect_analyze_loop_1): Ditto.
5118 (vect_analyze_loop_form): Ditto.
5119 (vect_analyze_loop_operations): Ditto.
5120 (vect_analyze_loop_2): Ditto.
5121 (vect_analyze_loop): Ditto.
5122 (report_vect_op): Ditto.
5123 (vect_is_slp_reduction): Ditto.
5124 (vect_is_simple_reduction_1): Ditto.
5125 (vect_get_known_peeling_cost): Ditto.
5126 (vect_estimate_min_profitable_iters): Ditto.
5127 (vect_model_reduction_cost): Ditto.
5128 (vect_model_induction_cost): Ditto.
5129 (get_initial_def_for_induction): Ditto.
5130 (vect_create_epilog_for_reduction): Ditto.
5131 (vectorizable_reduction): Ditto.
5132 (vectorizable_induction): Ditto.
5133 (vectorizable_live_operation): Ditto.
5134 (vect_loop_kill_debug_uses): Ditto.
5135 (vect_transform_loop): Ditto.
5136 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5137 (vect_stmt_relevant_p): Ditto.
5138 (process_use): Ditto.
5139 (vect_mark_stmts_to_be_vectorized): Ditto.
5140 (vect_model_simple_cost): Ditto.
5141 (vect_model_promotion_demotion_cost): Ditto.
5142 (vect_model_store_cost): Ditto.
5143 (vect_get_store_cost): Ditto.
5144 (vect_model_load_cost): Ditto.
5145 (vect_get_load_cost): Ditto.
5146 (vect_init_vector_1): Ditto.
5147 (vect_get_vec_def_for_operand): Ditto.
5148 (vect_finish_stmt_generation): Ditto.
5149 (vectorizable_call): Ditto.
5150 (vectorizable_conversion): Ditto.
5151 (vectorizable_assignment): Ditto.
5152 (vectorizable_shift): Ditto.
5153 (vectorizable_operation): Ditto.
5154 (vectorizable_store): Ditto.
5155 (vectorizable_load): Ditto.
5156 (vectorizable_condition): Ditto.
5157 (vect_analyze_stmt): Ditto.
5158 (vect_transform_stmt): Ditto.
5159 (vect_is_simple_use): Ditto.
5160 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5161 (vect_can_advance_ivs_p): Ditto.
5162 (vect_update_ivs_after_vectorizer): Ditto.
5163 (vect_do_peeling_for_loop_bound): Ditto.
5164 (vect_gen_niters_for_prolog_loop): Ditto.
5165 (vect_update_inits_of_drs): Ditto.
5166 (vect_create_cond_for_alias_checks): Ditto.
5167 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5168 (vect_build_slp_tree_1): Ditto.
5169 (vect_supported_load_permutation_p): Ditto.
5170 (vect_analyze_slp_instance): Ditto.
5171 (vect_analyze_slp): Ditto.
5172 (vect_make_slp_decision): Ditto.
5173 (vect_detect_hybrid_slp): Ditto.
5174 (vect_bb_vectorization_profitable_p): Ditto.
5175 (vect_slp_analyze_bb_1): Ditto.
5176 (vect_update_slp_costs_according_to_vf): Ditto.
5177 (vect_get_mask_element): Ditto.
5178 (vect_transform_slp_perm_load): Ditto.
5179 (vect_schedule_slp_instance): Ditto.
5180 (vect_schedule_slp): Ditto.
5181 (vect_slp_transform_bb): Ditto.
5182 * profile.c (read_profile_edge_counts): Ditto.
5183 (compute_branch_probabilities): Ditto.
5184 * coverage.c (get_coverage_counts): Ditto.
5186 2013-09-16 Diego Novillo <dnovillo@google.com>
5188 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5190 2013-09-16 Jan Hubicka <jh@suse.cz>
5192 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5193 abstract functions; for static functions check the presence of body.
5195 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5197 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5198 * config/aarch64/aarch64-simd.md
5199 (aarch64_mla_elt<mode>): New.
5200 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5201 (aarch64_mls_elt<mode>): Likewise.
5202 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5203 (aarch64_fma4_elt<mode>): Likewise.
5204 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5205 (aarch64_fma4_elt_to_128v2df): Likewise.
5206 (aarch64_fma4_elt_to_64df): Likewise.
5207 (fnma<mode>4): Likewise.
5208 (aarch64_fnma4_elt<mode>): Likewise.
5209 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5210 (aarch64_fnma4_elt_to_128v2df): Likewise.
5211 (aarch64_fnma4_elt_to_64df): Likewise.
5212 * config/aarch64/iterators.md (VDQSF): New.
5213 * config/aarch64/arm_neon.h
5214 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5215 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5217 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5219 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5220 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5221 (aarch64_mul3_elt_to_128df): Likewise.
5222 (aarch64_mul3_elt_to_64v2df): Likewise.
5223 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5225 (VMUL_CHANGE_NLANES) Likewise.
5228 * config/aarch64/arm_neon.h
5229 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5231 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5233 * config/aarch64/arm_neon.h
5234 (vcvtx_high_f32_f64): Fix parameters.
5236 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5237 Uros Bizjak <ubizjak@gmail.com>
5239 * config/alpha.c: Include tree-ssanames.h.
5241 2013-09-16 Richard Biener <rguenther@suse.de>
5243 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5244 (dot_rdg_1): Handle control_dd.
5245 (create_edge_for_control_dependence): New function.
5246 (create_rdg_edges): Add control dependences if asked for.
5247 (build_rdg): Likewise.
5248 (generate_loops_for_partition): If there are not necessary
5249 control stmts remove all their dependencies.
5250 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5251 (distribute_loop): Pass on control dependences.
5252 (tree_loop_distribution): Compute control dependences and remove
5253 restriction on number of loop nodes.
5255 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5257 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5260 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5262 * cse.c (try_const_anchors): Punt on CC modes.
5264 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5266 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5268 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5271 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5274 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5277 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5279 (darwin_objc1_section): Likewise.
5280 (darwin_file_end): Emit Image Info section when required.
5282 2013-09-14 Jan Hubicka <jh@suse.cz>
5284 * tree-into-ssa.c (gate_into_ssa): New.
5285 (pass_data_build_ssa): Use it.
5286 * cgraph.h (expand_thunk): Update prototype.
5287 * cgraphunit.c (analyze_function): Expand thunks early.
5288 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5289 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5290 set lowered flag; do not add new function.
5291 (assemble_thunks_and_aliases): Update.
5292 * tree-ssa.c (gate_init_datastructures): New gate.
5293 (pass_data_init_datastructures): Use it.
5295 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5298 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5299 xmm register set described in the psABI.
5301 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5303 * dwarf2out.c (should_emit_struct_debug): Add check
5304 for type_decl variable is not NULL.
5306 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5308 * config.gcc: Use new winnt-c.c target hooks
5309 * config/t-winnt: New file
5310 * config/winnt-c.c: New file
5311 * doc/tm.texi.in: Document new hook
5312 * doc/tm.texi: Regenerated
5314 2013-09-13 Jan Hubicka <jh@suse.cz>
5317 * ipa-inline.c (check_callers): New function.
5318 (check_caller_edge): Remove.
5319 (want_inline_function_to_all_callers_p): Also permit alises that are
5321 (inline_to_all_callers): Terminate the walk when devirtualization
5322 introduce new calls.
5324 2013-09-13 Jan Hubicka <jh@suse.cz>
5326 * ipa-inline-analysis.c (struct growth_data): Add node.
5327 (do_estimate_growth_1): Fix detection of recursion.
5329 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5331 PR tree-optimization/58392
5332 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5333 to avoid shadowing of outer loop variable. If
5334 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5335 replace_by_duplicate_decl simduid of loops that have it set and
5336 set dest_cfun->has_simduid_loops and/or
5337 dest_cfun->has_force_vect_loops.
5338 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5339 instead of maybe_lookup_decl.
5340 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5341 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5342 call. Copy over force_vect and copy and remap simduid. Set
5343 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5344 (copy_cfg_body): Remove blocks_to_copy argument. Use
5345 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5346 caller. Don't set cfun->has_simduid_loops and/or
5347 cfun->has_force_vect_loops here.
5348 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5350 (expand_call_inline, tree_function_versioning): Adjust copy_body
5353 2013-09-13 Martin Jambor <mjambor@suse.cz>
5356 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5357 the assert if the edge was a speculative one.
5359 2013-09-13 Richard Biener <rguenther@suse.de>
5361 * tree-data-ref.h (known_dependences_p): Move here ...
5362 * tree-loop-distribution.c (known_dependences_p): ... from here.
5363 (dump_rdg_component, debug_rdg_component): Remove.
5365 (generate_loops_for_partition): Use gimple_uid instead of
5366 relying on matching stmt visit order.
5367 (rdg_build_partitions): Take starting stmt vector.
5368 (ldist_gen): Merge into ...
5369 (distribute_loop): ... this function. Do not compute starting
5371 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5373 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5375 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5376 Set type attribute correctly. Set predicable_short_it attribute.
5377 (cmpsi_shiftsi): Remove %? from output template.
5379 2013-09-13 Richard Biener <rguenther@suse.de>
5381 * tree-loop-distribution.c (struct rdg_component,
5382 rdg_defs_used_in_other_loops_p, free_rdg_components,
5383 rdg_build_components): Remove.
5384 (stmts_from_loop): Do not record virtual PHIs.
5385 (generate_loops_for_partition): Skip virtual PHIs.
5386 (build_rdg_partition_for_component): Rename to ...
5387 (build_rdg_partition_for_vertex): ... this and adjust.
5388 (rdg_build_partitions): Take a vector of starting vertices
5389 instead of components. Remove unnecessary leftover handling.
5390 (ldist_gen): Do not build components or record other stores.
5391 (distribute_loop): Do not distribute loops containing stmts
5394 2013-09-13 Christian Bruel <christian.bruel@st.com>
5397 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5399 2013-09-13 Kai Tietz <ktietz@redhat.com>
5401 * config.gcc: Separate cases for mingw and cygwin targets,
5402 and add 64-bit cygwin target case.
5404 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5405 dll-export inline-functions.
5406 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5409 2013-09-13 Jeff Law <law@redhat.com>
5413 2013-09-06 Jeff Law <law@redhat.com>
5415 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5416 edge implied equivalences into successor phis.
5418 2013-09-12 DJ Delorie <dj@redhat.com>
5420 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5423 2013-09-12 Brooks Moses <bmoses@google.com>
5426 * Makefile.in: Do not install driver binaries in $(target)/bin.
5428 2013-09-12 DJ Delorie <dj@redhat.com>
5430 * config/rl78/rl78.opt (mrelax): New.
5431 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5432 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5434 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5436 (rl78_expand_epilogue): Likewise.
5438 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5441 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5442 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5443 pointer elimination with using remove_reg_equal_offset_note.
5445 2013-09-12 DJ Delorie <dj@redhat.com>
5447 * config/msp430/: New port.
5448 * config.gcc (msp430): Added.
5449 * doc/invoke.texi: Document MSP430 options.
5450 * doc/install.texi: Document msp430-elf
5451 * doc/md.texi: Document msp430-elf
5452 * doc/contrib.texi: Document msp430-elf
5454 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5457 2013-09-12 Martin Jambor <mjambor@suse.cz>
5460 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5461 reference descriptor is NULL.
5462 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5465 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5467 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5468 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5469 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5470 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5471 make_temp_ssa_name): move to tree-ssanames.h
5472 * tree-ssa-alias.h: Move prototype.
5473 * tree-ssa.h: Include tree-ssanames.h.
5474 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5475 * tree-ssanames.h: New. Move items from tree-flow*.h
5476 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5478 2013-09-12 Richard Biener <rguenther@suse.de>
5480 PR tree-optimization/58404
5481 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5482 propagate non-invariant addresses into dereferences wrapped
5483 in component references.
5485 2013-09-12 Richard Biener <rguenther@suse.de>
5487 PR tree-optimization/58402
5488 * passes.def: Move pass_late_warn_uninitialized later.
5490 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5492 * tree-ssa.h: New. Move content from tree-flow.h and
5494 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5495 Move prototypes belonging to tree-ssa.c.
5496 * tree-flow-inline.h (redirect_edge_var_map_def,
5497 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5499 * gimple.h: Adjust prototypes.
5500 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5502 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5503 * tree.h: Move prototype to tree-ssa.h.
5504 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5505 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5506 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5507 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5508 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5509 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5510 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5511 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5512 graphite-blocking.c, graphite-clast-to-gimple.c,
5513 graphite-dependences.c, graphite-interchange.c,
5514 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5515 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5516 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5517 ipa-reference.c, ipa-split.c, ipa-utils.c,
5518 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5519 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5520 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5521 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5522 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5523 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5524 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5525 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5526 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5527 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5528 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5529 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5530 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5531 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5532 value-prof.c, var-tracking.c,
5533 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5535 2013-09-12 Richard Biener <rguenther@suse.de>
5537 PR tree-optimization/58396
5538 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5539 (build_rdg): Take a loop-nest parameter, fix memleaks.
5540 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5542 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5544 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5547 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5549 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5550 The FMA instruction names should have a 'v' prefix.
5552 2013-09-12 Richard Biener <rguenther@suse.de>
5554 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5555 (dot_rdg): Use popen instead of system in optional code.
5556 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5557 (already_processed_vertex_p): Adjust.
5558 (has_anti_or_output_dependence, predecessor_has_mem_write,
5559 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5560 rdg_flag_uses): Remove.
5561 (rdg_flag_vertex): Simplify.
5562 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5564 (build_rdg_partition_for_component): Process the first vertex
5565 of a component only.
5566 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5568 2013-09-12 Alan Modra <amodra@gmail.com>
5570 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5572 2013-09-11 DJ Delorie <dj@redhat.com>
5573 Nick Clifton <nickc@redhat.com>
5575 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5576 (rl78_stack_based_mem): New.
5577 * config/rl78/constraints.md (Iv08): New.
5586 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5588 (movsi): Change from expand to insn-and-split.
5589 (ashrsi3): Clobber AX.
5593 * config/rl78/rl78.md (CC_REG): Fix.
5594 (addsi3): Allow memory and immediate operands.
5595 (addsi3_internal): Split into...
5596 (addsi3_internal_virt): ...new, and ...
5597 (addsi3_internal_real): ...new.
5599 (subsi3_internal_virt): New.
5600 (subsi3_internal_real): New.
5601 (mulsi3): Add memory operand.
5602 (mulsi3_rl78): Likewise.
5603 (mulsi3_g13): Likewise.
5604 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5605 (cbranchqi4_real): Add more constraint options.
5606 (cbranchhi4_real): Expand pattern.
5607 (cbranchhi4_real_signed): New.
5608 (cbranchhi4_real_inverted): New.
5609 (cbranchsi4_real_lt): New.
5610 (cbranchsi4_real_ge): New.
5611 (cbranchsi4_real_signed): New.
5612 (cbranchsi4_real): New.
5614 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5616 (lshrsi3_virt): Likewise.
5617 (ashlsi3_virt): Likewise.
5618 (cbranchqi4_virt_signed): New.
5619 (cbranchhi4_virt_signed): New.
5620 (cbranchsi4_virt): New.
5621 * config/rl78/rl78.c: Whitespace fixes throughout.
5622 (move_elim_pass): New.
5623 (pass_data_rl78_move_elim): New.
5624 (pass_rl78_move_elim): New.
5625 (make_pass_rl78_move_elim): New.
5626 (rl78_devirt_info): Run devirt earlier.
5627 (rl78_move_elim_info): New.
5628 (rl78_asm_file_start): Register it.
5629 (rl78_split_movsi): New.
5630 (rl78_as_legitimate_address): Allow virtual base registers when
5632 (rl78_addr_space_convert): Remove spurious debug stuff.
5633 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5634 (rl78_print_operand): More cases for not printing '#'.
5635 (rl78_expand_compare): Remove most of the logic.
5636 (content_memory): New.
5637 (clear_content_memory): New.
5638 (get_content_index): New.
5639 (get_content_name): New.
5640 (display_content_memory): New.
5641 (update_content): New.
5642 (record_content): New.
5643 (already_contains): New.
5644 (insn_ok_now): Re-recog insns with virtual registers.
5645 (add_postponed_content_update): New.
5646 (process_postponed_content_update): New.
5647 (gen_and_emit_move): New.
5648 (transcode_memory_rtx): Record new location content.
5649 Use gen_and_emit_move.
5650 (force_into_acc): New.
5651 (move_to_acc): Use gen_and_emit_move.
5652 (move_from_acc): Likewise.
5653 (move_acc_to_reg): Likewise.
5654 (move_to_x): Likewise.
5655 (move_to_hl): Likewise.
5656 (move_to_de): Likewise.
5657 (rl78_alloc_physical_registers_op1): Record location content.
5658 (has_constraint): New.
5659 (rl78_alloc_physical_registers_op2): Record location content.
5661 (rl78_alloc_physical_registers_ro1): Likewise.
5662 (rl78_alloc_physical_registers_cmp): Likewise.
5663 (rl78_alloc_physical_registers_umul): Likewise.
5664 (rl78_alloc_address_registers_macax): New.
5665 (rl78_alloc_physical_registers): Initialize and set location
5666 content memory as needed.
5667 (rl78_reorg): Make sure split2 is called.
5668 (rl78_rtx_costs): New.
5670 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5672 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5673 for (not (neg ...)) and (neg (not ...)) cases.
5675 2013-09-11 Richard Biener <rguenther@suse.de>
5678 * passes.def: Split critical edges before late uninit warning passes.
5679 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5681 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5683 PR tree-optimization/58385
5684 * fold-const.c (build_range_check): If both low and high are NULL,
5685 use omit_one_operand_loc to preserve exp side-effects.
5687 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5689 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5691 2013-09-11 Richard Biener <rguenther@suse.de>
5693 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5694 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5695 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5696 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5697 stmts_from_loop, known_dependences_p, build_empty_rdg,
5698 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5699 * tree-loop-distribution.c: ... here.
5700 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5701 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5702 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5703 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5704 * tree-loop-distribution.c: ... here.
5705 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5706 (struct partition_s): Add loops member.
5707 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5708 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5709 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5711 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5712 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5713 Sergey Lega <sergey.s.lega@intel.com>
5714 Anna Tikhonova <anna.tikhonova@intel.com>
5715 Ilya Tocar <ilya.tocar@intel.com>
5716 Andrey Turetskiy <andrey.turetskiy@intel.com>
5717 Ilya Verbin <ilya.verbin@intel.com>
5718 Kirill Yukhin <kirill.yukhin@intel.com>
5719 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5721 * config/i386/constraints.md (k): New.
5723 * config/i386/i386.c (const regclass_map): Add new mask registers.
5724 (dbx_register_map): Ditto.
5725 (dbx64_register_map): Ditto.
5726 (svr4_dbx_register_map): Ditto.
5727 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5729 (ix86_preferred_reload_class): Disable constants for mask registers.
5730 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5731 (ix86_hard_regno_mode_ok): Support new mask registers.
5732 (x86_order_regs_for_local_alloc): Ditto.
5733 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5734 (FIXED_REGISTERS): Add new mask registers.
5735 (CALL_USED_REGISTERS): Ditto.
5736 (REG_ALLOC_ORDER): Ditto.
5737 (VALID_MASK_REG_MODE): New.
5738 (FIRST_MASK_REG): Ditto.
5739 (LAST_MASK_REG): Ditto.
5740 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5741 (MAYBE_MASK_CLASS_P): New.
5742 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5743 (REG_CLASS_CONTENTS): Ditto.
5744 (MASK_REGNO_P): New.
5745 (ANY_MASK_REG_P): Ditto.
5746 (HI_REGISTER_NAMES): Add new mask registers.
5747 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5748 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5750 (attribute "type"): Add mskmov, msklog.
5751 (attribute "length_immediate"): Support them.
5752 (attribute "memory"): Ditto.
5753 (attribute "prefix_0f"): Ditto.
5754 (*movhi_internal): Support new mask registers.
5755 (*movqi_internal): Ditto.
5756 (define_split): Split out clobber pattern is a logic
5757 insn on mask registers.
5758 (*k<logic><mode>): New.
5759 (*andhi_1): Extend to support mask regs.
5760 (*andqi_1): Extend to support mask regs.
5762 (define_split): Split and-not to and and not if operands
5764 (*<code><mode>_1): Separate HI mode to new pattern...
5765 (*<code>hi_1): This.
5766 (*<code>qi_1): Extend to support mask regs.
5768 (kortestzhi): Ditto.
5769 (kortestchi): Ditto.
5771 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5772 (*one_cmplhi2_1): ...Here, now with mask registers support.
5773 (*one_cmplqi2_1): Support new mask registers.
5774 (HI/QImode arithmetics splitter): Don't split if mask registers
5776 (HI/QImode not splitter): Ditto.
5777 * config/i386/predicated.md (mask_reg_operand): New.
5778 (general_reg_operand): Ditto.
5780 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5782 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5783 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5785 2013-09-10 Jeff Law <law@redhat.com>
5787 PR tree-optimization/58380
5788 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5789 of threading through a buried loop header.
5791 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5792 return value for single successor case.
5794 2013-09-10 Jan Hubicka <jh@suse.cz>
5796 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5798 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5801 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5802 support conditional execution.
5803 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5805 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5807 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5809 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5810 Check destination too to check move insn.
5811 (undo_optional_reloads): Add check that the original peudo did not
5812 changed its allocation and the optional reload was inherited on last
5813 inheritance pass. Break loop after deciding to keep optional reload.
5814 (lra_undo_inheritance): Add check that inherited pseudo still in
5817 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5819 * config/aarch64/aarch64.md (generic_sched): New.
5820 * config/aarch64/aarch64-generic.md (load): Make conditional
5821 on generic_sched attribute.
5822 (nonload): Likewise.
5824 2013-09-10 Jan Hubicka <jh@suse.cz>
5826 * lto-cgraph.c: Include ipa-utils.h.
5827 (compute_ltrans_boundary): Also add possible targets into the boundary.
5829 2013-09-10 Jan Hubicka <jh@suse.cz>
5831 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5832 VAR_DECL of vtable rather than full expression.
5834 2013-09-10 Jan Hubicka <jh@suse.cz>
5835 Paolo Carlini <paolo.carlini@oracle.com>
5837 * cgraphunit.c (analyze_functions): Save input_location, set it
5838 to UNKNOWN_LOCATION and restore it at the end.
5840 2013-09-10 Martin Jambor <mjambor@suse.cz>
5842 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5843 represented by a thunk.
5845 2013-09-10 Jeff Law <law@redhat.com>
5847 PR tree-optimization/58343
5848 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5849 jump threading requests through loop headers buried in the middle
5850 of a jump threading path.
5852 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5853 in return value/type.
5855 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5857 PR rtl-optimization/58365
5858 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5859 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5862 2013-09-10 Richard Biener <rguenther@suse.de>
5864 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5865 * tree-data-ref.c (create_rdg_vertices): Collect all data
5866 references, signal failure to the caller, use data-ref API.
5867 (build_rdg): Compute data references only once. Maintain lifetime
5868 of data references and data dependences from within RDG.
5869 (free_rdg): Free dependence relations.
5870 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5871 inventing extra dependences.
5872 (distribute_loop): Update for RDG API changes.
5874 2013-09-10 Kai Tietz <ktietz@redhat.com>
5876 * doc/invoke.texi (fms-extensions): Document changed
5877 behavior for ms-abi targets.
5878 * config/i386/i386.c (ix86_option_override_internal):
5879 Set default value of option -fms-extension for ms-abi targets.
5881 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5883 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5885 2013-09-10 Alan Modra <amodra@gmail.com>
5888 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5890 2013-09-10 Alan Modra <amodra@gmail.com>
5892 * config/rs6000/predicates.md (add_cint_operand): New.
5893 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5894 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5895 using add_cint_operand.
5896 (largetoc_high_plus_aix): Likewise.
5898 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5900 PR tree-optimization/58364
5901 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5902 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5903 the current range can't be an unconditional true or false.
5905 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5907 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5909 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5911 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5913 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5916 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5918 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5920 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5921 NO_REGS for immediate that can't be moved directly into FP_REGS.
5923 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5925 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5926 comparison with negated operand.
5927 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5930 2013-09-09 Richard Biener <rguenther@suse.de>
5933 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5934 of a subloop record all its block as affecting loop-closed SSA form.
5936 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5938 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5939 of an rtx/bitpos pair.
5940 (store_fixed_bit_field): Update accordingly.
5942 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5944 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5946 * builtins.c (expand_errno_check): Likewise.
5947 * dwarf2cfi.c (init_return_column_size): Likewise.
5948 * except.c (sjlj_mark_call_sites): Likewise.
5949 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5950 * lra-constraints.c (emit_inc): Likewise.
5951 * ree.c (combine_set_extension): Likewise.
5952 * regmove.c (fixup_match_2): Likewise.
5953 * reload1.c (inc_for_reload): Likewise.
5955 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5957 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5958 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5959 Use gen_int_mode with the mode of the associated simplify_* call.
5960 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5961 * expmed.c (expand_shift_1): Likewise.
5962 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5963 * loop-iv.c (iv_number_of_iterations): Likewise.
5964 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5965 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5967 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5969 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5970 of the associated expand_* call.
5971 (asan_emit_stack_protection): Likewise.
5972 * builtins.c (round_trampoline_addr): Likewise.
5973 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5974 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5975 (emit_store_flag): Likewise.
5976 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5977 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5979 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5980 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5981 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5982 * modulo-sched.c (generate_prolog_epilog): Likewise.
5983 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5984 (expand_ctz, expand_ffs, expand_unop): Likewise.
5986 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5988 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5989 of the associated gen_rtx_* call.
5990 * caller-save.c (init_caller_save): Likewise.
5991 * combine.c (find_split_point, make_extraction): Likewise.
5992 (make_compound_operation): Likewise.
5993 * dwarf2out.c (mem_loc_descriptor): Likewise.
5994 * explow.c (plus_constant, probe_stack_range): Likewise.
5995 * expmed.c (expand_mult_const): Likewise.
5996 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5997 * reload1.c (init_reload): Likewise.
5998 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5999 * var-tracking.c (adjust_mems): Likewise.
6000 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6001 rather than gen_rtx_fmt_ee.
6003 2013-09-09 Jan Hubicka <jh@suse.cz>
6006 * value-prof.c (gimple_ic): Copy also abnormal edges.
6008 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6010 * asan.c (asan_shadow_cst): Use gen_int_mode.
6012 2013-09-08 Jan Hubicka <jh@suse.cz>
6014 * ipa-profile.c: Add toplevel comment.
6015 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6016 (contains_hot_call_p): New function.
6017 (ipa_propagate_frequency): Set frequencies based on counts when
6019 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6020 profile; do not tamper with profile after inlining if it is read.
6022 2013-09-08 Jan Hubicka <jh@suse.cz>
6024 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6027 2013-09-08 Jan Hubicka <jh@suse.cz>
6029 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6032 2013-09-08 Jeff Law <law@redhat.com>
6035 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6038 2013-09-08 Andi Kleen <ak@linux.intel.com>
6040 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6042 2013-09-08 Jan Hubicka <jh@suse.cz>
6044 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6046 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6048 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6049 for non-debug insns.
6050 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6051 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6052 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6055 2013-09-08 Jan Hubicka <jh@suse.cz>
6057 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6058 targets and devirtualize to BUILT_IN_UNREACHABLE.
6059 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6060 * ipa.c (walk_polymorphic_call_targets): New function.
6061 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6062 functions; use the new timevar.
6063 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6064 was removed from the program.
6065 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6066 not consider it in the walk when its vtable is dead.
6067 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6069 (devirt_variable_node_removal_hook): New function.
6070 (possible_polymorphic_call_targets): Also register
6071 devirt_variable_node_removal_hook.
6072 (ipa_devirt): Do not do non-speculative devirtualization.
6073 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6075 2013-09-08 Jan Hubicka <jh@suse.cz>
6077 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6078 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6079 varpool_remove_variable_insertion_hook): Declare.
6080 * varpool.c (varpool_node_hook_list): New structure.
6081 (first_varpool_node_removal_hook,
6082 first_varpool_variable_insertion_hook): New variables.
6083 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6084 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6085 varpool_remove_variable_insertion_hook,
6086 varpool_call_variable_insertion_hooks): New functions.
6087 (varpool_remove_node): Use it.
6089 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6092 * diagnostic.c (diagnostic_build_prefix): When s.file is
6093 "<built-in>" don't output line and column numbers.
6095 2013-09-06 Jan Hubicka <jh@suse.cz>
6097 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6098 * lto-streamer-out.c: Stream thunks, too.
6099 * lto-streamer-in.c (input_function): Pop cfun here
6100 (lto_read_body): Instead of here.
6102 2013-09-06 Caroline Tice <cmtice@google.com>
6104 * doc/install.texi: Add documentation for the --enable-vtable-verify
6105 and the --disable-libvtv configure options.
6107 2013-09-06 Jeff Law <law@redhat.com>
6109 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6110 edge implied equivalences into successor phis.
6112 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6114 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6116 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6118 * resource.c (mark_target_live_regs): Compute resources taking
6119 into account if a call is predicated or not.
6121 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6123 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6124 the compiler in the function names.
6126 2013-09-06 Jan Hubicka <jh@suse.cz>
6129 * ipa-inline.c (has_caller_p): New function.
6130 (want_inline_function_to_all_callers_p): Use it.
6131 (sum_callers, inline_to_all_callers): Break out from ...
6132 (ipa_inline): ... here.
6134 2013-09-06 Jan Hubicka <jh@suse.cz>
6136 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6137 only when AVX is enabled.
6139 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6141 * config/aarch64/aarch64.md
6142 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6143 is fpsimd_<load/store>2.
6144 (load_pair<mode>): Likewise.
6145 (store_pair<mode>): Likewise.
6147 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6149 * config/arm/types.md (type): Add "mrs" type.
6150 * config/aarch64/aarch64.md
6151 (aarch64_load_tp_hard): Make type "mrs".
6153 (load_tp_hard): Make type "mrs".
6154 * config/arm/cortex-a15.md: Update with new attributes.
6155 * config/arm/cortex-a5.md: Update with new attributes.
6156 * config/arm/cortex-a53.md: Update with new attributes.
6157 * config/arm/cortex-a7.md: Update with new attributes.
6158 * config/arm/cortex-a8.md: Update with new attributes.
6159 * config/arm/cortex-a9.md: Update with new attributes.
6160 * config/arm/cortex-m4.md: Update with new attributes.
6161 * config/arm/cortex-r4.md: Update with new attributes.
6162 * config/arm/fa526.md: Update with new attributes.
6163 * config/arm/fa606te.md: Update with new attributes.
6164 * config/arm/fa626te.md: Update with new attributes.
6165 * config/arm/fa726te.md: Update with new attributes.
6167 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6169 * config/aarch64/aarch64.md
6170 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6171 (*movtf_aarch64): Likewise.
6173 (thumb1_movdi_insn): Use "multiple" for type where more than one
6174 instruction is used for a move.
6175 (*arm32_movhf): Likewise.
6176 (*thumb_movdf_insn): Likewise.
6178 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6180 * config/arm/types.md (type): Rename fcpys to fmov.
6182 (*arm_movsi_vfp): Rename type fcpys as fmov.
6183 (*thumb2_movsi_vfp): Likewise
6184 (*movhf_vfp_neon): Likewise
6185 (*movhf_vfp): Likewise
6186 (*movsf_vfp): Likewise
6187 (*thumb2_movsf_vfp): Likewise
6188 (*movsfcc_vfp): Likewise
6189 (*thumb2_movsfcc_vfp): Likewise
6190 * config/aarch64/aarch64-simd.md
6191 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6192 * config/aarch64/aarch64.md
6193 (*movsi_aarch64): Replace type mov_reg with fmovs.
6194 (*movdi_aarch64): Likewise
6195 (*movsf_aarch64): Likewise
6196 (*movdf_aarch64): Likewise
6198 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6199 * config/arm/iwmmxt.md
6200 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6201 * config/arm/arm1020e.md: Update with new attributes.
6202 * config/arm/cortex-a15-neon.md: Update with new attributes.
6203 * config/arm/cortex-a5.md: Update with new attributes.
6204 * config/arm/cortex-a53.md: Update with new attributes.
6205 * config/arm/cortex-a7.md: Update with new attributes.
6206 * config/arm/cortex-a8-neon.md: Update with new attributes.
6207 * config/arm/cortex-a9.md: Update with new attributes.
6208 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6209 * config/arm/cortex-r4f.md: Update with new attributes.
6210 * config/arm/marvell-pj4.md: Update with new attributes.
6211 * config/arm/vfp11.md: Update with new attributes.
6213 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6215 * config/aarch64/aarch64.md
6216 (*madd<mode>): Fix type attribute.
6217 (*maddsi_uxtw): Likewise.
6218 (*msub<mode>): Likewise.
6219 (*msubsi_uxtw): Likewise.
6220 (<su_optab>maddsidi4): Likewise.
6221 (<su_optab>msubsidi4): Likewise.
6223 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6225 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6226 * config/arm/arm.md (core_cycles): Remove fdiv.
6227 * config/arm/vfp.md:
6228 (*sqrtsf2_vfp): Update for attribute changes.
6229 (*sqrtdf2_vfp): Likewise.
6230 * config/aarch64/aarch64.md:
6231 (sqrt<mode>2): Update for attribute changes.
6232 * config/arm/arm1020e.md: Update with new attributes.
6233 * config/arm/cortex-a15-neon.md: Update with new attributes.
6234 * config/arm/cortex-a5.md: Update with new attributes.
6235 * config/arm/cortex-a53.md: Update with new attributes.
6236 * config/arm/cortex-a7.md: Update with new attributes.
6237 * config/arm/cortex-a8-neon.md: Update with new attributes.
6238 * config/arm/cortex-a9.md: Update with new attributes.
6239 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6240 * config/arm/cortex-r4f.md: Update with new attributes.
6241 * config/arm/marvell-pj4.md: Update with new attributes.
6242 * config/arm/vfp11.md: Update with new attributes.
6244 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6246 * config/arm/types.md
6247 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6248 * config/aarch64/aarch64.md
6249 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6251 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6252 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6253 (float<GPI:mode><GPF:mode>2): Likewise.
6255 (*truncsisf2_vfp): Update with new attributes.
6256 (*truncsidf2_vfp): Likewise.
6257 (fixuns_truncsfsi2): Likewise.
6258 (fixuns_truncdfsi2): Likewise.
6259 (*floatsisf2_vfp): Likewise.
6260 (*floatsidf2_vfp): Likewise.
6261 (floatunssisf2): Likewise.
6262 (floatunssidf2): Likewise.
6263 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6264 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6265 * config/arm/arm1020e.md: Update with new attributes.
6266 * config/arm/cortex-a15-neon.md: Update with new attributes.
6267 * config/arm/cortex-a5.md: Update with new attributes.
6268 * config/arm/cortex-a53.md: Update with new attributes.
6269 * config/arm/cortex-a7.md: Update with new attributes.
6270 * config/arm/cortex-a8-neon.md: Update with new attributes.
6271 * config/arm/cortex-a9.md: Update with new attributes.
6272 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6273 * config/arm/cortex-r4f.md: Update with new attributes.
6274 * config/arm/marvell-pj4.md: Update with new attributes.
6275 * config/arm/vfp11.md: Update with new attributes.
6277 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6279 * config/aarch64/arm_neon.h
6280 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6281 (vqtbx<1,2,3,4><q>_s8): Likewise.
6283 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6285 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6286 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6287 (add<mode>3): Add type attribute.
6288 (add<mode>3): Likewise.
6289 (usadd<mode>3): Likewise.
6290 (ssadd<mode>3): Likewise.
6291 (sub<mode>3): Likewise.
6292 (sub<mode>3): Likewise.
6293 (ussub<mode>3): Likewise.
6294 (sssub<mode>3): Likewise.
6295 (ssmulsa3): Likewise.
6296 (usmulusa3): Likewise.
6297 (arm_usatsihi): Likewise.
6299 (*movdi_vfp): Add types for all instructions.
6300 (*movdi_vfp_cortexa8): Likewise.
6301 (*movhf_vfp_neon): Likewise.
6302 (*movhf_vfp): Likewise.
6303 (*movdf_vfp): Likewise.
6304 (*thumb2_movdf_vfp): Likewise.
6305 (*thumb2_movdfcc_vfp): Likewise.
6306 * config/arm/arm.md: Add type attribute to all insn patterns.
6307 (*thumb1_adddi3): Add type attribute.
6308 (*arm_adddi3): Likewise.
6309 (*adddi_sesidi_di): Likewise.
6310 (*adddi_zesidi_di): Likewise.
6311 (*thumb1_addsi3): Likewise.
6312 (addsi3_compare0): Likewise.
6313 (*addsi3_compare0_scratch): Likewise.
6314 (*compare_negsi_si): Likewise.
6315 (cmpsi2_addneg): Likewise.
6316 (*addsi3_carryin_<optab>): Likewise.
6317 (*addsi3_carryin_alt2_<optab>): Likewise.
6318 (*addsi3_carryin_clobercc_<optab>): Likewise.
6319 (*subsi3_carryin): Likewise.
6320 (*subsi3_carryin_const): Likewise.
6321 (*subsi3_carryin_compare): Likewise.
6322 (*subsi3_carryin_compare_const): Likewise.
6323 (*arm_subdi3): Likewise.
6324 (*thumb_subdi3): Likewise.
6325 (*subdi_di_zesidi): Likewise.
6326 (*subdi_di_sesidi): Likewise.
6327 (*subdi_zesidi_di): Likewise.
6328 (*subdi_sesidi_di): Likewise.
6329 (*subdi_zesidi_ze): Likewise.
6330 (thumb1_subsi3_insn): Likewise.
6331 (*arm_subsi3_insn): Likewise.
6332 (*anddi3_insn): Likewise.
6333 (*anddi_zesidi_di): Likewise.
6334 (*anddi_sesdi_di): Likewise.
6335 (*ne_zeroextracts): Likewise.
6336 (*ne_zeroextracts): Likewise.
6337 (*ite_ne_zeroextr): Likewise.
6338 (*ite_ne_zeroextr): Likewise.
6339 (*anddi_notdi_di): Likewise.
6340 (*anddi_notzesidi): Likewise.
6341 (*anddi_notsesidi): Likewise.
6342 (andsi_notsi_si): Likewise.
6343 (thumb1_bicsi3): Likewise.
6344 (*iordi3_insn): Likewise.
6345 (*iordi_zesidi_di): Likewise.
6346 (*iordi_sesidi_di): Likewise.
6347 (*thumb1_iorsi3_insn): Likewise.
6348 (*xordi3_insn): Likewise.
6349 (*xordi_zesidi_di): Likewise.
6350 (*xordi_sesidi_di): Likewise.
6351 (*arm_xorsi3): Likewise.
6352 (*andsi_iorsi3_no): Likewise.
6353 (*smax_0): Likewise.
6354 (*smax_m1): Likewise.
6355 (*arm_smax_insn): Likewise.
6356 (*smin_0): Likewise.
6357 (*arm_smin_insn): Likewise.
6358 (*arm_umaxsi3): Likewise.
6359 (*arm_uminsi3): Likewise.
6360 (*minmax_arithsi): Likewise.
6361 (*minmax_arithsi_): Likewise.
6362 (*satsi_<SAT:code>): Likewise.
6363 (arm_ashldi3_1bit): Likewise.
6364 (arm_ashrdi3_1bit): Likewise.
6365 (arm_lshrdi3_1bit): Likewise.
6366 (*arm_negdi2): Likewise.
6367 (*thumb1_negdi2): Likewise.
6368 (*arm_negsi2): Likewise.
6369 (*thumb1_negsi2): Likewise.
6370 (*negdi_extendsid): Likewise.
6371 (*negdi_zero_extend): Likewise.
6372 (*arm_abssi2): Likewise.
6373 (*thumb1_abssi2): Likewise.
6374 (*arm_neg_abssi2): Likewise.
6375 (*thumb1_neg_abss): Likewise.
6376 (one_cmpldi2): Likewise.
6377 (extend<mode>di2): Likewise.
6378 (*compareqi_eq0): Likewise.
6379 (*arm_extendhisi2addsi): Likewise.
6380 (*arm_movdi): Likewise.
6381 (*thumb1_movdi_insn): Likewise.
6382 (*arm_movt): Likewise.
6383 (*thumb1_movsi_insn): Likewise.
6384 (pic_add_dot_plus_four): Likewise.
6385 (pic_add_dot_plus_eight): Likewise.
6386 (tls_load_dot_plus_eight): Likewise.
6387 (*thumb1_movhi_insn): Likewise.
6388 (*thumb1_movsf_insn): Likewise.
6389 (*movdf_soft_insn): Likewise.
6390 (*thumb_movdf_insn): Likewise.
6391 (cbranchsi4_insn): Likewise.
6392 (cbranchsi4_scratch): Likewise.
6393 (*negated_cbranchsi4): Likewise.
6394 (*tbit_cbranch): Likewise.
6395 (*tlobits_cbranch): Likewise.
6396 (*tstsi3_cbranch): Likewise.
6397 (*cbranchne_decr1): Likewise.
6398 (*addsi3_cbranch): Likewise.
6399 (*addsi3_cbranch_scratch): Likewise.
6400 (*arm_cmpdi_insn): Likewise.
6401 (*arm_cmpdi_unsig): Likewise.
6402 (*arm_cmpdi_zero): Likewise.
6403 (*thumb_cmpdi_zero): Likewise.
6404 (*deleted_compare): Likewise.
6405 (*mov_scc): Likewise.
6406 (*mov_negscc): Likewise.
6407 (*mov_notscc): Likewise.
6408 (*cstoresi_eq0_thumb1_insn): Likewise.
6409 (cstoresi_nltu_thumb1): Likewise.
6410 (cstoresi_ltu_thu): Likewise.
6411 (thumb1_addsi3_addgeu): Likewise.
6412 (*arm_jump): Likewise.
6413 (*thumb_jump): Likewise.
6414 (*check_arch2): Likewise.
6415 (arm_casesi_internal): Likewise.
6416 (thumb1_casesi_dispatch): Likewise.
6417 (*arm_indirect_jump): Likewise.
6418 (*thumb1_indirect_jump): Likewise.
6420 (*and_scc): Likewise.
6421 (*ior_scc): Likewise.
6422 (*compare_scc): Likewise.
6423 (*cond_move): Likewise.
6424 (*cond_arith): Likewise.
6425 (*cond_sub): Likewise.
6426 (*cmp_ite0): Likewise.
6427 (*cmp_ite1): Likewise.
6428 (*cmp_and): Likewise.
6429 (*cmp_ior): Likewise.
6430 (*ior_scc_scc): Likewise.
6431 (*ior_scc_scc_cmp): Likewise.
6432 (*and_scc_scc): Likewise.
6433 (*and_scc_scc_cmp): Likewise.
6434 (*and_scc_scc_nod): Likewise.
6435 (*negscc): Likewise.
6436 (movcond_addsi): Likewise.
6437 (movcond): Likewise.
6438 (*ifcompare_plus_move): Likewise.
6439 (*if_plus_move): Likewise.
6440 (*ifcompare_move_plus): Likewise.
6441 (*if_move_plus): Likewise.
6442 (*ifcompare_arith_arith): Likewise.
6443 (*if_arith_arith): Likewise.
6444 (*ifcompare_arith_move): Likewise.
6445 (*if_arith_move): Likewise.
6446 (*ifcompare_move_arith): Likewise.
6447 (*if_move_arith): Likewise.
6448 (*ifcompare_move_not): Likewise.
6449 (*if_move_not): Likewise.
6450 (*ifcompare_not_move): Likewise.
6451 (*if_not_move): Likewise.
6452 (*ifcompare_shift_move): Likewise.
6453 (*if_shift_move): Likewise.
6454 (*ifcompare_move_shift): Likewise.
6455 (*if_move_shift): Likewise.
6456 (*ifcompare_shift_shift): Likewise.
6457 (*ifcompare_not_arith): Likewise.
6458 (*ifcompare_arith_not): Likewise.
6459 (*if_arith_not): Likewise.
6460 (*ifcompare_neg_move): Likewise.
6461 (*if_neg_move): Likewise.
6462 (*ifcompare_move_neg): Likewise.
6463 (*if_move_neg): Likewise.
6464 (prologue_thumb1_interwork): Likewise.
6465 (*cond_move_not): Likewise.
6466 (*sign_extract_onebit): Likewise.
6467 (*not_signextract_onebit): Likewise.
6468 (stack_tie): Likewise.
6469 (align_4): Likewise.
6470 (align_8): Likewise.
6471 (consttable_end): Likewise.
6472 (consttable_1): Likewise.
6473 (consttable_2): Likewise.
6474 (consttable_4): Likewise.
6475 (consttable_8): Likewise.
6476 (consttable_16): Likewise.
6477 (*thumb1_tablejump): Likewise.
6478 (prefetch): Likewise.
6479 (force_register_use): Likewise.
6480 (thumb_eh_return): Likewise.
6481 (load_tp_hard): Likewise.
6482 (load_tp_soft): Likewise.
6483 (tlscall): Likewise.
6484 (*arm_movtas_ze): Likewise.
6485 (*arm_rev): Likewise.
6486 (*arm_revsh): Likewise.
6487 (*arm_rev16): Likewise.
6488 * config/arm/thumb2.md
6489 (*thumb2_smaxsi3): Likewise.
6490 (*thumb2_sminsi3): Likewise.
6491 (*thumb32_umaxsi3): Likewise.
6492 (*thumb2_uminsi3): Likewise.
6493 (*thumb2_negdi2): Likewise.
6494 (*thumb2_abssi2): Likewise.
6495 (*thumb2_neg_abss): Likewise.
6496 (*thumb2_movsi_insn): Likewise.
6497 (tls_load_dot_plus_four): Likewise.
6498 (*thumb2_movhi_insn): Likewise.
6499 (*thumb2_mov_scc): Likewise.
6500 (*thumb2_mov_negs): Likewise.
6501 (*thumb2_mov_negs): Likewise.
6502 (*thumb2_mov_nots): Likewise.
6503 (*thumb2_mov_nots): Likewise.
6504 (*thumb2_movsicc_): Likewise.
6505 (*thumb2_movsfcc_soft_insn): Likewise.
6506 (*thumb2_indirect_jump): Likewise.
6507 (*thumb2_and_scc): Likewise.
6508 (*thumb2_ior_scc): Likewise.
6509 (*thumb2_ior_scc_strict_it): Likewise.
6510 (*thumb2_cond_move): Likewise.
6511 (*thumb2_cond_arith): Likewise.
6512 (*thumb2_cond_ari): Likewise.
6513 (*thumb2_cond_sub): Likewise.
6514 (*thumb2_negscc): Likewise.
6515 (*thumb2_movcond): Likewise.
6516 (thumb2_casesi_internal): Likewise.
6517 (thumb2_casesi_internal_pic): Likewise.
6518 (*thumb2_alusi3_short): Likewise.
6519 (*thumb2_mov<mode>_shortim): Likewise.
6520 (*thumb2_addsi_short): Likewise.
6521 (*thumb2_subsi_short): Likewise.
6522 (thumb2_addsi3_compare0): Likewise.
6523 (*thumb2_cbz): Likewise.
6524 (*thumb2_cbnz): Likewise.
6525 (*thumb2_one_cmplsi2_short): Likewise.
6526 (*thumb2_negsi2_short): Likewise.
6527 (*orsi_notsi_si): Likewise.
6528 * config/arm/arm1020e.md: Update with new attributes.
6529 * config/arm/arm1026ejs.md: Update with new attributes.
6530 * config/arm/arm1136jfs.md: Update with new attributes.
6531 * config/arm/arm926ejs.md: Update with new attributes.
6532 * config/arm/cortex-a15.md: Update with new attributes.
6533 * config/arm/cortex-a5.md: Update with new attributes.
6534 * config/arm/cortex-a53.md: Update with new attributes.
6535 * config/arm/cortex-a7.md: Update with new attributes.
6536 * config/arm/cortex-a8.md: Update with new attributes.
6537 * config/arm/cortex-a9.md: Update with new attributes.
6538 * config/arm/cortex-m4.md: Update with new attributes.
6539 * config/arm/cortex-r4.md: Update with new attributes.
6540 * config/arm/fa526.md: Update with new attributes.
6541 * config/arm/fa606te.md: Update with new attributes.
6542 * config/arm/fa626te.md: Update with new attributes.
6543 * config/arm/fa726te.md: Update with new attributes.
6545 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6547 * config/aarch64/aarch64-simd.md
6548 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6549 <vwx> iterator to ensure correct register choice.
6550 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6551 (aarch64_sqdmull_n<mode>): Likewise.
6552 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6553 * config/aarch64/arm_neon.h
6554 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6555 (vml<as><q>_n_<su>16): Likewise.
6556 (vml<as>l_high_lane<q>_<su>16): Likewise.
6557 (vml<as>l_high_n_<su>16): Likewise.
6558 (vml<as>l_lane<q>_<su>16): Likewise.
6559 (vml<as>l_n_<su>16): Likewise.
6560 (vmul<q>_lane<q>_<su>16): Likewise.
6561 (vmul<q>_n_<su>16): Likewise.
6562 (vmull_lane<q>_<su>16): Likewise.
6563 (vmull_n_<su>16): Likewise.
6564 (vmull_high_lane<q>_<su>16): Likewise.
6565 (vmull_high_n_<su>16): Likewise.
6566 (vqrdmulh<q>_n_s16): Likewise.
6568 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6570 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6571 have the correct lane parameter.
6573 2013-09-06 Richard Biener <rguenther@suse.de>
6575 * cfganal.c (control_dependences::~control_dependences):
6576 Properly free all of the vector.
6578 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6581 * config/i386/i386.c (ix86_conditional_register_usage):
6582 Proper initialize extended SSE registers.
6584 2013-09-06 Jan Hubicka <jh@suse.cz>
6586 PR tree-optimization/58311
6587 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6589 2013-09-06 Jan Hubicka <jh@suse.cz>
6591 * Makefile.in (tree-sra.o): Update dependencies.
6592 * tree-sra.c: Include ipa-utils.h
6593 (scan_function): Use recursive_call_p.
6594 (has_caller_p): New function.
6595 (cgraph_for_node_and_aliases): Count also callers of aliases.
6597 2013-09-06 Jan Hubicka <jh@suse.cz>
6600 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6601 * tree-tailcall.c: Include ipa-utils.h.
6602 (find_tail_calls): Use it.
6603 * ipa-pure-const.c (check_call): Likewise.
6604 * ipa-utils.c (recursive_call_p): New function.
6605 * ipa-utils.h (recursive_call_p): Dclare.
6606 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6607 (symtab_semantically_equivalent_p): New function.
6608 * Makefile.in (tree-tailcall.o): Update dependencies.
6610 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6612 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6615 2013-09-06 Richard Biener <rguenther@suse.de>
6617 * lto-streamer.h (lto_global_var_decls): Remove.
6618 * Makefile.in (OBJS): Remove lto-symtab.o.
6619 (lto-symtab.o): Remove.
6620 (GTFILES): Remove lto-symtab.c
6621 * lto-symtab.c: Move to lto/
6623 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6625 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6626 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6627 (UNSPEC_FPINT_RINT): New constant definitions.
6628 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6629 definition with 2 attributes.
6630 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6631 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6634 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6636 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6637 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6638 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6639 attribute to "z196_alone".
6640 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6643 2013-09-06 Richard Biener <rguenther@suse.de>
6645 * basic-block.h (class control_dependences): New.
6646 * tree-ssa-dce.c (control_dependence_map): Remove.
6648 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6649 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6650 find_pdom, find_control_dependence, find_all_control_dependences):
6652 (mark_control_dependent_edges_necessary,
6653 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6654 tree_dce_done, perform_tree_ssa_dce): Adjust.
6655 * cfganal.c (set_control_dependence_map_bit,
6656 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6657 find_all_control_dependences): Move from tree-ssa-dce.c and
6658 implement as methods of control_dependences class.
6659 (control_dependences::control_dependences): New.
6660 (control_dependences::~control_dependences): Likewise.
6661 (control_dependences::get_edges_dependent_on): Likewise.
6662 (control_dependences::get_edge): Likewise.
6664 2013-09-04 Jan Hubicka <jh@suse.cz>
6666 * tree.c (types_same_for_odr): Drop overactive check.
6667 * ipa-devirt.c (hash_type_name): Likewise.
6669 2013-09-04 Jan Hubicka <jh@suse.cz>
6671 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6672 (analyze_functions): ... here.
6674 2013-09-04 Jan Hubicka <jh@suse.cz>
6677 * cgraphunit.c (analyze_functions): Clear AUX fields
6678 after processing; initialize assembler name has.
6680 2013-09-05 Jeff Law <law@redhat.com>
6682 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6683 from thread_around_empty_block. Record threading path into PATH.
6684 Recurse if threading through the initial block is successful.
6685 (thread_across_edge): Corresponding changes to slightly simplify.
6687 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6689 * config/aarch64/aarch64.md
6690 (type): Remove frecpe, frecps, frecpx.
6691 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6692 fix to be a TARGET_SIMD instruction.
6693 (aarch64_frecps): Remove.
6694 * config/aarch64/aarch64-simd.md
6695 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6696 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6698 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6699 Sofiane Naci <sofiane.naci@arm.com>
6701 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6702 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6703 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6704 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6705 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6706 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6707 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6708 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6709 into "clz, "rbit". Rename "shift" to "shift_imm".
6710 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6711 changes. Update for attribute changes all occurrences of arlo_* and
6713 * config/arm/arm-fixed.md: Update for attribute changes
6714 all occurrences of arlo_* types.
6715 * config/arm/thumb2.md: Update for attribute changes all occurrences
6717 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6718 (cortexa7_older_only): Likewise.
6719 (cortexa7_younger): Likewise.
6720 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6721 (1020alu_shift_op): Likewise.
6722 (1020alu_shift_reg_op): Likewise.
6723 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6724 (alu_shift_op): Likewise.
6725 (alu_shift_reg_op): Likewise.
6726 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6727 (11_alu_shift_op): Likewise.
6728 (11_alu_shift_reg_op): Likewise.
6729 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6730 (9_alu_shift_reg_op): Likewise.
6731 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6733 (cortex_a15_alu_shift): Likewise.
6734 (cortex_a15_alu_shift_reg): Likewise.
6735 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6737 (cortex_a5_alu_shift): Likewise.
6738 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6740 (cortex_a53_alu_shift): Likewise.
6741 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6743 (cortex_a7_alu_reg): Likewise.
6744 (cortex_a7_alu_shift): Likewise.
6745 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6747 (cortex_a8_alu_shift): Likewise.
6748 (cortex_a8_alu_shift_reg): Likewise.
6749 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6750 (cortex_a9_dp_shift): Likewise.
6751 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6753 * config/arm/cortex-r4.md
6754 (cortex_r4_alu): Update for attribute changes.
6755 (cortex_r4_mov): Likewise.
6756 (cortex_r4_alu_shift_reg): Likewise.
6757 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6758 (526_alu_shift_op): Likewise.
6759 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6760 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6761 (626te_alu_shift_op): Likewise.
6762 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6763 (726te_alu_shift_op): Likewise.
6764 (726te_alu_shift_reg_op): Likewise.
6765 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6766 (mp626_alu_shift_op): Likewise.
6767 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6768 (pj4_alu_conds): Likewise.
6769 (pj4_shift): Likewise.
6770 (pj4_shift_conds): Likewise.
6771 (pj4_alu_shift): Likewise.
6772 (pj4_alu_shift_conds): Likewise.
6773 * config/aarch64/aarch64.md: Update for attribute change
6774 all occurrences of arlo_* and shift* types.
6776 2013-09-05 Mike Stump <mikestump@comcast.net>
6778 * tree.h: Move documentation for tree_function_decl to tree-core.h
6779 with the declaration.
6781 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6784 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6785 looking for widest mode.
6787 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6789 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6791 2013-09-05 Richard Biener <rguenther@suse.de>
6793 PR tree-optimization/58137
6794 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6795 Do not create vectors of pointers.
6796 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6797 types for the components of the vector initializer.
6798 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6799 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6801 2013-09-05 Martin Jambor <mjambor@suse.cz>
6803 * ipa-prop.c (remove_described_reference): Accept missing references,
6804 return false if that hppens, otherwise return true.
6805 (cgraph_node_for_jfunc): New function.
6806 (try_decrement_rdesc_refcount): Likewise.
6807 (try_make_edge_direct_simple_call): Use them.
6808 (ipa_edge_removal_hook): Remove references from rdescs.
6809 (ipa_edge_duplication_hook): Clone rdescs and their references
6810 when the new edge has the same caller as the old one.
6811 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6812 reference before removing any edges.
6814 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6816 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6818 * thumb2.md (thumb2_storewb_parisi): New pattern.
6820 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6822 * config/aarch64/aarch64-option-extensions.def: Add
6823 AARCH64_OPT_EXTENSION of 'crc'.
6824 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6825 (AARCH64_ISA_CRC): Ditto.
6826 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6827 description of the CRC extension.
6829 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6831 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6832 * config/rs6000/linux.h: Ditto.
6833 * alpha/linux.h: Ditto.
6834 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6835 no_c99_libc_has_function.
6836 * config/c6x/uclinux-elf.h: Ditto.
6837 * config/lm32/uclinux-elf.h: Ditto.
6838 * config/m68k/uclinux.h: Ditto.
6839 * config/moxie/uclinux.h: Ditto.
6840 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6841 (crisv32-*-linux*, cris-*-linux*): Ditto.
6842 * config/bfin/bfin.c: Include "tm_p.h".
6844 2013-09-05 Richard Biener <rguenther@suse.de>
6846 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6847 check for a definition without a basic-block.
6849 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6850 Sofiane Naci <sofiane.naci@arm.com>
6852 * config/aarch64/aarch64.md
6853 (*movti_aarch64): Rename r_2_f and f_2_r.
6854 (*movsf_aarch64): Likewise.
6855 (*movdf_aarch64): Likewise.
6856 (*movtf_aarch64): Likewise.
6857 (aarch64_movdi_<mode>low): Likewise.
6858 (aarch64_movdi_<mode>high): Likewise.
6859 (aarch64_mov<mode>high_di): Likewise.
6860 (aarch64_mov<mode>low_di): Likewise.
6861 (aarch64_movtilow_tilow): Likewise.
6862 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6863 values to config/arm/types.md
6864 (attribute "conds"): Update for attribute change.
6865 (anddi3_insn): Likewise.
6866 (iordi3_insn): Likewise.
6867 (xordi3_insn): Likewise.
6868 (one_cmpldi2): Likewise.
6869 * config/arm/types.md (type): Add Neon types.
6870 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6871 use "type" attribute.
6872 (movmisalign<mode>_neon_store): Likewise.
6873 (movmisalign<mode>_neon_load): Likewise.
6874 (vec_set<mode>_internal): Likewise.
6875 (vec_setv2di_internal): Likewise.
6876 (vec_extract<mode>): Likewise.
6877 (vec_extractv2di): Likewise.
6878 (add<mode>3_neon): Likewise.
6879 (adddi3_neon): Likewise.
6880 (sub<mode>3_neon): Likewise.
6881 (subdi3_neon): Likewise.
6882 (mul<mode>3_neon): Likewise.
6883 (mul<mode>3add<mode>_neon): Likewise.
6884 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6885 (fma<VCVTF:mode>4)): Likewise.
6886 (fma<VCVTF:mode>4_intrinsic): Likewise.
6887 (fmsub<VCVTF:mode>4)): Likewise.
6888 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6889 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6890 (ior<mode>3): Likewise.
6891 (and<mode>3): Likewise.
6892 (anddi3_neon): Likewise.
6893 (orn<mode>3_neon): Likewise.
6894 (orndi3_neon): Likewise.
6895 (bic<mode>3_neon): Likewise.
6896 (bicdi3_neon): Likewise.
6897 (xor<mode>3): Likewise.
6898 (one_cmpl<mode>2): Likewise.
6899 (abs<mode>2): Likewise.
6900 (neg<mode>2): Likewise.
6901 (umin<mode>3_neon): Likewise.
6902 (umax<mode>3_neon): Likewise.
6903 (smin<mode>3_neon): Likewise.
6904 (smax<mode>3_neon): Likewise.
6905 (vashl<mode>3): Likewise.
6906 (vashr<mode>3_imm): Likewise.
6907 (vlshr<mode>3_imm): Likewise.
6908 (ashl<mode>3_signed): Likewise.
6909 (ashl<mode>3_unsigned): Likewise.
6910 (neon_load_count): Likewise.
6911 (ashldi3_neon_noclobber): Likewise.
6912 (signed_shift_di3_neon): Likewise.
6913 (unsigned_shift_di3_neon): Likewise.
6914 (ashrdi3_neon_imm_noclobber): Likewise.
6915 (lshrdi3_neon_imm_noclobber): Likewise.
6916 (widen_ssum<mode>3): Likewise.
6917 (widen_usum<mode>3): Likewise.
6918 (quad_halves_<code>v4si): Likewise.
6919 (quad_halves_<code>v4sf): Likewise.
6920 (quad_halves_<code>v8hi): Likewise.
6921 (quad_halves_<code>v16qi): Likewise.
6922 (reduc_splus_v2di): Likewise.
6923 (neon_vpadd_internal<mode>): Likewise.
6924 (neon_vpsmin<mode>): Likewise.
6925 (neon_vpsmax<mode>): Likewise.
6926 (neon_vpumin<mode>): Likewise.
6927 (neon_vpumax<mode>): Likewise.
6928 (ss_add<mode>_neon): Likewise.
6929 (us_add<mode>_neon): Likewise.
6930 (ss_sub<mode>_neon): Likewise.
6931 (us_sub<mode>_neon): Likewise.
6932 (neon_vadd<mode>_unspec): Likewise.
6933 (neon_vaddl<mode>): Likewise.
6934 (neon_vaddw<mode>): Likewise.
6935 (neon_vhadd<mode>): Likewise.
6936 (neon_vqadd<mode>): Likewise.
6937 (neon_vaddhn<mode>): Likewise.
6938 (neon_vmul<mode>): Likewise.
6939 (neon_vmla<mode>): Likewise.
6940 (neon_vmlal<mode>): Likewise.
6941 (neon_vmls<mode>): Likewise.
6942 (neon_vmlsl<mode>): Likewise.
6943 (neon_vqdmulh<mode>): Likewise.
6944 (neon_vqdmlal<mode>): Likewise.
6945 (neon_vqdmlsl<mode>): Likewise.
6946 (neon_vmull<mode>): Likewise.
6947 (neon_vqdmull<mode>): Likewise.
6948 (neon_vsub<mode>_unspec): Likewise.
6949 (neon_vsubl<mode>): Likewise.
6950 (neon_vsubw<mode>): Likewise.
6951 (neon_vqsub<mode>): Likewise.
6952 (neon_vhsub<mode>): Likewise.
6953 (neon_vsubhn<mode>): Likewise.
6954 (neon_vceq<mode>): Likewise.
6955 (neon_vcge<mode>): Likewise.
6956 (neon_vcgeu<mode>): Likewise.
6957 (neon_vcgt<mode>): Likewise.
6958 (neon_vcgtu<mode>): Likewise.
6959 (neon_vcle<mode>): Likewise.
6960 (neon_vclt<mode>): Likewise.
6961 (neon_vcage<mode>): Likewise.
6962 (neon_vcagt<mode>): Likewise.
6963 (neon_vtst<mode>): Likewise.
6964 (neon_vabd<mode>): Likewise.
6965 (neon_vabdl<mode>): Likewise.
6966 (neon_vaba<mode>): Likewise.
6967 (neon_vabal<mode>): Likewise.
6968 (neon_vmax<mode>): Likewise.
6969 (neon_vmin<mode>): Likewise.
6970 (neon_vpaddl<mode>): Likewise.
6971 (neon_vpadal<mode>): Likewise.
6972 (neon_vpmax<mode>): Likewise.
6973 (neon_vpmin<mode>): Likewise.
6974 (neon_vrecps<mode>): Likewise.
6975 (neon_vrsqrts<mode>): Likewise.
6976 (neon_vqabs<mode>): Likewise.
6977 (neon_vqneg<mode>): Likewise.
6978 (neon_vcls<mode>): Likewise.
6979 (clz<mode>2): Likewise.
6980 (popcount<mode>2): Likewise.
6981 (neon_vrecpe): Likewise.
6982 (neon_vrsqrte): Likewise.
6983 (neon_vget_lane<mode>_sext_internal): Likewise.
6984 (neon_vget_lane<mode>_zext_internal): Likewise.
6985 (neon_vdup_n<mode>): Likewise.
6986 (neon_vdup_nv2di): Likewise.
6987 (neon_vdpu_lane<mode>_internal): Likewise.
6988 (neon_vswp<mode>): Likewise.
6989 (float<mode><V_cvtto>2): Likewise.
6990 (floatuns<mode><V_cvtto>2): Likewise.
6991 (fix_trunc<mode><V_cvtto>)2): Likewise
6992 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6993 (neon_vcvt<mode>): Likewise.
6994 (neon_vcvtv4sfv4hf): Likewise.
6995 (neon_vcvtv4hfv4sf): Likewise.
6996 (neon_vcvt_n<mode>): Likewise.
6997 (neon_vmovn<mode>): Likewise.
6998 (neon_vqmovn<mode>): Likewise.
6999 (neon_vqmovun<mode>): Likewise.
7000 (neon_vmovl<mode>): Likewise.
7001 (neon_vmul_lane<mode>): Likewise.
7002 (neon_vmull_lane<mode>): Likewise.
7003 (neon_vqdmull_lane<mode>): Likewise.
7004 (neon_vqdmulh_lane<mode>): Likewise.
7005 (neon_vmla_lane<mode>): Likewise.
7006 (neon_vmlal_lane<mode>): Likewise.
7007 (neon_vqdmlal_lane<mode>): Likewise.
7008 (neon_vmls_lane<mode>): Likewise.
7009 (neon_vmlsl_lane<mode>): Likewise.
7010 (neon_vqdmlsl_lane<mode>): Likewise.
7011 (neon_vext<mode>): Likewise.
7012 (neon_vrev64<mode>): Likewise.
7013 (neon_vrev32<mode>): Likewise.
7014 (neon_vrev16<mode>): Likewise.
7015 (neon_vbsl<mode>_internal): Likewise.
7016 (neon_vshl<mode>): Likewise.
7017 (neon_vqshl<mode>): Likewise.
7018 (neon_vshr_n<mode>): Likewise.
7019 (neon_vshrn_n<mode>): Likewise.
7020 (neon_vqshrn_n<mode>): Likewise.
7021 (neon_vqshrun_n<mode>): Likewise.
7022 (neon_vshl_n<mode>): Likewise.
7023 (neon_vqshl_n<mode>): Likewise.
7024 (neon_vqshlu_n<mode>): Likewise.
7025 (neon_vshll_n<mode>): Likewise.
7026 (neon_vsra_n<mode>): Likewise.
7027 (neon_vsri_n<mode>): Likewise.
7028 (neon_vsli_n<mode>): Likewise.
7029 (neon_vtbl1v8qi): Likewise.
7030 (neon_vtbl2v8qi): Likewise.
7031 (neon_vtbl3v8qi): Likewise.
7032 (neon_vtbl4v8qi): Likewise.
7033 (neon_vtbx1v8qi): Likewise.
7034 (neon_vtbx2v8qi): Likewise.
7035 (neon_vtbx3v8qi): Likewise.
7036 (neon_vtbx4v8qi): Likewise.
7037 (neon_vtrn<mode>_internal): Likewise.
7038 (neon_vzip<mode>_internal): Likewise.
7039 (neon_vuzp<mode>_internal): Likewise.
7040 (neon_vld1<mode>): Likewise.
7041 (neon_vld1_lane<mode>): Likewise.
7042 (neon_vld1_dup<mode>): Likewise.
7043 (neon_vld1_dupv2di): Likewise.
7044 (neon_vst1<mode>): Likewise.
7045 (neon_vst1_lane<mode>): Likewise.
7046 (neon_vld2<mode>): Likewise.
7047 (neon_vld2_lane<mode>): Likewise.
7048 (neon_vld2_dup<mode>): Likewise.
7049 (neon_vst2<mode>): Likewise.
7050 (neon_vst2_lane<mode>): Likewise.
7051 (neon_vld3<mode>): Likewise.
7052 (neon_vld3qa<mode>): Likewise.
7053 (neon_vld3qb<mode>): Likewise.
7054 (neon_vld3_lane<mode>): Likewise.
7055 (neon_vld3_dup<mode>): Likewise.
7056 (neon_vst3<mode>): Likewise.
7057 (neon_vst3qa<mode>): Likewise.
7058 (neon_vst3qb<mode>): Likewise.
7059 (neon_vst3_lane<mode>): Likewise.
7060 (neon_vld4<mode>): Likewise.
7061 (neon_vld4qa<mode>): Likewise.
7062 (neon_vld4qb<mode>): Likewise.
7063 (neon_vld4_lane<mode>): Likewise.
7064 (neon_vld4_dup<mode>): Likewise.
7065 (neon_vst4<mode>): Likewise.
7066 (neon_vst4qa<mode>): Likewise.
7067 (neon_vst4qb<mode>): Likewise.
7068 (neon_vst4_lane<mode>): Likewise.
7069 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7070 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7071 (neon_vec_<US>mult_lo_<mode>): Likewise.
7072 (neon_vec_<US>mult_hi_<mode>): Likewise.
7073 (neon_vec_<US>shiftl_<mode>): Likewise.
7074 (neon_unpack<US>_<mode>): Likewise.
7075 (neon_vec_<US>mult_<mode>): Likewise.
7076 (vec_pack_trunc_<mode>): Likewise.
7077 (neon_vec_pack_trunk_<mode>): Likewise.
7078 (neon_vabd<mode>_2): Likewise.
7079 (neon_vabd<mode>_3): Likewise.
7080 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7081 (thumb2_movsi_vfp): Likewise.
7082 (movdi_vfp): Likewise.
7083 (movdi_vfp_cortexa8): Likewise.
7084 (movhf_vfp_neon): Likewise.
7085 (movhf_vfp): Likewiwse.
7086 (movsf_vfp): Likewiwse.
7087 (thumb2_movsf_vfp): Likewiwse.
7088 (movdf_vfp): Likewise.
7089 (thumb2_movdf_vfp): Likewise.
7090 (movsfcc_vfp): Likewise.
7091 (thumb2_movsfcc_vfp): Likewise.
7092 (movdfcc_vfp): Likewise.
7093 (thumb2_movdfcc_vfp): Likewise.
7094 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7095 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7096 (v10_v2c): Likewise.
7097 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7099 (cortex_a15_neon_int_2): Likewise.
7100 (cortex_a15_neon_int_3): Likewise.
7101 (cortex_a15_neon_int_4): Likewise.
7102 (cortex_a15_neon_int_5): Likewise.
7103 (cortex_a15_neon_vqneg_vqabs): Likewise.
7104 (cortex_a15_neon_vmov): Likewise.
7105 (cortex_a15_neon_vaba): Likewise.
7106 (cortex_a15_neon_vaba_qqq): Likewise.
7107 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7108 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7109 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7110 scalar_64_32_long_scalar): Likewise.
7111 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7112 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7113 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7114 lotype_qdd_64_32_long): Likewise.
7115 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7116 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7117 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7118 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7119 (cortex_a15_neon_shift_1): Likewise.
7120 (cortex_a15_neon_shift_2): Likewise.
7121 (cortex_a15_neon_shift_3): Likewise.
7122 (cortex_a15_neon_vshl_ddd): Likewise.
7123 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7124 (cortex_a15_neon_vsra_vrsra): Likewise.
7125 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7126 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7127 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7128 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7129 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7130 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7131 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7132 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7133 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7134 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7135 (cortex_a15_neon_bp_simple): Likewise.
7136 (cortex_a15_neon_bp_2cycle): Likewise.
7137 (cortex_a15_neon_bp_3cycle): Likewise.
7138 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7139 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7140 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7141 (cortex_a15_neon_vld2_4_regs): Likewise.
7142 (cortex_a15_neon_vld3_vld4): Likewise.
7143 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7144 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7145 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7146 (cortex_a15_neon_vst3_vst4): Likewise.
7147 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7148 (cortex_a15_neon_vld3_vld4_lane" 10
7149 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7150 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7151 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7152 (cortex_a15_neon_ldm_2): Likewise.0
7153 (cortex_a15_neon_stm_2): Likewise.
7154 (cortex_a15_neon_mcr): Likewise.
7155 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7156 (cortex_a15_neon_mrc): Likewise.
7157 (cortex_a15_neon_mrrc): Likewise.
7158 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7160 (cortex_a15_alu_shift): Likewise.
7161 (cortex_a15_alu_shift_reg): Likewise.
7162 (cortex_a15_mult32): Likewise.
7163 (cortex_a15_mult64): Likewise.
7164 (cortex_a15_block): Likewise.
7165 (cortex_a15_branch): Likewise.
7166 (cortex_a15_load1): Likewise.
7167 (cortex_a15_load3): Likewise.
7168 (cortex_a15_store1): Likewise.
7169 (cortex_a15_store3): Likewise.
7170 (cortex_a15_call): Likewise.
7171 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7172 (cortex_a5_f2r): Likewise.
7173 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7175 (cortex_a53_f2r): Likewise.
7176 * config/arm/cortex-a7.md
7177 (cortex_a7_branch): Update for attribute change.
7178 (cortex_a7_call): Likewise.
7179 (cortex_a7_alu_imm): Likewise.
7180 (cortex_a7_alu_reg): Likewise.
7181 (cortex_a7_alu_shift): Likewise.
7182 (cortex_a7_mul): Likewise.
7183 (cortex_a7_load1): Likewise.
7184 (cortex_a7_store1): Likewise.
7185 (cortex_a7_load2): Likewise.
7186 (cortex_a7_store2): Likewise.
7187 (cortex_a7_load3): Likewise.
7188 (cortex_a7_store3): Likewise.
7189 (cortex_a7_load4): Likewise.
7190 (cortex_a7_store4): Likewise.
7191 (cortex_a7_fpalu): Likewise.
7192 (cortex_a7_fconst): Likewise.
7193 (cortex_a7_fpmuls): Likewise.
7194 (cortex_a7_neon_mul): Likewise.
7195 (cortex_a7_fpmacs): Likewise.
7196 (cortex_a7_neon_mla: Likewise.
7197 (cortex_a7_fpmuld: Likewise.
7198 (cortex_a7_fpmacd: Likewise.
7199 (cortex_a7_fpfmad: Likewise.
7200 (cortex_a7_fdivs: Likewise.
7201 (cortex_a7_fdivd: Likewise.
7202 (cortex_a7_r2f: Likewise.
7203 (cortex_a7_f2r: Likewise.
7204 (cortex_a7_f_flags: Likewise.
7205 (cortex_a7_f_loads: Likewise.
7206 (cortex_a7_f_loadd: Likewise.
7207 (cortex_a7_f_stores: Likewise.
7208 (cortex_a7_f_stored: Likewise.
7209 (cortex_a7_neon): Likewise.
7210 * config/arm/cortex-a8-neon.md
7211 (cortex_a8_neon_mrc): Update for attribute change.
7212 (cortex_a8_neon_mrrc): Likewise.
7213 (cortex_a8_neon_int_1): Likewise.
7214 (cortex_a8_neon_int_2): Likewise.
7215 (cortex_a8_neon_int_3): Likewise.
7216 (cortex_a8_neon_int_4): Likewise.
7217 (cortex_a8_neon_int_5): Likewise.
7218 (cortex_a8_neon_vqneg_vqabs): Likewise.
7219 (cortex_a8_neon_vmov): Likewise.
7220 (cortex_a8_neon_vaba): Likewise.
7221 (cortex_a8_neon_vaba_qqq): Likewise.
7222 (cortex_a8_neon_vsma): Likewise.
7223 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7224 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7225 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7226 long_scalar): Likewise.
7227 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7228 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7229 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7230 long_scalar_qdd_64_32_long): Likewise.
7231 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7232 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7233 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7234 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7235 (cortex_a8_neon_shift_1): Likewise.
7236 (cortex_a8_neon_shift_2): Likewise.
7237 (cortex_a8_neon_shift_3): Likewise.
7238 (cortex_a8_neon_vshl_ddd): Likewise.
7239 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7240 (cortex_a8_neon_vsra_vrsra): Likewise.
7241 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7242 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7243 (cortex_a8_neon_fp_vsum): Likewise.
7244 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7245 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7246 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7247 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7248 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7249 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7250 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7251 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7252 (cortex_a8_neon_bp_simple): Likewise.
7253 (cortex_a8_neon_bp_2cycle): Likewise.
7254 (cortex_a8_neon_bp_3cycle): Likewise.
7255 (cortex_a8_neon_ldr): Likewise.
7256 (cortex_a8_neon_str): Likewise.
7257 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7258 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7259 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7260 (cortex_a8_neon_vld2_4_regs): Likewise.
7261 (cortex_a8_neon_vld3_vld4): Likewise.
7262 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7263 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7264 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7265 (cortex_a8_neon_vst3_vst4): Likewise.
7266 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7267 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7268 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7269 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7270 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7271 (cortex_a8_neon_mcr): Likewise.
7272 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7273 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7274 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7276 (ca9_neon_mrrc): Likewise.
7277 (cortex_a9_neon_int_1): Likewise.
7278 (cortex_a9_neon_int_2): Likewise.
7279 (cortex_a9_neon_int_3): Likewise.
7280 (cortex_a9_neon_int_4): Likewise.
7281 (cortex_a9_neon_int_5): Likewise.
7282 (cortex_a9_neon_vqneg_vqabs): Likewise.
7283 (cortex_a9_neon_vmov): Likewise.
7284 (cortex_a9_neon_vaba): Likewise.
7285 (cortex_a9_neon_vaba_qqq): Likewise.
7286 (cortex_a9_neon_vsma): Likewise.
7287 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7288 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7289 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7290 long_scalar): Likewise.
7291 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7292 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7293 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7294 long_scalar_qdd_64_32_long): Likewise.
7295 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7296 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7297 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7298 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7299 (cortex_a9_neon_shift_1): Likewise.
7300 (cortex_a9_neon_shift_2): Likewise.
7301 (cortex_a9_neon_shift_3): Likewise.
7302 (cortex_a9_neon_vshl_ddd): Likewise.
7303 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7304 (cortex_a9_neon_vsra_vrsra): Likewise.
7305 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7306 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7307 (cortex_a9_neon_fp_vsum): Likewise.
7308 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7309 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7310 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7311 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7312 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7313 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7314 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7315 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7316 (cortex_a9_neon_bp_simple): Likewise.
7317 (cortex_a9_neon_bp_2cycle): Likewise.
7318 (cortex_a9_neon_bp_3cycle): Likewise.
7319 (cortex_a9_neon_ldr): Likewise.
7320 (cortex_a9_neon_str): Likewise.
7321 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7322 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7323 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7324 (cortex_a9_neon_vld2_4_regs): Likewise.
7325 (cortex_a9_neon_vld3_vld4): Likewise.
7326 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7327 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7328 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7329 (cortex_a9_neon_vst3_vst4): Likewise.
7330 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7331 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7332 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7333 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7334 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7335 (cortex_a9_neon_mcr): Likewise.
7336 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7337 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7338 (cortex_a9_fps): Likewise.
7339 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7341 (cortex_m4_fmuls): Likewise.
7342 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7344 (cortex_r4_mrc): Likewise.
7345 * config/arm/iterators.md: Update comment referring to neon_type.
7346 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7347 (iwmmxt_movsi_insn): Likewise.
7348 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7350 (pj4_core_to_vfp): Likewise.
7351 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7353 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7354 (vfp_fstore): Likewise.
7355 * doc/md.texi: Change references to neon_type to refer to type.
7357 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7359 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7361 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7363 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7366 2013-09-04 Easwaran Raman <eraman@google.com>
7369 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7370 (build_and_add_sum): Use it.
7371 (appears_later_in_bb): Simplify code.
7373 2013-09-04 Teresa Johnson <tejohnson@google.com>
7375 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7377 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7379 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7381 2013-09-04 Jan Hubicka <jh@suse.cz>
7383 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7384 * ipa-devirt.c: Include diganostic.h
7385 (odr_type_d): Add types and types_set.
7386 (hash_type_name): Work for types with vtables during LTO.
7387 (odr_hasher::remove): Fix comment; destroy types_set.
7388 (add_type_duplicate): New function,
7389 (get_odr_type): Use it.
7390 (dump_type_inheritance_graph): Dump type duplicates.
7391 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7393 * tree.c (types_same_for_odr): Give exact answers on types with
7396 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7398 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7399 explaining their differences.
7401 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7403 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7405 2013-09-03 Jeff Law <law@redhat.com>
7407 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7408 when not threading through a joiner block. Pass joiner/no joiner
7409 state to register_jump_thread.
7410 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7411 state from argument rather than implying on path length.
7412 Dump the entire jump thread path into debugging dump.
7413 * tree-flow.h (register_jump_thread): Update prototype.
7415 2013-08-29 Xinliang David Li <davidxl@google.com>
7417 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7418 Remove a trivial gcc_assert.
7420 2013-08-29 Xinliang David Li <davidxl@google.com>
7422 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7423 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7424 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7425 Delay base decl alignment adjustment.
7426 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7427 * tree-vectorizer.h: New data structure.
7428 (set_dr_misalignment): New function.
7429 (dr_misalignment): Ditto.
7430 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7431 (vectorizable_load): Ditto.
7432 (ensure_base_align): New function.
7433 (vectorize_loops): Add dbg_cnt support.
7434 (execute_vect_slp): Ditto.
7435 * dbgcnt.def: New debug counter.
7436 * Makefile: New dependency.
7438 2013-09-03 Meador Inge <meadori@codesourcery.com>
7442 2013-08-30 Meador Inge <meadori@codesourcery.com>
7444 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7446 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7448 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7449 function descriptor.
7451 2013-09-03 Richard Biener <rguenther@suse.de>
7453 * tree-affine.c (add_elt_to_tree): Fix association issue,
7454 avoid useless converts and make sure to always return a
7455 properly typed result.
7457 2013-09-03 Richard Biener <rguenther@suse.de>
7460 * fold-const.c (negate_expr_p): Fix division case.
7461 (negate_expr): Likewise.
7463 2013-09-03 Richard Biener <rguenther@suse.de>
7466 * tree-streamer-out.c: Include tm.h.
7467 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7469 2013-09-03 Jan Hubicka <jh@suse.cz>
7471 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7473 2013-09-03 Alan Modra <amodra@gmail.com>
7475 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7476 multilibs to big-endian target and vice versa.
7477 * config/rs6000/t-linux64: Use := assignment on all vars.
7478 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7479 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7480 * config/rs6000/t-linux64le: New file.
7481 * config/rs6000/t-linux64bele: New file.
7482 * config/rs6000/t-linux64lebe: New file.
7484 2013-09-02 Jan Hubicka <jh@suse.cz>
7486 * ipa-inline-transform.c (inline_transform): Do not
7487 optimize_inline_calls when not optimizing.
7489 2013-09-02 Jan Hubicka <jh@suse.cz>
7491 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7492 duplicated nodes for assembler names.
7493 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7494 hard registers from assembler name hash.
7496 2013-09-02 Jan Hubicka <jh@suse.cz>
7498 * ipa-split.c (execute_split_functions): Split externally visible
7499 functions called once.
7501 2013-09-02 Martin Jambor <mjambor@suse.cz>
7504 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7505 linked list. When finding the correct duplicate, also consider also
7506 the caller in additon to its inlined_to node.
7508 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7510 * config/aarch64/aarch64-simd-builtins.def
7511 (dup_lane_scalar): Remove.
7512 * config/aarch64/aarch64-simd.md
7513 (aarch64_simd_dup): Add 'w->w' alternative.
7514 (aarch64_dup_lane<mode>): Allow for VALL.
7515 (aarch64_dup_lane_scalar<mode>): Remove.
7516 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7517 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7518 (aarch64_get_lane_unsigned<mode>): Likewise.
7519 (aarch64_get_lane<mode>): Likewise.
7520 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7521 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7522 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7523 (VCON): Change container of V2SF.
7524 (vswap_width_name): Likewise.
7525 * config/aarch64/arm_neon.h
7526 (__aarch64_vdup_lane_any): New.
7527 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7528 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7529 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7531 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7534 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7535 the source or the destination is a single hard register.
7537 2013-09-02 Richard Biener <rguenther@suse.de>
7540 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7543 2013-09-02 Richard Biener <rguenther@suse.de>
7545 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7546 arithmetic to sizetype.
7548 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7550 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7551 Find auto-increment use both before and after candidate.
7553 2013-09-02 Marek Polacek <polacek@redhat.com>
7555 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7557 2013-09-01 Jan Hubicka <jh@suse.cz>
7559 * Makefile.in: Add ipa-profile.o
7560 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7561 * cgraph.c (struct cgraph_propagate_frequency_data,
7562 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7563 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7564 * cgraph.h (cgraph_propagate_frequency): Remove.
7565 * ipa-inline-analysis.c: Include ipa-utils.h;
7566 drop duplicated cfgloop.h.
7567 (inline_update_callee_summaries): Update.
7568 * ipa-profile.c: New file.
7569 * ipa-utils.h (ipa_propagate_frequency): Declare.
7570 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7571 data-streamer.h, value-prof.h.
7572 (symtab_remove_unreachable_nodes): Update profile.
7573 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7574 account_time_size, cmp_counts, dump_histogram,
7575 ipa_profile_generate_summary, ipa_profile_write_summary,
7576 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7577 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7578 Move to ipa-profile.c.
7580 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7582 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7584 2013-09-01 Jan Hubicka <jh@suse.cz>
7586 * common.opt (fdevirtualize-speculatively): New function.
7587 * invoke.texi (fdevirtualize-speculatively): Document.
7588 * ipa-devirt.c: Include ipa-inline.h
7589 (likely_target_p): New function.
7590 (ipa_devirt): New function.
7591 (gate_ipa_devirt): New function.
7592 (pass_data_ipa_devirt): New static var.
7593 (pass_ipa_devirt): Likewise.
7594 (make_pass_ipa_devirt): New function.
7595 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7596 (common_handle_option): Disable devirtualization when
7597 value range profiling is available.
7598 * passes.def (pass_ipa_devirt): Add.
7599 * timever.def (TV_IPA_DEVIRT): New timevar.
7600 * tree-pass.h (make_pass_ipa_devirt):
7602 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7604 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7605 include sanitize(undefined).
7607 2013-08-31 Diego Novillo <dnovillo@google.com>
7609 * Makefile.in (TREE_CORE_H): Define.
7611 (GTFILES): Add tree-core.h.
7612 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7614 * tree-core.h: New file.
7615 Move all data structures, enum, typedefs, global
7616 declarations and constants from ...
7619 2013-08-31 Jan Hubicka <jh@suse.cz>
7621 * bulitins.c (expand_builtin): Do not early exit for gcov
7622 instrumented functions.
7624 2013-08-31 Marek Polacek <polacek@redhat.com>
7626 * ubsan.c: Include tm_p.h.
7628 2013-08-31 Jan Hubicka <jh@suse.cz>
7630 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7633 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7634 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7635 * tree-cfg.c (verify_gimple_label): ... here.
7636 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7637 (ipa_merge_profiles): New function.
7638 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7639 (lto_input_function_body): Likewise.
7640 * ipa-utils.h (ipa_merge_profiles): Declare.
7641 * lto-streamer.h (lto_input_function_body): Update prototype.
7642 (emit_label_in_global_context_p): Remove.
7643 * lto-symtab.c: Include ipa-utils.h
7644 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7646 2013-08-31 Jan Hubicka <jh@suse.cz>
7648 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7650 2013-08-31 Jan Hubicka <jh@suse.cz>
7652 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7654 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7656 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7657 "cmp" RTX before signed_comparison_operator check to account
7660 2013-08-30 Jan Hubicka <jh@suse.cz>
7662 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7663 (detect_type_change_1): Rename to ...
7664 (detect_type_change): ... this one; early return on non-polymorphic
7666 (detect_type_change_ssa): Add comp_type parameter; update
7667 use of detect_type_change.
7668 (compute_complex_assign_jump_func): Add param_type parameter;
7669 update use of detect_type_change_ssa.
7670 (compute_complex_ancestor_jump_func): Likewise.
7671 (ipa_get_callee_param_type): New function.
7672 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7673 update calls to the jump function computation functions.
7675 2013-08-30 Teresa Johnson <tejohnson@google.com>
7676 Steven Bosscher <steven@gcc.gnu.org>
7678 * cfgrtl.c (fixup_new_cold_bb): New routine.
7679 (commit_edge_insertions): Invoke fixup_partitions.
7680 (find_partition_fixes): New routine.
7681 (fixup_partitions): Ditto.
7682 (verify_hot_cold_block_grouping): Update comments.
7683 (rtl_verify_edges): Invoke find_partition_fixes.
7684 (rtl_verify_bb_pointers): Update comments.
7685 (rtl_verify_bb_layout): Ditto.
7686 * basic-block.h (probably_never_executed_edge_p): Declare.
7687 (fixup_partitions): Ditto.
7688 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7689 * bb-reorder.c (sanitize_hot_paths): New function.
7690 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7692 * predict.c (probably_never_executed_edge_p): New routine.
7693 * cfg.c (check_bb_profile): Add partition insanity warnings.
7695 2013-08-30 Meador Inge <meadori@codesourcery.com>
7697 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7699 2013-08-30 Marek Polacek <polacek@redhat.com>
7701 * Makefile.in (ubsan.o): Add.
7702 (c-family/c-ubsan.o): Add.
7703 (builtins.o): Add ubsan.h dependency.
7704 * ubsan.h: New file.
7705 * ubsan.c: New file.
7706 * common.opt: Add -fsanitize=undefined option.
7707 (flag_sanitize): Add variable.
7708 (fsanitize=): Add option. Add Driver.
7709 (fsanitize=thread): Remove option.
7710 (fsanitize=address): Likewise.
7711 (static-libubsan): New option.
7712 * doc/invoke.texi: Document the new flag and -static-libubsan.
7713 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7714 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7715 * builtin-attrs.def (ATTR_COLD): Define.
7716 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7717 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7718 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7719 * flag-types.h (sanitize_code): New enum.
7720 * opts.c (common_handle_option): Parse command line arguments
7721 of -fsanitize=. Add -fsanitize=unreachable option.
7722 * varasm.c (get_variable_section): Adjust.
7723 (assemble_noswitch_variable): Likewise.
7724 (assemble_variable): Likewise.
7725 (output_constant_def_contents): Likewise.
7726 (categorize_decl_for_section): Likewise.
7727 (place_block_symbol): Likewise.
7728 (output_object_block): Likewise.
7729 * builtins.def: Likewise.
7730 * toplev.c (compile_file): Likewise.
7731 (process_options): Likewise.
7732 * cppbuiltin.c: Likewise.
7733 * tsan.c (tsan_pass): Likewise.
7734 (tsan_gate): Likewise.
7735 (tsan_gate_O0): Likewise.
7736 * cfgexpand.c (partition_stack_vars): Likewise.
7737 (expand_stack_vars): Likewise.
7738 (defer_stack_allocation): Likewise.
7739 (expand_used_vars): Likewise.
7740 * cfgcleanup.c (old_insns_match_p): Likewise.
7741 * asan.c (asan_finish_file): Likewise.
7742 (asan_instrument): Likewise.
7743 (gate_asan): Likewise.
7744 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7745 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7746 (asan_global_struct): Use pointer_sized_int_node instead
7747 calling build_nonstandard_integer_type.
7748 (initialize_sanitizer_builtins): Likewise.
7749 (asan_finish_file): Likewise.
7750 * gcc.c: Document %{%:function(args):X}.
7751 (static_spec_functions): Add sanitize.
7752 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7753 store funcval != NULL there.
7754 (do_spec_1): Adjust handle_spec_function caller.
7755 (handle_braces): Allow %:function(args) as condition.
7756 (sanitize_spec_function): New function.
7757 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7758 (LIBUBSAN_SPEC): Likewise.
7759 (LIBUBSAN_EARLY_SPEC): Likewise.
7760 (SANITIZER_SPEC): Handle libubsan.
7761 (SANITIZER_EARLY_SPEC): Likewise.
7762 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7763 instead of fsanitize=address.
7764 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7765 instead of fsanitize=address*.
7766 * builtins.c: Include ubsan.h.
7767 (fold_builtin_0): Instrument __builtin_unreachable.
7768 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7769 instead of flag_asan.
7770 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7771 (pointer_sized_int_node): Define.
7772 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7774 2013-08-30 Mike Stump <mikestump@comcast.net>
7776 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7779 2013-08-29 Jan Hubicka <jh@suse.cz>
7781 * cgraph.c (cgraph_function_body_availability): Handle weakref
7783 * passes.def: Remove pass_fixup_cfg.
7784 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7785 track when we need to remove functions.
7786 (gate_ipa_inline): Execute inlining always; add comment why.
7787 (pass_data_ipa_inline): Remove TODO_remove_functions.
7788 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7789 do not produce summaries.
7790 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7791 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7792 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7793 constant pool and vtable.
7795 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7797 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7798 New arm_neon.h's internal macros to specify 64-bit constants.
7799 Avoid using stdint.h's macros.
7801 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7803 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7804 before applying REG_P.
7806 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7808 PR tree-optimization/58277
7809 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7810 after seeing too many stmts with vdef in between dombb and current
7811 bb, invalidate everything.
7813 2013-08-30 Richard Biener <rguenther@suse.de>
7815 * fold-const.c (fold_single_bit_test): Fix overflow test.
7817 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7819 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7820 and which can live in a register, always retrieve the value on entry.
7821 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7822 passed by invisible reference specially.
7823 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7824 (vt_add_function_parameter): Correctly deal with a parameter passed by
7825 invisible reference.
7827 2013-08-30 Jan Hubicka <jh@suse.cz>
7829 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7831 2013-08-30 Richard Biener <rguenther@suse.de>
7833 PR tree-optimization/58228
7834 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7835 allow invariant loads in nested loop vectorization.
7837 2013-08-30 Richard Biener <rguenther@suse.de>
7839 PR tree-optimization/58223
7840 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7841 (has_anti_or_output_dependence): ... this and adjust to also
7842 look for output dependences.
7843 (mark_nodes_having_upstream_mem_writes): Adjust.
7844 (rdg_flag_uses): Likewise.
7846 2013-08-30 Richard Biener <rguenther@suse.de>
7848 PR tree-optimization/58010
7849 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7850 assert that we have a loop-closed PHI.
7852 2013-08-29 Jan Hubicka <jh@suse.cz>
7854 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7855 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7856 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7857 (lto_free_function_in_decl_state_for_node): New function.
7859 2013-08-29 Xinliang David Li <davidxl@google.com>
7861 * loop-unroll.c (report_unroll_peel): Minor message change.
7862 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7863 Emit alignment peeling message with default -fopt-info.
7864 (vect_loop_versioning): Emit loop version info message.
7865 * tree-vectorizer.c (vectorize_loops): Minor message change.
7866 (execute_vect_slp): Ditto.
7868 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7870 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7871 of the clone from the DECL_NAME of the original function.
7873 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7875 * passes.c (register_pass): Add overload.
7876 * tree-pass.h (register_pass): Forward declare it. Add comment.
7878 2013-08-29 Jan Hubicka <jh@suse.cz>
7880 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7881 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7882 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7883 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7884 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7885 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7886 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7887 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7889 2013-08-29 Teresa Johnson <tejohnson@google.com>
7891 * dumpfile.c (dump_loc): Output column number.
7892 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7893 * doc/invoke.texi: Document optall -fopt-info flag.
7894 * profile.c (read_profile_edge_counts): Use new dump framework.
7895 (compute_branch_probabilities): Ditto.
7896 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7897 when pass not in any opt group.
7898 * pass_manager.h (pass_manager::get_pass_profile): New method.
7899 * value-prof.c (check_counter): Use new dump framework.
7900 (check_ic_target): Ditto.
7901 * coverage.c (get_coverage_counts): Ditto.
7902 (coverage_init): Setup new dump framework.
7904 2013-08-29 Richard Biener <rguenther@suse.de>
7906 PR tree-optimization/58246
7907 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7908 handle the dominance check inside a basic-block.
7910 2013-08-29 Richard Biener <rguenther@suse.de>
7913 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7914 of default defs that appear in abnormal PHI nodes.
7916 2013-08-29 Richard Biener <rguenther@suse.de>
7918 PR tree-optimization/57685
7919 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7920 single-use operands to avoid exponential complexity.
7922 2013-08-28 Dehao Chen <dehao@google.com>
7924 * ipa-inline.c (edge_badness): Fix integer underflow.
7926 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7928 * gtm-builtins.def (_ITM_free): Declare leaf.
7930 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7933 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7934 (*tls_local_dynamic_base_64_largepic): Likewise.
7935 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7936 Remove predicate from call operand.
7937 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7938 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7940 2013-08-28 Jeff Law <law@redhat.com>
7942 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7943 checks for the number of predecessors and successors allowed.
7944 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7945 which require copying a joiner block if there is a request which
7946 is a subpath that requires no joiner block copying.
7948 2013-08-28 Jan Hubicka <jh@suse.cz>
7950 * lto-streamer-out.c (DFS_write_tree_body): Drop
7951 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7952 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7953 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7954 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7955 Do not read DECL_ERROR_ISSUED.
7956 (unpack_ts_decl_with_vis_value_fields): Do not read
7958 (lto_input_ts_binfo_tree_pointers): Do not read
7959 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7960 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7961 write DECL_ERROR_ISSUED..
7962 (pack_ts_decl_with_vis_value_fields): Do not write
7964 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7965 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7966 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7967 * tree.h (tree_decl_common): Update comment.
7968 (DECL_ERROR_ISSUED): Remove.
7970 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7973 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7975 2013-08-28 Jan Hubicka <jh@suse.cz>
7977 * builtins.def (free): Declare leaf.
7979 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7982 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7983 * configure: Regenerate.
7985 2013-08-27 Martin Jambor <mjambor@suse.cz>
7987 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7988 (ipa_ancestor_jf_data): Likewise.
7989 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7990 (ipa_get_jf_pass_through_type_preserved): New function.
7991 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7992 (ipa_get_jf_ancestor_type_preserved): New function.
7993 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7994 (ipa_get_jf_ancestor_result): Likewise.
7995 (propagate_vals_accross_pass_through): Use
7996 ipa_get_jf_pass_through_result to do all the value mappings.
7997 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7998 type_preserved flag.
7999 (ipa_set_jf_cst_copy): New function.
8000 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8001 (ipa_set_jf_arith_pass_through): Likewise.
8002 (ipa_set_ancestor_jf): Likewise.
8003 (compute_complex_assign_jump_func): Set type_preserved instead of
8005 (ipa_compute_jump_functions_for_edge): Likewise.
8006 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8007 (update_jump_functions_after_inlining): Update type_preserved.
8008 Explicitely create jump functions when combining one with pass_through.
8009 (ipa_write_jump_function): Stream the type_preserved flags.
8010 (ipa_read_jump_function): Likewise.
8012 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8013 Aldy Hernandez <aldyh@redhat.com>
8015 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8016 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8017 * function.h (struct function): Add has_force_vect_loops and
8019 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8020 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8022 * gimple.def: Update CLAUSES comments.
8023 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8024 (gimple_build_omp_for): Add argument to prototype.
8025 (gimple_omp_for_kind): New.
8026 (gimple_omp_for_set_kind): New.
8027 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8028 GOVD_DATA_SHARE_CLASS.
8029 (enum omp_region_type): Add ORT_SIMD.
8030 (gimple_add_tmp_var): Handle ORT_SIMD.
8031 (gimplify_var_or_parm_decl): Same.
8032 (is_gimple_stmt): Same.
8033 (omp_firstprivatize_variable): Same.
8034 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8035 (omp_notice_variable): Handle ORT_SIMD.
8036 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8037 (omp_check_private): Handle ORT_SIMD.
8038 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8040 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8041 Handle OMP_CLAUSE_LASTPRIVATE.
8042 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8044 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8045 (gimplify_expr): Handle OMP_SIMD.
8046 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8047 (expand_GOMP_SIMD_VF): New.
8048 (expand_GOMP_SIMD_LAST_LANE): New.
8049 * internal-fn.def (GOMP_SIMD_LANE): New.
8050 (GOMP_SIMD_VF): New.
8051 (GOMP_SIMD_LAST_LANE): New.
8052 * omp-low.c: Include target.h.
8053 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8055 (check_omp_nesting_restrictions): Same.
8057 (lower_rec_simd_input_clauses): New.
8058 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8060 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8061 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8062 (expand_omp_build_assign): New.
8063 (expand_omp_for_init_counts): New.
8064 (expand_omp_for_init_vars): New.
8065 (extract_omp_for_update_vars): New.
8066 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8067 and rewrite accordingly.
8068 (expand_omp_simd): New.
8069 (expand_omp_for): Use expand_omp_simd.
8070 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8071 (lower_omp_for): Do not lower the body.
8072 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8074 * tree-flow.h (find_omp_clause): Remove prototype.
8075 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8076 forcing vectorization of the loop, or if flag_tree_vectorize.
8077 (gate_tree_if_conversion): Similarly.
8078 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8079 gimple_build_omp_for.
8080 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8081 * tree-parloops (create_parallel_loop): Pass kind argument to
8082 gimple_build_omp_for.
8083 * tree-pretty-print.c (dump_omp_clause): Add cases for
8084 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8085 OMP_CLAUSE__SIMDUID_.
8086 (dump_generic_node): Handle OMP_SIMD.
8087 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8088 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8089 unroll OMP_SIMD loops here.
8090 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8091 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8093 (vect_analyze_data_refs): Handle simd loops.
8094 * tree-vect-loop.c (vectorizable_live_operation): Handle
8096 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8097 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8098 (vectorizable_load): Same.
8099 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8100 (struct simduid_to_vf): New.
8101 (simduid_to_vf::hash): New.
8102 (simduid_to-vf::equal): New.
8103 (struct simd_array_to_simduid): New.
8104 (simd_array_to_simduid::hash): New.
8105 (simd_array_to_simduid::equal): New.
8106 (adjust_simduid_builtins): New.
8107 (struct note_simd_array_uses_struct): New.
8108 (note_simd_array_uses_cb): New.
8109 (note_simd_array_uses): New.
8110 (vectorize_loops): Handle simd hints and adjust simd builtins
8112 * tree-vectorizer.h (struct _stmt_vec_info): Add
8113 simd_lane_access_p field.
8114 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8115 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8116 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8117 (omp_clause_code_name): Same.
8118 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8119 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8120 * tree.def (OMP_SIMD): New entry.
8121 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8122 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8123 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8124 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8125 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8126 (OMP_CLAUSE_LINEAR_STEP): New.
8127 (OMP_CLAUSE_SAFELEN_EXPR): New.
8128 (OMP_CLAUSE__SIMDUID__DECL): New.
8129 (find_omp_clause): New prototype.
8131 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8133 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8134 Haswell processor detection.
8136 2013-08-27 Christian Widmer <shadow@umbrox.de>
8139 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8140 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8141 AVX2 capable processors.
8143 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8145 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8146 of vget_low_* with implementations in terms of other intrinsics.
8148 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8151 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8152 values to -1, 0 and 1.
8154 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8156 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8157 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8158 (<optab><mode>3_insn): Remove.
8159 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8160 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8161 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8162 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8163 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8164 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8165 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8166 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8167 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8168 (ror<mode>3_insn): Likewise.
8169 * config/aarch64/predicates.md (aarch64_simd_register): New.
8171 2013-08-27 Richard Biener <rguenther@suse.de>
8173 PR tree-optimization/57521
8174 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8175 one edge is non-critical.
8176 (find_phi_replacement_condition): Make sure to use a non-critical
8177 edge. Cleanup and remove old bug workarounds.
8178 (bb_postdominates_preds): Remove.
8179 (if_convertible_loop_p_1): Do not compute post-dominators.
8180 (combine_blocks): Do not free post-dominators.
8181 (main_tree_if_conversion): Likewise.
8182 (pass_data_if_conversion): Add TODO_verify_ssa.
8184 2013-08-27 DJ Delorie <dj@redhat.com>
8186 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8188 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8190 * function.c (assign_parm_find_data_types): Set passed_mode and
8191 nominal_mode to the TYPE_MODE of nominal_type for the built
8192 pointer type in case of the struct-pass-by-reference.
8194 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8196 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8197 (UINT16_TYPE): Change default to "unsigned int".
8199 * config/avr/avr.opt (mfract-convert-truncate): New option.
8200 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8201 is set, round negative fractional integers according to n1169
8202 when converting to integer types.
8204 2013-08-26 Jan Hubicka <jh@suse.cz>
8206 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8207 methods can not be called indirectly when their address is not taken.
8209 2013-08-26 Jan Hubicka <jh@suse.cz>
8211 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8214 2013-08-26 Jan Hubicka <jh@suse.cz>
8216 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8219 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8221 * reload.c (find_valid_class): Allow classes that do not include
8222 FIRST_PSEUDO_REGISTER - 1.
8224 2013-08-26 Jan Hubicka <jh@suse.cz>
8226 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8227 fix edge count/frequency when speculation failed; fix type check
8228 for the direct call.
8230 2013-08-26 Jan Hubicka <jh@suse.cz>
8232 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8234 2013-08-26 Jan Hubicka <jh@suse.cz>
8236 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8237 to be changed by edge redirection.
8239 2013-08-26 Jan Hubicka <jh@suse.cz>
8241 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8242 formating; add sanity check.
8243 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8244 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8245 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8246 (resolve_noninline_speculation): Update callee keys, too.
8248 2013-08-26 Jan Hubicka <jh@suse.cz>
8250 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8251 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8253 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8255 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8258 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8260 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8261 (x86_64_elf_select_section): Ditto.
8262 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8264 (ix86_pass_by_reference): Ditto.
8265 (output_set_got): Ditto.
8266 (ix86_unary_operator_ok): Ditto.
8267 (ix86_expand_builtin): Ditto.
8269 2013-08-23 Jan Hubicka <jh@suse.cz>
8271 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8273 2013-08-23 Jan Hubicka <jh@suse.cz>
8275 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8276 (tree_decl_with_vis): Add FINAL field.
8278 2013-08-23 Jeff Law <law@redhat.com>
8280 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8281 the debugging dump when the expression is fully redundant.
8283 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8285 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8286 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8287 * pretty-print.c (pp_formatted_text_data): Likewise.
8288 (pp_write_text_to_stream): Likewise.
8289 (pp_write_text_as_dot_label_to_stream): Likewise.
8290 (pp_append_r): Likewise.
8291 (pp_format): Likewise.
8292 (pp_flush): Likewise.
8293 (pp_clear_output_area): Likewise.
8294 (pp_append_text): Likewise.
8295 (pp_formatted_text): Likewise.
8296 (pp_remaining_character_count_for_line): Likewise.
8297 (pp_newline): Likewise.
8298 (pp_character): Likewise.
8299 (output_buffer::~output_buffer): Define.
8300 (pretty_printer::~pretty_printer): Destruct output buffer.
8301 * pretty-print.h (output_buffer::~output_buffer): Declare.
8302 (pretty_printer::~pretty_printer): Declare virtual.
8304 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8307 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8308 HOST_WIDE_INT_M1U): New macros.
8309 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8310 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8311 unsigned -1 for lshift.
8312 * cse.c (cse_insn): Likewise.
8313 * double-int.c (rshift_double, lshift_double): Likewise.
8314 * builtins.c (fold_builtin_bitop): Likewise.
8315 * combine.c (force_to_mode): Likewise.
8316 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8317 * simplify-rtx.c (simplify_const_unary_operation,
8318 simplify_const_binary_operation): Likewise.
8319 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8320 check_va_list_escapes): Likewise.
8321 * rtlanal.c (nonzero_bits1): Likewise.
8322 * expmed.c (expand_smod_pow2): Likewise.
8323 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8325 2013-08-23 Jan Hubicka <jh@suse.cz>
8327 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8328 as having address taken.
8330 2013-08-23 Jan Hubicka <jh@suse.cz>
8332 * ipa-utils.h (method_class_type): Declare.
8333 * ipa-devirt.c (method_class_type): Export.
8335 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8336 do not walk base classes of anonymous types.
8338 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8340 PR rtl-optimization/58220
8342 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8343 handle SEQUENCE insns properly.
8345 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8347 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8349 (pp_newline_and_indent): Likewise.
8350 (pp_separate_with): Likewise.
8351 * pretty-print.c (pp_newline_and_flush): Define.
8352 (pp_newline_and_indent): Likewise.
8353 (pp_separate_with): Likewise.
8355 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8358 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8359 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8361 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8363 * config/i386/predicates.md (ext_sse_reg_operand): New.
8364 * config/i386/i386.md (*movti_internal): Use
8365 predicate to determine if EVEX is needed.
8366 (*movsi_internal): Ditto.
8367 (*movdf_internal): Ditto.
8368 (*movsf_internal): Ditto.
8369 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8371 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8373 PR tree-optimization/58209
8374 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8375 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8376 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8377 emit POINTER_PLUS_EXPR.
8378 (create_tailcall_accumulator): For pointer result type accumulate in
8381 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8383 * configure.ac: Add backslashes missing from the last change.
8384 * configure: Regenerate.
8386 2013-08-22 Jan Hubicka <jh@susue.cz>
8388 * ipa.c (function_and_variable_visibility): First remember function
8389 was global and then make it local.
8391 2013-08-22 Julian Brown <julian@codesourcery.com>
8393 * configure.ac: Add aarch64 to list of arches which use "nop" in
8395 * configure: Regenerate.
8397 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8399 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8400 gnu_libc_has_function.
8401 * config/s390/tpf.h: Likewise.
8403 2013-08-22 Jan Hubicka <jh@susue.cz>
8405 * timevar.c (validate_phases): Add cast.
8407 2013-08-22 Jan Hubicka <jh@susue.cz>
8409 * timevar.c (validate_phases): Use size_t for memory.
8410 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8412 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8414 * pretty-print.h (output_buffer::output_buffer): Declare.
8415 (pretty_printer::pretty_printer): Likewise.
8416 (pp_construct): Remove.
8417 * pretty-print.c (output_buffer::output_buffer): Define.
8418 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8419 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8420 (print_gimple_expr): Likewise.
8421 (print_gimple_seq): Likewise.
8422 (gimple_dump_bb): Likewise.
8423 * sched-vis.c (dump_value_slim): Likewise.
8424 (dump_insn_slim): Likewise.
8425 (dump_rtl_slim): Likewise.
8426 (str_pattern_slim): Likewise.
8427 * tree-mudflap.c (mf_varname_tree): Likewise.
8428 * graph.c (print_graph_cfg): Likewise.
8429 (start_graph_dump): Likewise.
8430 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8432 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8433 pretty printer initialization.
8434 * coretypes.h (diagnostic_context): Remove superflous type alias
8436 (pretty_printer): Likewise. Declare directly as a class.
8437 (pretty_print_info): Remove declaration as class.
8438 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8439 and pp_clear_output_area.
8440 (asan_add_global): Likewise.
8442 2013-08-22 Jan Hubicka <jh@suse.cz>
8444 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8445 * ipa-utils.h (update_type_inheritance_graph): Declare.
8446 (possible_polymorphic_call_target_p): Declare.
8447 (possible_polymorphic_call_target_p): New.
8448 * ipa-devirt.c: Update toplevel comments.
8449 (cached_polymorphic_call_targets): Move up.
8450 (odr_type_d): Move ID down.
8451 (polymorphic_type_binfo_p): Update comment.
8452 (odr_hasher::remove): Likewise;
8453 (get_odr_type): Set anonymous_namespace.
8454 (dump_odr_type): Dump it.
8455 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8456 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8457 (record_binfo): Add comment.
8458 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8460 (devirt_node_removal_hook): Do not iCE when cache is freed.
8461 (possible_polymorphic_call_target_p): New predicate.
8462 (update_type_inheritance_graph): New function.
8464 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8465 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8466 Sergey Lega <sergey.s.lega@intel.com>
8467 Anna Tikhonova <anna.tikhonova@intel.com>
8468 Ilya Tocar <ilya.tocar@intel.com>
8469 Andrey Turetskiy <andrey.turetskiy@intel.com>
8470 Ilya Verbin <ilya.verbin@intel.com>
8471 Kirill Yukhin <kirill.yukhin@intel.com>
8472 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8474 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8475 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8476 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8477 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8478 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8479 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8480 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8481 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8482 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8483 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8484 OPT_mavx512pf, OPT_mavx512er cases.
8485 * config/i386/constraints.md (v): New constraint.
8486 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8487 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8488 (bit_AVX512CD): New.
8489 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8490 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8491 * config/i386/i386-c.c (ix86_target_macros_internal):
8492 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8494 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8495 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8496 * config/i386/i386.c (regclass_map, dbx_register_map)
8497 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8498 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8499 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8501 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8502 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8503 -mavx512pf options. Fix formatting.
8504 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8505 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8506 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8507 -mavx512cd, -mavx512pf options.
8508 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8509 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8510 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8512 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8513 (ix86_set_reg_reg_cost): Ditto.
8514 (x86_order_regs_for_local_alloc): Ditto.
8515 (MAX_VECT_LEN): Extend to 64-byte.
8516 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8517 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8518 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8519 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8520 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8521 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8522 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8523 (SSE_REG_MODE_P): Support new modes.
8524 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8525 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8526 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8527 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8528 (REG_CLASS_CONTENTS): Add new registers.
8529 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8530 (EXT_REX_SSE_REGNO_P): New.
8531 (HI_REGISTER_NAMES): Add new registers.
8532 * config/i386/i386.md: Define constants for new registers.
8533 (mode): Add new 512-bit modes.
8534 (prefix): Support evex prefix.
8535 (isa): Support avx512f, noavx512f, fma_avx512f.
8536 (ssemodesuffix): Add new 512-bit modes.
8538 (*movxi_internal_avx512f): Ditto.
8539 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8541 (*movsi_internal): Likewise.
8542 (*movdf_internal): Likewise.
8543 (*movsf_internal): Likewise.
8544 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8545 (<code><mode>3): Likewise.
8546 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8548 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8549 with the new constraint "v".
8550 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8552 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8553 with the new constraint "v".
8554 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8555 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8556 (*<plusminus_insn><mode>3): Likewise.
8557 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8558 (*mul<mode>3): Likewise.
8559 (<sse>_vmmul<mode>3): Likewise.
8560 (<sse>_div<mode>3): Likewise.
8561 (<sse>_vmdiv<mode>3): Likewise.
8562 (<sse>_sqrt<mode>2): Likewise.
8563 (<sse>_vmsqrt<mode>2): Likewise.
8564 (*<code><mode>3_finite): Likewise.
8565 (*<code><mode>3) <smaxmin>: Likewise.
8566 (<sse>_vm<code><mode>3): Likewise.
8567 (*<code><mode>3) <any_logic>: Likewise.
8568 (*fma_fmadd_<mode>): Likewise.
8569 (*fma_fmsub_<mode>): Likewise.
8570 (*fma_fnmadd_<mode>): Likewise.
8571 (*fma_fnmsub_<mode>): Likewise.
8572 (*fma_fmaddsub_<mode>): Likewise.
8573 (*fma_fmsubadd_<mode>): Likewise.
8574 (*fmai_fmadd_<mode>): Likewise.
8575 (*fmai_fmsub_<mode>): Likewise.
8576 (*fmai_fnmadd_<mode>): Likewise.
8577 (*fmai_fnmsub_<mode>): Likewise.
8578 (sse_cvtsi2ss): Likewise.
8579 (sse_cvtsi2ssq): Likewise.
8580 (sse_cvtss2si): Likewise.
8581 (sse_cvtss2si_2): Likewise.
8582 (sse_cvtss2siq): Likewise.
8583 (sse_cvtss2siq_2): Likewise.
8584 (sse_cvttss2si): Likewise.
8585 (sse_cvtss2siq_2): Likewise.
8586 (float<sseintvecmodelower><mode>2): Likewise.
8587 (sse2_cvtsd2si_2): Likewise.
8588 (sse2_cvtsd2siq_2): Likewise.
8589 (*<plusminus_insn><mode>3): Likewise.
8590 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8591 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8592 (ashr<mode>3): Likewise.
8593 (<shift_insn><mode>3): Likewise.
8594 (avx2_<code><mode>3): Likewise.
8595 (*avx2_<code><mode>3): Likewise.
8596 (*andnot<mode>3): Likewise.
8597 (*<code><mode>3) <any_logic>: Likewise.
8598 (abs<mode>2): Likewise.
8599 (avx2_permvar<mode>): Likewise.
8600 (avx2_perm<mode>_1): Likewise.
8601 (*avx_vpermilp<mode>): Likewise.
8602 (avx_vpermilvar<mode>3): Likewise.
8603 (avx2_ashrv<mode>): Likewise.
8604 (avx2_<shift_insn>v<mode>): Likewise.
8605 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8607 * doc/rtl.texi: Document XImode.
8609 2013-08-21 Jeff Law <law@redhat.com>
8611 * tree-flow.h (register_jump_thread): Pass vector of edges
8612 instead of each important edge.
8613 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8614 thread path into a vector and pass that to register_jump_thread.
8615 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8616 passed in edge vector to the current 3-edge form.
8619 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8621 * dce.c (fini_dce): Call df_analyze again just in case
8622 delete_unmarked_insns removed anything.
8624 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8626 * reload.h (struct reg_equivs): Rename to ..
8627 (struct reg_equivs_s): .. this.
8629 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8631 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8633 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8635 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8637 2013-08-21 Jeff Law <law@redhat.com>
8639 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8640 simplify assignments too. If the RHS collapses to a singleton
8641 range, then return the value for the range.
8643 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8645 * config/i386/sse.md (V16): Rename to...
8647 (mov<mode>): Update iterator name.
8648 (*mov<mode>_internal): Ditto.
8649 (push<mode>1): Ditto.
8650 (movmisalign<mode>): Ditto.
8652 2013-08-20 Jan Hubicka <jh@suse.cz>
8655 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8656 entry for direct edges.
8657 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8659 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8661 Revert my last two changes, r201865 and r201864:
8664 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8666 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8667 instances can own GC refs.
8669 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8670 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8671 (gcc::context::gt_pch_nx): Likewise.
8672 (gcc::context::gt_pch_nx): Likewise.
8673 * ggc.h (gt_ggc_mx <T>): New.
8674 (gt_pch_nx_with_op <T>): New.
8675 (gt_pch_nx <T>): New.
8676 * passes.c (opt_pass::gt_ggc_mx): New.
8677 (opt_pass::gt_pch_nx): New.
8678 (opt_pass::gt_pch_nx_with_op): New.
8679 (pass_manager::gt_ggc_mx): New.
8680 (pass_manager::gt_pch_nx): New.
8681 (pass_manager::gt_pch_nx_with_op): New.
8682 (pass_manager::operator new): Use
8683 ggc_internal_cleared_alloc_stat rather than xcalloc.
8684 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8685 (pass_manager::gt_ggc_mx): New.
8686 (pass_manager::gt_pch_nx): New.
8687 (pass_manager::gt_pch_nx_with_op): New.
8688 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8689 (opt_pass::operator new): New.
8690 (opt_pass::gt_ggc_mx): New.
8691 (opt_pass::gt_pch_nx): New.
8692 (opt_pass::gt_pch_nx_with_op): New.
8695 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8697 * Makefile.in (GTFILES): Add context.h.
8698 * context.c (gcc::context::operator new): New.
8699 (gcc::context::gt_ggc_mx): New.
8700 (gcc::context::gt_pch_nx): New.
8701 (gcc::context::gt_pch_nx): New.
8702 * context.h (gcc::context): Add GTY((user)) marking.
8703 (gcc::context::operator new): New.
8704 (gcc::context::gt_ggc_mx): New.
8705 (gcc::context::gt_pch_nx): New.
8706 (gcc::context::gt_pch_nx): New.
8707 (g): Add GTY marking.
8708 (gt_ggc_mx (gcc::context *)): New.
8709 (gt_pch_nx (gcc::context *)): New.
8710 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8711 void *cookie)): New.
8712 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8714 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8716 * dce.c (fini_dce): Call df_analyze again just in case
8717 delete_unmarked_insns removed anything.
8719 2013-08-20 Teresa Johnson <tejohnson@google.com>
8721 PR rtl-optimizations/57451
8722 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8723 from crossing split section boundaries.
8725 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8727 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8728 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8729 with MULTLIB_DEFAULTS.
8731 2013-08-20 Nick Clifton <nickc@redhat.com>
8733 * target.def (narrow_volatile_bitfield): Note that the default
8734 value is false, not !TARGET_STRICT_ALIGN.
8735 * doc/tm.texi: Regenerate.
8737 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8739 Fix LIB_SPEC for systems without libpthread.
8741 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8742 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8744 * config/i386/linux-common.h: Likewise.
8745 * config/mips/linux-common.h: Likewise.
8747 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8749 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8752 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8754 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8755 instances can own GC refs.
8757 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8758 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8759 (gcc::context::gt_pch_nx): Likewise.
8760 (gcc::context::gt_pch_nx): Likewise.
8761 * ggc.h (gt_ggc_mx <T>): New.
8762 (gt_pch_nx_with_op <T>): New.
8763 (gt_pch_nx <T>): New.
8764 * passes.c (opt_pass::gt_ggc_mx): New.
8765 (opt_pass::gt_pch_nx): New.
8766 (opt_pass::gt_pch_nx_with_op): New.
8767 (pass_manager::gt_ggc_mx): New.
8768 (pass_manager::gt_pch_nx): New.
8769 (pass_manager::gt_pch_nx_with_op): New.
8770 (pass_manager::operator new): Use
8771 ggc_internal_cleared_alloc_stat rather than xcalloc.
8772 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8773 (pass_manager::gt_ggc_mx): New.
8774 (pass_manager::gt_pch_nx): New.
8775 (pass_manager::gt_pch_nx_with_op): New.
8776 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8777 (opt_pass::operator new): New.
8778 (opt_pass::gt_ggc_mx): New.
8779 (opt_pass::gt_pch_nx): New.
8780 (opt_pass::gt_pch_nx_with_op): New.
8782 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8784 * Makefile.in (GTFILES): Add context.h.
8785 * context.c (gcc::context::operator new): New.
8786 (gcc::context::gt_ggc_mx): New.
8787 (gcc::context::gt_pch_nx): New.
8788 (gcc::context::gt_pch_nx): New.
8789 * context.h (gcc::context): Add GTY((user)) marking.
8790 (gcc::context::operator new): New.
8791 (gcc::context::gt_ggc_mx): New.
8792 (gcc::context::gt_pch_nx): New.
8793 (gcc::context::gt_pch_nx): New.
8794 (g): Add GTY marking.
8795 (gt_ggc_mx (gcc::context *)): New.
8796 (gt_pch_nx (gcc::context *)): New.
8797 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8798 void *cookie)): New.
8799 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8801 2013-08-20 Alan Modra <amodra@gmail.com>
8804 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8805 (rs6000_emit_epilogue): Likewise.
8807 2013-08-19 Dehao Chen <dehao@google.com>
8809 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8811 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8812 Jakub Jelinek <jakub@redhat.com>
8814 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8815 (BUILT_IN_FABSD64): Likewise.
8816 (BUILT_IN_FABSD128): Likewise.
8817 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8818 (fold_builtin_1): Likewise.
8819 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8820 destination and source operands.
8821 (*abstd2_fpr): Likewise.
8822 (*nabstd2_fpr): Likewise.
8824 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8826 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8829 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8831 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8834 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8836 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8837 * builtins.c (default_libc_has_function): New.
8838 (gnu_libc_has_function): Ditto.
8839 (no_c99_libc_has_function): Ditto.
8840 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8841 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8842 (fold_builtin_sincos): Likewise.
8843 (fold_builtin_cexp): Likewise.
8844 * builtins.def (DEF_C94_BUILTIN): Likewise.
8845 (DEF_C99_BUILTIN): Likewise.
8846 (DEF_C99_C90RES_BUILTIN): Likewise.
8847 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8848 definitions to using this define.
8849 * config/darwin-protos.h (darwin_libc_has_function): New.
8850 * config/darwin.c (darwin_libc_has_function): Ditto.
8851 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8852 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8853 * config/darwin.h: Ditto.
8854 * config/elfos.h: Ditto.
8855 * config/freebsd.h: Ditto.
8856 * config/i386/cygming.h: Ditto.
8857 * config/i386/djgpp.h: Ditto.
8858 * config/i386/i386-interix.h: Ditto.
8859 * config/microblaze/microblaze.h: Ditto.
8860 * config/mmix/mmix.h: Ditto.
8861 * config/gnu-user.h: Ditto.
8862 * config/ia64/hpux.h: Ditto.
8863 * config/pa/pa-hpux.h: Ditto.
8864 * config/pdp11/pdp11.h: Ditto.
8865 * config/picochip/picochip.h: Ditto.
8866 * config/linux.h: Ditto.
8867 * config/netbsd.h: Ditto.
8868 * config/openbsd.h: Ditto.
8869 * config/rs6000/aix43.h: Ditto.
8870 * config/rs6000/aix51.h: Ditto.
8871 * config/rs6000/aix52.h: Ditto.
8872 * config/rs6000/aix53.h: Ditto.
8873 * config/rs6000/aix61.h: Ditto.
8874 * config/rs6000/darwin.h: Ditto.
8875 * config/rs6000/linux.h: Ditto.
8876 * config/rs6000/linux64.h: Ditto.
8877 * config/s390/tpf.h: Ditto.
8878 * config/sol2-10.h: Ditto.
8879 * config/sol2.h: Ditto.
8880 * config/vms/vms.h: Ditto.
8881 * config/vxworks.h: Ditto.
8882 * config/linux-android.c (linux_android_libc_has_function):
8883 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8884 * config/linux-protos.h (linux_android_libc_has_function):
8886 * config/i386/i386.c (ix86_libc_has_function): New.
8887 * config/i386/i386-protos.h
8888 (ix86_libc_has_function): New declaration.
8889 * config/i386/i386.md
8890 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8891 ("isinf<mode>2): Likewise.
8892 * convert.c (convert_to_integer): Using new target hook
8893 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8894 TARGET_C99_FUNCTIONS.
8895 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8896 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8897 * coretypes.h (function_class): New enum for different
8898 classes of functions.
8899 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8900 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8901 (TARGET_HAS_SINCOS): Likewise.
8902 (TARGET_LIBC_HAS_FUNCTION): New.
8903 * doc/tm.texi: Regenerated.
8904 * targhooks.h (default_libc_has_function): New declaration.
8905 (no_c99_libc_has_function): Ditto.
8906 (gnu_libc_has_function): Ditto.
8907 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8908 and TARGET_HAS_SINCOS.
8910 2013-08-18 Jan Hubicka <jh@suse.cz>
8912 * Makeifle-in (ipa-devirt.o): New.
8913 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8914 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8915 (analyze_functions): Look into possible targets of polymorphic call.
8916 * dumpfile.c (dump_files): Add type-inheritance dump.
8917 * dumpfile.h (TDI_inheritance): New.
8918 * ipa-devirt.c: New file.
8919 * ipa-utils.h (odr_type_d): Forward declare.
8920 (odr_type): New type.
8921 (build_type_inheritance_graph): Declare.
8922 (possible_polymorphic_call_targets): Declare and introduce inline
8923 variant when only edge is pased.
8924 (dump_possible_polymorphic_call_targets): Likewise.
8925 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8926 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8927 (types_same_for_odr): ... here.
8928 * tree.h (type_in_anonymous_namespace_p): Declare.
8930 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8932 PR tree-optimization/58006
8933 * tree-parloops.c (take_address_of): Don't ICE if get_name
8935 (eliminate_local_variables_stmt): Remove clobber stmts.
8937 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8939 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8940 error message is issued for an alias to undefined symbol.
8942 2013-08-18 Jan Hubicka <jh@suse.cz>
8944 * cgraph.c (cgraph_create_indirect_edge): Discover
8945 polymorphic calls and record basic info into indirect_info.
8946 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8947 devirtualization, ignore objc function calls.
8948 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8949 call with no parm index info.
8950 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8951 * tree.c (virtual_method_call_p): New function.
8952 * tree.h (virtual_method_call_p): Declare.
8954 2013-08-16 Jan Hubicka <jh@suse.cz>
8957 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8959 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8961 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8964 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8966 * gengtype.c (type_for_name): Add special-case support for
8967 locating types within the "gcc::" namespace.
8968 (open_base_files): Emit a "using namespace gcc" directive.
8970 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8973 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8974 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8976 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8977 array instead of each individual operand as a separate argument.
8978 (emit_fusion_gpr_load): Likewise.
8979 (expand_fusion_gpr_load): Add new function declaration.
8981 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8982 signature to have the operands passed as an array, instead of as
8983 separate arguments. Allow ZERO_EXTEND to be in the memory
8984 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8985 depend on the register live/dead flags when peepholes are run.
8986 (expand_fusion_gpr_load): New function to be called from the
8987 peephole2 pass, to change the register that addis sets to be the
8989 (emit_fusion_gpr_load): Change the calling signature to have the
8990 operands passed as an array, instead of as separate arguments.
8991 Allow ZERO_EXTEND to be in the memory address, and also
8992 SIGN_EXTEND if -mpower8-fusion-sign.
8994 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8996 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8997 adjust the register addis loads up in the peephole2 pass. Do not
8998 depend on the register live/dead state when the peephole pass is done.
9000 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9002 * gengtype.c (create_user_defined_type): Ensure that the kind
9003 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9004 declaration is seen before the GTY((user)) marking.
9006 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9009 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9011 2013-08-16 Jan Hubicka <jh@suse.cz>
9013 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9014 arugment expected_type.
9015 (gimple_fold_call): Use it.
9016 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9017 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9018 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9019 (try_make_edge_direct_virtual_call): Likewise.
9020 * tree.c (obj_type_ref_class): New.
9021 * tree.h (obj_type_ref_class): Use it.
9023 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9025 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9026 (rtl_slim_pp): Likewise.
9027 (init_rtl_slim_pretty_print): Likewise.
9028 (dump_value_slim): Don't call it. Use local pretty printer.
9029 (dump_insn_slim): Likewise.
9030 (dump_rtl_slim): Likewise.
9031 (str_pattern_slim): Likewise.
9032 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9035 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9037 PR tree-optimization/58164
9038 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9039 walk gimple_goto_dest of GIMPLE_GOTO.
9041 PR tree-optimization/58165
9042 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9043 bi_call must be the last stmt in a bb, don't split_block, instead
9044 use fallthru edge from it and give up if there is none.
9045 Release conds vector when returning early.
9047 2013-08-14 Xinliang David Li <davidxl@google.com>
9049 * config/i386/i386.c (ix86_option_override_internal):
9050 Remove unused variable and field.
9052 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9055 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9056 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9057 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9058 and Linux, correct BLKmode alignment when 128-bit alignment is
9059 required and compatibility flag is not set.
9060 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9061 for zero-size arguments when compatibility flag is not set.
9063 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9065 PR tree-optimization/58145
9066 * tree-sra.c (build_ref_for_offset): If prev_base has
9067 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9069 2013-08-14 Xinliang David Li <davidxl@google.com>
9071 * config/i386/i386.c (ix86_option_override_internal):
9072 Fix uninitialized variable error.
9074 2013-08-14 Xinliang David Li <davidxl@google.com>
9076 * config/i386/i386.opt: Define two new options.
9077 * config/i386/x86-tune.def: Add arch selector field in macros.
9078 * config/i386/i386.h: Adjust macro definition.
9079 * config/i386/i386.c (ix86_option_override_internal):
9081 (parse_mtune_ctrl_str): New function.
9082 (set_ix86_tune_features): New function.
9083 (ix86_function_specific_restore): Call the new helper function.
9085 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9087 PR rtl-optimization/57662
9088 * sel-sched.c (code_motion_process_successors): When the current insn
9089 is removed after the recursive traversal, break from the loop.
9090 Add comments and debug printouts.
9092 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9093 Alexandre Oliva <aoliva@redhat.com>
9096 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9097 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9098 there also UNSPEC_PLTOFF.
9100 2013-08-14 Marek Polacek <polacek@redhat.com>
9102 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9103 NUM_CONDITIONS bit positions.
9105 2013-08-13 Cary Coutant <ccoutant@google.com>
9107 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9108 (attr_checksum): Hash vector contents instead of pointer.
9109 (attr_checksum_ordered): Likewise.
9111 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9113 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9114 when Pmode != word_mode. Add length_address attribute.
9115 (sse3_monitor_<mode>): Merge from sse3_monitor and
9116 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9117 Pmode != word_mode. Update insn length attribute.
9118 * config/i386/i386.c (ix86_option_override_internal): Update
9119 ix86_gen_monitor selection for merged sse3_monitor insn.
9121 2013-08-13 Julian Brown <julian@codesourcery.com>
9123 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9124 perform invalid legitimization on greater-than-word-size modes for
9127 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9129 * ira.c (setup_class_translate_array): Use aclass instead of cl
9130 for classes not fully covered by allocno classes.
9132 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9134 PR tree-optimization/57661
9135 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9136 * tree-inline.c (tree_function_versioning): Initialize it.
9137 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9138 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9139 that is not being copied.
9142 * asan.c (instrument_strlen_call): Fix typo in comment.
9143 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9145 2013-08-13 Steve Ellcey <sellcey@mips.com>
9147 * config/mips/mips.md (prefetch): Use lw instead of ld on
9148 loongson in 32bit mode.
9150 2013-08-13 Nick Clifton <nickc@redhat.com>
9152 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9154 2013-08-13 Jan Hubicka <jh@suse.cz>
9156 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9157 introduced edge; fix typo in sanity check.
9158 (cgraph_resolve_speculation): Export; improve diagnostic.
9159 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9160 speculation at type mismatch.
9161 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9162 (cgraph_resolve_speculation): Declare.
9163 (symtab_can_be_discarded): New function.
9164 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9165 * ipa-inline-transform.c (speculation_removed): New global var.
9166 (clone_inlined_nodes): See if speculation can be removed.
9167 (inline_call): If speculations was removed, we growths may not match.
9168 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9169 (speculation_useful_p): New function.
9170 (resolve_noninline_speculation): New function.
9171 (inline_small_functions): Resolve useless speculations.
9172 * ipa-inline.h (speculation_useful_p): Declare
9173 * ipa.c (can_replace_by_local_alias): Simplify.
9174 (ipa_profile): Produce speculative calls in non-lto, too;
9175 add simple cost model; produce local aliases.
9177 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9179 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9182 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9184 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9186 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9188 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9189 iterator instead of X87MODEF.
9191 2013-08-12 Perez Read <netfirewall@gmail.com>
9194 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9195 operand 0 for intel asm alternative.
9196 (*movabs<mode>_2): Ditto for operand 1.
9198 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9200 * config/aarch64/arm_none.h
9201 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9203 2013-08-12 Nick Clifton <nickc@redhat.com>
9205 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9207 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9209 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9210 expand for QI/HImode operand to produce more effictive code for
9211 unsigned char(short) --> float(double) conversion.
9213 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9215 * doc/invoke.texi: Mention that -ftls-model does not force the final
9218 2013-08-12 Marek Polacek <polacek@redhat.com>
9219 Marc Glisse <marc.glisse@inria.fr>
9221 PR tree-optimization/57980
9222 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9223 when creating -1 constant.
9225 2013-08-10 Jan Hubicka <jh@suse.cz>
9227 Workaround binutils PR14342.
9228 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9229 (gimple_init_edge_profiler): Likewise.
9230 (gimple_gen_ic_func_profiler): Likewise.
9232 2013-08-09 Jan Hubicka <jh@suse.cz>
9234 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9236 2013-08-09 Xinliang David Li <davidxl@google.com>
9238 * config/i386/stringop.def: New file.
9239 * config/i386/stringop.opt: New file.
9240 * config/i386/i386-opts.h: Include stringopt.def.
9241 * config/i386/i386.opt: Include stringopt.opt.
9242 * config/i386/i386.c (ix86_option_override_internal):
9243 Override default size based stringop inline strategies with options.
9244 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9247 2013-08-09 Jan Hubicka <jh@suse.cz>
9249 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9251 2013-08-09 Jan Hubicka <jh@suse.cz>
9253 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9255 (dump_cgraph_node): Dump profile-id.
9256 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9257 and common_target_probability.
9258 * lto-cgraph.c (lto_output_edge): Stream common targets.
9259 (lto_output_node): Stream profile ids.
9260 (input_node): Stream profile ids.
9261 (input_edge): Stream common targets.
9262 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9263 * ipa.c: Include value-prof.h
9264 (ipa_profile_generate_summary): Turn indirect call statement histograms
9265 into common targets.
9266 (ipa_profile): Turn common targets into speculative edges.
9268 2013-08-09 Jan Hubicka <jh@suse.cz>
9270 * cgraph.h (cgraph_node): Add profile_id.
9271 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9272 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9273 (del_node_map): Update.
9274 (find_func_by_funcdef_no): Replace by ...
9275 (find_func_by_profile_id): ... this one.
9276 (gimple_ic_transform): Do not remove useful histograms when
9277 speculation is not done; dump info when indirect call removal
9279 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9280 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9281 (__gcov_indirect_call_profiler_v2): .. this one.
9282 * profile.h (init_node_map): Update.
9283 * coverage.c (coverage_compute_profile_id): New function.
9284 * coverage.h (coverage_compute_profile_id): Declare.
9285 * tree-profile.c (init_ic_make_global_vars): Make
9286 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9287 (gimple_init_edge_profiler): Update prototype of
9288 __gcov_indirect_call_profiler.
9289 (gimple_gen_ic_func_profiler): Simplify.
9290 (tree_profiling): Use init_node_map
9292 2013-08-09 Jan Hubicka <jh@suse.cz>
9294 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9295 non-speculative refs.
9296 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9297 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9298 (cgraph_set_call_stmt): Likewise.
9299 (cgraph_create_edge_1): Fix release checking compilatoin;
9301 (cgraph_free_edge): Free indirect info.
9302 (cgraph_turn_edge_to_speculative): New function.
9303 (cgraph_speculative_call_info): New function.
9304 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9305 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9306 (dump_cgraph_node): Dump speculation.
9307 (verify_edge_count_and_frequency): Accept speculative edges.
9308 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9309 (verify_cgraph_node): Handle speculation.
9310 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9311 (cgraph_set_call_stmt): Update prototype.
9312 (cgraph_make_edge_direct): Update prototype.
9313 (cgraph_speculative_call_info): Declare.
9314 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9315 to change; update call of ipa_find_references.
9316 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9317 lto_stmt_uid and speculative flags.
9318 (ipa_dump_references): Dump speculation.
9319 (ipa_clone_references): Clone speculative flag.
9320 (ipa_clone_referring): Likewise.
9321 (ipa_clone_ref): New function.
9322 (ipa_find_reference): Look into lto_stmt_uids
9323 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9324 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9325 (ipa_find_reference): Update declaration.
9326 (ipa_clone_ref): Declare.
9327 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9328 stream speculative flag.
9329 (lto_output_ref): Stream statements uids and speculation.
9330 (input_ref): Likewise.
9331 (input_edge): Stream speuclation.
9332 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9333 (cgraph_set_call_stmt_including_clones): Handle speculation.
9334 * ipa-inline.c (heap_edge_removal_hook): New function.
9335 (inline_small_functions): Register it.
9336 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9337 also initialize refs.
9338 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9340 (try_make_edge_direct_simple_call): Likewise.
9341 (try_make_edge_direct_simple_call): Likewise.
9342 (update_indirect_edges_after_inlining): Likewise.
9343 (remove_described_reference): Look proper lto_stmt_uid.
9344 (propagate_controlled_uses): Likewise.
9345 (propagate_controlled_uses): Liekwise.
9346 * tree-inline.c (copy_bb): Copy speculative edges.
9347 (redirect_all_calls): New function.
9348 (copy_cfg_body): Do redirection after loop info is updated.
9349 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9351 2013-08-09 Jan Hubicka <jh@suse.cz>
9353 * lto-streamer-out.c (output_function): Renumber PHIs.
9354 * lto-streamer-in.c (input_function): Likewise.
9356 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9358 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9359 (get_lane_unsigned): Likewise.
9360 (dup_lane_scalar): Likewise.
9361 (get_lane): enable for VALL.
9362 * config/aarch64/aarch64-simd.md
9363 (aarch64_dup_lane_scalar<mode>): Remove.
9364 (aarch64_get_lane_signed<mode>): Likewise.
9365 (aarch64_get_lane_unsigned<mode>): Likewise.
9366 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9367 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9368 (aarch64_get_lane<mode>): Enable for all vector modes.
9369 (aarch64_get_lanedi): Remove misleading constraints.
9370 * config/aarch64/arm_neon.h
9371 (__aarch64_vget_lane_any): Define.
9372 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9373 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9374 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9375 * config/aarch64/iterators.md (VDQQH): New.
9379 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9381 * configure.ac: Add GAS check for LEON instructions on SPARC.
9382 * configure: Regenerate.
9383 * config.in: Likewise.
9384 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9386 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9387 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9389 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9390 (AS_LEON_FLAG): New macro.
9391 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9392 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9393 Deal with LEON and LEON3 for the memory model.
9394 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9395 (atomic_compare_and_swap<mode>_1): Likewise.
9396 (*atomic_compare_and_swap<mode>_1): Likewise.
9398 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9400 * config/arm/neon.md (vcond): Fix floating-point vector
9401 comparisons against 0.
9403 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9405 * lra-constraints.c (emit_spill_move): Remove assert.
9406 (process_alt_operands): Add more debugging
9407 output. Increase reject for spilling into memory. Decrease
9408 reject for reloading scratch.
9409 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9411 2013-08-08 Steve Ellcey <sellcey@mips.com>
9413 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9414 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9415 micromips incompatible. Add nan2008.
9416 (MULTILIB_DIRNAMES): Add nan2008.
9417 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9418 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9419 and micromips incompatible. Add nan2008.
9420 (MULTILIB_DIRNAMES): Add nan2008.
9421 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9423 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9425 PR rtl-optimization/58079
9426 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9427 simplify_comparison has widened a comparison with an integer.
9429 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9431 * config/arm/neon.md (movmisalign<mode>): Disable when we
9432 don't allow unaligned accesses.
9433 (*movmisalign<mode>_neon_store): Likewise.
9434 (*movmisalign<mode>_neon_load): Likewise.
9435 (*movmisalign<mode>_neon_store): Likewise.
9436 (*movmisalign<mode>_neon_load): Likewise.
9438 2013-08-08 Jan Hubicka <jh@suse.cz>
9440 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9441 (make_pass_rebuild_cgraph_edges): Also clear references.
9442 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9443 * ipa-inline-transform.c (inline_transform): Remove all references
9445 * cgraphunit.c (expand_function): Remove all references after
9447 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9448 (ipa_find_reference): Rewrite to iterator.
9449 (remove_stmt_references): Likewise.
9450 (ipa_clear_stmts_in_references): New function.
9451 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9452 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9454 * ipa-split.c (split_function): Remove references in split function.
9456 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9459 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9460 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9462 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9465 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9466 suggested mode for the assignment isn't compatible with the
9469 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9472 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9474 2013-08-07 Xinliang David Li <davidxl@google.com>
9476 * config/i386/i386.opt: New option -mtune-ctrl=.
9477 * config/i386/x86-tune.def: New file.
9478 * config/i386/i386.h: include x86-tune.def.
9479 * config/i386/i386.c (ix86_option_override_internal):
9480 Parsing -mtune-ctrl= option and set tune features.
9482 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9485 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9486 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9489 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9491 * rtl.h (update_alignments): Declare.
9492 * final.c (grow_label_align): New function extracted from...
9493 (shorten_branches): ...here. Call it.
9494 (update_alignments): New function.
9495 * reorg.c (sibling_labels): New variable.
9496 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9497 the new label along with it onto the sibling_labels vector.
9498 (fill_simple_delay_slots): Adjust call to get_label_before.
9499 (fill_slots_from_thread): Likewise.
9500 (relax_delay_slots): Likewise.
9501 (make_return_insns): Likewise.
9502 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9504 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9506 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9507 document its semantics.
9508 (diagnostic_report_diagnostic): Adjust accordingly.
9510 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9512 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9513 (sparc_option_override): ...and port to new C++ pass API.
9514 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9516 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9518 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9520 2013-08-06 Caroline Tice <cmtice@google.com>
9522 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9523 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9524 * tree-pass.h: Add pass_vtable_verify.
9525 * varasm.c (assemble_variable): Add code to properly set the comdat
9526 section and name for the .vtable_map_vars section.
9527 (assemble_vtyv_preinit_initializer): New function.
9528 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9530 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9531 * vtable-verify.c: New file.
9532 * vtable-verify.h: New file.
9533 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9534 initialiation levels.
9535 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9536 * passes.def: Insert pass_vtable_verify.
9537 * aclocal.m4: Reorder includes.
9538 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9539 -fvtv-counts options.
9540 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9541 as appropriate, if -fvtable-verify=... is used.
9542 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9543 -fvtable-verify=... is used.
9544 * Makefile.in (OBJS): Add vtable-verify.o to list.
9545 (vtable-verify.o): Add new build rule.
9546 (GTFILES): Add vtable-verify.c to list.
9547 * common.opt (fvtable-verify=): New flag.
9548 (vtv_priority): Values for fvtable-verify= flag.
9549 (fvtv-counts): New flag.
9550 (fvtv-debug): New flag.
9551 * tree.h (save_vtable_map_decl): New extern function decl.
9553 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9555 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9556 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9557 (pass_data_rl78_devirt): ...new pass_data instance and...
9558 (make_pass_rl78_devirt): ...new function.
9559 (rl78_asm_file_start): Port pass registration to new C++ API.
9561 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9563 * coretypes.h (rtl_opt_pass): Add.
9564 (gcc::context): Add.
9565 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9566 (epiphany_init): Port to new C++ pass API.
9567 (epiphany_optimize_mode_switching): Likewise.
9568 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9569 (pass_manager::get_pass_mode_switching): New.
9570 (pass_manager::get_pass_peephole2): New.
9571 * mode-switching.c (pass_mode_switching): Add clone method.
9572 * recog.c (pass_peephole2): Add clone method.
9573 (pass_split_all_insns): Add clone method.
9575 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9577 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9578 (mips_option_override): ...here, porting to new C++ API for passes.
9580 2013-08-06 Jan Hubicka <jh@suse.cz>
9582 * cgraph.c (cgraph_get_body): New function based on lto.c
9584 * cgraph.h (cgraph_get_body): Declare.
9585 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9587 * cgraphunit.c (expand_function): Get body prior expanding.
9588 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9589 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9591 * passes.c (do_per_function_toporder): Get body.
9592 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9593 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9595 2013-08-06 Martin Jambor <mjambor@suse.cz>
9598 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9599 is not re-finalized. Rename second parameter to no_collect.
9601 2013-08-06 Martin Jambor <mjambor@suse.cz>
9604 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9605 MEM_REF has proper alignment information.
9607 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9610 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9612 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9613 rtx (*) (rtx, ...) with insn_gen_fn.
9614 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9615 insn_gen_fn::stored_funcptr. Add initializer braces.
9617 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9619 Rewrite how instances of passes are cloned to remove assumptions
9620 about their sizes (thus allowing pass subclasses to have
9621 additional data fields, albeit non-GC-managed ones at this point).
9623 * passes.c (make_pass_instance): Now that passes have clone
9624 methods, rewrite this function to eliminate XNEW and memcpy
9625 calls that used hardcoded sizes. Since this function no longer
9626 creates pass instances, rename it to...
9627 (add_pass_instance): ...this. Document the old way that passes were
9628 numbered and flagged, and rework this function to continue using it.
9629 (next_pass_1): Add an initial_pass argument for use by
9631 (position_pass): When adding multiple instances of a pass, use
9632 the pass's clone method, rather than relying on the XNEW/memcpy
9633 within the former make_pass_instance (now add_pass_instance).
9634 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9635 the initial instance of the current pass within the pass manager.
9637 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9639 This is the automated part of the conversion of passes from C
9640 structs to C++ classes.
9642 Patch autogenerated by refactor_passes.py from
9643 https://github.com/davidmalcolm/gcc-refactoring-scripts
9644 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9646 * asan.c (pass_asan): Convert from a global struct to a subclass of
9647 gimple_opt_pass along with...
9648 (pass_data_asan): ...new pass_data instance and...
9649 (make_pass_asan): ...new function.
9650 (pass_asan_O0): Convert from a global struct to a subclass of
9651 gimple_opt_pass along with...
9652 (pass_data_asan_O0): ...new pass_data instance and...
9653 (make_pass_asan_O0): ...new function.
9654 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9655 subclass of rtl_opt_pass along with...
9656 (pass_data_inc_dec): ...new pass_data instance and...
9657 (make_pass_inc_dec): ...new function.
9658 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9659 a subclass of rtl_opt_pass along with...
9660 (pass_data_reorder_blocks): ...new pass_data instance and...
9661 (make_pass_reorder_blocks): ...new function.
9662 (pass_duplicate_computed_gotos): Convert from a global struct to a
9663 subclass of rtl_opt_pass along with...
9664 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9665 (make_pass_duplicate_computed_gotos): ...new function.
9666 (pass_partition_blocks): Convert from a global struct to a subclass of
9667 rtl_opt_pass along with...
9668 (pass_data_partition_blocks): ...new pass_data instance and...
9669 (make_pass_partition_blocks): ...new function.
9670 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9671 struct to a subclass of rtl_opt_pass along with...
9672 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9674 (make_pass_branch_target_load_optimize1): ...new function.
9675 (pass_branch_target_load_optimize2): Convert from a global struct to a
9676 subclass of rtl_opt_pass along with...
9677 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9679 (make_pass_branch_target_load_optimize2): ...new function.
9680 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9681 of rtl_opt_pass along with...
9682 (pass_data_jump): ...new pass_data instance and...
9683 (make_pass_jump): ...new function.
9684 (pass_jump2): Convert from a global struct to a subclass of
9685 rtl_opt_pass along with...
9686 (pass_data_jump2): ...new pass_data instance and...
9687 (make_pass_jump2): ...new function.
9688 * cfgexpand.c (pass_expand): Convert from a global struct to a
9689 subclass of rtl_opt_pass along with...
9690 (pass_data_expand): ...new pass_data instance and...
9691 (make_pass_expand): ...new function.
9692 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9693 of rtl_opt_pass along with...
9694 (pass_data_free_cfg): ...new pass_data instance and...
9695 (make_pass_free_cfg): ...new function.
9696 (pass_into_cfg_layout_mode): Convert from a global struct to a
9697 subclass of rtl_opt_pass along with...
9698 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9699 (make_pass_into_cfg_layout_mode): ...new function.
9700 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9701 subclass of rtl_opt_pass along with...
9702 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9703 (make_pass_outof_cfg_layout_mode): ...new function.
9704 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9705 struct to a subclass of gimple_opt_pass along with...
9706 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9707 (make_pass_build_cgraph_edges): ...new function.
9708 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9709 subclass of gimple_opt_pass along with...
9710 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9711 (make_pass_rebuild_cgraph_edges): ...new function.
9712 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9713 subclass of gimple_opt_pass along with...
9714 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9716 (make_pass_remove_cgraph_callee_edges): ...new function.
9717 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9718 struct to a subclass of rtl_opt_pass along with...
9719 (pass_data_stack_adjustments): ...new pass_data instance and...
9720 (make_pass_stack_adjustments): ...new function.
9721 * combine.c (pass_combine): Convert from a global struct to a subclass
9722 of rtl_opt_pass along with...
9723 (pass_data_combine): ...new pass_data instance and...
9724 (make_pass_combine): ...new function.
9725 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9726 global struct to a subclass of rtl_opt_pass along with...
9727 (pass_data_compare_elim_after_reload): ...new pass_data instance
9729 (make_pass_compare_elim_after_reload): ...new function.
9730 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9731 of rtl_opt_pass along with...
9732 (pass_data_rtl_cprop): ...new pass_data instance and...
9733 (make_pass_rtl_cprop): ...new function.
9734 * cse.c (pass_cse): Convert from a global struct to a subclass of
9735 rtl_opt_pass along with...
9736 (pass_data_cse): ...new pass_data instance and...
9737 (make_pass_cse): ...new function.
9738 (pass_cse2): Convert from a global struct to a subclass of
9739 rtl_opt_pass along with...
9740 (pass_data_cse2): ...new pass_data instance and...
9741 (make_pass_cse2): ...new function.
9742 (pass_cse_after_global_opts): Convert from a global struct to a
9743 subclass of rtl_opt_pass along with...
9744 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9745 (make_pass_cse_after_global_opts): ...new function.
9746 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9747 of rtl_opt_pass along with...
9748 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9749 (make_pass_ud_rtl_dce): ...new function.
9750 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9751 rtl_opt_pass along with...
9752 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9753 (make_pass_fast_rtl_dce): ...new function.
9754 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9755 a subclass of rtl_opt_pass along with...
9756 (pass_data_df_initialize_opt): ...new pass_data instance and...
9757 (make_pass_df_initialize_opt): ...new function.
9758 (pass_df_initialize_no_opt): Convert from a global struct to a
9759 subclass of rtl_opt_pass along with...
9760 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9761 (make_pass_df_initialize_no_opt): ...new function.
9762 (pass_df_finish): Convert from a global struct to a subclass of
9763 rtl_opt_pass along with...
9764 (pass_data_df_finish): ...new pass_data instance and...
9765 (make_pass_df_finish): ...new function.
9766 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9767 rtl_opt_pass along with...
9768 (pass_data_rtl_dse1): ...new pass_data instance and...
9769 (make_pass_rtl_dse1): ...new function.
9770 (pass_rtl_dse2): Convert from a global struct to a subclass of
9771 rtl_opt_pass along with...
9772 (pass_data_rtl_dse2): ...new pass_data instance and...
9773 (make_pass_rtl_dse2): ...new function.
9774 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9775 subclass of rtl_opt_pass along with...
9776 (pass_data_dwarf2_frame): ...new pass_data instance and...
9777 (make_pass_dwarf2_frame): ...new function.
9778 * except.c (pass_set_nothrow_function_flags): Convert from a global
9779 struct to a subclass of rtl_opt_pass along with...
9780 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9782 (make_pass_set_nothrow_function_flags): ...new function.
9783 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9784 subclass of rtl_opt_pass along with...
9785 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9787 (make_pass_convert_to_eh_region_ranges): ...new function.
9788 * final.c (pass_compute_alignments): Convert from a global struct to a
9789 subclass of rtl_opt_pass along with...
9790 (pass_data_compute_alignments): ...new pass_data instance and...
9791 (make_pass_compute_alignments): ...new function.
9792 (pass_final): Convert from a global struct to a subclass of
9793 rtl_opt_pass along with...
9794 (pass_data_final): ...new pass_data instance and...
9795 (make_pass_final): ...new function.
9796 (pass_shorten_branches): Convert from a global struct to a subclass of
9797 rtl_opt_pass along with...
9798 (pass_data_shorten_branches): ...new pass_data instance and...
9799 (make_pass_shorten_branches): ...new function.
9800 (pass_clean_state): Convert from a global struct to a subclass of
9801 rtl_opt_pass along with...
9802 (pass_data_clean_state): ...new pass_data instance and...
9803 (make_pass_clean_state): ...new function.
9804 * function.c (pass_instantiate_virtual_regs): Convert from a global
9805 struct to a subclass of rtl_opt_pass along with...
9806 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9807 (make_pass_instantiate_virtual_regs): ...new function.
9808 (pass_leaf_regs): Convert from a global struct to a subclass of
9809 rtl_opt_pass along with...
9810 (pass_data_leaf_regs): ...new pass_data instance and...
9811 (make_pass_leaf_regs): ...new function.
9812 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9813 subclass of rtl_opt_pass along with...
9814 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9816 (make_pass_thread_prologue_and_epilogue): ...new function.
9817 (pass_match_asm_constraints): Convert from a global struct to a
9818 subclass of rtl_opt_pass along with...
9819 (pass_data_match_asm_constraints): ...new pass_data instance and...
9820 (make_pass_match_asm_constraints): ...new function.
9821 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9822 subclass of rtl_opt_pass along with...
9823 (pass_data_rtl_fwprop): ...new pass_data instance and...
9824 (make_pass_rtl_fwprop): ...new function.
9825 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9826 rtl_opt_pass along with...
9827 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9828 (make_pass_rtl_fwprop_addr): ...new function.
9829 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9830 rtl_opt_pass along with...
9831 (pass_data_rtl_pre): ...new pass_data instance and...
9832 (make_pass_rtl_pre): ...new function.
9833 (pass_rtl_hoist): Convert from a global struct to a subclass of
9834 rtl_opt_pass along with...
9835 (pass_data_rtl_hoist): ...new pass_data instance and...
9836 (make_pass_rtl_hoist): ...new function.
9837 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9838 subclass of gimple_opt_pass along with...
9839 (pass_data_lower_cf): ...new pass_data instance and...
9840 (make_pass_lower_cf): ...new function.
9841 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9842 from a global struct to a subclass of gimple_opt_pass along with...
9843 (pass_data_strength_reduction): ...new pass_data instance and...
9844 (make_pass_strength_reduction): ...new function.
9845 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9846 of rtl_opt_pass along with...
9847 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9848 (make_pass_rtl_ifcvt): ...new function.
9849 (pass_if_after_combine): Convert from a global struct to a subclass of
9850 rtl_opt_pass along with...
9851 (pass_data_if_after_combine): ...new pass_data instance and...
9852 (make_pass_if_after_combine): ...new function.
9853 (pass_if_after_reload): Convert from a global struct to a subclass of
9854 rtl_opt_pass along with...
9855 (pass_data_if_after_reload): ...new pass_data instance and...
9856 (make_pass_if_after_reload): ...new function.
9857 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9858 a subclass of rtl_opt_pass along with...
9859 (pass_data_initialize_regs): ...new pass_data instance and...
9860 (make_pass_initialize_regs): ...new function.
9861 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9862 of ipa_opt_pass_d along with...
9863 (pass_data_ipa_cp): ...new pass_data instance and...
9864 (make_pass_ipa_cp): ...new function.
9865 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9866 global struct to a subclass of gimple_opt_pass along with...
9867 (pass_data_inline_parameters): ...new pass_data instance and...
9868 (make_pass_inline_parameters): ...new function.
9869 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9870 subclass of gimple_opt_pass along with...
9871 (pass_data_early_inline): ...new pass_data instance and...
9872 (make_pass_early_inline): ...new function.
9873 (pass_ipa_inline): Convert from a global struct to a subclass of
9874 ipa_opt_pass_d along with...
9875 (pass_data_ipa_inline): ...new pass_data instance and...
9876 (make_pass_ipa_inline): ...new function.
9877 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9878 struct to a subclass of gimple_opt_pass along with...
9879 (pass_data_local_pure_const): ...new pass_data instance and...
9880 (make_pass_local_pure_const): ...new function.
9881 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9882 ipa_opt_pass_d along with...
9883 (pass_data_ipa_pure_const): ...new pass_data instance and...
9884 (make_pass_ipa_pure_const): ...new function.
9885 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9886 to a subclass of ipa_opt_pass_d along with...
9887 (pass_data_ipa_reference): ...new pass_data instance and...
9888 (make_pass_ipa_reference): ...new function.
9889 * ipa-split.c (pass_split_functions): Convert from a global struct to
9890 a subclass of gimple_opt_pass along with...
9891 (pass_data_split_functions): ...new pass_data instance and...
9892 (make_pass_split_functions): ...new function.
9893 (pass_feedback_split_functions): Convert from a global struct to a
9894 subclass of gimple_opt_pass along with...
9895 (pass_data_feedback_split_functions): ...new pass_data instance and...
9896 (make_pass_feedback_split_functions): ...new function.
9897 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9898 global struct to a subclass of simple_ipa_opt_pass along with...
9899 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9901 (make_pass_ipa_function_and_variable_visibility): ...new function.
9902 (pass_ipa_free_inline_summary): Convert from a global struct to a
9903 subclass of simple_ipa_opt_pass along with...
9904 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9905 (make_pass_ipa_free_inline_summary): ...new function.
9906 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9907 subclass of ipa_opt_pass_d along with...
9908 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9910 (make_pass_ipa_whole_program_visibility): ...new function.
9911 (pass_ipa_profile): Convert from a global struct to a subclass of
9912 ipa_opt_pass_d along with...
9913 (pass_data_ipa_profile): ...new pass_data instance and...
9914 (make_pass_ipa_profile): ...new function.
9915 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9916 ipa_opt_pass_d along with...
9917 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9918 (make_pass_ipa_cdtor_merge): ...new function.
9919 * ira.c (pass_ira): Convert from a global struct to a subclass of
9920 rtl_opt_pass along with...
9921 (pass_data_ira): ...new pass_data instance and...
9922 (make_pass_ira): ...new function.
9923 (pass_reload): Convert from a global struct to a subclass of
9924 rtl_opt_pass along with...
9925 (pass_data_reload): ...new pass_data instance and...
9926 (make_pass_reload): ...new function.
9927 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9928 subclass of rtl_opt_pass along with...
9929 (pass_data_cleanup_barriers): ...new pass_data instance and...
9930 (make_pass_cleanup_barriers): ...new function.
9931 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9932 of rtl_opt_pass along with...
9933 (pass_data_loop2): ...new pass_data instance and...
9934 (make_pass_loop2): ...new function.
9935 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9936 rtl_opt_pass along with...
9937 (pass_data_rtl_loop_init): ...new pass_data instance and...
9938 (make_pass_rtl_loop_init): ...new function.
9939 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9940 rtl_opt_pass along with...
9941 (pass_data_rtl_loop_done): ...new pass_data instance and...
9942 (make_pass_rtl_loop_done): ...new function.
9943 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9944 subclass of rtl_opt_pass along with...
9945 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9946 (make_pass_rtl_move_loop_invariants): ...new function.
9947 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9948 rtl_opt_pass along with...
9949 (pass_data_rtl_unswitch): ...new pass_data instance and...
9950 (make_pass_rtl_unswitch): ...new function.
9951 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9952 subclass of rtl_opt_pass along with...
9953 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9955 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9956 (pass_rtl_doloop): Convert from a global struct to a subclass of
9957 rtl_opt_pass along with...
9958 (pass_data_rtl_doloop): ...new pass_data instance and...
9959 (make_pass_rtl_doloop): ...new function.
9960 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9961 a subclass of rtl_opt_pass along with...
9962 (pass_data_lower_subreg): ...new pass_data instance and...
9963 (make_pass_lower_subreg): ...new function.
9964 (pass_lower_subreg2): Convert from a global struct to a subclass of
9965 rtl_opt_pass along with...
9966 (pass_data_lower_subreg2): ...new pass_data instance and...
9967 (make_pass_lower_subreg2): ...new function.
9968 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9969 struct to a subclass of ipa_opt_pass_d along with...
9970 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9971 (make_pass_ipa_lto_gimple_out): ...new function.
9972 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9973 of ipa_opt_pass_d along with...
9974 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9975 (make_pass_ipa_lto_finish_out): ...new function.
9976 * mode-switching.c (pass_mode_switching): Convert from a global struct
9977 to a subclass of rtl_opt_pass along with...
9978 (pass_data_mode_switching): ...new pass_data instance and...
9979 (make_pass_mode_switching): ...new function.
9980 * modulo-sched.c (pass_sms): Convert from a global struct to a
9981 subclass of rtl_opt_pass along with...
9982 (pass_data_sms): ...new pass_data instance and...
9983 (make_pass_sms): ...new function.
9984 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9985 subclass of gimple_opt_pass along with...
9986 (pass_data_expand_omp): ...new pass_data instance and...
9987 (make_pass_expand_omp): ...new function.
9988 (pass_lower_omp): Convert from a global struct to a subclass of
9989 gimple_opt_pass along with...
9990 (pass_data_lower_omp): ...new pass_data instance and...
9991 (make_pass_lower_omp): ...new function.
9992 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9993 of gimple_opt_pass along with...
9994 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9995 (make_pass_diagnose_omp_blocks): ...new function.
9996 * passes.c (pass_early_local_passes): Convert from a global struct to
9997 a subclass of simple_ipa_opt_pass along with...
9998 (pass_data_early_local_passes): ...new pass_data instance and...
9999 (make_pass_early_local_passes): ...new function.
10000 (pass_all_early_optimizations): Convert from a global struct to a
10001 subclass of gimple_opt_pass along with...
10002 (pass_data_all_early_optimizations): ...new pass_data instance and...
10003 (make_pass_all_early_optimizations): ...new function.
10004 (pass_all_optimizations): Convert from a global struct to a subclass
10005 of gimple_opt_pass along with...
10006 (pass_data_all_optimizations): ...new pass_data instance and...
10007 (make_pass_all_optimizations): ...new function.
10008 (pass_all_optimizations_g): Convert from a global struct to a subclass
10009 of gimple_opt_pass along with...
10010 (pass_data_all_optimizations_g): ...new pass_data instance and...
10011 (make_pass_all_optimizations_g): ...new function.
10012 (pass_rest_of_compilation): Convert from a global struct to a subclass
10013 of rtl_opt_pass along with...
10014 (pass_data_rest_of_compilation): ...new pass_data instance and...
10015 (make_pass_rest_of_compilation): ...new function.
10016 (pass_postreload): Convert from a global struct to a subclass of
10017 rtl_opt_pass along with...
10018 (pass_data_postreload): ...new pass_data instance and...
10019 (make_pass_postreload): ...new function.
10020 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10021 subclass of rtl_opt_pass along with...
10022 (pass_data_gcse2): ...new pass_data instance and...
10023 (make_pass_gcse2): ...new function.
10024 * postreload.c (pass_postreload_cse): Convert from a global struct to
10025 a subclass of rtl_opt_pass along with...
10026 (pass_data_postreload_cse): ...new pass_data instance and...
10027 (make_pass_postreload_cse): ...new function.
10028 * predict.c (pass_profile): Convert from a global struct to a subclass
10029 of gimple_opt_pass along with...
10030 (pass_data_profile): ...new pass_data instance and...
10031 (make_pass_profile): ...new function.
10032 (pass_strip_predict_hints): Convert from a global struct to a subclass
10033 of gimple_opt_pass along with...
10034 (pass_data_strip_predict_hints): ...new pass_data instance and...
10035 (make_pass_strip_predict_hints): ...new function.
10036 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10037 of rtl_opt_pass along with...
10038 (pass_data_peephole2): ...new pass_data instance and...
10039 (make_pass_peephole2): ...new function.
10040 (pass_split_all_insns): Convert from a global struct to a subclass of
10041 rtl_opt_pass along with...
10042 (pass_data_split_all_insns): ...new pass_data instance and...
10043 (make_pass_split_all_insns): ...new function.
10044 (pass_split_after_reload): Convert from a global struct to a subclass
10045 of rtl_opt_pass along with...
10046 (pass_data_split_after_reload): ...new pass_data instance and...
10047 (make_pass_split_after_reload): ...new function.
10048 (pass_split_before_regstack): Convert from a global struct to a
10049 subclass of rtl_opt_pass along with...
10050 (pass_data_split_before_regstack): ...new pass_data instance and...
10051 (make_pass_split_before_regstack): ...new function.
10052 (pass_split_before_sched2): Convert from a global struct to a subclass
10053 of rtl_opt_pass along with...
10054 (pass_data_split_before_sched2): ...new pass_data instance and...
10055 (make_pass_split_before_sched2): ...new function.
10056 (pass_split_for_shorten_branches): Convert from a global struct to a
10057 subclass of rtl_opt_pass along with...
10058 (pass_data_split_for_shorten_branches): ...new pass_data instance
10060 (make_pass_split_for_shorten_branches): ...new function.
10061 * ree.c (pass_ree): Convert from a global struct to a subclass of
10062 rtl_opt_pass along with...
10063 (pass_data_ree): ...new pass_data instance and...
10064 (make_pass_ree): ...new function.
10065 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10066 subclass of rtl_opt_pass along with...
10067 (pass_data_stack_regs): ...new pass_data instance and...
10068 (make_pass_stack_regs): ...new function.
10069 (pass_stack_regs_run): Convert from a global struct to a subclass of
10070 rtl_opt_pass along with...
10071 (pass_data_stack_regs_run): ...new pass_data instance and...
10072 (make_pass_stack_regs_run): ...new function.
10073 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10074 subclass of rtl_opt_pass along with...
10075 (pass_data_cprop_hardreg): ...new pass_data instance and...
10076 (make_pass_cprop_hardreg): ...new function.
10077 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10078 subclass of rtl_opt_pass along with...
10079 (pass_data_reginfo_init): ...new pass_data instance and...
10080 (make_pass_reginfo_init): ...new function.
10081 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10082 of rtl_opt_pass along with...
10083 (pass_data_regmove): ...new pass_data instance and...
10084 (make_pass_regmove): ...new function.
10085 * regrename.c (pass_regrename): Convert from a global struct to a
10086 subclass of rtl_opt_pass along with...
10087 (pass_data_regrename): ...new pass_data instance and...
10088 (make_pass_regrename): ...new function.
10089 * reorg.c (pass_delay_slots): Convert from a global struct to a
10090 subclass of rtl_opt_pass along with...
10091 (pass_data_delay_slots): ...new pass_data instance and...
10092 (make_pass_delay_slots): ...new function.
10093 (pass_machine_reorg): Convert from a global struct to a subclass of
10094 rtl_opt_pass along with...
10095 (pass_data_machine_reorg): ...new pass_data instance and...
10096 (make_pass_machine_reorg): ...new function.
10097 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10098 of rtl_opt_pass along with...
10099 (pass_data_sched): ...new pass_data instance and...
10100 (make_pass_sched): ...new function.
10101 (pass_sched2): Convert from a global struct to a subclass of
10102 rtl_opt_pass along with...
10103 (pass_data_sched2): ...new pass_data instance and...
10104 (make_pass_sched2): ...new function.
10105 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10106 to a subclass of rtl_opt_pass along with...
10107 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10108 (make_pass_stack_ptr_mod): ...new function.
10109 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10110 to a subclass of rtl_opt_pass along with...
10111 (pass_data_rtl_store_motion): ...new pass_data instance and...
10112 (make_pass_rtl_store_motion): ...new function.
10113 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10114 of gimple_opt_pass along with...
10115 (pass_data_tracer): ...new pass_data instance and...
10116 (make_pass_tracer): ...new function.
10117 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10118 to a subclass of gimple_opt_pass along with...
10119 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10120 (make_pass_diagnose_tm_blocks): ...new function.
10121 (pass_lower_tm): Convert from a global struct to a subclass of
10122 gimple_opt_pass along with...
10123 (pass_data_lower_tm): ...new pass_data instance and...
10124 (make_pass_lower_tm): ...new function.
10125 (pass_tm_init): Convert from a global struct to a subclass of
10126 gimple_opt_pass along with...
10127 (pass_data_tm_init): ...new pass_data instance and...
10128 (make_pass_tm_init): ...new function.
10129 (pass_tm_mark): Convert from a global struct to a subclass of
10130 gimple_opt_pass along with...
10131 (pass_data_tm_mark): ...new pass_data instance and...
10132 (make_pass_tm_mark): ...new function.
10133 (pass_tm_edges): Convert from a global struct to a subclass of
10134 gimple_opt_pass along with...
10135 (pass_data_tm_edges): ...new pass_data instance and...
10136 (make_pass_tm_edges): ...new function.
10137 (pass_tm_memopt): Convert from a global struct to a subclass of
10138 gimple_opt_pass along with...
10139 (pass_data_tm_memopt): ...new pass_data instance and...
10140 (make_pass_tm_memopt): ...new function.
10141 (pass_ipa_tm): Convert from a global struct to a subclass of
10142 simple_ipa_opt_pass along with...
10143 (pass_data_ipa_tm): ...new pass_data instance and...
10144 (make_pass_ipa_tm): ...new function.
10145 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10146 subclass of gimple_opt_pass along with...
10147 (pass_data_call_cdce): ...new pass_data instance and...
10148 (make_pass_call_cdce): ...new function.
10149 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10150 subclass of gimple_opt_pass along with...
10151 (pass_data_build_cfg): ...new pass_data instance and...
10152 (make_pass_build_cfg): ...new function.
10153 (pass_split_crit_edges): Convert from a global struct to a subclass of
10154 gimple_opt_pass along with...
10155 (pass_data_split_crit_edges): ...new pass_data instance and...
10156 (make_pass_split_crit_edges): ...new function.
10157 (pass_warn_function_return): Convert from a global struct to a
10158 subclass of gimple_opt_pass along with...
10159 (pass_data_warn_function_return): ...new pass_data instance and...
10160 (make_pass_warn_function_return): ...new function.
10161 (pass_warn_function_noreturn): Convert from a global struct to a
10162 subclass of gimple_opt_pass along with...
10163 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10164 (make_pass_warn_function_noreturn): ...new function.
10165 (pass_warn_unused_result): Convert from a global struct to a subclass
10166 of gimple_opt_pass along with...
10167 (pass_data_warn_unused_result): ...new pass_data instance and...
10168 (make_pass_warn_unused_result): ...new function.
10169 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10170 a subclass of gimple_opt_pass along with...
10171 (pass_data_merge_phi): ...new pass_data instance and...
10172 (make_pass_merge_phi): ...new function.
10173 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10174 a subclass of gimple_opt_pass along with...
10175 (pass_data_lower_complex): ...new pass_data instance and...
10176 (make_pass_lower_complex): ...new function.
10177 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10178 gimple_opt_pass along with...
10179 (pass_data_lower_complex_O0): ...new pass_data instance and...
10180 (make_pass_lower_complex_O0): ...new function.
10181 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10182 subclass of gimple_opt_pass along with...
10183 (pass_data_lower_eh): ...new pass_data instance and...
10184 (make_pass_lower_eh): ...new function.
10185 (pass_refactor_eh): Convert from a global struct to a subclass of
10186 gimple_opt_pass along with...
10187 (pass_data_refactor_eh): ...new pass_data instance and...
10188 (make_pass_refactor_eh): ...new function.
10189 (pass_lower_resx): Convert from a global struct to a subclass of
10190 gimple_opt_pass along with...
10191 (pass_data_lower_resx): ...new pass_data instance and...
10192 (make_pass_lower_resx): ...new function.
10193 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10194 of gimple_opt_pass along with...
10195 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10196 (make_pass_lower_eh_dispatch): ...new function.
10197 (pass_cleanup_eh): Convert from a global struct to a subclass of
10198 gimple_opt_pass along with...
10199 (pass_data_cleanup_eh): ...new pass_data instance and...
10200 (make_pass_cleanup_eh): ...new function.
10201 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10202 to a subclass of simple_ipa_opt_pass along with...
10203 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10204 (make_pass_ipa_lower_emutls): ...new function.
10205 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10206 a subclass of gimple_opt_pass along with...
10207 (pass_data_if_conversion): ...new pass_data instance and...
10208 (make_pass_if_conversion): ...new function.
10209 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10210 subclass of gimple_opt_pass along with...
10211 (pass_data_build_ssa): ...new pass_data instance and...
10212 (make_pass_build_ssa): ...new function.
10213 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10214 global struct to a subclass of gimple_opt_pass along with...
10215 (pass_data_loop_distribution): ...new pass_data instance and...
10216 (make_pass_loop_distribution): ...new function.
10217 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10218 subclass of gimple_opt_pass along with...
10219 (pass_data_mudflap_1): ...new pass_data instance and...
10220 (make_pass_mudflap_1): ...new function.
10221 (pass_mudflap_2): Convert from a global struct to a subclass of
10222 gimple_opt_pass along with...
10223 (pass_data_mudflap_2): ...new pass_data instance and...
10224 (make_pass_mudflap_2): ...new function.
10225 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10226 subclass of gimple_opt_pass along with...
10227 (pass_data_mudflap_1): ...new pass_data instance and...
10228 (make_pass_mudflap_1): ...new function.
10229 (pass_mudflap_2): Convert from a global struct to a subclass of
10230 gimple_opt_pass along with...
10231 (pass_data_mudflap_2): ...new pass_data instance and...
10232 (make_pass_mudflap_2): ...new function.
10233 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10234 gimple_opt_pass along with...
10235 (pass_data_nrv): ...new pass_data instance and...
10236 (make_pass_nrv): ...new function.
10237 (pass_return_slot): Convert from a global struct to a subclass of
10238 gimple_opt_pass along with...
10239 (pass_data_return_slot): ...new pass_data instance and...
10240 (make_pass_return_slot): ...new function.
10241 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10242 to a subclass of gimple_opt_pass along with...
10243 (pass_data_object_sizes): ...new pass_data instance and...
10244 (make_pass_object_sizes): ...new function.
10245 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10246 global struct to a subclass of gimple_opt_pass along with...
10247 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10249 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10250 (pass_fixup_cfg): Convert from a global struct to a subclass of
10251 gimple_opt_pass along with...
10252 (pass_data_fixup_cfg): ...new pass_data instance and...
10253 (make_pass_fixup_cfg): ...new function.
10254 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10255 (make_pass_mudflap_1): ...new function.
10256 (pass_mudflap_2): Replace declaration with that of...
10257 (make_pass_mudflap_2): ...new function.
10258 (pass_asan): Replace declaration with that of...
10259 (make_pass_asan): ...new function.
10260 (pass_asan_O0): Replace declaration with that of...
10261 (make_pass_asan_O0): ...new function.
10262 (pass_tsan): Replace declaration with that of...
10263 (make_pass_tsan): ...new function.
10264 (pass_tsan_O0): Replace declaration with that of...
10265 (make_pass_tsan_O0): ...new function.
10266 (pass_lower_cf): Replace declaration with that of...
10267 (make_pass_lower_cf): ...new function.
10268 (pass_refactor_eh): Replace declaration with that of...
10269 (make_pass_refactor_eh): ...new function.
10270 (pass_lower_eh): Replace declaration with that of...
10271 (make_pass_lower_eh): ...new function.
10272 (pass_lower_eh_dispatch): Replace declaration with that of...
10273 (make_pass_lower_eh_dispatch): ...new function.
10274 (pass_lower_resx): Replace declaration with that of...
10275 (make_pass_lower_resx): ...new function.
10276 (pass_build_cfg): Replace declaration with that of...
10277 (make_pass_build_cfg): ...new function.
10278 (pass_early_tree_profile): Replace declaration with that of...
10279 (make_pass_early_tree_profile): ...new function.
10280 (pass_cleanup_eh): Replace declaration with that of...
10281 (make_pass_cleanup_eh): ...new function.
10282 (pass_sra): Replace declaration with that of...
10283 (make_pass_sra): ...new function.
10284 (pass_sra_early): Replace declaration with that of...
10285 (make_pass_sra_early): ...new function.
10286 (pass_early_ipa_sra): Replace declaration with that of...
10287 (make_pass_early_ipa_sra): ...new function.
10288 (pass_tail_recursion): Replace declaration with that of...
10289 (make_pass_tail_recursion): ...new function.
10290 (pass_tail_calls): Replace declaration with that of...
10291 (make_pass_tail_calls): ...new function.
10292 (pass_tree_loop): Replace declaration with that of...
10293 (make_pass_tree_loop): ...new function.
10294 (pass_tree_loop_init): Replace declaration with that of...
10295 (make_pass_tree_loop_init): ...new function.
10296 (pass_lim): Replace declaration with that of...
10297 (make_pass_lim): ...new function.
10298 (pass_tree_unswitch): Replace declaration with that of...
10299 (make_pass_tree_unswitch): ...new function.
10300 (pass_predcom): Replace declaration with that of...
10301 (make_pass_predcom): ...new function.
10302 (pass_iv_canon): Replace declaration with that of...
10303 (make_pass_iv_canon): ...new function.
10304 (pass_scev_cprop): Replace declaration with that of...
10305 (make_pass_scev_cprop): ...new function.
10306 (pass_empty_loop): Replace declaration with that of...
10307 (make_pass_empty_loop): ...new function.
10308 (pass_record_bounds): Replace declaration with that of...
10309 (make_pass_record_bounds): ...new function.
10310 (pass_graphite): Replace declaration with that of...
10311 (make_pass_graphite): ...new function.
10312 (pass_graphite_transforms): Replace declaration with that of...
10313 (make_pass_graphite_transforms): ...new function.
10314 (pass_if_conversion): Replace declaration with that of...
10315 (make_pass_if_conversion): ...new function.
10316 (pass_loop_distribution): Replace declaration with that of...
10317 (make_pass_loop_distribution): ...new function.
10318 (pass_vectorize): Replace declaration with that of...
10319 (make_pass_vectorize): ...new function.
10320 (pass_slp_vectorize): Replace declaration with that of...
10321 (make_pass_slp_vectorize): ...new function.
10322 (pass_complete_unroll): Replace declaration with that of...
10323 (make_pass_complete_unroll): ...new function.
10324 (pass_complete_unrolli): Replace declaration with that of...
10325 (make_pass_complete_unrolli): ...new function.
10326 (pass_parallelize_loops): Replace declaration with that of...
10327 (make_pass_parallelize_loops): ...new function.
10328 (pass_loop_prefetch): Replace declaration with that of...
10329 (make_pass_loop_prefetch): ...new function.
10330 (pass_iv_optimize): Replace declaration with that of...
10331 (make_pass_iv_optimize): ...new function.
10332 (pass_tree_loop_done): Replace declaration with that of...
10333 (make_pass_tree_loop_done): ...new function.
10334 (pass_ch): Replace declaration with that of...
10335 (make_pass_ch): ...new function.
10336 (pass_ccp): Replace declaration with that of...
10337 (make_pass_ccp): ...new function.
10338 (pass_phi_only_cprop): Replace declaration with that of...
10339 (make_pass_phi_only_cprop): ...new function.
10340 (pass_build_ssa): Replace declaration with that of...
10341 (make_pass_build_ssa): ...new function.
10342 (pass_build_alias): Replace declaration with that of...
10343 (make_pass_build_alias): ...new function.
10344 (pass_build_ealias): Replace declaration with that of...
10345 (make_pass_build_ealias): ...new function.
10346 (pass_dominator): Replace declaration with that of...
10347 (make_pass_dominator): ...new function.
10348 (pass_dce): Replace declaration with that of...
10349 (make_pass_dce): ...new function.
10350 (pass_dce_loop): Replace declaration with that of...
10351 (make_pass_dce_loop): ...new function.
10352 (pass_cd_dce): Replace declaration with that of...
10353 (make_pass_cd_dce): ...new function.
10354 (pass_call_cdce): Replace declaration with that of...
10355 (make_pass_call_cdce): ...new function.
10356 (pass_merge_phi): Replace declaration with that of...
10357 (make_pass_merge_phi): ...new function.
10358 (pass_split_crit_edges): Replace declaration with that of...
10359 (make_pass_split_crit_edges): ...new function.
10360 (pass_pre): Replace declaration with that of...
10361 (make_pass_pre): ...new function.
10362 (pass_profile): Replace declaration with that of...
10363 (make_pass_profile): ...new function.
10364 (pass_strip_predict_hints): Replace declaration with that of...
10365 (make_pass_strip_predict_hints): ...new function.
10366 (pass_lower_complex_O0): Replace declaration with that of...
10367 (make_pass_lower_complex_O0): ...new function.
10368 (pass_lower_complex): Replace declaration with that of...
10369 (make_pass_lower_complex): ...new function.
10370 (pass_lower_vector): Replace declaration with that of...
10371 (make_pass_lower_vector): ...new function.
10372 (pass_lower_vector_ssa): Replace declaration with that of...
10373 (make_pass_lower_vector_ssa): ...new function.
10374 (pass_lower_omp): Replace declaration with that of...
10375 (make_pass_lower_omp): ...new function.
10376 (pass_diagnose_omp_blocks): Replace declaration with that of...
10377 (make_pass_diagnose_omp_blocks): ...new function.
10378 (pass_expand_omp): Replace declaration with that of...
10379 (make_pass_expand_omp): ...new function.
10380 (pass_expand_omp_ssa): Replace declaration with that of...
10381 (make_pass_expand_omp_ssa): ...new function.
10382 (pass_object_sizes): Replace declaration with that of...
10383 (make_pass_object_sizes): ...new function.
10384 (pass_strlen): Replace declaration with that of...
10385 (make_pass_strlen): ...new function.
10386 (pass_fold_builtins): Replace declaration with that of...
10387 (make_pass_fold_builtins): ...new function.
10388 (pass_stdarg): Replace declaration with that of...
10389 (make_pass_stdarg): ...new function.
10390 (pass_early_warn_uninitialized): Replace declaration with that of...
10391 (make_pass_early_warn_uninitialized): ...new function.
10392 (pass_late_warn_uninitialized): Replace declaration with that of...
10393 (make_pass_late_warn_uninitialized): ...new function.
10394 (pass_cse_reciprocals): Replace declaration with that of...
10395 (make_pass_cse_reciprocals): ...new function.
10396 (pass_cse_sincos): Replace declaration with that of...
10397 (make_pass_cse_sincos): ...new function.
10398 (pass_optimize_bswap): Replace declaration with that of...
10399 (make_pass_optimize_bswap): ...new function.
10400 (pass_optimize_widening_mul): Replace declaration with that of...
10401 (make_pass_optimize_widening_mul): ...new function.
10402 (pass_warn_function_return): Replace declaration with that of...
10403 (make_pass_warn_function_return): ...new function.
10404 (pass_warn_function_noreturn): Replace declaration with that of...
10405 (make_pass_warn_function_noreturn): ...new function.
10406 (pass_cselim): Replace declaration with that of...
10407 (make_pass_cselim): ...new function.
10408 (pass_phiopt): Replace declaration with that of...
10409 (make_pass_phiopt): ...new function.
10410 (pass_forwprop): Replace declaration with that of...
10411 (make_pass_forwprop): ...new function.
10412 (pass_phiprop): Replace declaration with that of...
10413 (make_pass_phiprop): ...new function.
10414 (pass_tree_ifcombine): Replace declaration with that of...
10415 (make_pass_tree_ifcombine): ...new function.
10416 (pass_dse): Replace declaration with that of...
10417 (make_pass_dse): ...new function.
10418 (pass_nrv): Replace declaration with that of...
10419 (make_pass_nrv): ...new function.
10420 (pass_rename_ssa_copies): Replace declaration with that of...
10421 (make_pass_rename_ssa_copies): ...new function.
10422 (pass_sink_code): Replace declaration with that of...
10423 (make_pass_sink_code): ...new function.
10424 (pass_fre): Replace declaration with that of...
10425 (make_pass_fre): ...new function.
10426 (pass_check_data_deps): Replace declaration with that of...
10427 (make_pass_check_data_deps): ...new function.
10428 (pass_copy_prop): Replace declaration with that of...
10429 (make_pass_copy_prop): ...new function.
10430 (pass_vrp): Replace declaration with that of...
10431 (make_pass_vrp): ...new function.
10432 (pass_uncprop): Replace declaration with that of...
10433 (make_pass_uncprop): ...new function.
10434 (pass_return_slot): Replace declaration with that of...
10435 (make_pass_return_slot): ...new function.
10436 (pass_reassoc): Replace declaration with that of...
10437 (make_pass_reassoc): ...new function.
10438 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10439 (make_pass_rebuild_cgraph_edges): ...new function.
10440 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10441 (make_pass_remove_cgraph_callee_edges): ...new function.
10442 (pass_build_cgraph_edges): Replace declaration with that of...
10443 (make_pass_build_cgraph_edges): ...new function.
10444 (pass_local_pure_const): Replace declaration with that of...
10445 (make_pass_local_pure_const): ...new function.
10446 (pass_tracer): Replace declaration with that of...
10447 (make_pass_tracer): ...new function.
10448 (pass_warn_unused_result): Replace declaration with that of...
10449 (make_pass_warn_unused_result): ...new function.
10450 (pass_diagnose_tm_blocks): Replace declaration with that of...
10451 (make_pass_diagnose_tm_blocks): ...new function.
10452 (pass_lower_tm): Replace declaration with that of...
10453 (make_pass_lower_tm): ...new function.
10454 (pass_tm_init): Replace declaration with that of...
10455 (make_pass_tm_init): ...new function.
10456 (pass_tm_mark): Replace declaration with that of...
10457 (make_pass_tm_mark): ...new function.
10458 (pass_tm_memopt): Replace declaration with that of...
10459 (make_pass_tm_memopt): ...new function.
10460 (pass_tm_edges): Replace declaration with that of...
10461 (make_pass_tm_edges): ...new function.
10462 (pass_split_functions): Replace declaration with that of...
10463 (make_pass_split_functions): ...new function.
10464 (pass_feedback_split_functions): Replace declaration with that of...
10465 (make_pass_feedback_split_functions): ...new function.
10466 (pass_strength_reduction): Replace declaration with that of...
10467 (make_pass_strength_reduction): ...new function.
10468 (pass_ipa_lower_emutls): Replace declaration with that of...
10469 (make_pass_ipa_lower_emutls): ...new function.
10470 (pass_ipa_function_and_variable_visibility): Replace declaration with
10472 (make_pass_ipa_function_and_variable_visibility): ...new function.
10473 (pass_ipa_tree_profile): Replace declaration with that of...
10474 (make_pass_ipa_tree_profile): ...new function.
10475 (pass_early_local_passes): Replace declaration with that of...
10476 (make_pass_early_local_passes): ...new function.
10477 (pass_ipa_whole_program_visibility): Replace declaration with that
10479 (make_pass_ipa_whole_program_visibility): ...new function.
10480 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10481 (make_pass_ipa_lto_gimple_out): ...new function.
10482 (pass_ipa_increase_alignment): Replace declaration with that of...
10483 (make_pass_ipa_increase_alignment): ...new function.
10484 (pass_ipa_inline): Replace declaration with that of...
10485 (make_pass_ipa_inline): ...new function.
10486 (pass_ipa_free_lang_data): Replace declaration with that of...
10487 (make_pass_ipa_free_lang_data): ...new function.
10488 (pass_ipa_free_inline_summary): Replace declaration with that of...
10489 (make_pass_ipa_free_inline_summary): ...new function.
10490 (pass_ipa_cp): Replace declaration with that of...
10491 (make_pass_ipa_cp): ...new function.
10492 (pass_ipa_reference): Replace declaration with that of...
10493 (make_pass_ipa_reference): ...new function.
10494 (pass_ipa_pure_const): Replace declaration with that of...
10495 (make_pass_ipa_pure_const): ...new function.
10496 (pass_ipa_pta): Replace declaration with that of...
10497 (make_pass_ipa_pta): ...new function.
10498 (pass_ipa_lto_finish_out): Replace declaration with that of...
10499 (make_pass_ipa_lto_finish_out): ...new function.
10500 (pass_ipa_tm): Replace declaration with that of...
10501 (make_pass_ipa_tm): ...new function.
10502 (pass_ipa_profile): Replace declaration with that of...
10503 (make_pass_ipa_profile): ...new function.
10504 (pass_ipa_cdtor_merge): Replace declaration with that of...
10505 (make_pass_ipa_cdtor_merge): ...new function.
10506 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10508 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10509 (pass_init_datastructures): Replace declaration with that of...
10510 (make_pass_init_datastructures): ...new function.
10511 (pass_fixup_cfg): Replace declaration with that of...
10512 (make_pass_fixup_cfg): ...new function.
10513 (pass_expand): Replace declaration with that of...
10514 (make_pass_expand): ...new function.
10515 (pass_instantiate_virtual_regs): Replace declaration with that of...
10516 (make_pass_instantiate_virtual_regs): ...new function.
10517 (pass_rtl_fwprop): Replace declaration with that of...
10518 (make_pass_rtl_fwprop): ...new function.
10519 (pass_rtl_fwprop_addr): Replace declaration with that of...
10520 (make_pass_rtl_fwprop_addr): ...new function.
10521 (pass_jump): Replace declaration with that of...
10522 (make_pass_jump): ...new function.
10523 (pass_jump2): Replace declaration with that of...
10524 (make_pass_jump2): ...new function.
10525 (pass_lower_subreg): Replace declaration with that of...
10526 (make_pass_lower_subreg): ...new function.
10527 (pass_cse): Replace declaration with that of...
10528 (make_pass_cse): ...new function.
10529 (pass_fast_rtl_dce): Replace declaration with that of...
10530 (make_pass_fast_rtl_dce): ...new function.
10531 (pass_ud_rtl_dce): Replace declaration with that of...
10532 (make_pass_ud_rtl_dce): ...new function.
10533 (pass_rtl_dce): Replace declaration with that of...
10534 (make_pass_rtl_dce): ...new function.
10535 (pass_rtl_dse1): Replace declaration with that of...
10536 (make_pass_rtl_dse1): ...new function.
10537 (pass_rtl_dse2): Replace declaration with that of...
10538 (make_pass_rtl_dse2): ...new function.
10539 (pass_rtl_dse3): Replace declaration with that of...
10540 (make_pass_rtl_dse3): ...new function.
10541 (pass_rtl_cprop): Replace declaration with that of...
10542 (make_pass_rtl_cprop): ...new function.
10543 (pass_rtl_pre): Replace declaration with that of...
10544 (make_pass_rtl_pre): ...new function.
10545 (pass_rtl_hoist): Replace declaration with that of...
10546 (make_pass_rtl_hoist): ...new function.
10547 (pass_rtl_store_motion): Replace declaration with that of...
10548 (make_pass_rtl_store_motion): ...new function.
10549 (pass_cse_after_global_opts): Replace declaration with that of...
10550 (make_pass_cse_after_global_opts): ...new function.
10551 (pass_rtl_ifcvt): Replace declaration with that of...
10552 (make_pass_rtl_ifcvt): ...new function.
10553 (pass_into_cfg_layout_mode): Replace declaration with that of...
10554 (make_pass_into_cfg_layout_mode): ...new function.
10555 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10556 (make_pass_outof_cfg_layout_mode): ...new function.
10557 (pass_loop2): Replace declaration with that of...
10558 (make_pass_loop2): ...new function.
10559 (pass_rtl_loop_init): Replace declaration with that of...
10560 (make_pass_rtl_loop_init): ...new function.
10561 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10562 (make_pass_rtl_move_loop_invariants): ...new function.
10563 (pass_rtl_unswitch): Replace declaration with that of...
10564 (make_pass_rtl_unswitch): ...new function.
10565 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10566 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10567 (pass_rtl_doloop): Replace declaration with that of...
10568 (make_pass_rtl_doloop): ...new function.
10569 (pass_rtl_loop_done): Replace declaration with that of...
10570 (make_pass_rtl_loop_done): ...new function.
10571 (pass_web): Replace declaration with that of...
10572 (make_pass_web): ...new function.
10573 (pass_cse2): Replace declaration with that of...
10574 (make_pass_cse2): ...new function.
10575 (pass_df_initialize_opt): Replace declaration with that of...
10576 (make_pass_df_initialize_opt): ...new function.
10577 (pass_df_initialize_no_opt): Replace declaration with that of...
10578 (make_pass_df_initialize_no_opt): ...new function.
10579 (pass_reginfo_init): Replace declaration with that of...
10580 (make_pass_reginfo_init): ...new function.
10581 (pass_inc_dec): Replace declaration with that of...
10582 (make_pass_inc_dec): ...new function.
10583 (pass_stack_ptr_mod): Replace declaration with that of...
10584 (make_pass_stack_ptr_mod): ...new function.
10585 (pass_initialize_regs): Replace declaration with that of...
10586 (make_pass_initialize_regs): ...new function.
10587 (pass_combine): Replace declaration with that of...
10588 (make_pass_combine): ...new function.
10589 (pass_if_after_combine): Replace declaration with that of...
10590 (make_pass_if_after_combine): ...new function.
10591 (pass_ree): Replace declaration with that of...
10592 (make_pass_ree): ...new function.
10593 (pass_partition_blocks): Replace declaration with that of...
10594 (make_pass_partition_blocks): ...new function.
10595 (pass_match_asm_constraints): Replace declaration with that of...
10596 (make_pass_match_asm_constraints): ...new function.
10597 (pass_regmove): Replace declaration with that of...
10598 (make_pass_regmove): ...new function.
10599 (pass_split_all_insns): Replace declaration with that of...
10600 (make_pass_split_all_insns): ...new function.
10601 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10602 (make_pass_fast_rtl_byte_dce): ...new function.
10603 (pass_lower_subreg2): Replace declaration with that of...
10604 (make_pass_lower_subreg2): ...new function.
10605 (pass_mode_switching): Replace declaration with that of...
10606 (make_pass_mode_switching): ...new function.
10607 (pass_sms): Replace declaration with that of...
10608 (make_pass_sms): ...new function.
10609 (pass_sched): Replace declaration with that of...
10610 (make_pass_sched): ...new function.
10611 (pass_ira): Replace declaration with that of...
10612 (make_pass_ira): ...new function.
10613 (pass_reload): Replace declaration with that of...
10614 (make_pass_reload): ...new function.
10615 (pass_clean_state): Replace declaration with that of...
10616 (make_pass_clean_state): ...new function.
10617 (pass_branch_prob): Replace declaration with that of...
10618 (make_pass_branch_prob): ...new function.
10619 (pass_value_profile_transformations): Replace declaration with that
10621 (make_pass_value_profile_transformations): ...new function.
10622 (pass_postreload_cse): Replace declaration with that of...
10623 (make_pass_postreload_cse): ...new function.
10624 (pass_gcse2): Replace declaration with that of...
10625 (make_pass_gcse2): ...new function.
10626 (pass_split_after_reload): Replace declaration with that of...
10627 (make_pass_split_after_reload): ...new function.
10628 (pass_branch_target_load_optimize1): Replace declaration with that
10630 (make_pass_branch_target_load_optimize1): ...new function.
10631 (pass_thread_prologue_and_epilogue): Replace declaration with that
10633 (make_pass_thread_prologue_and_epilogue): ...new function.
10634 (pass_stack_adjustments): Replace declaration with that of...
10635 (make_pass_stack_adjustments): ...new function.
10636 (pass_peephole2): Replace declaration with that of...
10637 (make_pass_peephole2): ...new function.
10638 (pass_if_after_reload): Replace declaration with that of...
10639 (make_pass_if_after_reload): ...new function.
10640 (pass_regrename): Replace declaration with that of...
10641 (make_pass_regrename): ...new function.
10642 (pass_cprop_hardreg): Replace declaration with that of...
10643 (make_pass_cprop_hardreg): ...new function.
10644 (pass_reorder_blocks): Replace declaration with that of...
10645 (make_pass_reorder_blocks): ...new function.
10646 (pass_branch_target_load_optimize2): Replace declaration with that
10648 (make_pass_branch_target_load_optimize2): ...new function.
10649 (pass_leaf_regs): Replace declaration with that of...
10650 (make_pass_leaf_regs): ...new function.
10651 (pass_split_before_sched2): Replace declaration with that of...
10652 (make_pass_split_before_sched2): ...new function.
10653 (pass_compare_elim_after_reload): Replace declaration with that of...
10654 (make_pass_compare_elim_after_reload): ...new function.
10655 (pass_sched2): Replace declaration with that of...
10656 (make_pass_sched2): ...new function.
10657 (pass_stack_regs): Replace declaration with that of...
10658 (make_pass_stack_regs): ...new function.
10659 (pass_stack_regs_run): Replace declaration with that of...
10660 (make_pass_stack_regs_run): ...new function.
10661 (pass_df_finish): Replace declaration with that of...
10662 (make_pass_df_finish): ...new function.
10663 (pass_compute_alignments): Replace declaration with that of...
10664 (make_pass_compute_alignments): ...new function.
10665 (pass_duplicate_computed_gotos): Replace declaration with that of...
10666 (make_pass_duplicate_computed_gotos): ...new function.
10667 (pass_variable_tracking): Replace declaration with that of...
10668 (make_pass_variable_tracking): ...new function.
10669 (pass_free_cfg): Replace declaration with that of...
10670 (make_pass_free_cfg): ...new function.
10671 (pass_machine_reorg): Replace declaration with that of...
10672 (make_pass_machine_reorg): ...new function.
10673 (pass_cleanup_barriers): Replace declaration with that of...
10674 (make_pass_cleanup_barriers): ...new function.
10675 (pass_delay_slots): Replace declaration with that of...
10676 (make_pass_delay_slots): ...new function.
10677 (pass_split_for_shorten_branches): Replace declaration with that of...
10678 (make_pass_split_for_shorten_branches): ...new function.
10679 (pass_split_before_regstack): Replace declaration with that of...
10680 (make_pass_split_before_regstack): ...new function.
10681 (pass_convert_to_eh_region_ranges): Replace declaration with that
10683 (make_pass_convert_to_eh_region_ranges): ...new function.
10684 (pass_shorten_branches): Replace declaration with that of...
10685 (make_pass_shorten_branches): ...new function.
10686 (pass_set_nothrow_function_flags): Replace declaration with that of...
10687 (make_pass_set_nothrow_function_flags): ...new function.
10688 (pass_dwarf2_frame): Replace declaration with that of...
10689 (make_pass_dwarf2_frame): ...new function.
10690 (pass_final): Replace declaration with that of...
10691 (make_pass_final): ...new function.
10692 (pass_rtl_seqabstr): Replace declaration with that of...
10693 (make_pass_rtl_seqabstr): ...new function.
10694 (pass_release_ssa_names): Replace declaration with that of...
10695 (make_pass_release_ssa_names): ...new function.
10696 (pass_early_inline): Replace declaration with that of...
10697 (make_pass_early_inline): ...new function.
10698 (pass_inline_parameters): Replace declaration with that of...
10699 (make_pass_inline_parameters): ...new function.
10700 (pass_update_address_taken): Replace declaration with that of...
10701 (make_pass_update_address_taken): ...new function.
10702 (pass_convert_switch): Replace declaration with that of...
10703 (make_pass_convert_switch): ...new function.
10704 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10705 to a subclass of simple_ipa_opt_pass along with...
10706 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10707 (make_pass_ipa_tree_profile): ...new function.
10708 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10709 subclass of gimple_opt_pass along with...
10710 (pass_data_sra_early): ...new pass_data instance and...
10711 (make_pass_sra_early): ...new function.
10712 (pass_sra): Convert from a global struct to a subclass of
10713 gimple_opt_pass along with...
10714 (pass_data_sra): ...new pass_data instance and...
10715 (make_pass_sra): ...new function.
10716 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10717 gimple_opt_pass along with...
10718 (pass_data_early_ipa_sra): ...new pass_data instance and...
10719 (make_pass_early_ipa_sra): ...new function.
10720 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10721 subclass of gimple_opt_pass along with...
10722 (pass_data_ccp): ...new pass_data instance and...
10723 (make_pass_ccp): ...new function.
10724 (pass_fold_builtins): Convert from a global struct to a subclass of
10725 gimple_opt_pass along with...
10726 (pass_data_fold_builtins): ...new pass_data instance and...
10727 (make_pass_fold_builtins): ...new function.
10728 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10729 subclass of gimple_opt_pass along with...
10730 (pass_data_copy_prop): ...new pass_data instance and...
10731 (make_pass_copy_prop): ...new function.
10732 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10733 global struct to a subclass of gimple_opt_pass along with...
10734 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10735 (make_pass_rename_ssa_copies): ...new function.
10736 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10737 subclass of gimple_opt_pass along with...
10738 (pass_data_dce): ...new pass_data instance and...
10739 (make_pass_dce): ...new function.
10740 (pass_dce_loop): Convert from a global struct to a subclass of
10741 gimple_opt_pass along with...
10742 (pass_data_dce_loop): ...new pass_data instance and...
10743 (make_pass_dce_loop): ...new function.
10744 (pass_cd_dce): Convert from a global struct to a subclass of
10745 gimple_opt_pass along with...
10746 (pass_data_cd_dce): ...new pass_data instance and...
10747 (make_pass_cd_dce): ...new function.
10748 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10749 subclass of gimple_opt_pass along with...
10750 (pass_data_dominator): ...new pass_data instance and...
10751 (make_pass_dominator): ...new function.
10752 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10753 gimple_opt_pass along with...
10754 (pass_data_phi_only_cprop): ...new pass_data instance and...
10755 (make_pass_phi_only_cprop): ...new function.
10756 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10757 subclass of gimple_opt_pass along with...
10758 (pass_data_dse): ...new pass_data instance and...
10759 (make_pass_dse): ...new function.
10760 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10761 a subclass of gimple_opt_pass along with...
10762 (pass_data_forwprop): ...new pass_data instance and...
10763 (make_pass_forwprop): ...new function.
10764 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10765 struct to a subclass of gimple_opt_pass along with...
10766 (pass_data_tree_ifcombine): ...new pass_data instance and...
10767 (make_pass_tree_ifcombine): ...new function.
10768 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10769 subclass of gimple_opt_pass along with...
10770 (pass_data_ch): ...new pass_data instance and...
10771 (make_pass_ch): ...new function.
10772 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10773 subclass of gimple_opt_pass along with...
10774 (pass_data_tree_loop): ...new pass_data instance and...
10775 (make_pass_tree_loop): ...new function.
10776 (pass_tree_loop_init): Convert from a global struct to a subclass of
10777 gimple_opt_pass along with...
10778 (pass_data_tree_loop_init): ...new pass_data instance and...
10779 (make_pass_tree_loop_init): ...new function.
10780 (pass_lim): Convert from a global struct to a subclass of
10781 gimple_opt_pass along with...
10782 (pass_data_lim): ...new pass_data instance and...
10783 (make_pass_lim): ...new function.
10784 (pass_tree_unswitch): Convert from a global struct to a subclass of
10785 gimple_opt_pass along with...
10786 (pass_data_tree_unswitch): ...new pass_data instance and...
10787 (make_pass_tree_unswitch): ...new function.
10788 (pass_predcom): Convert from a global struct to a subclass of
10789 gimple_opt_pass along with...
10790 (pass_data_predcom): ...new pass_data instance and...
10791 (make_pass_predcom): ...new function.
10792 (pass_vectorize): Convert from a global struct to a subclass of
10793 gimple_opt_pass along with...
10794 (pass_data_vectorize): ...new pass_data instance and...
10795 (make_pass_vectorize): ...new function.
10796 (pass_graphite): Convert from a global struct to a subclass of
10797 gimple_opt_pass along with...
10798 (pass_data_graphite): ...new pass_data instance and...
10799 (make_pass_graphite): ...new function.
10800 (pass_graphite_transforms): Convert from a global struct to a subclass
10801 of gimple_opt_pass along with...
10802 (pass_data_graphite_transforms): ...new pass_data instance and...
10803 (make_pass_graphite_transforms): ...new function.
10804 (pass_check_data_deps): Convert from a global struct to a subclass of
10805 gimple_opt_pass along with...
10806 (pass_data_check_data_deps): ...new pass_data instance and...
10807 (make_pass_check_data_deps): ...new function.
10808 (pass_iv_canon): Convert from a global struct to a subclass of
10809 gimple_opt_pass along with...
10810 (pass_data_iv_canon): ...new pass_data instance and...
10811 (make_pass_iv_canon): ...new function.
10812 (pass_scev_cprop): Convert from a global struct to a subclass of
10813 gimple_opt_pass along with...
10814 (pass_data_scev_cprop): ...new pass_data instance and...
10815 (make_pass_scev_cprop): ...new function.
10816 (pass_record_bounds): Convert from a global struct to a subclass of
10817 gimple_opt_pass along with...
10818 (pass_data_record_bounds): ...new pass_data instance and...
10819 (make_pass_record_bounds): ...new function.
10820 (pass_complete_unroll): Convert from a global struct to a subclass of
10821 gimple_opt_pass along with...
10822 (pass_data_complete_unroll): ...new pass_data instance and...
10823 (make_pass_complete_unroll): ...new function.
10824 (pass_complete_unrolli): Convert from a global struct to a subclass of
10825 gimple_opt_pass along with...
10826 (pass_data_complete_unrolli): ...new pass_data instance and...
10827 (make_pass_complete_unrolli): ...new function.
10828 (pass_parallelize_loops): Convert from a global struct to a subclass
10829 of gimple_opt_pass along with...
10830 (pass_data_parallelize_loops): ...new pass_data instance and...
10831 (make_pass_parallelize_loops): ...new function.
10832 (pass_loop_prefetch): Convert from a global struct to a subclass of
10833 gimple_opt_pass along with...
10834 (pass_data_loop_prefetch): ...new pass_data instance and...
10835 (make_pass_loop_prefetch): ...new function.
10836 (pass_iv_optimize): Convert from a global struct to a subclass of
10837 gimple_opt_pass along with...
10838 (pass_data_iv_optimize): ...new pass_data instance and...
10839 (make_pass_iv_optimize): ...new function.
10840 (pass_tree_loop_done): Convert from a global struct to a subclass of
10841 gimple_opt_pass along with...
10842 (pass_data_tree_loop_done): ...new pass_data instance and...
10843 (make_pass_tree_loop_done): ...new function.
10844 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10845 struct to a subclass of gimple_opt_pass along with...
10846 (pass_data_cse_reciprocals): ...new pass_data instance and...
10847 (make_pass_cse_reciprocals): ...new function.
10848 (pass_cse_sincos): Convert from a global struct to a subclass of
10849 gimple_opt_pass along with...
10850 (pass_data_cse_sincos): ...new pass_data instance and...
10851 (make_pass_cse_sincos): ...new function.
10852 (pass_optimize_bswap): Convert from a global struct to a subclass of
10853 gimple_opt_pass along with...
10854 (pass_data_optimize_bswap): ...new pass_data instance and...
10855 (make_pass_optimize_bswap): ...new function.
10856 (pass_optimize_widening_mul): Convert from a global struct to a
10857 subclass of gimple_opt_pass along with...
10858 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10859 (make_pass_optimize_widening_mul): ...new function.
10860 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10861 subclass of gimple_opt_pass along with...
10862 (pass_data_phiopt): ...new pass_data instance and...
10863 (make_pass_phiopt): ...new function.
10864 (pass_cselim): Convert from a global struct to a subclass of
10865 gimple_opt_pass along with...
10866 (pass_data_cselim): ...new pass_data instance and...
10867 (make_pass_cselim): ...new function.
10868 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10869 subclass of gimple_opt_pass along with...
10870 (pass_data_phiprop): ...new pass_data instance and...
10871 (make_pass_phiprop): ...new function.
10872 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10873 subclass of gimple_opt_pass along with...
10874 (pass_data_pre): ...new pass_data instance and...
10875 (make_pass_pre): ...new function.
10876 (pass_fre): Convert from a global struct to a subclass of
10877 gimple_opt_pass along with...
10878 (pass_data_fre): ...new pass_data instance and...
10879 (make_pass_fre): ...new function.
10880 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10881 subclass of gimple_opt_pass along with...
10882 (pass_data_reassoc): ...new pass_data instance and...
10883 (make_pass_reassoc): ...new function.
10884 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10885 subclass of gimple_opt_pass along with...
10886 (pass_data_sink_code): ...new pass_data instance and...
10887 (make_pass_sink_code): ...new function.
10888 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10889 subclass of gimple_opt_pass along with...
10890 (pass_data_strlen): ...new pass_data instance and...
10891 (make_pass_strlen): ...new function.
10892 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10893 struct to a subclass of gimple_opt_pass along with...
10894 (pass_data_build_alias): ...new pass_data instance and...
10895 (make_pass_build_alias): ...new function.
10896 (pass_build_ealias): Convert from a global struct to a subclass of
10897 gimple_opt_pass along with...
10898 (pass_data_build_ealias): ...new pass_data instance and...
10899 (make_pass_build_ealias): ...new function.
10900 (pass_ipa_pta): Convert from a global struct to a subclass of
10901 simple_ipa_opt_pass along with...
10902 (pass_data_ipa_pta): ...new pass_data instance and...
10903 (make_pass_ipa_pta): ...new function.
10904 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10905 subclass of gimple_opt_pass along with...
10906 (pass_data_uncprop): ...new pass_data instance and...
10907 (make_pass_uncprop): ...new function.
10908 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10909 global struct to a subclass of gimple_opt_pass along with...
10910 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10911 (make_pass_late_warn_uninitialized): ...new function.
10912 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10913 to a subclass of gimple_opt_pass along with...
10914 (pass_data_init_datastructures): ...new pass_data instance and...
10915 (make_pass_init_datastructures): ...new function.
10916 (pass_early_warn_uninitialized): Convert from a global struct to a
10917 subclass of gimple_opt_pass along with...
10918 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10919 (make_pass_early_warn_uninitialized): ...new function.
10920 (pass_update_address_taken): Convert from a global struct to a
10921 subclass of gimple_opt_pass along with...
10922 (pass_data_update_address_taken): ...new pass_data instance and...
10923 (make_pass_update_address_taken): ...new function.
10924 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10925 struct to a subclass of gimple_opt_pass along with...
10926 (pass_data_release_ssa_names): ...new pass_data instance and...
10927 (make_pass_release_ssa_names): ...new function.
10928 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10929 subclass of gimple_opt_pass along with...
10930 (pass_data_stdarg): ...new pass_data instance and...
10931 (make_pass_stdarg): ...new function.
10932 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10933 global struct to a subclass of gimple_opt_pass along with...
10934 (pass_data_convert_switch): ...new pass_data instance and...
10935 (make_pass_convert_switch): ...new function.
10936 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10937 to a subclass of gimple_opt_pass along with...
10938 (pass_data_tail_recursion): ...new pass_data instance and...
10939 (make_pass_tail_recursion): ...new function.
10940 (pass_tail_calls): Convert from a global struct to a subclass of
10941 gimple_opt_pass along with...
10942 (pass_data_tail_calls): ...new pass_data instance and...
10943 (make_pass_tail_calls): ...new function.
10944 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10945 struct to a subclass of gimple_opt_pass along with...
10946 (pass_data_lower_vector): ...new pass_data instance and...
10947 (make_pass_lower_vector): ...new function.
10948 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10949 gimple_opt_pass along with...
10950 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10951 (make_pass_lower_vector_ssa): ...new function.
10952 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10953 to a subclass of gimple_opt_pass along with...
10954 (pass_data_slp_vectorize): ...new pass_data instance and...
10955 (make_pass_slp_vectorize): ...new function.
10956 (pass_ipa_increase_alignment): Convert from a global struct to a
10957 subclass of simple_ipa_opt_pass along with...
10958 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10959 (make_pass_ipa_increase_alignment): ...new function.
10960 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10961 gimple_opt_pass along with...
10962 (pass_data_vrp): ...new pass_data instance and...
10963 (make_pass_vrp): ...new function.
10964 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10965 subclass of simple_ipa_opt_pass along with...
10966 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10967 (make_pass_ipa_free_lang_data): ...new function.
10968 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10969 gimple_opt_pass along with...
10970 (pass_data_tsan): ...new pass_data instance and...
10971 (make_pass_tsan): ...new function.
10972 (pass_tsan_O0): Convert from a global struct to a subclass of
10973 gimple_opt_pass along with...
10974 (pass_data_tsan_O0): ...new pass_data instance and...
10975 (make_pass_tsan_O0): ...new function.
10976 * var-tracking.c (pass_variable_tracking): Convert from a global
10977 struct to a subclass of rtl_opt_pass along with...
10978 (pass_data_variable_tracking): ...new pass_data instance and...
10979 (make_pass_variable_tracking): ...new function.
10980 * web.c (pass_web): Convert from a global struct to a subclass of
10981 rtl_opt_pass along with...
10982 (pass_data_web): ...new pass_data instance and...
10983 (make_pass_web): ...new function.
10984 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10985 declaration with that of...
10986 (make_pass_mode_switch_use): ...new function.
10987 (pass_resolve_sw_modes): Replace declaration with that of...
10988 (make_pass_resolve_sw_modes): ...new function.
10989 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10990 from a global struct to a subclass of rtl_opt_pass along with...
10991 (pass_data_mode_switch_use): ...new pass_data instance and...
10992 (make_pass_mode_switch_use): ...new function.
10993 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10994 from a global struct to a subclass of rtl_opt_pass along with...
10995 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10996 (make_pass_resolve_sw_modes): ...new function.
10997 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10998 struct to a subclass of rtl_opt_pass along with...
10999 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11000 (make_pass_insert_vzeroupper): ...new function.
11001 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11002 global struct to a subclass of rtl_opt_pass along with...
11003 (pass_data_work_around_errata): ...new pass_data instance and...
11004 (make_pass_work_around_errata): ...new function.
11005 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11006 struct to a subclass of rtl_opt_pass along with...
11007 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11008 (make_pass_mips_machine_reorg2): ...new function.
11010 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11012 * passes.c (pass_manager::operator new): New.
11014 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11016 Handwritten part of conversion of passes to C++ classes.
11018 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11019 (toplev.o): Add dep on PASS_MANAGER_H.
11020 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11021 of early local pases to reflect this moving from a global to a
11022 member of gcc::pass_manager.
11023 (cgraph_add_new_function): Likewise.
11024 * lto-cgraph.c (lto_output_node): Update for conversion of
11025 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11026 * passes.c (opt_pass::clone): New.
11027 (opt_pass::gate): New.
11028 (opt_pass::execute): New.
11029 (opt_pass::opt_pass): New.
11030 (pass_manager::execute_early_local_passes): New.
11031 (pass_manager::execute_pass_mode_switching): new.
11032 (finish_optimization_passes): Convert to...
11033 (pass_manager::finish_optimization_passes): ...this.
11034 (finish_optimization_passes): Update for conversion of passes to
11036 (register_dump_files_1): Use has_gate since we cannot portably
11037 check a vtable entry against NULL.
11038 (dump_one_pass): Likewise.
11039 (ipa_write_summaries_2): Likewise.
11040 (ipa_write_optimization_summaries_1): Likewise.
11041 (ipa_read_summaries_1): Likewise.
11042 (ipa_read_optimization_summaries_1): Likewise.
11043 (execute_ipa_stmt_fixups): Likewise.
11044 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11045 pass-creation functions rather than wiring up globals, and
11046 storing the results in fields of pass_manager generated using
11047 pass-instances.def.
11048 (pass_manager::dump_profile_report): Update for conversion of
11049 passes to C++ classes.
11050 (pass_manager::execute_ipa_summary_passes): Likewise.
11051 (execute_one_ipa_transform_pass): Likewise.
11052 (execute_one_pass): Use has_gate and has_execute since we cannot
11053 portably check a vtable entry against NULL.
11054 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11055 (pass_manager): Use pass-instances.def to add fields for the
11056 various pass instances.
11057 * toplev.c (finalize): Update for move of
11058 finish_optimization_passes to a method of gcc::pass_manager.
11059 * toplev.h (finish_optimization_passes): Move to method of class
11061 * tree-pass.h (struct pass_data): New.
11062 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11063 (opt_pass::gate): Convert to virtual function.
11064 (opt_pass::~opt_pass): New.
11065 (opt_pass::clone): New.
11066 (opt_pass::execute): Convert to virtual function.
11067 (opt_pass::opt_pass): New.
11068 (opt_pass::ctxt_): new.
11069 (gimple_opt_pass): Convert to subclass of opt_pass.
11070 (gimple_opt_pass::gimple_opt_pass): New.
11071 (rtl_opt_pass): Convert to subclass of opt_pass.
11072 (rtl_opt_pass::rtl_opt_pass): New.
11073 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11074 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11075 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11076 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11077 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11078 invocation of pass_mode_switching to reflect this moving from a
11079 global to a member of gcc::pass_manager.
11080 (ix86_option_override): Rework how pass_insert_vzeroupper is
11081 added to the pass_manager to reflect autogenerated changes.
11082 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11084 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11086 PR rtl-optimization/57708
11087 * recog.c (peep2_find_free_register): Validate all regs in a
11090 2013-08-05 Jan Hubicka <jh@suse.cz>
11093 * cgraph.c (verify_cgraph_node): Accept local flags from other
11095 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11096 (function_and_variable_visibility): Likewise.
11097 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11099 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11101 * graph.c (init_graph_slim_pretty_print): Remove.
11102 (print_graph_cfg): Do not call it. Use local pretty printer.
11103 (start_graph_dump): Likewise.
11105 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11107 * gimple-pretty-print.c (buffer): Remove.
11108 (initialized): Likewise.
11109 (maybe_init_pretty_print): Likewise.
11110 (print_gimple_stmt): Do not call it. Use non-static local
11111 pretty_printer variable.
11112 (print_gimple_expr): Likewise.
11113 (print_gimple_seq): Likewise.
11114 (gimple_dump_bb): Likewise.
11116 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11118 * asan.c (asan_pp): Remove.
11119 (asan_pp_initialized): Likewise.
11120 (asan_pp_initialize): Likewise.
11121 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11122 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11123 (asan_add_global): Likewise.
11125 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11127 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11128 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11129 * pretty-print.c (pp_formatted_text_data): Likewise.
11130 (pp_indent): Rename from pp_base_indent.
11131 (pp_format): Rename from pp_base_format.
11132 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11133 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11134 (pp_flush): Rename from pp_base_flush.
11135 (pp_set_line_maximum_length): Rename from
11136 pp_base_set_line_maximum_length.
11137 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11138 (pp_set_prefix): Rename from pp_base_set_prefix.
11139 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11140 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11141 (pp_append_text): Rename from pp_base_append_text.
11142 (pp_formatted_text): Rename from pp_base_formatted_text.
11143 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11144 (pp_remaining_character_count_for_line): Rename from
11145 pp_base_remaining_character_count_for_line.
11146 (pp_newline): Rename from pp_base_newline.
11147 (pp_character): Rename from pp_base_character.
11148 (pp_string): Rename from pp_base_string.
11149 (pp_maybe_space): Rename from pp_base_maybe_space.
11150 * asan.c (asan_pp_string): Adjust.
11151 (asan_emit_stack_protection): Likewise.
11152 (asan_add_global): Likewise.
11153 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11154 * tree-mudflap.c (mf_varname_tree): Likewise.
11155 * tree-pretty-print.c (pp_tree_identifier): Rename from
11156 pp_base_tree_identifier.
11157 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11158 Declare as function.
11160 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11162 * pretty-print.h (pp_bar_bar): New.
11163 (pp_ampersand_ampersand): Likewise.
11164 (pp_less_equal): Likewise.
11165 (pp_greater_equal): Likewise.
11166 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11167 printer functions instead of pp_string or operators and punctuators.
11168 (dump_gimple_call): Likewise.
11169 (dump_gimple_omp_for): Likewise.
11170 (dump_gimple_transaction): Likewise.
11171 (dump_gimple_phi): Likewise.
11172 (pp_gimple_stmt_1): Likewise.
11173 * sched-vis.c (print_insn): Likewise.
11174 * tree-mudflap.c (mf_varname_tree): Likewise.
11175 * tree-pretty-print.c (dump_block_node): Likewise.
11176 (dump_generic_node): Likewise.
11178 2013-08-02 Jan Hubicka <jh@suse.cz>
11180 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11182 * lto-streamer-out.c (tree_is_indexable): Results decls and
11183 parm decls are not indexable.
11184 (DFS_write_tree_body): Do not follow args and results.
11185 (hash_tree): Likewise.
11186 (output_functions): Rearrange so struct function is needed
11187 only when real body is output; be able to also ouptut abstract
11188 functions; output DECL_ARGUMENTS and DECL_RESULT.
11189 (lto_output): When not in WPA, ale store abstract functions.
11190 (write_symbol): Do not care about RESULT_DECL.
11191 (output_symbol_p): Handle correctly sbtract decls.
11192 * lto-streamer-in.c (input_function): Rearrange so struct
11193 function can be NULL at entry; allow streaming of
11194 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11195 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11196 sanity check during LTO.
11197 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11198 RESULT_DECl and DECL_ARGUMENTS.
11199 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11202 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11204 * pretty-print.h (pp_underscore): New.
11206 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11207 printer functions instead of pp_character.
11208 (dump_binary_rhs): Likewise.
11209 (dump_ternary_rhs): Likewise.
11210 (dump_gimple_call_args): Likewise.
11211 (pp_points_to_solution): Likewise.
11212 (dump_gimple_call): Likewise.
11213 (dump_gimple_switch): Likewise.
11214 (dump_gimple_cond): Likewise.
11215 (dump_gimple_bind): Likewise.
11216 (dump_gimple_try): Likewise.
11217 (dump_gimple_omp_for): Likewise.
11218 (dump_gimple_omp_continue): Likewise.
11219 (dump_gimple_omp_single): Likewise.
11220 (dump_gimple_omp_sections): Likewise.
11221 (dump_gimple_omp_block): Likewise.
11222 (dump_gimple_omp_critical): Likewise.
11223 (dump_gimple_transaction): Likewise.
11224 (dump_gimple_asm): Likewise.
11225 (dump_gimple_phi): Likewise.
11226 (dump_gimple_omp_parallel): Likewise.
11227 (dump_gimple_omp_task): Likewise.
11228 (dump_gimple_omp_atomic_load): Likewise.
11229 (dump_gimple_omp_atomic_store): Likewise.
11230 (dump_gimple_mem_ops): Likewise.
11231 (pp_gimple_stmt_1): Likewise.
11232 (pp_cfg_jump): Likewise.
11233 (dump_implicit_edges): Likewise.
11234 (gimple_dump_bb_for_graph): Likewise.
11235 * graph.c (draw_cfg_node): Likewise.
11236 * langhooks.c (lhd_print_error_function): Likewise.
11237 * sched-vis.c (print_exp): Likewise.
11238 (print_value): Likewise.
11239 (print_pattern): Likewise.
11240 (print_insn): Likewise.
11241 (rtl_dump_bb_for_graph): Likewise.
11242 * tree-pretty-print.c (dump_function_declaration): Likewise.
11243 (dump_array_domain): Likewise.
11244 (dump_omp_clause): Likewise.
11245 (dump_location): Likewise.
11246 (dump_generic_node): Likewise.
11247 (print_struct_decl): Likewise.
11248 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11250 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11252 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11253 candidate table when replacing a candidate statement.
11254 (replace_rhs_if_not_dup): Likewise.
11255 (replace_one_candidate): Likewise.
11257 2013-08-02 Jan Hubicka <jh@suse.cz>
11258 Martin Liska <marxin.liska@gmail.com>
11260 * cgraphunit.c (add_new_function): Fix logic when adding from
11262 (assemble_thunk): Rename to ...
11263 (expand_thunk); .. this one; export; get it working with
11264 general functions; make produced gimple valid.
11265 * cgraph.h (expand_thunk): Declare.
11267 2013-08-02 Jan Hubicka <jh@suse.cz>
11269 * ipa-cp.c (gather_context_independent_values): Use
11270 ipa_get_param_move_cost.
11271 (get_replacement_map): Remove PARAM; move parameter folding
11273 (create_specialized_node): Update.
11274 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11275 assert that we have gimple body; update move_cost.
11276 (count_formal_params): Assert that we have gimple body.
11277 (ipa_dump_param): New function.
11278 (ipa_alloc_node_params): Break out from ...
11279 (ipa_initialize_node_params): ... here.
11280 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11281 (ipa_write_node_info): Stream move costs.
11282 (ipa_read_node_info): Read move costs.
11283 (ipa_update_after_lto_read): Do not recompute node params.
11284 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11285 (ipa_get_param): Check we are not in WPA.
11286 (ipa_get_param_move_cost): New.
11287 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11288 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11289 parm numbers to be present.
11291 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11293 PR rtl-optimization/58048
11294 * lra-constraints.c (process_alt_operands): Don't check asm
11295 operand on register.
11297 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11299 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11300 the implied StoreLoad barrier for atomic operations if before.
11302 2013-08-02 Jan Hubicka <jh@suse.cz>
11303 Martin Liska <marxin.liska@gmail.com>
11305 * cgraph.c (cgraph_function_body_availability): Do not check
11307 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11308 symtab_node_availability): Declare.
11309 * ipa.c (can_replace_by_local_alias): New.
11310 (function_and_variable_visibility): Use it.
11311 * symtab.c (symtab_for_node_and_aliases,
11312 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11314 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11316 PR rtl-optimization/57963
11317 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11318 (lra_constraints): Use them.
11320 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11322 * config/arm/types.md (define_attr "type"): Add "load_acq"
11324 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11326 (cortex_a53_store1): Likewise.
11328 2013-08-01 Jan Hubicka <jh@suse.cz>
11330 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11331 partitions are not needed.
11333 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11335 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11336 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11337 MAYBE_NON_Q_CLASS_P where appropriate.
11339 2013-08-01 Jan Hubicka <jh@suse.cz>
11341 * cgraph.h (release_function_body): Declare.
11342 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11343 of unused delcarations.
11345 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11347 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11348 RTL form when subtracting a constant.
11350 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11352 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11353 Generate canonical plus rtx with negated immediate instead of minus
11355 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11357 2013-08-01 Jan Hubicka <jh@suse.cz>
11359 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11360 (cgraph_release_function_body): Likewise.
11361 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11362 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11363 to used_as_abstract_origin.
11364 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11365 symbols used as abstract origins.
11366 * cgraphunit.c (analyze_functions): Update.
11367 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11368 used_as_abstract_origin.
11369 * tree-inline.c (tree_function_versioning): Update
11370 used_as_abstract_origin; be ready for DECL_RESULT and
11371 DECL_ARGUMENTS to be NULL.
11373 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11374 for abstract functions.
11375 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11378 2013-08-01 Jan Hubicka <jh@suse.cz>
11380 * profile.c (compute_value_histograms): Fix thinko.
11382 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11384 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11385 aarch-common-protos.h to extra_headers.
11386 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11387 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11388 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11390 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11392 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11393 Include "../arm/types.md". Define "type" attribute for all patterns.
11394 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11397 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11399 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11400 to support power8 load fusion.
11401 (fusion_gpr_mem_load): Likewise.
11403 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11405 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11406 declarations for power8 load fusion.
11407 (emit_fusion_gpr_load): Likewise.
11409 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11410 tuning for power8, turn on fusion mode by default. Turn on sign
11411 extending fusion mode if normal fusion mode is on, and we are at
11413 (fusion_gpr_load_p): New function, return true if we can fuse an
11414 addis instruction with a dependent load to a GPR.
11415 (emit_fusion_gpr_load): Emit the instructions for power8 load
11418 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11419 (VSX load fusion peepholes): New peepholes to fuse together an
11420 addi instruction with a VSX load instruction.
11422 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11423 peepholes to fuse an addis instruction with a load to a GPR base
11424 register. If we are supporting sign extending fusions, convert
11425 sign extending loads to zero extending loads and add an explicit
11428 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11430 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11431 aarch-common-protos.h to extra_headers.
11432 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11433 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11434 (arm_early_store_addr_dep): Likewise.
11435 (arm_no_early_alu_shift_dep): Likewise.
11436 (arm_no_early_alu_shift_value_dep): Likewise.
11437 (arm_no_early_mul_dep): Likewise.
11438 (arm_no_early_store_addr_dep): Likewise.
11439 (arm_mac_accumulator_is_mul_result): Likewise.
11440 (arm_mac_accumulator_is_result): Likewise.
11441 * config/arm/aarch-common.c: ... here. New file.
11442 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11444 (arm_early_store_addr_dep): Likewise.
11445 (arm_no_early_alu_shift_dep): Likewise.
11446 (arm_no_early_alu_shift_value_dep): Likewise.
11447 (arm_no_early_mul_dep): Likewise.
11448 (arm_no_early_store_addr_dep): Likewise.
11449 (arm_mac_accumulator_is_mul_result): Likewise.
11450 (arm_mac_accumulator_is_result): Likewise.
11451 * config/arm/aarch-common-protos.h: ... here. New file.
11452 * config/arm/t-arm (aarch-common.o): Define.
11454 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11456 * config/arm/arm.md: Include new file "types.md".
11457 (define_attr "type"): Move from here to ...
11458 (define_attr "mul32"): Likewise.
11459 (define_attr "mul64"): Likewise.
11460 * config/arm/types.md: ... here. New file.
11462 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11464 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11465 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11467 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11469 * gen-pass-instances.awk: Fix offset of substr().
11471 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11473 * Makefile.in (pass-instances.def): New.
11474 (passes.o): Replace dependency on passes.def with one on
11477 * gen-pass-instances.awk: New.
11479 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11480 rather than passes.def, updating local definition of NEXT_PASS
11481 macro to add an extra NUM parameter (currently unused).
11483 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11485 * Makefile.in (PASS_MANAGER_H): New.
11486 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11487 (passes.o): Likewise.
11488 (statistics.o): Likewise.
11489 (cgraphunit.o): Likewise.
11490 (context.o): Depend on PASS_MANAGER_H.
11492 * pass_manager.h: New.
11494 * cgraphunit.c (cgraph_add_new_function): Update for moves
11495 of globals to fields of pass_manager.
11496 (analyze_function): Likewise.
11497 (expand_function): Likewise.
11498 (ipa_passes): Likewise.
11499 (compile): Likewise.
11501 * context.c (context::context): New.
11502 * context.h (context::context): New.
11503 (context::get_passes): New.
11504 (context::passes_): New.
11506 * lto-cgraph.c (input_node): Update for moves of globals to
11507 fields of pass_manager.
11509 * passes.c (all_passes): Remove, in favor of a field of the
11510 same name within the new class pass_manager.
11511 (all_small_ipa_passes): Likewise.
11512 (all_lowering_passes): Likewise.
11513 (all_regular_ipa_passes): Likewise.
11514 (all_late_ipa_passes): Likewise.
11515 (all_lto_gen_passes): Likewise.
11516 (passes_by_id): Likewise.
11517 (passes_by_id_size): Likewise.
11518 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11519 the new class pass_manager.
11520 (set_pass_for_id): Convert to...
11521 (pass_manager::set_pass_for_id): ...method.
11522 (get_pass_for_id): Convert to...
11523 (pass_manager::get_pass_for_id): ...method.
11524 (register_one_dump_file): Move body of implementation into...
11525 (pass_manager::register_one_dump_file): ...here.
11526 (register_dump_files_1): Convert to...
11527 (pass_manager::register_dump_files_1): ...method.
11528 (register_dump_files): Convert to...
11529 (pass_manager::register_dump_files): ...method.
11530 (create_pass_tab): Update for moves of globals to fields of
11532 (dump_passes): Move body of implementation into...
11533 (pass_manager::dump_passes): ...here.
11534 (register_pass): Move body of implementation into...
11535 (pass_manager::register_pass): ...here.
11536 (init_optimization_passes): Convert into...
11537 (pass_manager::pass_manager): ...constructor for new
11538 pass_manager class, and initialize the pass_lists array.
11539 (check_profile_consistency): Update for moves of globals to
11540 fields of pass_manager.
11541 (dump_profile_report): Move body of implementation into...
11542 (pass_manager::dump_profile_report): ...here.
11543 (ipa_write_summaries_1): Update for moves of pass lists from
11544 being globals to fields of pass_manager.
11545 (ipa_write_optimization_summaries): Likewise.
11546 (ipa_read_summaries): Likewise.
11547 (ipa_read_optimization_summaries): Likewise.
11548 (execute_all_ipa_stmt_fixups): Likewise.
11550 * statistics.c (statistics_fini): Update for moves of globals to
11551 fields of pass_manager.
11553 * toplev.c (general_init): Replace call to
11554 init_optimization_passes with construction of the pass_manager
11557 * tree-pass.h (all_passes): Remove, in favor of a field of the
11558 same name within the new class pass_manager.
11559 (all_small_ipa_passes): Likewise.
11560 (all_lowering_passes): Likewise.
11561 (all_regular_ipa_passes): Likewise.
11562 (all_lto_gen_passes): Likewise.
11563 (all_late_ipa_passes): Likewise.
11564 (passes_by_id): Likewise.
11565 (passes_by_id_size): Likewise.
11566 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11567 the new class pass_manager.
11568 (get_pass_for_id): Remove.
11570 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11572 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11575 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11577 * arm.md (mulhi3): New expand pattern.
11579 2013-07-30 Jan Hubicka <jh@suse.cz>
11580 Martin Liska <marxin.liska@gmail.com>
11582 * profile.c (compute_value_histograms): Do not ICE when
11583 there is mismatch only on some counters.
11585 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11587 PR rtl-optimization/57637
11588 * function.c (move_insn_for_shrink_wrap): Also check the
11589 GEN set of the LIVE problem for the liveness analysis
11590 if it exists, otherwise give up.
11592 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11594 PR tree-optimization/57993
11595 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11596 replaced statement in the candidate table.
11597 (phi_add_costs): Return infinite cost when the hidden basis does
11598 not dominate all phis on which the candidate is dependent.
11599 (replace_one_candidate): Record replaced statement in the
11602 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11604 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11605 (ashlv2si3): New expander.
11606 (*ashlv2si3_i): New define_insn_and_split.
11607 * predicates.md (float_operation): Allow patterns with three
11608 basic sub-patterns.
11610 PR rtl-optimization/58021
11611 * mode-switching.c (create_pre_exit): Always split off preceding
11612 insns if we are not at the basic block head.
11614 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11616 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11617 (UCLIBC_DYNAMIC_LINKER): New macro.
11618 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11620 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11621 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11623 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11624 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11625 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11626 for SF and DF modes. Use ieee_quad_format for TF mode.
11627 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11628 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11629 (mips_option_override): Handle `-mnan=legacy'.
11630 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11631 `-mabs=2008' and `-mnan=2008'.
11632 (OPTION_DEFAULT_SPECS): Add "nan" default.
11633 (ASM_SPEC): Handle `-mnan='.
11634 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11635 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11636 comment accordingly.
11637 (neg<mode>2): Likewise.
11638 * config/mips/mips.opt (mabs, mnan): New options.
11639 * doc/install.texi (Configuration): Document `--with-nan=' option.
11640 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11642 (MIPS Options): Document them.
11643 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11644 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11645 * configure: Regenerate.
11646 * config.in: Regenerate.
11648 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11650 * config/i386/i386.md (float post-reload splitters): Do not check
11651 for subregs of SSE registers.
11653 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11654 H.J. Lu <hongjiu.lu@intel.com>
11658 * config/i386/i386.md (post-reload splitter
11659 to avoid partial SSE reg dependency stalls): New pattern.
11661 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11663 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11665 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11667 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11668 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11670 2013-07-26 Cary Coutant <ccoutant@google.com>
11672 * dwarf2out.c (die_checksum_ordered): Don't include template
11673 instantiations in signature.
11674 (is_template_parameter): New function.
11675 (is_template_instantiation): New function.
11676 (generate_skeleton_bottom_up): Don't include template instantiations
11678 (generate_skeleton): Likewise.
11679 (break_out_comdat_types): Move recursive call to break out nested
11681 (prune_unused_types_mark_generic_parms_dies): Call
11682 is_template_parameter.
11684 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11686 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11687 uses vector registers.
11688 * config/aarch64/iterators.md: Add attributes rtn and vas.
11690 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11691 Richard Earnshaw <richard.earnshaw@arm.com>
11693 * combine.c (simplify_comparison): Re-canonicalize operands
11695 * config/arm/arm.md (movcond_addsi): New splitter.
11697 2013-07-25 Sterling Augustine <saugustine@google.com>
11699 * dwarf2out.c (size_of_pubnames): Move code to...
11700 (include_pubname_in_output): ...here. New.
11701 (want_pubnames): Rearrange.
11702 (output_pubnames): Call include_pubname_in_output. Move assertion.
11704 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11706 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11708 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11711 * doc/extend.texi: Remove obsolete builtins. Fix
11712 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11714 2013-07-25 Jan Hubicka <jh@suse.cz>
11716 * cgraph.c (release_function_body): Break out from ...
11717 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11718 and DECL_ARGUMENTS.
11719 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11720 old_tree in the map.
11721 (create_specialized_node): Update.
11722 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11724 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11725 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11726 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11727 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11730 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11732 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11733 addsi3_carryin_alt2_<optab>): Correct output template.
11735 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11737 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11738 Adjust for arm_restrict_it.
11739 Remove trailing whitespace.
11741 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11743 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11744 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11746 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11748 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11750 PR rtl-optimization/57960
11751 * lra-constraints.c (process_alt_operands): Use the right mode
11752 when checking strict_low.
11754 2013-07-25 Jan Hubicka <jh@suse.cz>
11756 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11757 * cgraph.c (cgraph_remove_node): Do not release function body
11758 when in cgraph streaming.
11759 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11760 in other partitions are not considered reachable; fix handling of
11763 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11765 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11767 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11772 * config/arm/arm.md ("*sibcall_insn): Replace use of
11773 Ss with US. Adjust output for v5 and v4t.
11774 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11776 * config/arm/constraints.md ("Ss"): Rename to US.
11778 2013-07-25 Terry Guo <terry.guo@arm.com>
11780 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11781 shift_add/shift_sub0/shift_sub1 RTXs.
11783 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11784 Anton Blanchard <anton@au1.ibm.com>
11786 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11787 (altivec_vpks<VI_char>ss): Likewise.
11788 (altivec_vpks<VI_char>us): Likewise.
11789 (altivec_vpku<VI_char>us): Likewise.
11790 (altivec_vpku<VI_char>um): Likewise.
11792 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11794 Introduce context class.
11796 * Makefile.in (CONTEXT_H): New.
11797 (OBJS): Add context.o.
11798 (toplev.o): Add CONTEXT_H to dependencies.
11801 * toplev.c (general_init): Create the singleton gcc::context instance.
11807 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11809 PR rtl-optimization/57968
11810 * mode-switching.c (create_pre_exit): Allow instructions that
11811 don't set a return register to need a non-exit mode.
11813 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11814 Anton Blanchard <anton@au1.ibm.com>
11816 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11817 operands to vperm for little endian.
11818 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11819 of lvsl to create the control mask for a vperm for little endian.
11821 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11822 Anton Blanchard <anton@au1.ibm.com>
11824 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11825 two operands for little-endian.
11827 2013-07-23 Steve Ellcey <sellcey@mips.com>
11829 * config/mips/mips.c (mips_case_values_threshold): New.
11830 (TARGET_CASE_VALUES_THRESHOLD): Define.
11832 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11833 Anton Blanchard <anton@au1.ibm.com>
11835 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11836 selection of field for vector splat in little endian mode.
11838 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11840 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11841 expanders to rs6000.md.
11842 (ior<mode>3): Likewise.
11843 (and<mode>3): Likewise.
11844 (one_cmpl<mode>2): Likewise.
11845 (nor<mode>3): Likewise.
11846 (andc<mode>3): Likewise.
11847 (eqv<mode>3): Likewise.
11848 (nand<mode>3): Likewise.
11849 (orc<mode>3): Likewise.
11851 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11854 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11855 to split multi-word logical operations.
11856 (rs6000_split_logical_di): Likewise.
11857 (rs6000_split_logical): Likewise.
11859 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11860 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11861 and allow TImode operations in 32-bit.
11862 (vsx_and<mode>3_64bit): Likewise.
11863 (vsx_ior<mode>3_32bit): Likewise.
11864 (vsx_ior<mode>3_64bit): Likewise.
11865 (vsx_xor<mode>3_32bit): Likewise.
11866 (vsx_xor<mode>3_64bit): Likewise.
11867 (vsx_one_cmpl<mode>2_32bit): Likewise.
11868 (vsx_one_cmpl<mode>2_64bit): Likewise.
11869 (vsx_nor<mode>3_32bit): Likewise.
11870 (vsx_nor<mode>3_64bit): Likewise.
11871 (vsx_andc<mode>3_32bit): Likewise.
11872 (vsx_andc<mode>3_64bit): Likewise.
11873 (vsx_eqv<mode>3_32bit): Likewise.
11874 (vsx_eqv<mode>3_64bit): Likewise.
11875 (vsx_nand<mode>3_32bit): Likewise.
11876 (vsx_nand<mode>3_64bit): Likewise.
11877 (vsx_orc<mode>3_32bit): Likewise.
11878 (vsx_orc<mode>3_64bit): Likewise.
11880 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11881 logical types in GPRs.
11883 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11884 logical insns to rs6000.md, and allow TImode operations in
11886 (altivec_ior<mode>3): Likewise.
11887 (altivec_xor<mode>3): Likewise.
11888 (altivec_one_cmpl<mode>2): Likewise.
11889 (altivec_nor<mode>3): Likewise.
11890 (altivec_andc<mode>3): Likewise.
11892 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11893 attributes for moving the 128-bit logical operations into
11895 (BOOL_REGS_OUTPUT): Likewise.
11896 (BOOL_REGS_OP1): Likewise.
11897 (BOOL_REGS_OP2): Likewise.
11898 (BOOL_REGS_UNARY): Likewise.
11899 (BOOL_REGS_AND_CR0): Likewise.
11900 (one_cmpl<mode>2): Add support for DI logical operations on
11901 32-bit, splitting the operations to 32-bit.
11902 (anddi3): Likewise.
11903 (iordi3): Likewise.
11904 (xordi3): Likewise.
11905 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11906 changes to combine the 32/64-bit code, allow logical operations on
11907 TI mode in 32-bit, and to use similar match_operator patterns like
11908 scalar mode uses. Combine the Altivec and VSX code for logical
11909 operations, and move it here.
11910 (ior<mode>3, 128-bit types): Likewise.
11911 (xor<mode>3, 128-bit types): Likewise.
11912 (one_cmpl<mode>3, 128-bit types): Likewise.
11913 (nor<mode>3, 128-bit types): Likewise.
11914 (andc<mode>3, 128-bit types): Likewise.
11915 (eqv<mode>3, 128-bit types): Likewise.
11916 (nand<mode>3, 128-bit types): Likewise.
11917 (orc<mode>3, 128-bit types): Likewise.
11918 (and<mode>3_internal): Likewise.
11919 (bool<mode>3_internal): Likewise.
11920 (boolc<mode>3_internal1): Likewise.
11921 (boolc<mode>3_internal2): Likewise.
11922 (boolcc<mode>3_internal1): Likewise.
11923 (boolcc<mode>3_internal2): Likewise.
11924 (eqv<mode>3_internal1): Likewise.
11925 (eqv<mode>3_internal2): Likewise.
11926 (one_cmpl1<mode>3_internal): Likewise.
11928 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11930 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11931 Rename flag_stack_usage to flag_stack_usage_info.
11933 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11935 * config/microblaze/sync.md: New file.
11936 * config/microblaze/microblaze.md: Include sync.md
11937 * config/microblaze/microblaze.c: Add print_operand 'y'.
11938 * config/microblaze/constraints.md: Add memory_contraint
11939 'Q' which is a single register.
11941 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11943 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11945 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11947 * reload.c (find_reloads): Exit loop once we find this operand
11948 cannot be reloaded somehow for this alternative.
11950 * reload.c (find_reloads): Exit loop once we find a hard register.
11952 * rtlanal.c (computed_jump_p): Exit loop once we find label
11955 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11957 * cfgloopmanip.c (remove_path): Exit loop after setting
11960 * gensupport.c (subst_dup): Avoid loop if code is not
11961 MATCH_DUP nor MATCH_OP_DUP.
11963 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11965 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11967 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11969 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11970 true for SP_REGNUM if mode == ptr_mode.
11971 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11972 with value R0_REGNUM + 31.
11974 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11976 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11977 pad pointer-typed argument downward.
11979 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11981 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11982 and __ILP32__ when the ILP32 model is in use.
11984 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11986 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11987 (aarch64_load_symref_appropriately): In the case of
11988 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11989 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11990 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11991 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11992 if 'mode' doesn't equal to 'ptr_mode'.
11993 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11994 'vcall_offset'; change to call aarch64_emit_move differently depending
11995 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11996 to calculate the upper bound of 'vcall_offset'.
11997 (aarch64_cannot_force_const_mem): Change to also return true if
11999 (aarch64_legitimize_reload_address): In the case of large
12000 displacements, add new local variable 'xmode' and an assertion
12001 based on it; change to use 'xmode' to generate the new rtx and
12003 (aarch64_asm_trampoline_template): Change to generate the template
12004 differently depending on TARGET_ILP32 or not; change to use
12005 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12006 (aarch64_trampoline_size): Removed.
12007 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12008 and replace immediate literals with it. Change to use 'ptr_mode'
12009 instead of 'DImode' and call convert_memory_address if the mode
12010 of 'fnaddr' doesn't equal to 'ptr_mode'.
12011 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12013 (aarch64_elf_asm_destructor): Likewise.
12014 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12015 on TARGET_ILP32 instead of aarch64_trampoline_size.
12016 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12017 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12018 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12019 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12020 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12021 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12022 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12023 depending on the value of 'mode'.
12024 (add_losym_<mode>): New.
12025 (ldr_got_small_<mode>): New, based on ldr_got_small.
12026 (ldr_got_small): Remove.
12027 (ldr_got_small_sidi): New.
12028 * config/aarch64/iterators.md (P): New.
12029 (PTR): Change to 'ptr_mode' in the condition.
12031 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12033 * config.gcc (aarch64*-*-*): Support --with-abi.
12034 (aarch64*-*-elf): Support --with-multilib-list.
12035 (aarch64*-*-linux*): Likewise.
12036 (supported_defaults): Add abi to aarch64*-*-*.
12037 * configure.ac: Mention AArch64 for --with-multilib-list.
12038 * configure: Re-generated.
12039 * config/aarch64/biarchilp32.h: New file.
12040 * config/aarch64/biarchlp64.h: New file.
12041 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12043 (MULTILIB_DEFAULTS): Ditto.
12044 (DRIVER_SELF_SPECS): Ditto.
12045 (ASM_SPEC): Update to also substitute -mabi.
12046 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12047 file whose name depends on -mabi= and -mbig-endian.
12048 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12050 (POINTER_SIZE): New define.
12051 (POINTERS_EXTEND_UNSIGNED): Ditto.
12052 (enum aarch64_abi_type): New enumeration tag.
12053 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12054 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12055 (TARGET_ILP32): New define.
12056 * config/aarch64/aarch64.opt (mabi): New.
12057 (aarch64_abi): New.
12058 (ilp32, lp64): New values for -mabi.
12059 * config/aarch64/t-aarch64 (comma): New define.
12060 (MULTILIB_OPTIONS): Ditto.
12061 (MULTILIB_DIRNAMES): Ditto.
12062 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12063 * doc/invoke.texi: Document -mabi for AArch64.
12065 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12067 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12069 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12070 Anton Blanchard <anton@au1.ibm.com>
12072 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12073 endianness when selecting field to splat.
12075 2013-07-22 Eric Christopher <echristo@gmail.com>
12077 * dwarf2out.c (die_odr_checksum): New function to use
12078 CHECKSUM_ macros and ULEB128 for DIE tag.
12079 (generate_type_signature): Use.
12081 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12083 * config.gcc (sparc*-*-*): Accept leon3 processor.
12084 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12085 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12086 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12087 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12088 (mfix-ut699): Adjust comment.
12089 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12090 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12091 (CPP_CPU_SPEC): Likewise.
12092 (ASM_CPU_SPEC): Likewise.
12093 * config/sparc/sparc.c (leon3_cost): New constant.
12094 (sparc_option_override): Add leon3 support.
12095 (mem_ref): New function.
12096 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12097 (sparc_do_work_around_errata): Look into the instruction in the delay
12098 slot and adjust accordingly. Add fix for the data cache nullify issues
12099 of the UT699. Change insertion position for the NOP.
12100 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12101 (leon3_load): New reservation.
12102 (leon_store): Bump latency to 2.
12103 (grfpu): New automaton.
12104 (grfpu_alu): New unit.
12105 (grfpu_ds): Likewise.
12106 (leon_fp_alu): Adjust.
12107 (leon_fp_mult): Delete.
12108 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12109 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12110 * config/sparc/sparc.md (cpu): Add leon3.
12111 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12112 (swapsi): Likewise.
12113 (atomic_test_and_set): Likewise.
12114 (ldstub): Likewise.
12116 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12118 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12119 default for R5900 targets.
12120 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12121 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12122 * config/mips/mips.c (mips_option_override): Report an error for
12123 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12124 for -march=r5900 -mhard-float.
12126 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12128 * df-problems.c (can_move_insns_across): Exit loop once we
12129 find a non-fixed, non-global register.
12131 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12134 * omega.c (omega_eliminate_red): Break after setting red_found.
12135 (omega_problem_has_red_equations): Similarly after setting found.
12136 (omega_query_variable): Similarly after setting coupled.
12138 2013-07-22 Marek Polacek <polacek@redhat.com>
12140 * gimplify.c: Don't include gimple.h twice.
12142 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12144 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12145 instead of TARGET_THUMB1.
12146 (Pz): New constraint.
12147 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12149 (compare_negsi_si): Likewise.
12150 (compare_addsi2_op0): Likewise.
12151 (compare_addsi2_op1): Likewise.
12152 (addsi3_carryin_<optab>): Likewise.
12153 (addsi3_carryin_alt2_<optab>): Likewise.
12154 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12155 for arm_restrict_it.
12156 (subsi3_carryin): Likewise.
12157 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12158 (minmax_arithsi): Disable for arm_restrict_it.
12159 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12160 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12161 (satsi_<SAT:code>_shift): Likewise.
12162 (arm_shiftsi3): Add alternative for 16-bit encoding.
12163 (arm32_movhf): Disable for arm_restrict_it.
12164 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12165 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12167 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12169 * config/arm/arm.md (attribute "insn"): Delete.
12170 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12171 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12172 (not_shiftsi): Update for attribute change.
12173 (not_shiftsi_compare0): Likewise.
12174 (not_shiftsi_compare0_scratch): Likewise.
12175 (arm_one_cmplsi2): Likewise.
12176 (thumb1_one_cmplsi2): Likewise.
12177 (notsi_compare0): Likewise.
12178 (notsi_compare0_scratch): Likewise.
12179 (thumb1_movdi_insn): Likewise.
12180 (arm_movsi_insn): Likewise.
12181 (movhi_insn_arch4): Likewise.
12182 (movhi_bytes): Likewise.
12183 (arm_movqi_insn): Likewise.
12184 (thumb1_movqi_insn): Likewise.
12185 (arm32_movhf): Likewise.
12186 (thumb1_movhf): Likewise.
12187 (arm_movsf_soft_insn): Likewise.
12188 (thumb1_movsf_insn): Likewise.
12189 (thumb_movdf_insn): Likewise.
12190 (movsicc_insn): Likewise.
12191 (movsfcc_soft_insn): Likewise.
12192 (and_scc): Likewise.
12193 (cond_move): Likewise.
12194 (if_move_not): Likewise.
12195 (if_not_move): Likewise.
12196 (if_shift_move): Likewise.
12197 (if_move_shift): Likewise.
12198 (if_shift_shift): Likewise.
12199 (if_not_arith): Likewise.
12200 (if_arith_not): Likewise.
12201 (cond_move_not): Likewise.
12202 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12203 (neon_mov<mode>): Likewise.
12204 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12205 (thumb2_movsi_vfp): Likewise.
12206 (movsf_vfp): Likewise.
12207 (thumb2_movsf_vfp): Likewise.
12208 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12210 (cortexa7_older_only): Likewise.
12211 (cortexa7_younger): Likewise.
12212 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12213 (1020alu_shift_op): Likewise.
12214 (1020alu_shift_reg_op): Likewise.
12215 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12216 (alu_shift_op): Likewise.
12217 (alu_shift_reg_op): Likewise.
12218 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12219 (11_alu_shift_op): Likewise.
12220 (11_alu_shift_reg_op): Likewise.
12221 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12222 (9_alu_shift_reg_op): Likewise.
12223 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12225 (cortex_a15_alu_shift): Likewise.
12226 (cortex_a15_alu_shift_reg): Likewise.
12227 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12228 (cortex_a5_alu_shift): Likewise.
12229 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12231 (cortex_a53_alu_shift): Likewise.
12232 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12234 (cortex_a7_alu_reg): Likewise.
12235 (cortex_a7_alu_shift): Likewise.
12236 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12237 (cortex_a8_alu_shift): Likewise.
12238 (cortex_a8_alu_shift_reg): Likewise.
12239 (cortex_a8_mov): Likewise.
12240 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12241 (cortex_a9_dp_shift): Likewise.
12242 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12243 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12244 (cortex_r4_mov): Likewise.
12245 (cortex_r4_alu_shift): Likewise.
12246 (cortex_r4_alu_shift_reg): Likewise.
12247 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12248 (526_alu_shift_op): Likewise.
12249 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12250 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12251 (626te_alu_shift_op): Likewise.
12252 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12253 (726te_alu_op): Likewise.
12254 (726te_alu_shift_op): Likewise.
12255 (726te_alu_shift_reg_op): Likewise.
12256 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12257 (mp626_alu_shift_op): Likewise.
12258 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12259 (pj4_alu_e1_conds): Likewise.
12260 (pj4_alu): Likewise.
12261 (pj4_alu_conds): Likewise.
12262 (pj4_shift): Likewise.
12263 (pj4_shift_conds): Likewise.
12264 (pj4_alu_shift): Likewise.
12265 (pj4_alu_shift_conds): Likewise.
12267 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12269 * config/arm/predicates.md (shiftable_operator_strict_it):
12271 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12272 Disable cond_exec version for arm_restrict_it.
12273 (thumb2_smaxsi3): Convert to generate cond_exec.
12274 (thumb2_sminsi3): Likewise.
12275 (thumb32_umaxsi3): Likewise.
12276 (thumb2_uminsi3): Likewise.
12277 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12278 (thumb2_neg_abssi2): Likewise.
12279 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12280 (thumb2_movsicc_insn): Adjust alternatives.
12281 (thumb2_mov_negscc): Disable for arm_restrict_it.
12282 (thumb2_mov_negscc_strict_it): New pattern.
12283 (thumb2_mov_notscc_strict_it): New pattern.
12284 (thumb2_mov_notscc): Disable for arm_restrict_it.
12285 (thumb2_ior_scc): Likewise.
12286 (thumb2_ior_scc_strict_it): New pattern.
12287 (thumb2_cond_move): Adjust for arm_restrict_it.
12288 (thumb2_cond_arith): Disable for arm_restrict_it.
12289 (thumb2_cond_arith_strict_it): New pattern.
12290 (thumb2_cond_sub): Adjust for arm_restrict_it.
12291 (thumb2_movcond): Likewise.
12292 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12293 (thumb2_zero_extendhisi2_v6): Likewise.
12294 (thumb2_zero_extendqisi2_v6): Likewise.
12295 (orsi_notsi_si): Likewise.
12296 (orsi_not_shiftsi_si): Likewise.
12298 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12300 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12301 instruction sequence is 1 byte shorter.
12303 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12305 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12306 it is not needed after split.
12308 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12311 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12312 second label for nonlocal goto receivers. Don't output pic base labels
12313 unless we're producing PIC; mark that action unreachable().
12314 (ix86_save_reg): If the function contains a nonlocal label, save the
12316 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12317 * config/darwin.c (emitted_pic_label_num): New GTY.
12318 (update_pic_label_number_if_needed): New.
12319 (machopic_output_function_base_name): Adjust for nonlocal receiver
12321 (machopic_should_output_picbase_label): New.
12322 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12323 (nonlocal_goto_receiver): New insn and split.
12325 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12327 * config/aarch64/aarch64-builtins.c
12328 (aarch64_fold_builtin): Fold abs in all modes.
12329 * config/aarch64/aarch64-simd-builtins.def
12330 (abs): Enable for all modes.
12331 * config/aarch64/arm_neon.h
12332 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12333 (vabs_f64): Add missing intrinsic.
12335 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12337 * config/aarch64/arm_neon.h (vabs_s64): New function
12339 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12342 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12343 * config/avr/avr.md (adjust_len): Add `round'.
12344 * config/avr/avr-protos.h (avr_out_round): New prototype.
12345 (avr_out_plus): Add `out_label' argument.
12346 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12347 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12348 Handle the case where `insn' is just a pattern.
12349 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12350 (avr_out_round): New function.
12351 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12353 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12355 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12356 Add check for flag_stack_usage to handle -fstack-usage support
12358 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12360 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12361 interaction for new Power8 flags and VSX.
12363 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12365 PR middle-end/57698
12366 * tree-inline.c (expand_call_inline): Emit errors during
12367 early_inlining only if optimization is not turned on.
12369 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12373 * passes.c (init_optimization_passes): Move the construction of
12374 the pass hierarchy into a new passes.def file.
12376 * Makefile.in (passes.o): Add dependency on passes.def.
12378 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12380 * passes.c (init_optimization_passes): Introduce macros for
12381 constructing the tree of passes (INSERT_PASSES_AFTER,
12382 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12383 TERMINATE_PASS_LIST).
12385 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12386 Wei Mi <wmi@google.com>
12388 PR rtl-optimization/57878
12389 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12391 (reload_pseudo_compare_func): Check nregs first for reload
12394 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12396 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12398 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12400 * read-rtl.c (validate_const_int): Once an invalid character is
12401 seen, quit the loop.
12403 * gengtype.c (write_roots): Similarly once we find the "deletable"
12404 or "if_marked" option.
12406 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12408 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12409 "xtab" and "sat". Move value "clz" from here to ...
12410 (attriubte "type"): ... here.
12411 (satsi_<SAT:code>): Delete "insn" attribute.
12412 (satsi_<SAT:code>_shift): Likewise.
12413 (arm_zero_extendqisi2addsi): Likewise.
12414 (arm_extendqisi2addsi): Likewise.
12415 (clzsi2): Update for attribute changes.
12416 (rbitsi2): Likewise.
12417 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12419 (arm_usatsihi): Likewise.
12420 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12422 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12424 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12425 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12426 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12427 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12428 in alphabetical order.
12429 (attribute "core_cycles"): Update for attribute changes.
12430 (arm_addsi3): Likewise.
12431 (addsi3_compare0): Likewise.
12432 (addsi3_compare0_scratch): Likewise.
12433 (addsi3_compare_op1): Likewise.
12434 (addsi3_compare_op2): Likewise.
12435 (compare_addsi2_op0): Likewise.
12436 (compare_addsi2_op1): Likewise.
12437 (addsi3_carryin_shift_<optab>): Likewise.
12438 (subsi3_carryin_shift): Likewise.
12439 (rsbsi3_carryin_shift): Likewise.
12440 (arm_subsi3_insn): Likewise.
12441 (subsi3_compare0): Likewise.
12442 (subsi3_compare): Likewise.
12443 (arm_andsi3_insn): Likewise.
12444 (thumb1_andsi3_insn): Likewise.
12445 (andsi3_compare0): Likewise.
12446 (andsi3_compare0_scratch): Likewise.
12447 (zeroextractsi_compare0_scratch
12448 (andsi_not_shiftsi_si): Likewise.
12449 (iorsi3_insn): Likewise.
12450 (iorsi3_compare0): Likewise.
12451 (iorsi3_compare0_scratch): Likewise.
12452 (arm_xorsi3): Likewise.
12453 (thumb1_xorsi3_insn): Likewise.
12454 (xorsi3_compare0): Likewise.
12455 (xorsi3_compare0_scratch): Likewise.
12456 (satsi_<SAT:code>_shift): Likewise.
12458 (arm_shiftsi3): Likewise.
12459 (shiftsi3_compare0): Likewise.
12460 (not_shiftsi): Likewise.
12461 (not_shiftsi_compare0): Likewise.
12462 (not_shiftsi_compare0_scratch): Likewise.
12463 (arm_one_cmplsi2): Likewise.
12464 (thumb_one_complsi2): Likewise.
12465 (notsi_compare0): Likewise.
12466 (notsi_compare0_scratch): Likewise.
12467 (thumb1_zero_extendhisi2): Likewise.
12468 (arm_zero_extendhisi2): Likewise.
12469 (arm_zero_extendhisi2_v6): Likewise.
12470 (arm_zero_extendhisi2addsi): Likewise.
12471 (thumb1_zero_extendqisi2): Likewise.
12472 (thumb1_zero_extendqisi2_v6): Likewise.
12473 (arm_zero_extendqisi2): Likewise.
12474 (arm_zero_extendqisi2_v6): Likewise.
12475 (arm_zero_extendqisi2addsi): Likewise.
12476 (thumb1_extendhisi2): Likewise.
12477 (arm_extendhisi2): Likewise.
12478 (arm_extendhisi2_v6): Likewise.
12479 (arm_extendqisi): Likewise.
12480 (arm_extendqisi_v6): Likewise.
12481 (arm_extendqisi2addsi): Likewise.
12482 (thumb1_extendqisi2): Likewise.
12483 (thumb1_movdi_insn): Likewise.
12484 (arm_movsi_insn): Likewise.
12485 (movsi_compare0): Likewise.
12486 (movhi_insn_arch4): Likewise.
12487 (movhi_bytes): Likewise.
12488 (arm_movqi_insn): Likewise.
12489 (thumb1_movqi_insn): Likewise.
12490 (arm32_movhf): Likewise.
12491 (thumb1_movhf): Likewise.
12492 (arm_movsf_soft_insn): Likewise.
12493 (thumb1_movsf_insn): Likewise.
12494 (movdf_soft_insn): Likewise.
12495 (thumb_movdf_insn): Likewise.
12496 (arm_cmpsi_insn): Likewise.
12497 (cmpsi_shiftsi): Likewise.
12498 (cmpsi_shiftsi_swp): Likewise.
12499 (arm_cmpsi_negshiftsi_si): Likewise.
12500 (movsicc_insn): Likewise.
12501 (movsfcc_soft_insn): Likewise.
12502 (arith_shiftsi): Likewise.
12503 (arith_shiftsi_compare0
12504 (arith_shiftsi_compare0_scratch
12505 (sub_shiftsi): Likewise.
12506 (sub_shiftsi_compare0
12507 (sub_shiftsi_compare0_scratch
12508 (and_scc): Likewise.
12509 (cond_move): Likewise.
12510 (if_plus_move): Likewise.
12511 (if_move_plus): Likewise.
12512 (if_move_not): Likewise.
12513 (if_not_move): Likewise.
12514 (if_shift_move): Likewise.
12515 (if_move_shift): Likewise.
12516 (if_shift_shift): Likewise.
12517 (if_not_arith): Likewise.
12518 (if_arith_not): Likewise.
12519 (cond_move_not): Likewise.
12520 (thumb1_ashlsi3): Set type attribute.
12521 (thumb1_ashrsi3): Likewise.
12522 (thumb1_lshrsi3): Likewise.
12523 (thumb1_rotrsi3): Likewise.
12524 (shiftsi3_compare0_scratch): Likewise.
12525 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12526 (neon_mov<mode>): Likewise.
12527 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12529 (thumb2_movsi_insn): Likewise.
12530 (thumb2_cmpsi_neg_shiftsi): Likewise.
12531 (thumb2_extendqisi_v6): Likewise.
12532 (thumb2_zero_extendhisi2_v6): Likewise.
12533 (thumb2_zero_extendqisi2_v6): Likewise.
12534 (thumb2_shiftsi3_short): Likewise.
12535 (thumb2_addsi3_compare0_scratch): Likewise.
12536 (orsi_not_shiftsi_si): Likewise.
12537 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12538 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12540 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12541 (1020alu_shift_op): Likewise.
12542 (1020alu_shift_reg_op): Likewise.
12543 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12544 (alu_shift_op): Likewise.
12545 (alu_shift_reg_op): Likewise.
12546 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12547 (11_alu_shift_op): Likewise.
12548 (11_alu_shift_reg_op): Likewise.
12549 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12550 (9_alu_shift_reg_op): Likewise.
12551 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12553 (cortex_a15_alu_shift): Likewise.
12554 (cortex_a15_alu_shift_reg): Likewise.
12555 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12557 (cortex_a5_alu_shift): Likewise.
12558 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12560 (cortex_a53_alu_shift): Likewise.
12561 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12563 (cortex_a7_alu_reg): Likewise.
12564 (cortex_a7_alu_shift): Likewise.
12565 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12567 (cortex_a8_alu_shift): Likewise.
12568 (cortex_a8_alu_shift_reg): Likewise.
12569 (cortex_a8_mov): Likewise.
12570 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12571 (cortex_a9_dp_shift): Likewise.
12572 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12574 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12576 (cortex_r4_mov): Likewise.
12577 (cortex_r4_alu_shift): Likewise.
12578 (cortex_r4_alu_shift_reg): Likewise.
12579 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12580 (526_alu_shift_op): Likewise.
12581 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12582 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12583 (626te_alu_shift_op): Likewise.
12584 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12585 (726te_alu_op): Likewise.
12586 (726te_alu_shift_op): Likewise.
12587 (726te_alu_shift_reg_op): Likewise.
12588 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12589 (mp626_alu_shift_op): Likewise.
12590 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12591 (pj4_alu_e1_conds): Likewise.
12592 (pj4_alu): Likewise.
12593 (pj4_alu_conds): Likewise.
12594 (pj4_shift): Likewise.
12595 (pj4_shift_conds): Likewise.
12596 (pj4_alu_shift): Likewise.
12597 (pj4_alu_shift_conds): Likewise.
12598 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12600 (cortexa7_older_only): Likewise.
12601 (cortexa7_younger): Likewise.
12603 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12605 * ipa-pure-const.c (generate_summary): Rename to...
12606 (pure_const_generate_summary): ... this.
12608 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12610 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12612 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12615 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12618 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12620 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12621 enabled without -march=zEC12.
12622 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12625 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12627 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12628 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12629 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12630 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12631 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12632 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12633 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12635 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12636 and ISA_HAS_NMADD3_NMSUB3.
12637 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12638 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12639 (nmsub4<mode>, nmsub3<mode>): Likewise.
12640 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12642 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12644 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12645 TARGET_MIPS5400 checking.
12647 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12648 Peter Bergner <bergner@vnet.ibm.com>
12650 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12651 registers in the comment.
12652 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12653 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12654 rather than FIRST_PSEUDO_REGISTERS.
12656 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12658 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12659 enable extra ISA flags with TARGET_HTM.
12661 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12663 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12666 2013-07-15 Cong Hou <congh@google.com>
12668 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12669 in compare function for sorting.
12671 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12673 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12674 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12675 * config/rs6000/rs6000.opt: Add -mhtm option.
12676 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12677 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12678 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12679 __HTM__ if the HTM instructions are available.
12680 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12681 htm_spr_reg_operand): New define_predicates.
12682 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12683 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12685 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12686 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12687 HTM builtin functions.
12688 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12689 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12690 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12691 (rs6000_builtin_mask_calculate): Likewise.
12692 (rs6000_option_override_internal): Likewise.
12693 (bdesc_htm): Add new HTM builtin support.
12694 (htm_spr_num): New function.
12695 (htm_spr_regno): Likewise.
12696 (rs6000_htm_spr_icode): Likewise.
12697 (htm_expand_builtin): Likewise.
12698 (htm_init_builtins): Likewise.
12699 (rs6000_expand_builtin): Add support for HTM builtin functions.
12700 (rs6000_init_builtins): Likewise.
12701 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12703 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12704 (TARGET_HTM, MASK_HTM): Define macros.
12705 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12706 (FIXED_REGISTERS): Likewise.
12707 (CALL_USED_REGISTERS): Likewise.
12708 (CALL_REALLY_USED_REGISTERS): Likewise.
12709 (REG_ALLOC_ORDER): Likewise.
12710 (enum reg_class): Likewise.
12711 (REG_CLASS_NAMES): Likewise.
12712 (REG_CLASS_CONTENTS): Likewise.
12713 (REGISTER_NAMES): Likewise.
12714 (ADDITIONAL_REGISTER_NAMES): Likewise.
12715 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12716 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12717 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12718 * config/rs6000/htm.md: New file.
12719 * config/rs6000/htmintrin.h: New file.
12720 * config/rs6000/htmxlintrin.h: New file.
12722 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12724 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12725 Define SYMBOL_TINY_GOT, update comment.
12726 * config/aarch64/aarch64.c
12727 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12728 (aarch64_expand_mov_immediate): Likewise.
12729 (aarch64_print_operand): Likewise.
12730 (aarch64_classify_symbol): Likewise.
12731 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12732 (ldr_got_tiny): Define.
12734 2013-07-13 Tobias Grosser <tobias@grosser.es>
12736 PR tree-optimization/54094
12737 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12738 scheduling dimension for the parallelism check from the polyhedral
12739 information in the AST.
12740 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12743 2013-07-13 Jason Merrill <jason@redhat.com>
12745 * print-tree.c (debug_vec_tree): Use debug_raw.
12746 (debug_raw (vec<tree, va_gc> &)): New.
12747 (debug_raw (vec<tree, va_gc> *)): New.
12748 * tree.h: Declare them.
12750 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12752 * ifcvt.c (ifcvt_after_combine): New static variable.
12753 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12755 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12756 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12757 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12759 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12761 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12763 2013-07-12 Michael Matz <matz@suse.de>
12765 PR middle-end/55771
12766 * convert.c (convert_to_real): Reject non-float inner types.
12768 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12770 * config/aarch64/aarch64-protos.h
12771 (aarch64_simd_immediate_valid_for_move): Remove.
12772 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12773 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12775 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12777 2013-07-11 Steve Ellcey <sellcey@mips.com>
12779 * config/mips/mips.c (mips_conditional_register_usage): Do not
12780 use t[0-7] registers in MIPS16 mode when optimizing for size.
12782 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12784 * config/i386/i386.c (dispatch_function_versions): Fix array
12785 indexing of function_version_info to match actual_versions.
12787 2013-07-11 Teresa Johnson <tejohnson@google.com>
12789 * vec.h (struct va_gc): Move release out-of-line.
12790 (va_gc::release): Call ggc_free on released vec.
12792 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12794 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12795 Require GOT register as additional operand in UNSPEC.
12796 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12797 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12798 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12799 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12800 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12801 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12802 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12804 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12807 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12808 name seen by assembler/linker rather if available.
12810 2013-07-11 Andreas Schwab <schwab@suse.de>
12812 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12814 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12816 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12818 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12820 * doc/tm.texi.in: Move hook documentation to ....
12821 * target.def: ... here.
12823 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12824 text on @hook line.
12825 * doc/tm.texi: Regenerate.
12827 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12830 * doc/invoke.texi: Document -Wconditionally-supported.
12832 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12835 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12838 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12841 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12842 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12843 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12844 Remove duplicate devices.
12845 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12846 * config/avr/t-multilib: Regenerate.
12847 * config/avr/avr-tables.opt: Regenerate.
12848 * doc/avr-mmcu.texi: Regenerate.
12850 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12853 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12855 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12857 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12858 the cost of MULT when optimizing for size.
12860 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12862 * config/cr16/cr16-protos.h: Don't include target.h.
12864 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12866 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12867 adjust register size for TDmode and TFmode for VSX registers.
12869 2013-07-08 Kai Tietz <ktietz@redhat.com>
12872 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12873 hook_bool_const_tree_true.
12875 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12877 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12878 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12879 * config/s390/s390.md: Define FPR*_REGNUM constants.
12880 Fix FPR2_REGNUM constant (18 -> 17).
12881 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12882 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12883 ("trunc<BFP:mode><DFP_ALL:mode>2")
12884 ("trunc<DFP_ALL:mode><BFP:mode>2")
12885 ("*extend<BFP:mode><DFP_ALL:mode>2")
12886 ("*extend<DFP_ALL:mode><BFP:mode>2")
12887 ("extend<BFP:mode><DFP_ALL:mode>2")
12888 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12891 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12893 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12895 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12897 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12898 and cfun_fpr_bit_p to cfun_fpr_save_p.
12899 (s390_frame_area, s390_register_info, s390_frame_info)
12900 (s390_emit_prologue, s390_emit_epilogue)
12901 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12903 * config/s390/s390.h: Define *_REGNUM macros for floating point
12906 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12908 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12910 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12912 PR rtl-optimization/57786
12913 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12914 and break out of the loop when it is set to false.
12916 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12919 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12920 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12921 (const_int 63)) 0)).
12922 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12923 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12924 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12926 PR rtl-optimization/57829
12927 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12928 mask bits outside of mode are just sign-extension from mode to HWI.
12930 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12932 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12933 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12934 adjust_address instead of change_address to keep info about alignment.
12935 (emit_strmov): Remove.
12936 (emit_memmov): New function.
12937 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12938 (expand_movmem_epilogue): Likewise and return updated rtx for
12940 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12941 destination and source.
12942 (decide_alignment): Refactor, handle vector_loop.
12943 (ix86_expand_movmem): Likewise.
12944 (ix86_expand_setmem): Likewise.
12945 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12947 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12949 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12950 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12952 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12954 * config/i386/sse.md (sse_movlhps): Change alternative 3
12955 of operand 2 to "m".
12957 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12960 * config/i386/sse.md (iptr): New mode attribute.
12961 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12962 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12963 (<sse>_vmmul<mode>3): Ditto.
12964 (<sse>_vmdiv<mode>3): Ditto.
12965 (sse_vmrcpv4sf2): Ditto.
12966 (<sse>_vmsqrt<mode>2): Ditto.
12967 (sse_vmrsqrtv4sf2): Ditto.
12968 (<sse>_vm<code><mode>3): Ditto.
12969 (avx_vmcmp<mode>3): Ditto.
12970 (<sse>_vmmaskcmp<mode>3): Ditto.
12971 (<sse>_comi): Ditto.
12972 (<sse>_ucomi): Ditto.
12973 (*xop_vmfrcz_<mode>): Ditto.
12974 (*fmai_fmadd_<mode>): Ditto.
12975 (*fmai_fmsub_<mode>): Ditto.
12976 (*fmai_fnmadd_<mode>): Ditto.
12977 (*fmai_fnmsub_<mode>): Ditto.
12978 (*fma4i_vmfmadd_<mode>): Ditto.
12979 (*fma4i_vmfmsub_<mode>): Ditto.
12980 (*fma4i_vmfnmadd_<mode>): Ditto.
12981 (*fma4i_vmfnmsub_<mode>): Ditto.
12982 (*xop_vmfrcz_<mode>): Ditto.
12983 (sse_cvtps2pi): Ditto.
12984 (sse_cvttps2pi): Ditto.
12985 (sse_cvtss2si): Ditto.
12986 (sse_cvtss2si_2): Ditto.
12987 (sse_cvtss2siq_2): Ditto.
12988 (sse_cvttss2si): Ditto.
12989 (sse_cvttss2siq): Ditto.
12990 (sse_cvtsd2si): Ditto.
12991 (sse_cvtsd2si_2): Ditto.
12992 (sse_cvtsd2siq_2): Ditto.
12993 (sse_cvttsd2si): Ditto.
12994 (sse_cvttsd2siq): Ditto.
12995 (sse_cvtsd2ss): Ditto.
12996 (sse_cvtss2sd): Ditto.
12997 (avx2_pbroadcast<mode>): Ditto.
12998 (avx2_pbroadcast<mode>_1): Ditto.
12999 (*avx_vperm_broadcast_v4sf): Ditto.
13001 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13002 (sse_movlhps): Ditto.
13003 (sse_storehps): Ditto.
13004 (sse_loadhps): Ditto.
13005 (sse_storelps): Ditto.
13006 (sse_loadlps): Ditto.
13007 (*vec_concatv4sf): Ditto.
13008 (*vec_interleave_highv2df): Ditto.
13009 (*vec_interleave_lowv2df): Ditto.
13010 (*vec_extractv2df_1_sse): Ditto.
13011 (*vec_extractv2df_0_sse): Ditto.
13012 (sse2_storelpd): Ditto.
13013 (sse2_loadlpd): Ditto.
13014 (sse2_movsd): Ditto.
13015 (*vec_concatv4si): Ditto.
13016 (vec_concatv2di): Ditto.
13018 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13019 for Intel asm dialect.
13020 (mmx_punpcklwd): Ditto.
13021 (mmx_punpckldq): Ditto.
13023 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13024 for intel assembler dialect.
13026 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13029 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13030 for BUILT_IN_C{LZ,LRSB}*.
13031 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13032 * tree-vrp.c (extract_range_basic): Handle
13033 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13034 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13035 fall thru to code calling set_value*.
13036 * builtins.c (expand_builtin): Remove *IMAX cases.
13037 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13038 if width is bigger than 2*HWI.
13040 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13042 PR rtl-optimization/55342
13043 * lra-int.h (lra_subreg_reload_pseudos): New.
13044 * lra.c: Add undoing optional reloads to the block diagram.
13045 (lra_subreg_reload_pseudos): New.
13046 (lra_optional_reload_pseudos): Change comments.
13047 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13048 lra_optional_reload_pseudos after undo transformations.
13049 * lra-assigns.c (pseudo_prefix_title): New.
13050 (lra_setup_reg_renumber): Use it.
13051 (spill_for): Ditto. Check subreg reload pseudos too.
13052 (assign_by_spills): Consider subreg reload pseudos too.
13053 * lra-constraints.c (simplify_operand_subreg): Use
13054 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13055 (curr_insn_transform): Recognize and do optional reloads.
13056 (undo_optional_reloads): New.
13057 (lra_undo_inheritance): Call undo_optional_reloads.
13059 2013-07-05 Thomas Quinot <quinot@adacore.com>
13061 * tree-complex.c (expand_complex_operations_1): Fix typo.
13063 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13065 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13066 (tune_params): New member 'const vec_costs'.
13067 * config/aarch64/aarch64.c (generic_vector_cost): New.
13068 (generic_tunings): New member 'generic_vector_cost'.
13069 (aarch64_builtin_vectorization_cost): New.
13070 (aarch64_add_stmt_cost): New.
13071 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13072 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13074 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13077 * config/i386/predicates.md (vsib_address_operand): Disallow
13078 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13080 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13082 PR middle-end/55030
13083 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13084 expand_builtin_setjmp_receiver.
13085 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13086 with NULL for the label parameter.
13087 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13088 the frame-pointer. Adjust comments.
13089 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13090 only if LABEL is non-NULL.
13092 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13094 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13095 (ARM_ABI_AAPCS64): Ditto.
13097 (ARM_DEFAULT_ABI): Ditto.
13099 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13101 * config/aarch64/aarch64-builtins.c
13102 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13103 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13105 * config/aarch64/aarch64-simd.md
13106 (aarch64_ld1<VALL:mode>): New.
13107 (aarch64_st1<VALL:mode>): Likewise.
13108 * config/aarch64/arm_neon.h
13109 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13111 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13113 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13115 (ix86_option_override_internal):Turn on all -mavx target flags by
13116 default as they are dependent on AVX anyway.
13118 2013-07-02 Cary Coutant <ccoutant@google.com>
13120 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13121 deterministic hash.
13122 (loc_checksum_ordered): Likewise.
13123 (hash_loc_operands): Remove inline keyword.
13125 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13127 PR tree-optimization/57741
13128 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13129 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13130 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13131 Allow REAL_CST step_exprs if flag_associative_math.
13132 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13134 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13136 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13138 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13140 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13142 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13144 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13146 (iorsi3_insn): Likewise.
13147 (arm_xorsi3): Likewise.
13149 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13151 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13153 (attribute "type"): ... here, and prefix with "wmmx_".
13154 (attribute "core_cycles"): Update for attribute changes.
13155 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13156 (tbcstv4hi): Likewise.
13157 (tbcstv2si): Likewise.
13158 (iwmmxt_iordi3): Likewise.
13159 (iwmmxt_xordi3): Likewise.
13160 (iwmmxt_anddi3): Likewise.
13161 (iwmmxt_nanddi3): Likewise.
13162 (iwmmxt_arm_movdi): Likewise.
13163 (iwmmxt_movsi_insn): Likewise.
13164 (mov<mode>_internal): Likewise.
13165 (and<mode>3_iwmmxt): Likewise.
13166 (ior<mode>3_iwmmxt): Likewise.
13167 (xor<mode>3_iwmmxt): Likewise.
13168 (add<mode>3_iwmmxt): Likewise.
13169 (ssaddv8qi3): Likewise.
13170 (ssaddv4hi3): Likewise.
13171 (ssaddv2si3): Likewise.
13172 (usaddv8qi3): Likewise.
13173 (usaddv4hi3): Likewise.
13174 (usaddv2si3): Likewise.
13175 (sub<mode>3_iwmmxt): Likewise.
13176 (sssubv8qi3): Likewise.
13177 (sssubv4hi3): Likewise.
13178 (sssubv2si3): Likewise.
13179 (ussubv8qi3): Likewise.
13180 (ussubv4hi3): Likewise.
13181 (ussubv2si3): Likewise.
13182 (mulv4hi3_iwmmxt): Likewise.
13183 (smulv4hi3_highpart): Likewise.
13184 (umulv4hi3_highpart): Likewise.
13185 (iwmmxt_wmacs): Likewise.
13186 (iwmmxt_wmacsz): Likewise.
13187 (iwmmxt_wmacu): Likewise.
13188 (iwmmxt_wmacuz): Likewise.
13189 (iwmmxt_clrdi): Likewise.
13190 (iwmmxt_clrv8qi): Likewise.
13191 (iwmmxt_clr4hi): Likewise.
13192 (iwmmxt_clr2si): Likewise.
13193 (iwmmxt_uavgrndv8qi3): Likewise.
13194 (iwmmxt_uavgrndv4hi3): Likewise.
13195 (iwmmxt_uavgv8qi3): Likewise.
13196 (iwmmxt_uavgv4hi3): Likewise.
13197 (iwmmxt_tinsrb): Likewise.
13198 (iwmmxt_tinsrh): Likewise.
13199 (iwmmxt_tinsrw): Likewise.
13200 (iwmmxt_textrmub): Likewise.
13201 (iwmmxt_textrmsb): Likewise.
13202 (iwmmxt_textrmuh): Likewise.
13203 (iwmmxt_textrmsh): Likewise.
13204 (iwmmxt_textrmw): Likewise.
13205 (iwmxxt_wshufh): Likewise.
13206 (eqv8qi3): Likewise.
13207 (eqv4hi3): Likewise.
13208 (eqv2si3): Likewise.
13209 (gtuv8qi3): Likewise.
13210 (gtuv4hi3): Likewise.
13211 (gtuv2si3): Likewise.
13212 (gtv8qi3): Likewise.
13213 (gtv4hi3): Likewise.
13214 (gtv2si3): Likewise.
13215 (smax<mode>3_iwmmxt): Likewise.
13216 (umax<mode>3_iwmmxt): Likewise.
13217 (smin<mode>3_iwmmxt): Likewise.
13218 (umin<mode>3_iwmmxt): Likewise.
13219 (iwmmxt_wpackhss): Likewise.
13220 (iwmmxt_wpackwss): Likewise.
13221 (iwmmxt_wpackdss): Likewise.
13222 (iwmmxt_wpackhus): Likewise.
13223 (iwmmxt_wpackwus): Likewise.
13224 (iwmmxt_wpackdus): Likewise.
13225 (iwmmxt_wunpckihb): Likewise.
13226 (iwmmxt_wunpckihh): Likewise.
13227 (iwmmxt_wunpckihw): Likewise.
13228 (iwmmxt_wunpckilb): Likewise.
13229 (iwmmxt_wunpckilh): Likewise.
13230 (iwmmxt_wunpckilw): Likewise.
13231 (iwmmxt_wunpckehub): Likewise.
13232 (iwmmxt_wunpckehuh): Likewise.
13233 (iwmmxt_wunpckehuw): Likewise.
13234 (iwmmxt_wunpckehsb): Likewise.
13235 (iwmmxt_wunpckehsh): Likewise.
13236 (iwmmxt_wunpckehsw): Likewise.
13237 (iwmmxt_wunpckelub): Likewise.
13238 (iwmmxt_wunpckeluh): Likewise.
13239 (iwmmxt_wunpckeluw): Likewise.
13240 (iwmmxt_wunpckelsb): Likewise.
13241 (iwmmxt_wunpckelsh): Likewise.
13242 (iwmmxt_wunpckelsw): Likewise.
13243 (ror<mode>3): Likewise.
13244 (ashr<mode>3_iwmmxt): Likewise.
13245 (lshr<mode>3_iwmmxt): Likewise.
13246 (ashl<mode>3_iwmmxt): Likewise.
13247 (ror<mode>3_di): Likewise.
13248 (ashr<mode>3_di): Likewise.
13249 (lshr<mode>3_di): Likewise.
13250 (ashl<mode>3_di): Likewise.
13251 (iwmmxt_wmadds): Likewise.
13252 (iwmmxt_wmaddu): Likewise.
13253 (iwmmxt_tmia): Likewise.
13254 (iwmmxt_tmiaph): Likewise.
13255 (iwmmxt_tmiabb): Likewise.
13256 (iwmmxt_tmiatb): Likewise.
13257 (iwmmxt_tmiabt): Likewise.
13258 (iwmmxt_tmiatt): Likewise.
13259 (iwmmxt_tmovmskb): Likewise.
13260 (iwmmxt_tmovmskh): Likewise.
13261 (iwmmxt_tmovmskw): Likewise.
13262 (iwmmxt_waccb): Likewise.
13263 (iwmmxt_wacch): Likewise.
13264 (iwmmxt_waccw): Likewise.
13265 (iwmmxt_waligni): Likewise.
13266 (iwmmxt_walignr): Likewise.
13267 (iwmmxt_walignr0): Likewise.
13268 (iwmmxt_walignr1): Likewise.
13269 (iwmmxt_walignr2): Likewise.
13270 (iwmmxt_walignr3): Likewise.
13271 (iwmmxt_wsadb): Likewise.
13272 (iwmmxt_wsadh): Likewise.
13273 (iwmmxt_wsadbz): Likewise.
13274 (iwmmxt_wsadhz): Likewise.
13275 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13276 (iwmmxt_wabsdiffb): Likewise.
13277 (iwmmxt_wabsdiffh): Likewise.
13278 (iwmmxt_wabsdiffw): Likewise.
13279 (iwmmxt_waddsubhx): Likewise
13280 (iwmmxt_wsubaddhx): Likewise.
13281 (addc<mode>3): Likewise.
13282 (iwmmxt_avg4): Likewise.
13283 (iwmmxt_avg4r): Likewise.
13284 (iwmmxt_wmaddsx): Likewise.
13285 (iwmmxt_wmaddux): Likewise.
13286 (iwmmxt_wmaddsn): Likewise.
13287 (iwmmxt_wmaddun): Likewise.
13288 (iwmmxt_wmulwsm): Likewise.
13289 (iwmmxt_wmulwum): Likewise.
13290 (iwmmxt_wmulsmr): Likewise.
13291 (iwmmxt_wmulumr): Likewise.
13292 (iwmmxt_wmulwsmr): Likewise.
13293 (iwmmxt_wmulwumr): Likewise.
13294 (iwmmxt_wmulwl): Likewise.
13295 (iwmmxt_wqmulm): Likewise.
13296 (iwmmxt_wqmulwm): Likewise.
13297 (iwmmxt_wqmulmr): Likewise.
13298 (iwmmxt_wqmulwmr): Likewise.
13299 (iwmmxt_waddbhusm): Likewise.
13300 (iwmmxt_waddbhusl): Likewise.
13301 (iwmmxt_wqmiabb): Likewise.
13302 (iwmmxt_wqmiabt): Likewise.
13303 (iwmmxt_wqmiatb): Likewise.
13304 (iwmmxt_wqmiatt): Likewise.
13305 (iwmmxt_wqmiabbn): Likewise.
13306 (iwmmxt_wqmiabtn): Likewise.
13307 (iwmmxt_wqmiatbn): Likewise.
13308 (iwmmxt_wqmiattn): Likewise.
13309 (iwmmxt_wmiabb): Likewise.
13310 (iwmmxt_wmiabt): Likewise.
13311 (iwmmxt_wmiatb): Likewise.
13312 (iwmmxt_wmiatt): Likewise.
13313 (iwmmxt_wmiabbn): Likewise.
13314 (iwmmxt_wmiabtn): Likewise.
13315 (iwmmxt_wmiatbn): Likewise.
13316 (iwmmxt_wmiattn): Likewise.
13317 (iwmmxt_wmiawbb): Likewise.
13318 (iwmmxt_wmiawbt): Likewise.
13319 (iwmmxt_wmiawtb): Likewise.
13320 (iwmmxt_wmiawtt): Likewise.
13321 (iwmmxt_wmiawbbn): Likewise.
13322 (iwmmxt_wmiawbtn): Likewise.
13323 (iwmmxt_wmiawtbn): Likewise.
13324 (iwmmxt_wmiawttn): Likewise.
13325 (iwmmxt_wmerge): Likewise.
13326 (iwmmxt_tandc<mode>3): Likewise.
13327 (iwmmxt_torc<mode>3): Likewise.
13328 (iwmmxt_torvsc<mode>3): Likewise.
13329 (iwmmxt_textrc<mode>3): Likewise.
13330 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13331 (wmmxt_pack): Likewise.
13332 (wmmxt_mult_c1): Likewise.
13333 (wmmxt_mult_c2): Likewise.
13334 (wmmxt_alu_c1): Likewise.
13335 (wmmxt_alu_c2): Likewise.
13336 (wmmxt_alu_c3): Likewise.
13337 (wmmxt_transfer_c1): Likewise.
13338 (wmmxt_transfer_c2): Likewise.
13339 (wmmxt_transfer_c3): Likewise.
13340 (marvell_f_iwmmxt_wstr): Likewise.
13341 (marvell_f_iwmmxt_wldr): Likewise.
13343 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13345 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13347 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13350 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13351 * lra-constraints.c (need_for_split_p): Check call used hard regs
13352 living through calls.
13354 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13355 call used regs for call insn.
13357 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13360 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13361 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13363 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13365 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13367 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13369 * lra-constraints.c (need_for_split_p): Check call used hard regs
13370 living through calls.
13372 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13375 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13376 to tie with any other modes. Eliminate Altivec vector mode tests,
13377 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13378 code, to return 0 if testing MODE2 for a condition, if we've
13379 already tested MODE1 for the same condition.
13381 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13383 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13386 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13388 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13389 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13391 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13393 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13395 (aarch64_symbolic_constant_p): Remove.
13396 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13397 static. Fix line length and white space.
13398 (aarch64_symbolic_constant_p): Remove.
13399 * config/aarch64/predicates.md (aarch64_valid_symref):
13400 Use aarch64_classify_symbol_expression.
13402 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13404 * config/arm/constraints.md (Ts): New constraint.
13405 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13407 (compare_scc): Use "Ts" constraint for operand 0.
13408 (ior_scc_scc): Likewise.
13409 (and_scc_scc): Likewise.
13410 (and_scc_scc_nodom): Likewise.
13411 (ior_scc_scc_cmp): Likewise for operand 7.
13412 (and_scc_scc_cmp): Likewise.
13413 * config/arm/thumb2.md (thumb2_movsi_insn):
13414 Add alternatives for 16-bit encodings.
13415 (thumb2_movhi_insn): Likewise.
13416 (thumb2_movsicc_insn): Likewise.
13417 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13418 (thumb2_negscc): Use "Ts" constraint.
13419 Move mvn instruction outside cond_exec block.
13420 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13421 for 16-bit encodings.
13423 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13425 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13427 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13428 (mulsi3subsi): Likewise.
13429 (mulsidi3adddi): Likewise.
13430 (mulsidi3_v6): Likewise.
13431 (umulsidi3_v6): Likewise.
13432 (umulsidi3adddi_v6): Likewise.
13433 (smulsi3_highpart_v6): Likewise.
13434 (umulsi3_highpart_v6): Likewise.
13435 (mulhisi3tb): Likewise.
13436 (mulhisi3bt): Likewise.
13437 (mulhisi3tt): Likewise.
13438 (maddhisi4): Likewise.
13439 (maddhisi4tb): Likewise.
13440 (maddhisi4tt): Likewise.
13441 (maddhidi4): Likewise.
13442 (maddhidi4tb): Likewise.
13443 (maddhidi4tt): Likewise.
13444 (zeroextractsi_compare0_scratch): Likewise.
13445 (insv_zero): Likewise.
13446 (insv_t2): Likewise.
13447 (anddi_notzesidi_di): Likewise.
13448 (anddi_notsesidi_di): Likewise.
13449 (andsi_notsi_si): Likewise.
13450 (iordi_zesidi_di): Likewise.
13451 (xordi_zesidi_di): Likewise.
13452 (andsi_iorsi3_notsi): Likewise.
13453 (smax_0): Likewise.
13454 (smax_m1): Likewise.
13455 (smin_0): Likewise.
13456 (not_shiftsi): Likewise.
13457 (unaligned_loadsi): Likewise.
13458 (unaligned_loadhis): Likewise.
13459 (unaligned_loadhiu): Likewise.
13460 (unaligned_storesi): Likewise.
13461 (unaligned_storehi): Likewise.
13462 (extv_reg): Likewise.
13463 (extzv_t2): Likewise.
13464 (divsi3): Likewise.
13465 (udivsi3): Likewise.
13466 (arm_zero_extendhisi2addsi): Likewise.
13467 (arm_zero_extendqisi2addsi): Likewise.
13468 (compareqi_eq0): Likewise.
13469 (arm_extendhisi2_v6): Likewise.
13470 (arm_extendqisi2addsi): Likewise.
13471 (arm_movt): Likewise.
13472 (thumb2_ldrd): Likewise.
13473 (thumb2_ldrd_base): Likewise.
13474 (thumb2_ldrd_base_neg): Likewise.
13475 (thumb2_strd): Likewise.
13476 (thumb2_strd_base): Likewise.
13477 (thumb2_strd_base_neg): Likewise.
13478 (arm_negsi2): Add alternative for 16-bit encoding.
13479 (arm_one_cmplsi2): Likewise.
13481 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13483 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13484 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13485 (movdfcc): Likewise.
13486 * config/arm/vfp.md (*thumb2_movsf_vfp):
13487 Disable predication for arm_restrict_it.
13488 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13489 (*thumb2_movdfcc_vfp): Likewise.
13490 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13491 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13492 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13493 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13494 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13495 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13496 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13497 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13498 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13499 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13500 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13501 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13502 Disable predication for arm_restrict_it.
13504 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13506 * config/i386/bmiintrin.h (_bextr_u32): New.
13507 (_bextr_u64): Ditto.
13509 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13511 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13512 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13513 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13514 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13515 * config/mips/n32-elf.h: ...this new file.
13517 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13520 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13521 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13523 2013-06-27 Catherine Moore <clm@codesourcery.com>
13525 * config/mips/mips-tables.opt: Regenerate.
13526 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13527 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13528 * doc/invoke.texi: Add -m14kc.
13530 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13533 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13534 constraints of operand 1 and 2.
13537 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13538 to match RTL canonicalization. Swap predicates and
13539 constraints of operand 1 and 2.
13541 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13543 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13544 Process OP_INOUT regs for splitting too.
13546 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13548 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13549 decl before the loop, initialize to NULL.
13550 (vectorizable_load): Initialize ptr_incr to NULL.
13552 2013-06-27 Martin Jambor <mjambor@suse.cz>
13555 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13556 * ipa-ref.c (ipa_maybe_record_reference): New function.
13557 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13558 * ipa-cp.c (create_specialized_node): Record potential references from
13560 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13562 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13564 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13565 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13566 'mode' to force_reg.
13567 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13568 (aarch64_expand_mov_immediate): Likewise.
13570 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13572 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13573 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13575 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13577 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13578 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13579 (struct machine_function): Add tbegin_p.
13580 (s390_canonicalize_comparison): Fold CC mode compares to
13581 conditional jump if possible.
13582 (s390_emit_jump): Return the emitted jump.
13583 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13584 Handle CCRAWmode compares.
13585 (s390_option_override): Default to -mhtm if available.
13586 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13587 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13588 FPRs instead of df_regs_ever_live_p.
13589 (s390_optimize_nonescaping_tx): New function.
13590 (s390_init_frame_layout): Extend clobbered_regs array to cover
13592 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13593 (s390_expand_tbegin): New function.
13594 (enum s390_builtin): New enum definition.
13595 (code_for_builtin): New array definition.
13596 (s390_init_builtins): New function.
13597 (s390_expand_builtin): New function.
13598 (TARGET_INIT_BUILTINS): Define.
13599 (TARGET_EXPAND_BUILTIN): Define.
13600 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13601 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13602 (s390_alc_comparison): Likewise.
13603 * config/s390/s390-modes.def: Add CCRAWmode.
13604 * config/s390/s390.h (processor_flags): Add PF_TX.
13605 (TARGET_CPU_HTM): Define macro.
13606 (TARGET_HTM): Define macro.
13607 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13608 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13609 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13610 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13611 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13612 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13613 ("*cc_to_int"): Move up.
13614 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13615 constants other than 0.
13616 ("*ccraw_to_int"): New insn and splitter definition.
13617 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13618 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13619 ("tx_assist"): New expander.
13620 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13621 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13622 * config/s390/s390.opt: Add -mhtm option.
13623 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13624 * config/s390/htmxlintrin.h: New file.
13625 * config/s390/htmintrin.h: New file.
13626 * config/s390/s390intrin.h: New file.
13627 * doc/extend.texi: Document htm builtins.
13628 * config.gcc: Add the new header files to extra_headers.
13630 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13632 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13633 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13635 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13636 Pat Haugen <pthaugen@us.ibm.com>
13637 Peter Bergner <bergner@vnet.ibm.com>
13639 * config/rs6000/power8.md: New.
13640 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13641 setting for power8 entry.
13642 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13643 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13644 test for Power4/Power5 only.
13645 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13647 (force_new_group): Adjust comment.
13648 * config/rs6000/rs6000.md: Include power8.md.
13650 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13652 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13653 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13655 (tune_params): Update comment.
13656 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13657 (arm_max_conditional_execute): New function.
13658 (thumb2_final_prescan_insn): Use max_insn_skipped and
13659 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13661 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13663 PR tree-optimization/57705
13664 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13665 SSA_NAME step, provided that it is not defined inside the loop.
13666 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13667 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13669 2013-06-25 Martin Jambor <mjambor@suse.cz>
13671 PR middle-end/57670
13672 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13673 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13675 (ipa_note_param_call): Initialize member_ptr flag.
13676 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13677 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13678 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13679 (ipa_read_indirect_edge_info): Likewise.
13681 2013-06-25 Richard Biener <rguenther@suse.de>
13683 PR middle-end/56977
13684 * passes.c (init_optimization_passes): Move pass_fold_builtins
13685 and pass_dce earlier with -Og.
13687 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13689 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13690 <BIT_FIELD_REF>: Remove trailing TAB.
13691 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13694 2013-06-24 Martin Jambor <mjambor@suse.cz>
13696 PR tree-optimization/57358
13697 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13698 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13699 (ipa_analyze_params_uses): Generate pessimistic info when true.
13701 2013-06-24 Martin Jambor <mjambor@suse.cz>
13703 PR tree-optimization/57539
13704 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13705 global.inlined_to of the new node to it. All callers changed.
13706 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13707 inlining_into, pass it to cgraph_clone_node.
13708 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13709 ipa_free_edge_args_substructures.
13710 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13711 rdesc linked list. Do not assert rdesc edges have inlined caller.
13712 Assert we have found an rdesc in the rdesc list.
13714 2013-06-24 Richard Biener <rguenther@suse.de>
13716 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13717 (pointer_set_lookup): Declare.
13718 (class pointer_map): New template class implementing a
13719 generic pointer to T map.
13720 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13721 pointer_map<T>::contains, pointer_map<T>::insert,
13722 pointer_map<T>::traverse): New functions.
13723 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13724 (pointer_set_lookup): New function.
13725 (pointer_set_contains): Use pointer_set_lookup.
13726 (pointer_set_insert): Likewise.
13727 (insert_aux): Remove.
13728 (struct pointer_map_t): Embed a pointer_set_t.
13729 (pointer_map_create): Adjust.
13730 (pointer_map_destroy): Likewise.
13731 (pointer_map_contains): Likewise.
13732 (pointer_map_insert): Likewise.
13733 (pointer_map_traverse): Likewise.
13734 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13735 pointer_map<unsigned> instead of a pointer_map_t.
13736 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13737 (streamer_tree_cache_lookup): Likewise.
13738 (streamer_tree_cache_create): Likewise.
13739 (streamer_tree_cache_delete): Likewise.
13740 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13741 pointer_map<unsigned> instead of a pointer_map_t.
13742 (lto_init_tree_ref_encoder): Adjust.
13743 (lto_destroy_tree_ref_encoder): Likewise.
13744 * lto-section-out.c (lto_output_decl_index): Likewise.
13745 (lto_record_function_out_decl_state): Likewise.
13746 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13748 2013-06-24 Richard Biener <rguenther@suse.de>
13750 PR tree-optimization/57488
13751 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13753 2013-06-24 Alan Modra <amodra@gmail.com>
13755 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13756 (gen_easy_altivec_constant): Likewise.
13757 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13758 easy_vector_constant_msb): Likewise.
13760 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13763 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13764 add missing return true.
13766 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13769 * config/sh/predicates.md (general_extend_operand): Invoke
13770 general_movsrc_operand for memory operands.
13771 (general_movsrc_operand): Allow reg+reg addressing, do not use
13772 general_operand for memory operands.
13774 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13776 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13777 when current target options does not apply.
13778 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13779 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13780 * config/i386/bmiintrin.h: Pass appropriate target
13781 attributes to header.
13782 * config/i386/mmintrin.h: Ditto.
13783 * config/i386/nmmintrin.h: Ditto.
13784 * config/i386/avx2intrin.h: Ditto.
13785 * config/i386/fxsrintrin.h: Ditto.
13786 * config/i386/tbmintrin.h: Ditto.
13787 * config/i386/xsaveintrin.h: Ditto.
13788 * config/i386/f16cintrin.h: Ditto.
13789 * config/i386/xtestintrin.h: Ditto.
13790 * config/i386/xsaveoptintrin.h: Ditto.
13791 * config/i386/bmi2intrin.h: Ditto.
13792 * config/i386/lzcntintrin.h: Ditto.
13793 * config/i386/smmintrin.h: Ditto.
13794 * config/i386/wmmintrin.h: Ditto.
13795 * config/i386/x86intrin.h: Remove all header include guards.
13796 * config/i386/prfchwintrin.h: Ditto.
13797 * config/i386/pmmintrin.h: Ditto.
13798 * config/i386/tmmintrin.h: Ditto.
13799 * config/i386/xmmintrin.h: Ditto.
13800 * config/i386/popcntintrin.h: Ditto.
13801 * config/i386/rdseedintrin.h: Ditto.
13802 * config/i386/ammintrin.h: Ditto.
13803 * config/i386/emmintrin.h: Ditto.
13804 * config/i386/immintrin.h: Remove all header include guards.
13805 * config/i386/fma4intrin.h: Ditto.
13806 * config/i386/lwpintrin.h: Ditto.
13807 * config/i386/xopintrin.h: Ditto.
13808 * config/i386/ia32intrin.h: Ditto.
13809 * config/i386/avxintrin.h: Ditto.
13810 * config/i386/rtmintrin.h: Ditto.
13811 * config/i386/fmaintrin.h: Ditto.
13812 * config/i386/mm3dnow.h: Ditto.
13814 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13816 * common/config/i386/i386-common.c: Handle LZCNT.
13818 2013-06-22 Andi Kleen <ak@linux.intel.com>
13820 * doc/extend.texi: Use __atomic_store_n instead of
13821 __atomic_store in HLE example.
13823 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13825 * config/sh/sh.c: Remove <cstdlib> workaround.
13827 2013-06-21 Andi Kleen <ak@linux.intel.com>
13829 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13831 2013-06-21 Andi Kleen <ak@linux.intel.com>
13833 * doc/extend.texi: Document that __atomic_clear and
13834 __atomic_test_and_set should only be used with bool.
13836 2013-06-20 Jan Hubicka <jh@suse.cz>
13838 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13839 types_same_for_odr.
13840 * tree.c (decls_same_for_odr): New function.
13841 (same_for_edr): New function.
13842 (types_same_for_odr): New function.
13843 (get_binfo_at_offset): Use it.
13844 * tree.h (types_same_for_odr): Declare.
13846 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13847 Jason Merrill <jason@redhat.com>
13849 * system.h: Include <cstdlib> as well as <stdlib.h>.
13851 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13854 * config/i386/i386.c (construct_container): Report error if
13855 long double is used with disabled x87 float returns.
13857 2013-06-20 Jan Hubicka <jh@suse.cz>
13859 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13861 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13863 PR rtl-optimization/57425
13864 PR rtl-optimization/57569
13865 * alias.c (write_dependence_p): Remove parameters mem_mode and
13866 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13867 Changed all callers.
13868 (canon_anti_dependence): Get comments and semantics in sync.
13869 Add parameter mem_canonicalized. Changed all callers.
13870 * rtl.h (canon_anti_dependence): Update prototype.
13872 2013-06-20 Richard Biener <rguenther@suse.de>
13874 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13875 byte case, inline streamer_read_uchar and defer section
13878 2013-06-20 Richard Biener <rguenther@suse.de>
13880 PR tree-optimization/57584
13881 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13882 SSA names into the expanded expression that take part in
13883 abnormal coalescing.
13885 2013-06-19 Sharad Singhai <singhai@google.com>
13887 * gcov.c (print_usage): Handle new option.
13888 (process_args): Ditto.
13889 (get_gcov_intermediate_filename): New function.
13890 (output_intermediate_file): New function.
13891 (output_gcov_file): New function
13892 (generate_results): Handle new option.
13893 (release_function): Relase demangled name.
13894 (read_graph_file): Handle demangled name.
13895 (output_lines): Ditto.
13896 * doc/gcov.texi: Document gcov intermediate format.
13898 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13901 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13902 (lra_emit_add): Use the functions. Add comment about Y as an
13905 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13908 * collect2.c (collect_atexit): New.
13909 (collect_exit): Delete.
13910 (main): Register collect_atexit with atexit.
13911 (collect_wait): Change collect_exit to exit.
13913 * collect2.h (collect_exit): Delete.
13914 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13916 2013-06-19 Wei Mi <wmi@google.com>
13918 PR rtl-optimization/57518
13919 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13920 if regno is used in paradoxical subreg.
13921 (update_equiv_regs): Check pdx_subregs[regno] before
13922 set a reg to be equivalent with a mem.
13924 2013-06-19 Matthias Klose <doko@ubuntu.com>
13927 * file-find.h (find_a_file): Add a mode parameter.
13928 * file-find.c (find_a_file): Likewise.
13929 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13930 with X_OK for the executables.
13931 * collect2.c (main): Call find_a_file with X_OK.
13933 2013-06-19 Steve Ellcey <sellcey@mips.com>
13936 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13937 Use NEXT_INSN instead of next_real_insn.
13939 2013-06-19 Jan Hubicka <jh@suse.cz>
13941 * cgraph.h (const_value_known_p): Replace by ...
13942 (ctor_for_folding): .. this one.
13943 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13944 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13945 * expr.c (expand_expr_real_1): Likewise.
13946 (string_constant): Likewise.
13947 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13948 * ipa.c (process_references): Likewise.
13949 (symtab_remove_unreachable_nodes): Likewise.
13950 * ipa-inline-analysis.c (param_change_prob): Likewise.
13951 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13952 (get_base_constructor): Likwise.
13953 * varpool.c (varpool_remove_node): Likewise.
13954 (varpool_remove_initializer): LIkewise.
13955 (dump_varpool_node): LIkwise.
13956 (const_value_known_p): Rewrite to ...
13957 (ctor_for_folding): ... this one.
13959 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13962 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13963 PERSONALITY in $PATH derived prefixes.
13965 2013-06-19 Jeff Law <law@redhat.com>
13967 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13970 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13971 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13974 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13976 * config/arm/vfp.md: Move VFP instruction classification documentation
13978 * config/arm/arm.md: ... here. Update instruction classification
13981 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13983 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13984 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13985 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13987 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13990 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13991 with NEXT_INSN. Use tablejump_p to check for jump table data
13994 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13997 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13998 that now in C++ the value is correct per the C++ standards.
14000 2013-06-19 Richard Biener <rguenther@suse.de>
14002 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14003 for global context.
14005 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14008 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14011 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14012 with next_active_insn.
14014 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14016 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14017 functions are inlined during failures to flag an error.
14018 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14019 in early inline pass.
14021 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14023 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14026 2013-06-18 Julian Brown <julian@codesourcery.com>
14028 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14029 Permit virtual register pre-reload if !strict.
14030 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14032 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14034 * config/arm/neon.md (movmisalign<mode>): Use
14035 neon_perm_struct_or_reg_operand instead of
14036 neon_struct_or_register_operand.
14037 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14038 neon_permissive_struct_operand instead of neon_struct_operand.
14039 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14040 neon_vector_mem_operand.
14041 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14042 neon_vector_mem_operand.
14043 (neon_permissive_struct_operand): New.
14044 (neon_struct_or_register_operand): Rename to...
14045 (neon_perm_struct_or_reg_operand): This. Adjust call to
14046 neon_vector_mem_operand.
14048 2013-06-18 Richard Biener <rguenther@suse.de>
14050 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14051 * lto-streamer.h: Include pointer-set.h.
14052 (struct lto_decl_slot): Remove.
14053 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14054 Remove next_index entry.
14055 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14056 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14057 (lto_init_tree_ref_encoder): Adjust.
14058 (lto_destroy_tree_ref_encoder): Likewise.
14059 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14060 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14061 (lto_output_decl_index): Adjust.
14062 (lto_new_out_decl_state): Likewise.
14063 (lto_record_function_out_decl_state): Likewise.
14064 * lto-streamer-out.c (copy_function): Likewise.
14066 2013-06-18 Richard Biener <rguenther@suse.de>
14068 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14069 * cgraphunit.c: Include cfgloop.h.
14070 (init_lowered_empty_function): Initialize the loop tree.
14071 (assemble_thunk): Insert new BBs into loops.
14073 2013-06-18 Richard Biener <rguenther@suse.de>
14075 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14076 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14077 the map from cache entry to cache index optional.
14078 (streamer_tree_cache_replace_tree): Adjust accordingly.
14079 (streamer_tree_cache_append): Likewise.
14080 (streamer_tree_cache_delete): Likewise.
14081 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14082 streamer cache map from cache entry to cache index.
14083 * lto-streamer-out.c (create_output_block): Adjust.
14085 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14087 * config/arm/arm.md (attribute "insn"): Move multiplication and
14088 division attributes to...
14089 (attribute "type"): ... here. Remove mult.
14090 (attribute "mul32"): New attribute.
14091 (attribute "mul64"): Add umaal.
14092 (*arm_mulsi3): Update attributes.
14093 (*arm_mulsi3_v6): Likewise.
14094 (*thumb_mulsi3): Likewise.
14095 (*thumb_mulsi3_v6): Likewise.
14096 (*mulsi3_compare0): Likewise.
14097 (*mulsi3_compare0_v6): Likewise.
14098 (*mulsi_compare0_scratch): Likewise.
14099 (*mulsi_compare0_scratch_v6): Likewise.
14100 (*mulsi3addsi): Likewise.
14101 (*mulsi3addsi_v6): Likewise.
14102 (*mulsi3addsi_compare0): Likewise.
14103 (*mulsi3addsi_compare0_v6): Likewise.
14104 (*mulsi3addsi_compare0_scratch): Likewise.
14105 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14106 (*mulsi3subsi): Likewise.
14107 (*mulsidi3adddi): Likewise.
14108 (*mulsi3addsi_v6): Likewise.
14109 (*mulsidi3adddi_v6): Likewise.
14110 (*mulsidi3_nov6): Likewise.
14111 (*mulsidi3_v6): Likewise.
14112 (*umulsidi3_nov6): Likewise.
14113 (*umulsidi3_v6): Likewise.
14114 (*umulsidi3adddi): Likewise.
14115 (*umulsidi3adddi_v6): Likewise.
14116 (*smulsi3_highpart_nov6): Likewise.
14117 (*smulsi3_highpart_v6): Likewise.
14118 (*umulsi3_highpart_nov6): Likewise.
14119 (*umulsi3_highpart_v6): Likewise.
14120 (mulhisi3): Likewise.
14121 (*mulhisi3tb): Likewise.
14122 (*mulhisi3bt): Likewise.
14123 (*mulhisi3tt): Likewise.
14124 (maddhisi4): Likewise.
14125 (*maddhisi4tb): Likewise.
14126 (*maddhisi4tt): Likewise.
14127 (maddhidi4): Likewise.
14128 (*maddhidi4tb): Likewise.
14129 (*maddhidi4tt): Likewise.
14130 (divsi3): Likewise.
14131 (udivsi3): Likewise.
14132 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14133 (thumb2_mulsi_short_compare0): Likewise.
14134 (thumb2_mulsi_short_compare0_scratch): Likewise.
14135 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14136 (1020mult2): Likewise.
14137 (1020mult3): Likewise.
14138 (1020mult4): Likewise.
14139 (1020mult5): Likewise.
14140 (1020mult6): Likewise.
14141 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14143 (cortex_a15_mult64): Likewise.
14144 (cortex_a15_sdiv): Likewise.
14145 (cortex_a15_udiv): Likewise.
14146 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14152 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14153 (pj4_ir_div): Likewise.
14154 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14155 (11_mult2): Likewise.
14156 (11_mult3): Likewise.
14157 (11_mult4): Likewise.
14158 (11_mult5): Likewise.
14159 (11_mult6): Likewise.
14160 (11_mult7): Likewise.
14161 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14162 (cortex_a8_mla): Likewise.
14163 (cortex_a8_mull): Likewise.
14164 (cortex_a8_smulwy): Likewise.
14165 (cortex_a8_smlald): Likewise.
14166 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14167 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14168 (cortex_r4_mul_3): Likewise.
14169 (cortex_r4_mla_4): Likewise.
14170 (cortex_r4_mla_3): Likewise.
14171 (cortex_r4_smlald): Likewise.
14172 (cortex_r4_mull): Likewise.
14173 (cortex_r4_sdiv): Likewise.
14174 (cortex_r4_udiv): Likewise.
14175 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14176 (cortex_a7_idiv): Likewise.
14177 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14178 (9_mult2): Likewise.
14179 (9_mult3): Likewise.
14180 (9_mult4): Likewise.
14181 (9_mult5): Likewise.
14182 (9_mult6): Likewise.
14183 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14184 (cortex_a53_sdiv): Likewise.
14185 (cortex_a53_udiv): Likewise.
14186 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14187 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14188 (mp626_mult2): Likewise.
14189 (mp626_mult3): Likewise.
14190 (mp626_mult4): Likewise.
14191 * config/arm/fa526.md (526_mult1): Update attribute change.
14192 (526_mult2): Likewise.
14193 * config/arm/arm-generic.md (mult): Update attribute change.
14194 (mult_ldsched_strongarm): Likewise.
14195 (mult_ldsched): Likewise.
14196 (multi_cycle): Likewise.
14197 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14198 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14199 (606te_mult2): Likewise.
14200 (606te_mult3): Likewise.
14201 (606te_mult4): Likewise.
14202 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14203 (cortex_a9_mac16): Likewise.
14204 (cortex_a9_multiply): Likewise.
14205 (cortex_a9_mac): Likewise.
14206 (cortex_a9_multiply_long): Likewise.
14207 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14208 (626te_mult2): Likewise.
14209 (626te_mult3): Likewise.
14210 (626te_mult4): Likewise.
14212 2013-06-18 Richard Biener <rguenther@suse.de>
14215 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14217 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14220 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14221 with next_active_insn.
14223 2013-06-18 Alan Modra <amodra@gmail.com>
14225 * config/rs6000/rs6000.h (enum data_align): New.
14226 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14227 (DATA_ABI_ALIGNMENT): Define.
14228 (CONSTANT_ALIGNMENT): Correct comment.
14229 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14230 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14232 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14234 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14235 ATTRIBUTE_UNUSED marking.
14237 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14239 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14240 alternative and update.
14241 (aarch64_dup_lanedi): Delete.
14242 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14243 * config/aarch64/aarch64-simd-builtins.def: Update.
14245 2013-06-17 Richard Biener <rguenther@suse.de>
14247 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14248 (lto_input_scc): Declare.
14249 (lto_input_tree_1): Likewise.
14250 (struct lto_stats_d): Add num_tree_bodies_output and
14251 num_pickle_refs_output.
14252 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14253 (lto_read_tree_1): Split out from ...
14254 (lto_read_tree): ... this.
14255 (lto_input_scc): New function.
14256 (lto_input_tree_1): Split out from ...
14257 (lto_input_tree): ... this. Handle LTO_tree_scc.
14258 (lto_data_in_create): Create the streamer cache without hashes.
14259 * lto-streamer-out.c (create_output_block): Create the streamer
14260 cache with hashes when not doing WPA.
14261 (lto_write_tree_1): Split out from ...
14262 (lto_write_tree): ... this.
14263 (get_symbol_initial_value): New function.
14264 (lto_output_tree_1): Split out from ...
14265 (lto_output_tree): ... this. Write trees as series of SCCs
14266 using a DFS walk via DFS_write_tree.
14267 (struct sccs, struct scc_entry): New types.
14268 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14269 (DFS_write_tree_body): New function.
14270 (DFS_write_tree): Likewise.
14271 (hash_tree): Likewise.
14272 (scc_entry_compare): Likewise.
14273 (hash_scc): Likewise.
14274 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14275 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14276 TREE_CHAIN as regular reference.
14277 (streamer_read_integer_cst): Remove.
14278 (streamer_get_pickled_tree): Adjust.
14279 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14280 of DECL_EXTERNALs in BLOCK_VARS for now.
14281 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14283 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14284 Add hash value argument and record that if hashes are recorded
14286 (streamer_tree_cache_insert_1): Adjust.
14287 (streamer_tree_cache_insert): Likewise.
14288 (streamer_tree_cache_insert_at): Rename to ...
14289 (streamer_tree_cache_replace_tree): ... this and adjust.
14290 (streamer_tree_cache_append): Adjust.
14291 (record_common_node): Likewise.
14292 (streamer_tree_cache_create): Add argument whether to
14293 record hash values together with trees.
14294 (streamer_tree_cache_delete): Adjust.
14295 * tree-streamer.h (struct streamer_tree_cache_d): Add
14297 (streamer_read_integer_cst): Remove.
14298 (streamer_tree_cache_insert): Adjust.
14299 (streamer_tree_cache_append): Likewise.
14300 (streamer_tree_cache_insert_at): Rename to ...
14301 (streamer_tree_cache_replace_tree): ... this and adjust.
14302 (streamer_tree_cache_create): Add argument whether to record hashes.
14303 (streamer_tree_cache_get): Rename to ...
14304 (streamer_tree_cache_get_tree): ... this.
14305 (streamer_tree_cache_get_hash): New function.
14306 * tree.c (cache_integer_cst): New function.
14307 * tree.h (cache_integer_cst): Declare.
14308 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14309 * lto-symtab.c (lto_varpool_replace_node): Only release
14310 DECL_INITIAL of non-prevailing decls.
14311 * varpool.c (varpool_remove_initializer): Do not release
14312 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14314 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14316 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14317 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14318 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14319 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14320 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14321 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14322 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14323 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14324 instead of TARGET_64BIT.
14325 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14326 Require ISA_HAS_<D>DIV.
14328 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14330 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14331 (mips*-*-linux*): Move default with_llsc setting to where other
14333 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14335 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14336 Likewise. Remove default with_tune setting. Move default float
14337 setting to its own block. Handle with_llsc in the same block as above.
14339 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14341 PR rtl-optimization/57425
14342 PR rtl-optimization/57569
14343 * alias.c (write_dependence_p): Add new parameters mem_mode,
14344 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14345 Changed all callers.
14346 (canon_anti_dependence): New function.
14347 * cse.c (check_dependence): Use canon_anti_dependence.
14348 * cselib.c (cselib_invalidate_mem): Likewise.
14349 * rtl.h (canon_anti_dependence): Declare.
14351 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14353 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14354 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14355 ".set mips3" for 64-bit targets.
14357 2013-06-15 Dehao Chen <dehao@google.com>
14359 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14360 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14361 (gimple_check_call_args): Likewise.
14362 * value-prof.c (check_ic_target): Likewise.
14363 * ipa-inline.c (early_inliner): Likewise.
14364 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14365 * cgraph.c (cgraph_create_edge_1): Likewise.
14366 (cgraph_make_edge_direct): Likewise.
14368 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14371 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14372 rs6000_output_move_128bit to handle emitting quad memory
14373 operations. Set attribute length to 8 bytes.
14375 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14377 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14379 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14380 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14381 (aarch64_<su>mlsl<mode>): Likewise.
14383 2013-06-14 Mike Stump <mikestump@comcast.net>
14385 * Makefile.in (TARGET_H): Add insn-codes.h.
14387 2013-06-14 Alan Modra <amodra@gmail.com>
14389 PR middle-end/57134
14390 PR middle-end/57586
14391 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14392 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14393 bitfield expansion when EXPAND_MEMORY.
14394 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14396 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14398 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14399 test for clearing quad memory on 32-bit later.
14401 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14403 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14404 (fold_negate_expr): Likewise.
14405 (fold_real_zero_addition_p): Handle vectors.
14406 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14408 2013-06-14 Alan Modra <amodra@gmail.com>
14410 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14412 2013-06-13 Jan Hubicka <jh@suse.cz>
14414 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14415 Local comdats are not externally visible.
14416 * symtab.c (dump_symtab_base): Dump externally visible.
14417 (verify_symtab_base): Verify back links in the symtab hash.
14419 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14421 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14422 CONVERT_EXPR as equal nodes.
14424 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14426 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14428 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14430 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14431 Generalize to complex and vector.
14432 * tree.c (build_all_ones_cst): New function.
14433 * tree.h (build_all_ones_cst): Declare it.
14435 2013-06-13 Alan Modra <amodra@gmail.com>
14437 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14438 * config/rs6000/rs6000.md (signbittf2): New insn.
14439 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14440 (abstf2_internal, cmptf_internal2): Likewise.
14441 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14443 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14444 Pat Haugen <pthaugen@us.ibm.com>
14445 Peter Bergner <bergner@vnet.ibm.com>
14447 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14448 power8 byte, half-word, and quad-word atomic instructions.
14449 (emit_store_conditional): Likewise.
14450 (rs6000_expand_atomic_compare_and_swap): Likewise.
14451 (rs6000_expand_atomic_op): Likewise.
14453 * config/rs6000/sync.md (larx): Add new modes for power8.
14455 (AINT): New mode iterator to include TImode as well as normal
14456 integer modes on power8.
14457 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14458 that VSX registers are not considered. Use AINT mode iterator
14459 instead of INT1 to allow inclusion of quad word atomic operations
14461 (load_locked<mode>): Likewise.
14462 (store_conditional<mode>): Likewise.
14463 (atomic_compare_and_swap<mode>): Likewise.
14464 (atomic_exchange<mode>): Likewise.
14465 (atomic_nand<mode>): Likewise.
14466 (atomic_fetch_<fetchop_name><mode>): Likewise.
14467 (atomic_nand_fetch<mode>): Likewise.
14468 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14470 (ATOMIC): On power8, add QImode, HImode modes.
14471 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14472 modes that promote to SImode.
14473 (load_lockedti): Convert TImode arguments to PTImode, so that we
14474 get a guaranteed even/odd register pair.
14475 (load_lockedpti): Likewise.
14476 (store_conditionalti): Likewise.
14477 (store_conditionalpti): Likewise.
14479 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14480 atomic load/store instructions.
14483 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14485 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14487 (insn_count): New attribute, with most cases extracted from...
14488 (length): ...here. Redefine most cases in terms of insn_count.
14489 (single_insn): Delete.
14490 (can_delay): Use insn_count to check for single instructions.
14491 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14492 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14493 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14494 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14495 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14496 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14497 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14498 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14499 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14500 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14501 rather than "length".
14502 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14503 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14504 Use "insn_count" rather than "length".
14505 * config/mips/mips-dsp.md
14506 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14507 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14510 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14512 PR tree-optimization/57361
14513 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14515 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14517 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14519 (aarch64_simd_combine<mode>): New instruction expansion.
14520 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14521 function prototype.
14522 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14523 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14525 2013-06-12 Jan Hubicka <jh@suse.cz>
14527 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14528 decl has when in streaming stage.
14529 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14530 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14532 2013-06-12 Roland Stigge <stigge@antcom.de>
14535 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14537 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14539 PR tree-optimization/57537
14540 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14541 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14543 2013-06-12 Richard Biener <rguenther@suse.de>
14545 * data-streamer.h (streamer_write_char_stream): CSE
14546 obs->current_pointer.
14547 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14548 streamer_write_char_stream manually and optimize the resulting loop.
14549 (streamer_write_hwi_stream): Likewise.
14551 2013-06-12 Jan Hubicka <jh@suse.cz>
14553 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14554 * cgraph.h (varpool_create_empty_node): Declare.
14555 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14557 * symtab.c (symtab_unregister_node): Be lax about missin entries
14559 (symtab_get_node): Update comment.
14560 * varpool.c (varpool_create_empty_node): Break out from ...
14561 (varpool_node_for_decl): ... here.
14562 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14564 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14566 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14567 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14568 part. Use straight-line flow at the end.
14569 <COMPONENT_REF>: Remove superfluous else.
14570 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14572 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14575 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14576 target hook even for !TREE_PUBLIC decls. If no resolution info
14577 is available, return false for common and external decls.
14579 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14581 * config/rl78/constraints.md (U): New constraint.
14582 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14585 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14588 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14589 to allow returning address to AT_PLATFORM name.
14591 2013-06-11 Jan Hubicka <jh@suse.cz>
14593 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14594 * cgraph.h (symtab_node_base): Add weakref flag.
14595 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14596 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14597 (output_weakrefs): Use weakref flag.
14598 * fold-const.c (simple_operand_p): Handle WEAK.
14599 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14600 * ipa.c (varpool_externally_visible_p): Drop weakref.
14601 (function_and_variable_visibility): Update comment; fix weakref
14602 sanity checks; do not clear DECL_WEAK on them.
14603 * lto-cgraph.c (lto_output_node): update.
14604 (lto_output_varpool_node): Update.
14605 (input_overwrite_node): Update.
14606 (input_node): Update.
14607 (input_varpool_node): Update.
14608 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14609 (lto_symtab_merge_symbols): Add sanity check.
14610 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14611 * passes.c (rest_of_decl_compilation): Set static flag, too.
14612 * symtab.c (dump_symtab_base): Dump weakref.
14613 (verify_symtab_base): Sanity check weakrefs.
14614 (symtab_make_decl_local): Remove duplicated code.
14615 (symtab_alias_ultimate_target): Simplify.
14616 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14618 2013-06-11 Tom de Vries <tom@codesourcery.com>
14620 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14621 sequence_vect == NULL.
14623 2013-06-11 DJ Delorie <dj@redhat.com>
14625 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14626 (rl78_unwind_word_mode): New.
14628 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14630 * final.c (debug_prefix_maps): Make static.
14632 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14634 * function.c (initial_trampoline): Remove stray copy.
14636 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14638 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14640 2013-06-11 Martin Jambor <mjambor@suse.cz>
14642 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14643 within bounds at the beginning of the function.
14645 2013-06-11 Alan Modra <amodra@gmail.com>
14647 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14649 (get_named_section): Don't NULL !DECL_P decl.
14651 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14653 * doc/invoke.texi (core-avx2): Document.
14655 (atom): Updated with MOVBE.
14657 2013-06-11 Richard Biener <rguenther@suse.de>
14659 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14661 2013-06-11 Anton Blanchard <anton@samba.org>
14663 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14664 correct shift value in little-endian mode.
14666 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14669 * varasm.c (get_variable_align): Move #endif to the right place.
14671 2013-06-10 Cary Coutant <ccoutant@google.com>
14673 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14674 for hash so that hash table traversal order is deterministic.
14676 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14677 Pat Haugen <pthaugen@us.ibm.com>
14678 Peter Bergner <bergner@vnet.ibm.com>
14680 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14681 of vectors in GPRS if they are direct moves or quad word load or
14684 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14686 (direct_move_p): Likewise.
14687 (quad_load_store_p): Likewise.
14689 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14690 classes into bins based on the physical register type.
14691 (reg_class_to_reg_type): Likewise.
14692 (IS_STD_REG_TYPE): Likewise.
14693 (IS_FP_VECT_REG_TYPE): Likewise.
14694 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14695 use direct move instructions.
14696 (reload_gpr_vsx): Likewise.
14697 (reload_vsx_gpr): Likewise.
14698 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14699 information that is a simplification of register classes. Also
14700 precalculate direct move reload helpers.
14701 (direct_move_p): New function to return true if the operation can
14702 be done as a direct move instruciton.
14703 (quad_load_store_p): New function to return true if the operation
14704 is a quad memory operation.
14705 (rs6000_legitimize_address): If quad memory, only allow register
14706 indirect for TImode addresses.
14707 (rs6000_legitimate_address_p): Likewise.
14708 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14709 (rs6000_reload_register_type): Likewise.
14710 (register_to_reg_type): Return register type.
14711 (rs6000_secondary_reload_simple_move): New helper function for
14712 secondary reload and secondary memory needed to identify anything
14713 that is a simple move, and does not need reloading.
14714 (rs6000_secondary_reload_direct_move): New helper function for
14715 secondary reload to identify cases that can be done with several
14716 instructions via the direct move instructions.
14717 (rs6000_secondary_reload_move): New helper function for secondary
14718 reload to identify moves between register types that can be done.
14719 (rs6000_secondary_reload): Add support for quad memory operations
14720 and for direct move.
14721 (rs6000_secondary_memory_needed): Likewise.
14722 (rs6000_debug_secondary_memory_needed): Change argument names.
14723 (rs6000_output_move_128bit): New function to return the move to
14724 use for 128-bit moves, including knowing about the various
14725 limitations of quad memory operations.
14727 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14728 memory operations. call rs6000_output_move_128bit for the actual
14729 instruciton(s) to generate.
14730 (vsx_movti_64bit): Likewise.
14732 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14733 (UNSPEC_P8V_MTVSRWZ): Likewise.
14734 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14735 (UNSPEC_P8V_MTVSRD): Likewise.
14736 (UNSPEC_P8V_XXPERMDI): Likewise.
14737 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14738 (UNSPEC_FUSION_GPR): Likewise.
14739 (FMOVE128_GPR): New iterator for direct move.
14740 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14741 (f32_sv): Likewise.
14742 (f32_dm): Likewise.
14743 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14744 loads and direct move instructions.
14745 (zero_extendsidi2_lfiwzx): Likewise.
14746 (extendsidi2_lfiwax): Likewise.
14747 (extendsidi2_nocell): Likewise.
14748 (floatsi<mode>2_lfiwax): Likewise.
14749 (lfiwax): Likewise.
14750 (floatunssi<mode>2_lfiwzx): Likewise.
14751 (lfiwzx): Likewise.
14752 (fix_trunc<mode>_stfiwx): Likewise.
14753 (fixuns_trunc<mode>_stfiwx): Likewise.
14754 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14755 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14756 (parity<mode>2_cmpb): Set length/type attr.
14757 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14758 for 'mr.' to fast_compare.
14759 (bpermd_<mode>): Change type attr to popcnt.
14760 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14761 (p8_mtvsrwz_1): Likewise.
14762 (p8_mtvsrwz_2): Likewise.
14763 (reload_fpr_from_gpr<mode>): Likewise.
14764 (p8_mtvsrd_1): Likewise.
14765 (p8_mtvsrd_2): Likewise.
14766 (p8_xxpermdi_<mode>): Likewise.
14767 (reload_vsx_from_gpr<mode>): Likewise.
14768 (reload_vsx_from_gprsf): Likewise.
14769 (p8_mfvsrd_3_<mode>): LIkewise.
14770 (reload_gpr_from_vsx<mode>): Likewise.
14771 (reload_gpr_from_vsxsf): Likewise.
14772 (p8_mfvsrd_4_disf): Likewise.
14773 (multi-word GPR splits): Do not split direct moves or quad memory
14776 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14778 * tree-into-ssa.c (interesting_blocks): Make static.
14780 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14783 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14784 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14785 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14786 (get_variable_align): New function.
14787 (get_variable_section, emit_bss, emit_common,
14788 assemble_variable_contents, place_block_symbol): Use
14789 get_variable_align instead of DECL_ALIGN.
14790 (assemble_noswitch_variable): Add align argument, use it
14791 instead of DECL_ALIGN.
14792 (assemble_variable): Adjust caller. Use get_variable_align
14793 instead of DECL_ALIGN.
14794 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14796 (DATA_ABI_ALIGNMENT): Define.
14797 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14798 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14799 opt is false, only return the psABI mandated alignment increase.
14800 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14801 (DATA_ABI_ALIGNMENT): ... this.
14802 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14803 (DATA_ABI_ALIGNMENT): ... this.
14804 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14805 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14806 (DATA_ABI_ALIGNMENT): ... this.
14807 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14808 * doc/tm.texi: Regenerated.
14810 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14812 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14813 cmp_code to construct REG_EQUAL note.
14815 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14818 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14819 that operands[2] doesn't overlap with operands[0].
14821 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14822 Jan Hubicka <jh@suse.cz>
14824 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14825 hack to mark symbols as used.
14827 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14829 PR rtl-optimization/57559
14830 * lra-constraints.c (process_alt_operands): Don't discourage
14831 memory with known offset for offsetable memory constraint.
14832 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14834 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14836 * varasm.c (struct oc_local_state): Reorder fields.
14837 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14838 and adjust accordingly.
14839 (output_constructor): Reorder initialization code and adjust call to
14840 output_constructor_bitfield.
14842 2013-06-07 Jan Hubicka <jh@suse.cz>
14844 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14846 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14848 * tree-object-size.c (unknown): Make const.
14850 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14852 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14853 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14854 for last alternative in the cpu_facility attribute.
14856 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14859 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14860 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14861 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14862 * config/arm/constraints.md (Dg): New constraint.
14863 * config/arm/neon.md (xordi3_neon): Remove.
14864 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14865 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14867 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14869 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14870 Clean up alternatives.
14872 2013-06-07 Alan Modra <amodra@gmail.com>
14874 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14877 2013-06-07 Alan Modra <amodra@gmail.com>
14879 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14881 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14883 * config/arm/constraints.md (Df): New constraint.
14884 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14885 Correct length attribute for last two alternatives.
14887 2013-06-07 Alan Modra <amodra@gmail.com>
14889 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14890 override user -mfp-in-toc.
14891 (offsettable_ok_by_alignment): Consider just the current access
14892 rather than the whole object, unless BLKmode. Handle
14893 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14894 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14895 for -mcmodel=medium.
14896 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14897 override user -mfp-in-toc or -msum-in-toc. Default to
14898 -mno-fp-in-toc for -mcmodel=medium.
14900 2013-06-06 DJ Delorie <dj@redhat.com>
14902 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14903 TARGET_VALID_POINTER_MODE.
14905 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14906 Pat Haugen <pthaugen@us.ibm.com>
14907 Peter Bergner <bergner@vnet.ibm.com>
14909 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14910 Document new power8 builtins.
14912 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14913 condition code register, to allow 128-bit logical operations to be
14914 done in the VSX or GPR registers.
14915 (nor<mode>3): Use the canonical form for nor.
14916 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14917 vclz*, and vpopcnt* vector instructions.
14918 (nand<mode>3): Likewise.
14919 (orc<mode>3): Likewise.
14920 (clz<mode>2): LIkewise.
14921 (popcount<mode>2): Likewise.
14923 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14924 that only the GPRs are recognized.
14926 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14927 support for new power8 builtins.
14929 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14931 (xscvdpspn): Likewise.
14937 (vpopcnt): Likewise.
14938 (vpopcntb): Likewise.
14939 (vpopcnth): Likewise.
14940 (vpopcntw): Likewise.
14941 (vpopcntd): Likewise.
14943 (vmrgew): Likewise.
14944 (vmrgow): Likewise.
14946 (eqv_v16qi3): Likewise.
14947 (eqv_v8hi3): Likewise.
14948 (eqv_v4si3): Likewise.
14949 (eqv_v2di3): Likewise.
14950 (eqv_v4sf3): Likewise.
14951 (eqv_v2df3): Likewise.
14953 (nand_v16qi3): Likewise.
14954 (nand_v8hi3): Likewise.
14955 (nand_v4si3): Likewise.
14956 (nand_v2di3): Likewise.
14957 (nand_v4sf3): Likewise.
14958 (nand_v2df3): Likewise.
14960 (orc_v16qi3): Likewise.
14961 (orc_v8hi3): Likewise.
14962 (orc_v4si3): Likewise.
14963 (orc_v2di3): Likewise.
14964 (orc_v4sf3): Likewise.
14965 (orc_v2df3): Likewise.
14967 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14968 allow power8 quad mode in 64-bit.
14969 (rs6000_builtin_vectorized_function): Add support to vectorize
14970 ISA 2.07 count leading zeros, population count builtins.
14971 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14972 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14973 (builtin_function_type): Add vgbbd builtin function which takes an
14975 (altivec_expand_vec_perm_const): Add support for new power8 merge
14978 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14979 that does not include TImdoe for use with 32-bit.
14980 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14982 (UNSPEC_VSX_CVDPSPN): Likewise.
14983 (vsx_xscvdpspn): Likewise.
14984 (vsx_xscvspdpn): Likewise.
14985 (vsx_xscvdpspn_scalar): Likewise.
14986 (vsx_xscvspdpn_directmove): Likewise.
14987 (vsx_and<mode>3): Split logical operations into 32-bit and
14988 64-bit. Add support to do logical operations on TImode as well as
14989 VSX vector types. Allow logical operations to be done in either
14990 VSX registers or in general purpose registers in 64-bit mode. Add
14991 splitters if GPRs were used. For AND, add clobber of CCmode to
14992 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14994 (vsx_and<mode>3_32bit): Likewise.
14995 (vsx_and<mode>3_64bit): Likewise.
14996 (vsx_ior<mode>3): Likewise.
14997 (vsx_ior<mode>3_32bit): Likewise.
14998 (vsx_ior<mode>3_64bit): Likewise.
14999 (vsx_xor<mode>3): Likewise.
15000 (vsx_xor<mode>3_32bit): Likewise.
15001 (vsx_xor<mode>3_64bit): Likewise.
15002 (vsx_one_cmpl<mode>2): Likewise.
15003 (vsx_one_cmpl<mode>2_32bit): Likewise.
15004 (vsx_one_cmpl<mode>2_64bit): Likewise.
15005 (vsx_nor<mode>3): Likewise.
15006 (vsx_nor<mode>3_32bit): Likewise.
15007 (vsx_nor<mode>3_64bit): Likewise.
15008 (vsx_andc<mode>3): Likewise.
15009 (vsx_andc<mode>3_32bit): Likewise.
15010 (vsx_andc<mode>3_64bit): Likewise.
15011 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15012 and xxlorc instructions.
15013 (vsx_eqv<mode>3_64bit): Likewise.
15014 (vsx_nand<mode>3_32bit): Likewise.
15015 (vsx_nand<mode>3_64bit): Likewise.
15016 (vsx_orc<mode>3_32bit): Likewise.
15017 (vsx_orc<mode>3_64bit): Likewise.
15019 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15021 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15023 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15024 (p8_vmrgow): Likewise.
15025 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15026 GPRs to be split under VSX.
15027 (p8v_clz<mode>2): Add power8 count leading zero support.
15028 (p8v_popcount<mode>2): Add power8 population count support.
15029 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15032 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15035 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15037 (vec_nand): Likewise.
15038 (vec_vclz): Likewise.
15039 (vec_vclzb): Likewise.
15040 (vec_vclzd): Likewise.
15041 (vec_vclzh): Likewise.
15042 (vec_vclzw): Likewise.
15043 (vec_vgbbd): Likewise.
15044 (vec_vmrgew): Likewise.
15045 (vec_vmrgow): Likewise.
15046 (vec_vpopcnt): Likewise.
15047 (vec_vpopcntb): Likewise.
15048 (vec_vpopcntd): Likewise.
15049 (vec_vpopcnth): Likewise.
15050 (vec_vpopcntw): Likewise.
15052 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15054 PR rtl-optimization/57468
15055 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15058 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15060 PR rtl-optimization/57459
15061 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15062 type when setting live regs.
15064 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15066 * config/s390/s390.opt (mlra): New option.
15067 * config/s390/s390.c (s390_decompose_address): Check displacement
15068 for all registers for LRA.
15069 (s390_secondary_reload): Don't used secondary reloads for LRA.
15070 (s390_lra_p): New function.
15071 (TARGET_LRA_P): Define.
15072 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15073 of attribute cpu_facility to zarch for the last alternative.
15074 (*cmpmem_short): Ditto.
15076 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15078 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15079 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15080 (arm_expand_prologue): Likewise.
15082 2013-06-06 Teresa Johnson <tejohnson@google.com>
15085 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15086 as this is now done by redirect_edge_and_branch_force.
15087 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15088 barriers, and fix interaction with splitting.
15089 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15090 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15091 reflect changes made in the routine.
15092 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15093 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15094 since this is called in cfglayout mode, and replace partition fixup
15095 with assert as that is now done by force_nonfallthru_and_redirect.
15096 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15097 already be marked with region crossing note.
15098 (insert_section_boundary_note): Make non-static, gate on flag
15099 has_bb_partition, rewrite to also check for multiple partitions.
15100 (rest_of_handle_reorder_blocks): Remove call to
15101 insert_section_boundary_note, now done later during free_cfg.
15102 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15103 * bb-reorder.h (insert_section_boundary_note): Declare.
15104 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15105 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15106 invoke insert_section_boundary_note.
15107 (try_redirect_by_replacing_jump): Remove unnecessary
15108 check for region crossing note.
15109 (fixup_partition_crossing): New function.
15110 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15111 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15112 in non-cfglayout mode.
15113 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15114 remove old code that tried to do this. Emit barrier correctly
15115 when we are in cfglayout mode.
15116 (last_bb_in_partition): New function.
15117 (rtl_split_edge): Correctly fixup partition boundaries.
15118 (commit_one_edge_insertion): Remove old code that tried to
15119 fixup region crossing edge since this is now handled in
15120 split_block, and set up insertion point correctly since
15121 block may now end in a jump.
15122 (verify_hot_cold_block_grouping): Guard against checking when not in
15123 linearized RTL mode.
15124 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15126 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15127 rtl_verify_flow_info, so not called in cfglayout mode.
15128 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15129 (fixup_reorder_chain): Remove old code that attempted to fixup region
15130 crossing note as this is now handled in force_nonfallthru_and_redirect.
15131 (duplicate_insn_chain): Don't duplicate switch section notes.
15132 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15134 * basic-block.h (emit_barrier_after_bb): Declare.
15136 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15138 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15139 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15140 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15142 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15144 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15146 * config/arm/ldmstm.md: Regenerate.
15148 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15150 * config/arm/sync.md (atomic_loaddi_1):
15151 Disable predication for arm_restrict_it.
15152 (arm_load_exclusive<mode>): Likewise.
15153 (arm_load_exclusivesi): Likewise.
15154 (arm_load_exclusivedi): Likewise.
15155 (arm_load_acquire_exclusive<mode>): Likewise.
15156 (arm_load_acquire_exclusivesi): Likewise.
15157 (arm_load_acquire_exclusivedi): Likewise.
15158 (arm_store_exclusive<mode>): Likewise.
15159 (arm_store_exclusive<mode>): Likewise.
15160 (arm_store_release_exclusivedi): Likewise.
15161 (arm_store_release_exclusive<mode>): Likewise.
15163 2013-06-06 Richard Biener <rguenther@suse.de>
15165 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15167 (lto_tag_is_tree_code_p): Adjust.
15168 (lto_tag_is_gimple_code_p): Likewise.
15169 (lto_gimple_code_to_tag): Likewise.
15170 (lto_tag_to_gimple_code): Likewise.
15171 (lto_tree_code_to_tag): Likewise.
15172 (lto_tag_to_tree_code): Likewise.
15173 * data-streamer.h (streamer_write_hwi_in_range): Use
15174 uhwi streaming to stream the normalized range.
15175 (streamer_read_hwi_in_range): Likewise.
15177 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15179 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15180 (predicable_short_it): Likewise.
15181 (predicated): Likewise.
15182 (enabled): Handle above.
15183 (define_cond_exec): Set predicated attribute to yes.
15185 2013-06-05 Mike Stump <mikestump@comcast.net>
15187 * gdbinit.in (__FUNCTION__): Add.
15189 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15191 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15192 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15194 2013-06-05 Jan Hubicka <jh@suse.cz>
15196 * varasm.c (mark_decl_referenced): Revert the removal until targets
15199 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15201 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15202 instead of mark_decl_referenced.
15204 2013-06-05 Jan Hubicka <jh@suse.cz>
15206 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15207 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15208 and symtab_used_from_object_file_p.
15209 (cgraph_make_node_local_1): Clear forced_by_abi.
15210 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15211 * cgraph.h (symtab_node_base): Add forced_by_abi.
15212 (decide_is_variable_needed): Remove.
15213 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15214 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15215 (decide_is_symbol_needed): ... this one; handle symbols in general;
15216 always analyze virtuals; honnor forced_by_abi.
15217 (cgraph_finalize_function): Update.
15218 (varpool_finalize_decl): Update.
15219 (symbol_defined_and_needed): Remove.
15220 (analyze_functions): Update.
15221 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15222 output_refs, input_overwrite_node): Handle forced_by_abi.
15223 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15224 (address_taken_from_non_vtable_p): ... this one.
15225 (comdat_can_be_unshared_p_1): New function.
15226 (cgraph_comdat_can_be_unshared_p): Rename to ...
15227 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15228 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15229 (function_and_variable_visibility): Clear forced_by_abi as needed.
15230 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15231 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15232 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15233 * varpool.c (decide_is_variable_needed): Remove.
15235 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15237 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15238 (arm_option_override): Override arm_restrict_it where appropriate.
15239 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15240 * config/arm/arm.opt (mrestrict-it): New command-line option.
15241 * doc/invoke.texi: Document -mrestrict-it.
15243 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15245 * tsan.c (tsan_atomic_table): Make const.
15247 2013-06-05 Richard Biener <rguenther@suse.de>
15249 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15250 index associated with the tree we are supposed to replace.
15251 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15252 TREE_ASM_WRITTEN as zero for everything but SSA names.
15254 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15256 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15258 2013-06-04 Jan Hubicka <jh@suse.cz>
15260 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15261 (input_node, input_varpool_node): Handle correctly external same
15263 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15264 nodes at ltrans stage.
15266 2013-06-04 Jan Hubicka <jh@suse.cz>
15268 * ipa-inline.c (update_caller_keys): Fix availability test.
15269 (update_callee_keys): Likewise.
15270 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15271 to follow ELF standard.
15273 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15275 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15276 (mips64r5900el-*-elf*): New configurations.
15277 * config/mips/mips-cpus.def (r5900): New processor.
15278 * config/mips/mips-tables.opt: Regenerate.
15279 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15280 (mips_issue_rate): Handle PROCESSOR_R5900.
15281 (mips_reorg_process_insns): Force reorder mode for the R5900.
15282 * config/mips/mips.h (TARGET_MIPS5900): Define.
15283 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15285 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15287 * config/mips/mips.md (processor): Add r5900.
15288 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15290 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15292 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15293 into function to generate MOVI instruction.
15294 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15295 (aarch64_preferred_simd_mode): Turn into wrapper.
15296 (aarch64_output_scalar_simd_mov_immediate): New function.
15297 * config/aarch64/aarch64-protos.h: Add prototype for above.
15299 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15301 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15302 element_char member.
15303 (sizetochar): Return signed char.
15304 (aarch64_simd_valid_immediate): Remove elchar and other
15305 unnecessary variables.
15306 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15307 Calculate element_char as required.
15308 * config/aarch64/aarch64-protos.h: Update and move prototype
15309 for aarch64_output_simd_mov_immediate.
15310 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15313 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15315 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15316 information completed by aarch64_simd_valid_immediate.
15317 (aarch64_legitimate_constant_p): Update arguments.
15318 (aarch64_simd_valid_immediate): Work with struct rather than many
15320 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15321 (aarch64_simd_make_constant): Update arguments.
15322 (aarch64_output_simd_mov_immediate): Work with struct rather than
15323 many pointers. Output immediate directly rather than as operand.
15324 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15326 * config/aarch64/constraints.md (Dn): Update arguments.
15328 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15330 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15332 (aarch64_simd_immediate_valid_for_move): Remove.
15333 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15334 (aarch64_simd_make_constant): Update call.
15335 (aarch64_output_simd_mov_immediate): Update call.
15336 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15338 * config/aarch64/constraints.md (Dn): Update call.
15340 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15342 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15343 return type to bool for prototype.
15344 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15345 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15346 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15348 2013-06-04 Catherine Moore <clm@codesourcery.com>
15350 * config/mips/mips.opt (meva): New.
15351 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15352 (ASM_SPEC): Handle -meva.
15353 * doc/invoke.texi (meva): Document.
15355 2013-06-04 Alan Modra <amodra@gmail.com>
15357 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15360 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15362 * rtl.def: Add extra fourth optional field to define_cond_exec.
15363 * gensupport.c (process_one_cond_exec): Process attributes from
15365 * doc/md.texi: Document fourth field in define_cond_exec.
15367 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15369 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15370 out the processing order as in store_bit_field_1.
15372 2013-06-04 Jan Hubicka <jh@suse.cz>
15374 PR middle-end/57500
15375 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15376 non-VAR_DECL node if it does not exist yet.
15378 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15380 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15381 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15382 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15383 target_cpu_default setting.
15385 2013-06-03 Teresa Johnson <tejohnson@google.com>
15387 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15388 default to -fopt-info=optimized instead of all.
15389 * doc/invoke.texi: Ditto.
15390 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15391 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15392 (execute_vect_slp): Emit BB vectorization success under
15393 MSG_OPTIMIZED_LOCATIONS.
15394 * tree-vect-slp.c (vect_slp_transform_bb): Change
15395 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15396 * tree-vect-loop.c (vect_transform_loop): Ditto.
15398 2013-06-03 Jason Merrill <jason@redhat.com>
15401 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15402 Use TARGET_EXPR for C++.
15404 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15406 PR rtl-optimization/57268
15407 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15408 if DEBUG_INSN_P (insn).
15411 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15413 PR rtl-optimization/57268
15414 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15415 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15417 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15419 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15420 (ix86_avoid_lea_for_addr): Likewise.
15421 (exact_dependency_1): Likewise.
15422 (ix86_adjust_cost): Likewise.
15423 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15425 (do_reorder_for_imul): Fix typo, formatting and
15426 !reload_completed check removed.
15427 (ix86_sched_reorder): Fix typo and formatting.
15428 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15431 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15433 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15435 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15437 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15438 <STRING_CST>: Likewise.
15439 <VECTOR_CST>: Likewise.
15441 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15442 Mikael Morin <mikael@gcc.gnu.org>
15444 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15445 * config.in: Regenerated.
15446 * configure: Regenerated.
15448 2013-06-01 Jan Hubicka <jh@suse.cz>
15450 PR middle-end/57366
15451 * cgraphunit.c (compile): When weakref is not supported,
15452 set up transparent aliases before final output pass.
15453 * varasm.c (assemble_alias): Do not try to do it here.
15455 2013-06-01 Jan Hubicka <jh@suse.cz>
15457 PR middle-end/57467
15458 * passes.c (for_per_function): Skip unanalyzed functions.
15460 2013-06-01 Jan Hubicka <jh@suse.cz>
15462 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15463 (lto_symtab_merge_symbols_1): ... this one.
15464 (lto_symtab_merge_cgraph_nodes): Rename to ...
15465 (lto_symtab_merge_symbols): ... this one; simplify.
15466 * cgraph.c (same_body_aliases_done): Rename to ...
15467 (cpp_implicit_aliases_done): ... this one.
15468 (cgraph_create_function_alias): Update.
15469 (cgraph_same_body_alias): Update.
15470 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15471 (verify_edge_corresponds_to_fndecl): Simplify.
15472 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15473 (cgraph_node): Remove same_body_alias.
15474 (varpool_node): Remove alias_of and extra_name_alias.
15475 (same_body_aliases_done): Rename to ..
15476 (cpp_implicit_aliases_done): ... this one.
15477 (symtab_alias_ultimate_target): Add default parameter.
15478 (symtab_resolve_alias): New function.
15479 (fixup_same_cpp_alias_visibility): Declare.
15480 (cgraph_function_node): Add default parameter.
15481 (cgraph_node_asm_name): Likewise.
15482 (cgraph_function_or_thunk_node): Add default parameter; do
15483 not ICE when it is NULL.
15484 (varpool_variable_node): Likewise.
15485 * tree-emutls.c (create_emultls_var): Update.
15486 (ipa_lower_emutls): Update.
15487 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15488 (cgraph_reset_node): Reset alias info.
15489 (cgraph_finalize_function): Update.
15490 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15491 (analyze_function): Simplify.
15492 (cgraph_process_same_body_aliases): Simplify.
15493 (analyze_functions): Fixup same body aliases.
15494 (handle_alias_pairs): Simplify.
15495 (assemble_thunk): Update.
15496 (assemble_thunks_and_aliases): Update.
15497 (output_weakrefs): Rewrite.
15498 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15499 (lto_output_varpool_node): Likewise.
15500 (compute_ltrans_boundary): Remve assert.
15501 (get_alias_symbol): New functoin.
15502 (input_node): Rewrite alias handling.
15503 (input_varpool_node): Likewise.
15504 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15505 * ipa.c (process_references): Handle weakrefs correctly.
15506 (symtab_remove_unreachable_nodes): Likewise.
15507 * trans-mem.c (get_cg_data): Update.
15508 (ipa_tm_create_version_alias): Update.
15509 (ipa_tm_execute): Update.
15510 * symtab.c (dump_symtab_base): Dump aliases.
15511 (verify_symtab_base): Verify aliases.
15512 (symtab_node_availability): New function.
15513 (symtab_alias_ultimate_target): Simplify.
15514 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15515 handle all the fixup cases.
15516 (symtab_resolve_alias): New function.
15517 * passes.c (ipa_write_summaries): Handle weakrefs.
15518 * varpool.c (varpool_analyze_node): Simplify.
15519 (assemble_aliases): Update.
15520 (varpool_create_variable_alias): Simplify.
15521 (varpool_extra_name_alias): Simplify.
15522 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15523 (lto_symtab_merge_symbols): ... this one.
15525 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15528 PR rtl-optimization/57268
15529 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15530 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15532 2013-06-01 Tobias Burnus <burnus@net-b.de>
15534 Partially reverted:
15535 2013-05-31 Tobias Burnus <burnus@net-b.de>
15537 PR middle-end/57073
15538 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15541 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15543 PR rtl-optimization/57268
15544 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15545 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15547 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15549 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15550 unordered comparison operators when -fno-trapping-math is in effect
15552 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15553 and implement unordered comparison operators properly on the e500.
15555 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15557 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15558 for constant scalar integers.
15559 (simplify_relational_operation_1): Likewise.
15561 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15563 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15564 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15567 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15568 Igor Zamyatin <igor.zamyatin@intel.com>
15570 Silvermont (SLM) architecture performance tuning.
15571 * config/i386/i386.h (enum ix86_tune_indices): Add
15572 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15573 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15575 * config/i386/i386.c (initial_ix86_tune_features)
15576 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15577 (ix86_lea_outperforms): Handle Silvermont tuning.
15578 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15580 (ix86_use_lea_for_mov): Likewise.
15581 (ix86_avoid_lea_for_addr): Likewise.
15582 (ix86_lea_for_add_ok): Likewise.
15583 (exact_dependency_1): New function.
15584 (exact_store_load_dependency): Likewise.
15585 (ix86_adjust_cost): Handle Silvermont tuning.
15586 (do_reoder_for_imul): Likewise.
15587 (swap_top_of_ready_list): New function.
15588 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15590 * config/i386/i386.md (peepholes that split memory operand in fp
15593 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15595 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15596 Remove un-necessary braces.
15598 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15600 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15601 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15603 2013-05-31 Tobias Burnus <burnus@net-b.de>
15605 PR middle-end/57073
15606 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15608 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15611 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15612 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15613 * config/arm/neon.md (iordi3_neon): Remove.
15614 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15615 * config/arm/predicates.md (imm_for_neon_logic_operand):
15616 Move to earlier in the file.
15617 (neon_logic_op2): Likewise.
15618 (arm_iordi_operand_neon): New predicate.
15620 2013-05-31 Richard Biener <rguenther@suse.de>
15622 PR tree-optimization/57478
15623 PR tree-optimization/57453
15624 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15627 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15629 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15630 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15632 2013-05-30 Tobias Burnus <burnus@net-b.de>
15633 Thomas Koenig <tkoenig@gcc.gnu.org>
15635 PR middle-end/57073
15636 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15637 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15639 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15641 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15643 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15645 * target.def (register_usage_leveling_p): New hook.
15646 * targhooks.c (default_register_usage_leveling_p): New.
15647 * targhooks.h (default_register_usage_leveling_p): New prototype.
15648 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15649 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15650 * doc/tm.texi: Update.
15651 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15653 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15655 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15656 (*insv_reg<mode>): New define_insn.
15658 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15660 PR rtl-optimization/57439
15661 * postreload.c (move2add_valid_value_p): Check that we have
15662 a zero subreg_regno_offset when accessing the register in
15663 the requested mode.
15665 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15666 Igor Zamyatin <igor.zamyatin@intel.com>
15668 Silvermont (SLM) architecture pipeline model, tuning and
15670 * config.gcc: Add slm config options and target.
15672 * config/i386/slm.md: New.
15674 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15676 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15678 (ix86_target_macros_internal): Likewise.
15680 * config/i386/i386.c (slm_cost): New cost.
15681 (m_SLM): New macro flag.
15682 (initial_ix86_tune_features): Set m_SLM.
15683 (x86_accumulate_outgoing_args): Likewise.
15684 (x86_arch_always_fancy_math_387): Likewise.
15685 (processor_target_table): Add slm cost.
15686 (cpu_names): Add slm cpu name.
15687 (x86_option_override_internal): Set SLM ISA.
15688 (ix86_issue_rate): New case PROCESSOR_SLM.
15689 (ia32_multipass_dfa_lookahead): Likewise.
15690 (fold_builtin_cpu): Add slm.
15692 * config/i386/i386.h (TARGET_SLM): New target macro.
15693 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15694 (processor_type): Add PROCESSOR_SLM.
15696 * config/i386/i386.md (cpu): Add new value "slm".
15697 (slm.md): Include slm.md.
15699 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15700 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15702 * config/arm/arm-protos.h: Add and update function protos.
15703 * config/arm/arm.c (use_simple_return_p): New added.
15704 (thumb2_expand_return): Check simple_return flag.
15705 * config/arm/arm.md: Add simple_return and conditional simple_return.
15706 * config/arm/iterators.md: Add iterator for return and simple_return.
15708 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15710 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15711 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15712 (arm_emit_vfp_multi_reg_pop): Likewise.
15713 (thumb2_emit_ldrd_pop): Likewise.
15714 (arm_expand_epilogue): Add misc REG_CFA notes.
15715 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15717 2013-05-29 Lawrence Crowl <crowl@google.com>
15719 * config/arm/t-arm: Update for below.
15721 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15722 Change type to hash_table. Update dependent calls and types.
15724 * config/i386/t-cygming: Update for below.
15726 * config/i386/t-interix: Update for below.
15728 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15729 Change type to hash_table. Update dependent calls and types.
15730 (i386_find_on_wrapper_list::wrappers): Likewise.
15732 * config/ia64/t-ia64: Update for below.
15734 * config/ia64/ia64.c (bundle_state_table):
15735 Change type to hash_table. Update dependent calls and types.
15737 * config/mips/mips.c (mips_reorg_process_insns::htab):
15738 Change type to hash_table. Update dependent calls and types.
15740 * config/sol2.c (solaris_comdat_htab):
15741 Change type to hash_table. Update dependent calls and types.
15743 * config/t-sol2: Update for above.
15745 2013-05-29 Teresa Johnson <tejohnson@google.com>
15747 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15748 functions are not yet marked as defined.
15750 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15751 Pat Haugen <pthaugen@us.ibm.com>
15752 Peter Bergner <bergner@vnet.ibm.com>
15754 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15758 (vrotl<mode>3): Likewise.
15759 (vashl<mode>3): Likewise.
15760 (vlshr<mode>3): Likewise.
15761 (vashr<mode>3): Likewise.
15763 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15764 support for power8 V2DI builtins.
15766 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15767 power8 V2DI builtins.
15768 (vupkhsw): Likewise.
15769 (vupklsw): Likewise.
15770 (vaddudm): Likewise.
15771 (vminsd): Likewise.
15772 (vmaxsd): Likewise.
15773 (vminud): Likewise.
15774 (vmaxud): Likewise.
15775 (vpkudum): Likewise.
15776 (vpksdss): Likewise.
15777 (vpkudus): Likewise.
15778 (vpksdus): Likewise.
15783 (vsubudm): Likewise.
15784 (vcmpequd): Likewise.
15785 (vcmpgtsd): Likewise.
15786 (vcmpgtud): Likewise.
15787 (vcmpequd_p): Likewise.
15788 (vcmpgtsd_p): Likewise.
15789 (vcmpgtud_p): Likewise.
15790 (vupkhsw): Likewise.
15791 (vupklsw): Likewise.
15792 (vaddudm): Likewise.
15793 (vmaxsd): Likewise.
15794 (vmaxud): Likewise.
15795 (vminsd): Likewise.
15796 (vminud): Likewise.
15797 (vpksdss): Likewise.
15798 (vpksdus): Likewise.
15799 (vpkudum): Likewise.
15800 (vpkudus): Likewise.
15805 (vsubudm): Likewise.
15807 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15808 support for power8 V2DI instructions.
15810 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15811 power8 V2DI instructions. Combine pack and unpack insns to use an
15812 iterator for each mode. Check whether a particular mode supports
15813 Altivec instructions instead of just checking TARGET_ALTIVEC.
15814 (UNSPEC_VPKUWUM): Likewise.
15815 (UNSPEC_VPKSHSS): Likewise.
15816 (UNSPEC_VPKSWSS): Likewise.
15817 (UNSPEC_VPKUHUS): Likewise.
15818 (UNSPEC_VPKSHUS): Likewise.
15819 (UNSPEC_VPKUWUS): Likewise.
15820 (UNSPEC_VPKSWUS): Likewise.
15821 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15822 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15823 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15824 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15825 (UNSPEC_VUPKHSB): Likewise.
15826 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15827 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15828 (UNSPEC_VUPKHSH): Likewise.
15829 (UNSPEC_VUPKLSB): Likewise.
15830 (UNSPEC_VUPKLSH): Likewise.
15832 (VI_char): Likewise.
15833 (VI_scalar): Likewise.
15834 (VI_unit): Likewise.
15836 (VP_small): Likewise.
15837 (VP_small_lc): Likewise.
15838 (VU_char): Likewise.
15839 (add<mode>3): Likewise.
15840 (altivec_vaddcuw): Likewise.
15841 (altivec_vaddu<VI_char>s): Likewise.
15842 (altivec_vadds<VI_char>s): Likewise.
15843 (sub<mode>3): Likewise.
15844 (altivec_vsubcuw): Likewise.
15845 (altivec_vsubu<VI_char>s): Likewise.
15846 (altivec_vsubs<VI_char>s): Likewise.
15847 (altivec_vavgs<VI_char>): Likewise.
15848 (altivec_vcmpbfp): Likewise.
15849 (altivec_eq<mode>): Likewise.
15850 (altivec_gt<mode>): Likewise.
15851 (altivec_gtu<mode>): Likewise.
15852 (umax<mode>3): Likewise.
15853 (smax<mode>3): Likewise.
15854 (umin<mode>3): Likewise.
15855 (smin<mode>3): Likewise.
15856 (altivec_vpkuhum): Likewise.
15857 (altivec_vpkuwum): Likewise.
15858 (altivec_vpkshss): Likewise.
15859 (altivec_vpkswss): Likewise.
15860 (altivec_vpkuhus): Likewise.
15861 (altivec_vpkshus): Likewise.
15862 (altivec_vpkuwus): Likewise.
15863 (altivec_vpkswus): Likewise.
15864 (altivec_vpks<VI_char>ss): Likewise.
15865 (altivec_vpks<VI_char>us): Likewise.
15866 (altivec_vpku<VI_char>us): Likewise.
15867 (altivec_vpku<VI_char>um): Likewise.
15868 (altivec_vrl<VI_char>): Likewise.
15869 (altivec_vsl<VI_char>): Likewise.
15870 (altivec_vsr<VI_char>): Likewise.
15871 (altivec_vsra<VI_char>): Likewise.
15872 (altivec_vsldoi_<mode>): Likewise.
15873 (altivec_vupkhsb): Likewise.
15874 (altivec_vupkhs<VU_char>): Likewise.
15875 (altivec_vupkls<VU_char>): Likewise.
15876 (altivec_vupkhsh): Likewise.
15877 (altivec_vupklsb): Likewise.
15878 (altivec_vupklsh): Likewise.
15879 (altivec_vcmpequ<VI_char>_p): Likewise.
15880 (altivec_vcmpgts<VI_char>_p): Likewise.
15881 (altivec_vcmpgtu<VI_char>_p): Likewise.
15882 (abs<mode>2): Likewise.
15883 (vec_unpacks_hi_v16qi): Likewise.
15884 (vec_unpacks_hi_v8hi): Likewise.
15885 (vec_unpacks_lo_v16qi): Likewise.
15886 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15887 (vec_unpacks_lo_v8hi): Likewise.
15888 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15889 (vec_pack_trunc_v8h): Likewise.
15890 (vec_pack_trunc_v4si): Likewise.
15891 (vec_pack_trunc_<mode>): Likewise.
15893 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15895 (vec_vmaxsd): Likewise.
15896 (vec_vmaxud): Likewise.
15897 (vec_vminsd): Likewise.
15898 (vec_vminud): Likewise.
15899 (vec_vpksdss): Likewise.
15900 (vec_vpksdus): Likewise.
15901 (vec_vpkudum): Likewise.
15902 (vec_vpkudus): Likewise.
15903 (vec_vrld): Likewise.
15904 (vec_vsld): Likewise.
15905 (vec_vsrad): Likewise.
15906 (vec_vsrd): Likewise.
15907 (vec_vsubudm): Likewise.
15908 (vec_vupkhsw): Likewise.
15909 (vec_vupklsw): Likewise.
15911 2013-05-29 Jan Hubicka <jh@suse.cz>
15913 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15914 flags; reorder rest of fields in more consistent way.
15915 (varpool_node): Remove analyzed, finalized and alias.
15916 (cgraph_ndoe): Likewise.
15917 (symtab_alias_ultimate_target): New function.
15918 (cgraph_function_node): Move offline.
15919 (cgraph_reset_node): Declare.
15920 (cgraph_comdat_can_be_unshared_p): Remove.
15921 (varpool_remove_initializer): Declare.
15922 (varpool_first_defined_variable, varpool_next_defined_variable
15923 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15924 (cgraph_function_with_gimple_body_p): Update.
15925 (varpool_all_refs_explicit_p): Update.
15926 (symtab_alias_target): New function.
15927 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15928 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15929 (cgraph_function_or_thunk_node): Simplify using
15930 symtab_alias_ultimate_target.
15931 (varpool_variable_node): Likewise.
15932 * cgraph.c (cgraph_create_function_alias): Update.
15933 (cgraph_add_thunk): Update.
15934 (cgraph_remove_node): Update.
15935 (dump_cgraph_node): Do not dump removed flags.
15936 (cgraph_function_body_availability): Update.
15937 (cgraph_propagate_frequency): Update.
15938 (verify_cgraph_node): Check sanity of local flag.
15939 (cgraph_function_node): Move here from cgraph.h; revamp for
15940 cgraph_function_or_thunk_node.
15941 * lto-symtab.c (lto_varpool_replace_node): Update.
15942 (lto_symtab_resolve_can_prevail_p): Update.
15943 (lto_symtab_merge_cgraph_nodes): Update.
15944 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15945 propagate_constants_accross_call, devirtualization_time_bonus,
15946 ipcp_propagate_stage): Update.
15947 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15948 * ipa-inline-transform.c (clone_inlined_nodes,
15949 preserve_function_body_p): Update.
15950 * ipa-reference.c (propagate): Update.
15951 (write_node_summary_p): Update.
15952 * toplev.c (wrapup_global_declaration_2): Update.
15953 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15954 (analyze_function) ... this one.
15955 (cgraph_process_new_functions): Update.
15956 (cgraph_reset_node): Export.
15957 (cgraph_finalize_function): Update.
15958 (cgraph_add_new_function): Update.
15959 (process_function_and_variable_attributes): Update.
15960 (varpool_finalize_decl): Update.
15961 (symbol_finalized): Remove.
15962 (symbol_finalized_and_needed): Rename to ...
15963 (symbol_defined_and_needed): ... update.
15964 (cgraph_analyze_functions): Update.
15965 (handle_alias_pairs): Update.
15966 (mark_functions_to_output): Update.
15967 (assemble_thunk): Update.
15968 (output_in_order): Update.
15969 (output_weakrefs): Update.
15970 (finalize_compilation_unit): Update.
15971 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15972 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15973 input_node, input_varpool_node): Update.
15974 * dbxout.c (dbxout_expand_expr): Update.
15975 * cgraphclones.c (cgraph_clone_node): Update.
15976 (cgraph_copy_node_for_versioning): Update.
15977 (cgraph_materialize_clone): Update.
15978 (cgraph_materialize_all_clones): Update.
15979 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15980 propagate_pure_const, propagate_nothrow): Update.
15981 * lto-streamer-out.c (lto_output, write_symbol): Update.
15982 * ipa-utils.c (ipa_reverse_postorder): Update.
15983 * ipa-inline.c (can_inline_edge_p): Update.
15984 (update_caller_keys, ipa_inline): Update.
15985 * dwarf2out.c (reference_to_unused,
15986 premark_types_used_by_global_vars_helper): Update.
15987 * tree-eh.c (tree_could_trap_p): Update.
15988 * ipa-split.c (consider_split, execute_split_functions): Update.
15989 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15990 has_addr_references_p): Update; move ahead in file for better
15992 (process_references): Simplify.
15993 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15994 bodies are removed.
15995 (cgraph_comdat_can_be_unshared_p): Make static.
15996 (cgraph_externally_visible_p): Update.
15997 (varpool_externally_visible_p): Update.
15998 (function_and_variable_visibility): Update.
15999 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16000 ipa_tm_mark_force_output_node): Update.
16001 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16002 estimate_edge_devirt_benefit, inline_generate_summary,
16003 inline_write_summary): Update.
16004 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16005 * ipa-prop.c (ipa_compute_jump_functions): Update.
16006 (ipa_print_node_params, ipa_prop_read_section,
16007 ipa_update_after_lto_read, read_replacements_section): Update.
16008 * varasm.c (mark_decl_referenced): Update.
16009 (assemble_alias, dump_tm_clone_pairs): Update.
16010 * tree-inline.c (copy_bb): Update.
16011 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16013 * symtab.c (dump_symtab_base): Print new flags.
16014 (verify_symtab_base): Verify new flags.
16015 (symtab_alias_ultimate_target): New function.
16016 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16017 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16019 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16021 * i386.c (ix86_get_function_versions_dispatcher,
16022 ix86_generate_version_dispatcher_body): Update.
16023 (fold_builtin_cpu): Use varpool_add_new_variable.
16024 * varpool.c (varpool_remove_initializer): Break out from ...
16025 (varpool_remove_node): ... this one.
16026 (dump_varpool_node, varpool_node_for_asm,
16027 cgraph_variable_initializer_availability, varpool_analyze_node,
16028 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16029 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16032 2013-05-29 Jan Hubicka <jh@suse.cz>
16034 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16036 2013-05-29 Easwaran Raman <eraman@google.com>
16038 PR tree-optimization/57442
16039 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16040 when control exits the main loop.
16042 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16044 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16046 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16047 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16048 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16051 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16053 PR tree-optimization/57441
16054 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16055 Don't limit size of incr_vec to number of candidates.
16057 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16059 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16060 and mips16 directories.
16061 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16062 (MULTILIB_DIRNAMES): Ditto.
16063 (MULTILIB_EXCEPTIONS): Add new exceptions.
16064 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16065 (MULTILIB_DIRNAMES): Ditto.
16066 (MULTILIB_EXCEPTIONS): Add new exceptions.
16068 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16069 Marcus Shawcroft <marcus.shawcroft@arm.com>
16071 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16072 SYMBOL_TINY_ABSOLUTE.
16073 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16074 SYMBOL_TINY_ABSOLUTE.
16075 (aarch64_expand_mov_immediate): Likewise.
16076 (aarch64_classify_symbol): Likewise.
16077 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16078 Permit SYMBOL_TINY_ABSOLUTE.
16079 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16081 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16082 Marcus Shawcroft <marcus.shawcroft@arm.com>
16084 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16085 Refactor if/switch. Replace gcc_assert with if.
16087 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16089 * config/i386/i386.c (initial_ix86_tune_features): Enable
16090 FP Reassociation for AMD bdver1 and bdver2.
16092 2013-05-29 Martin Jambor <mjambor@suse.cz>
16094 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16095 and IMAGPART_EXPR do not occur within other handled_components.
16097 2013-05-29 Richard Biener <rguenther@suse.de>
16099 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16100 access on whether the use is in the BB we currently try to
16102 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16103 vectorize to vect_bb_slp_scalar_cost.
16105 2013-05-29 Richard Biener <rguenther@suse.de>
16107 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16108 computing scalar cost offsetted by stmts that are kept live
16110 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16111 for computation of scalar cost.
16113 2013-05-28 Steve Ellcey <sellcey@mips.com>
16115 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16117 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16119 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16120 array notation built-in reduction functions.
16121 * doc/passes.texi (Passes): Added documentation about changes done
16123 * doc/invoke.texi (C Dialect Options): Added documentation about
16124 the -fcilkplus flag.
16125 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16126 (BUILTINS_DEF): Depend on cilkplus.def.
16127 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16128 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16129 * cilkplus.def: New file.
16131 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16133 PR rtl-optimization/57439
16134 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16136 2013-05-28 Easwaran Raman <eraman@google.com>
16138 PR tree-optimization/57337
16139 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16140 (find_insert_point): Correctly identify the insertion point
16141 when two statements with the same UID is compared.
16143 2013-05-28 Richard Biener <rguenther@suse.de>
16145 PR tree-optimization/56787
16146 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16147 from the list of data references.
16148 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16150 (vect_analyze_loop_operations): Likewise.
16151 (vect_transform_loop): Remove clobbers.
16153 2013-05-28 Martin Jambor <mjambor@suse.cz>
16155 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16156 and REALPART_EXPRs have scalar type.
16158 2013-05-28 Richard Biener <rguenther@suse.de>
16160 PR tree-optimization/57411
16161 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16163 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16164 virtual operand propagation.
16166 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16168 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16169 destination register for bmasksi_vis.
16170 (vector_init_bshuffle): Likewise.
16171 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16173 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16175 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16176 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16177 mode if the instruction isn't available in the original mode.
16178 * config/sparc/sparc.opt (mfix-ut699): New option.
16179 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16180 (divdf3): Turn into expander.
16181 (divdf3_nofix): New insn.
16182 (divdf3_fix): Likewise.
16183 (divsf3): Disable if -mfix-ut699.
16184 (sqrtdf2): Turn into expander.
16185 (sqrtdf2_nofix): New insn.
16186 (sqrtdf2_fix): Likewise.
16187 (sqrtsf2): Disable if -mfix-ut699.
16189 2013-05-27 Richard Biener <rguenther@suse.de>
16191 PR middle-end/57412
16192 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16193 block for the new loop.
16195 2013-05-27 Richard Biener <rguenther@suse.de>
16197 PR tree-optimization/57343
16198 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16199 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16200 (number_of_iterations_cond): Do not build the folded tree.
16202 2013-05-27 Richard Biener <rguenther@suse.de>
16205 PR middle-end/57381
16206 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16207 OEP_CONSTANT_ADDRESS_OF retained.
16209 PR tree-optimization/57417
16210 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16211 for unchanged base.
16212 (set_ssa_val_to): Compare addresses using
16213 get_addr_base_and_unit_offset.
16215 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16217 PR rtl-optimization/56833
16218 * postreload.c (move2add_record_mode): New function.
16219 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16220 (move2add_use_add2_insn): Use move2add_record_sym_value.
16221 (move2add_use_add3_insn): Likewise.
16222 (reload_cse_move2add): Use move2add_valid_value_p and
16223 move2add_record_mode. Invalidate call-clobbered and REG_INC
16224 affected regs by setting reg_mode to VOIDmode.
16225 (move2add_note_store): Don't pretend the inside of a SUBREG is
16226 the actual destination. Invalidate single/leading registers by
16227 setting reg_mode to VOIDmode.
16228 Use move2add_record_sym_value, move2add_valid_value_p and
16229 move2add_record_mode.
16231 2013-05-27 Richard Biener <rguenther@suse.de>
16233 PR tree-optimization/57396
16234 * tree-affine.c (double_int_constant_multiple_p): Properly
16235 return false for val == 0 and div != 0.
16237 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16239 * config/mips/mips.h: Use #elif in preprocessor conditions.
16241 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16244 * config/mips/constraints.md (kl): New constraint.
16245 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16246 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16247 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16248 from LO for MIPS16.
16249 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16251 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16254 * config/mips/mips.c (mips_can_inline_p): New function.
16255 (TARGET_CAN_INLINE_P): Define.
16257 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16259 * sched-int.h (ds_t, dw_t): Make unsigned int.
16260 Fix documentation that describes how all the ds_t bits are used.
16261 Reserve the last bit for delayed-branch scheduling.
16262 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16263 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16264 (gen_dep_weak_1): Remove prototype.
16265 * sched-deps.c (get_dep_weak_1): Make static.
16266 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16267 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16268 * doc/tm.texi: Regenerate.
16269 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16271 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16274 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16276 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16277 Sandra Loosemore <sandra@codesourcery.com>
16279 * config.gcc (powerpc-*): Allow native for with-cpu.
16281 2013-05-24 Jeff Law <law@redhat.com>
16283 PR tree-optimization/57124
16284 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16285 conversion feeding a condition if the range has an overflow
16286 if -fstrict-overflow. Add warnings for when we do make the
16289 2013-05-24 Dehao Chen <dehao@google.com>
16291 * tree-cfg.c (locus_discrim_map): Fix the typo.
16292 (locus_discrim_hasher): Likewise.
16293 (locus_discrim_hasher::hash): Likewise.
16294 (locus_discrim_hasher::equal): Likewise.
16296 2013-05-24 Martin Jambor <mjambor@suse.cz>
16298 PR tree-optimization/57294
16299 * cgraph.h (ipa_record_stmt_references): Declare.
16300 * cgraphbuild.c (ipa_record_stmt_references): New function.
16301 (build_cgraph_edges): Use ipa_record_stmt_references.
16302 (rebuild_cgraph_edges): Likewise.
16303 (cgraph_rebuild_references): Likewise.
16304 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16305 associated with the old statement and build references from the
16306 newly built statements.
16307 * ipa-ref.c (ipa_remove_stmt_references): New function.
16308 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16310 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16312 * lra-constraints.c (emit_spill_move): Use smaller mode for
16314 (check_and_process_move): Consider mem-reg moves for secondary
16316 (curr_insn_transform): Don't lose insns emitted before for
16317 secondary memory moves.
16318 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16319 reg set up in the current insn.
16321 2013-05-24 Dehao Chen <dehao@google.com>
16323 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16325 (locus_descrim_hasher::equal): Likewise.
16326 (build_gimple_cfg): New discriminator assignment algorithm.
16327 (make_edges): Likewise.
16328 (next_discriminator_for_locus): Likewise.
16329 (same_line_p): Likewise.
16330 (assign_discriminators): Likewise.
16331 (make_cond_expr_edges): Likewise.
16332 (make_gimple_switch_edges): Likewise.
16333 (make_goto_expr_edges): Likewise.
16334 (make_gimple_asm_edges): Likewise.
16336 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16338 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16339 X format specifier to only display bottom 16 bits.
16340 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16341 immediate to match for operand 2, since it will be masked.
16343 2013-05-24 Richard Biener <rguenther@suse.de>
16345 PR tree-optimization/57287
16346 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16347 all SSA names that occur in abnormal PHIs.
16349 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16352 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16353 that index is not negative.
16355 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16357 PR rtl-optimization/55177
16358 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16359 (simplify_byte_swapping_operation): New.
16360 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16361 (simplify_relational_operation_1): Deal with BSWAP.
16363 2013-05-23 Richard Henderson <rth@redhat.com>
16366 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16367 (ix86_reorg): Call it.
16369 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16372 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16373 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16374 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16376 2013-05-23 Christian Bruel <christian.bruel@st.com>
16379 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16381 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16382 Marcus Shawcroft <marcus.shawcroft@arm.com>
16384 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16385 * config/aarch64/constraints.md (Usa): Remove.
16386 * doc/md.texi (AArch64 Usa): Remove.
16388 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16389 Marcus Shawcroft <marcus.shawcroft@arm.com>
16391 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16392 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16393 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16394 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16396 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16398 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16399 instruction (AdvSIMD).
16400 * config/aarch64/aarch64-builtins.c
16401 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16402 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16404 2013-05-23 Martin Jambor <mjambor@suse.cz>
16406 PR middle-end/57347
16407 * tree.h (contains_bitfld_component_ref_p): Declare.
16408 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16409 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16411 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16412 not access a bit-field. Assert all final offsets are byte-aligned.
16414 2013-05-23 Richard Biener <rguenther@suse.de>
16416 PR tree-optimization/57380
16417 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16418 least one invariant or re-used load.
16419 * passes.c (init_optimization_passes): Move pass_phiprop before
16422 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16424 * config/aarch64/aarch64-simd.md
16425 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16427 2013-05-23 Richard Biener <rguenther@suse.de>
16429 PR middle-end/57381
16430 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16431 OEP_CONSTANT_ADDRESS_OF retained.
16433 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16435 PR middle-end/57344
16436 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16437 don't lower unit. Handle unit not being always BITS_PER_WORD.
16439 2013-05-23 Richard Biener <rguenther@suse.de>
16441 PR rtl-optimization/57341
16442 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16443 instead of true_dependence.
16445 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16447 * bb-reorder.c (branch_threshold): Make const.
16448 (exec_threshold): Ditto.
16450 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16451 Pat Haugen <pthaugen@us.ibm.com>
16452 Peter Bergner <bergner@vnet.ibm.com>
16454 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16455 documentation for the power8 crypto builtins.
16457 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16459 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16460 macros for defining power8 builtin functions.
16461 (BU_P8V_AV_2): Likewise.
16462 (BU_P8V_AV_P): Likewise.
16463 (BU_P8V_VSX_1): Likewise.
16464 (BU_P8V_OVERLOAD_1): Likewise.
16465 (BU_P8V_OVERLOAD_2): Likewise.
16466 (BU_CRYPTO_1): Likewise.
16467 (BU_CRYPTO_2): Likewise.
16468 (BU_CRYPTO_3): Likewise.
16469 (BU_CRYPTO_OVERLOAD_1): Likewise.
16470 (BU_CRYPTO_OVERLOAD_2): Likewise.
16471 (XSCVSPDP): Fix typo, point to the correct instruction.
16472 (VCIPHER): Add power8 crypto builtins.
16473 (VCIPHERLAST): Likewise.
16474 (VNCIPHER): Likewise.
16475 (VNCIPHERLAST): Likewise.
16476 (VPMSUMB): Likewise.
16477 (VPMSUMH): Likewise.
16478 (VPMSUMW): Likewise.
16479 (VPERMXOR_V2DI): Likewise.
16480 (VPERMXOR_V4SI: Likewise.
16481 (VPERMXOR_V8HI: Likewise.
16482 (VPERMXOR_V16QI: Likewise.
16483 (VSHASIGMAW): Likewise.
16484 (VSHASIGMAD): Likewise.
16485 (VPMSUM): Likewise.
16486 (VPERMXOR): Likewise.
16487 (VSHASIGMA): Likewise.
16489 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16490 __CRYPTO__ if the crypto instructions are available.
16491 (altivec_overloaded_builtins): Add support for overloaded power8
16494 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16495 support for power8 crypto builtins.
16496 (builtin_function_type): Likewise.
16497 (altivec_init_builtins): Add support for builtins that take vector
16498 long long (V2DI) arguments.
16500 * config/rs6000/crypto.md: New file, define power8 crypto
16503 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16504 Pat Haugen <pthaugen@us.ibm.com>
16505 Peter Bergner <bergner@vnet.ibm.com>
16507 * doc/invoke.texi (Option Summary): Add power8 options.
16508 (RS/6000 and PowerPC Options): Likewise.
16510 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16511 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16512 wm, wn, wr documentation.
16514 * config/rs6000/constraints.md (wm): New constraint for VSX
16515 registers if direct move instructions are enabled.
16516 (wn): New constraint for no registers.
16517 (wq): New constraint for quad word even GPR registers.
16518 (wr): New constraint if 64-bit instructions are enabled.
16519 (wv): New constraint if power8 vector instructions are enabled.
16520 (wQ): New constraint for quad word memory locations.
16522 * config/rs6000/predicates.md (const_0_to_15_operand): New
16523 constraint for 0..15 for crypto instructions.
16524 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16525 as GPR and floating point registers.
16526 (int_reg_operand): New predicate to match only GPR registers.
16527 (base_reg_operand): New predicate to match base registers.
16528 (quad_int_reg_operand): New predicate to match even GPR registers
16529 for quad memory operations.
16530 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16531 operations in both GPR and VSX registers.
16532 (quad_memory_operand): New predicate for quad memory operations.
16533 (reg_or_indexed_operand): New predicate for direct move support.
16535 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16536 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16537 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16538 (POWERPC_MASKS): Add power8 options.
16539 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16542 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16543 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16545 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16546 (-mpower8-fusion): New power8 options.
16547 (-mpower8-fusion-sign): Likewise.
16548 (-mpower8-vector): Likewise.
16549 (-mcrypto): Likewise.
16550 (-mdirect-move): Likewise.
16551 (-mquad-memory): Likewise.
16553 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16554 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16556 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16557 (rs6000_debug_vector_unit): Add p8_vector.
16558 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16559 definitions. Also print fusion state.
16560 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16561 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16562 (rs6000_option_override_internal): Add support for power8.
16563 (rs6000_common_init_builtins): Add debugging for skipped builtins
16564 if -mdebug=builtin.
16565 (rs6000_adjust_cost): Add power8 support.
16566 (rs6000_issue_rate): Likewise.
16567 (insn_must_be_first_in_group): Likewise.
16568 (insn_must_be_last_in_group): Likewise.
16569 (force_new_group): Likewise.
16570 (rs6000_register_move_cost): Likewise.
16571 (rs6000_opt_masks): Likewise.
16573 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16574 power8 capable assembler, default to power7 options.
16575 (TARGET_DIRECT_MOVE): Likewise.
16576 (TARGET_CRYPTO): Likewise.
16577 (TARGET_P8_VECTOR): Likewise.
16578 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16579 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16580 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16581 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16582 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16583 (TARGET_XSCVDPSPN): Likewise.
16584 (TARGET_XSCVSPDPN): Likewsie.
16585 (TARGET_SYNC_HI_QI): Likewise.
16586 (TARGET_SYNC_TI): Likewise.
16587 (MASK_CRYPTO): Likewise.
16588 (MASK_DIRECT_MOVE): Likewise.
16589 (MASK_P8_FUSION): Likewise.
16590 (MASK_P8_VECTOR): Likewise.
16591 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16592 temporary used by some of the direct move instructions to get two FP
16593 temporary registers does not force creation of a stack frame.
16594 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16595 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16596 that any VSX registers are tieable, even if they are also an
16597 Altivec vector mode.
16598 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16599 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16600 (RS6000_BTM_CRYPTO): Likewise.
16601 (RS6000_BTM_COMMON): Likewise.
16603 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16604 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16605 (enum rs6000_vector): Add power8 vector support.
16607 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16611 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16612 (any_sibcall_could_use_r3): this and handle indirect calls.
16613 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16615 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16617 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16619 2013-05-22 Richard Biener <rguenther@suse.de>
16621 PR middle-end/57349
16622 * profile.c (branch_prob): Do not split blocks that are
16623 abnormally receiving from ECF_RETURNS_TWICE functions.
16625 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16627 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16628 address mode. Move pointer mode initialization to the same place.
16630 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16632 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16633 while it has any effect.
16635 2013-05-21 Easwaran Raman <eraman@google.com>
16637 PR tree-optimization/57322
16638 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16639 UID of the statement added to the BB to be 1.
16641 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16643 PR tree-optimization/57331
16644 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16645 of conversion from pointer type to integral type with integer.
16647 2013-05-21 Martin Jambor <mjambor@suse.cz>
16650 * ipa-prop.c (ipa_read_node_info): Process param_used and
16651 controlled_uses in the same order as when writing.
16653 2013-05-21 Magnus Granberg <baldrick@free.fr>
16656 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16658 2013-05-21 Richard Biener <rguenther@suse.de>
16660 PR tree-optimization/57318
16661 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16662 estimate stmts with side-effects as likely eliminated.
16664 2013-05-21 Richard Biener <rguenther@suse.de>
16666 PR tree-optimization/57330
16667 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16668 preserve the call stmts fntype.
16670 2013-05-21 Richard Biener <rguenther@suse.de>
16672 PR tree-optimization/57303
16673 * tree-ssa-sink.c (statement_sink_location): Improve killing
16674 stmt detection and properly handle self-assignments.
16676 2013-05-21 Christian Bruel <christian.bruel@st.com>
16678 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16679 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16680 registers. Set register size out of the PARALLEL loop.
16682 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16685 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16686 (*fmasf4, *fmasf4_media): New insns.
16688 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16690 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16691 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16692 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16693 (mips_idiv_insns): Update the comments to say that the returned
16694 instruction counts are in units of BASE_INSN_LENGTH.
16695 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16696 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16697 using 2 rather than 4 as the length of indirect MIPS16 and
16698 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16699 length of a NOP. Don't divide MIPS16 lengths by 2.
16700 (mips16_split_long_branches): Assume a branch is long if the
16701 length is greater than 4 rather than 8.
16702 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16703 rather than multiplying them by 2. Multiply instruction counts
16704 by BASE_INSN_LENGTH rather than 4.
16705 (*jump_mips16, tls_get_tp_mips16_<mode>)
16706 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16708 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16710 * config/mips/mips.md (extended_mips16): Remove branch case.
16711 (length): Remove duplicated extended_mips16 test.
16713 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16715 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16717 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16719 * recog.h (Recog_data): Rename to...
16720 (recog_data_d): ...this.
16721 (recog_data): Update accordingly.
16722 * recog.c (recog_data): Likewise.
16723 * reload.c (save_recog_data): Likewise.
16724 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16725 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16727 2013-05-17 Julian Brown <julian@codesourcery.com>
16729 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16730 found in a REG_EQUAL note, invalidate it.
16732 2013-05-17 Easwaran Raman <eraman@google.com>
16734 * tree-ssa-reassoc.c (find_insert_point): New function.
16735 (insert_stmt_after): Likewise.
16736 (get_def_stmt): Likewise.
16737 (ensure_ops_are_available): Likewise.
16738 (not_dominated_by): Likewise.
16739 (rewrite_expr_tree): Do not move statements beyond what is
16740 necessary. Remove call to swap_ops_for_binary_stmt...
16741 (reassociate_bb): ... and move it here.
16742 (build_and_add_sum): Assign UIDs for new statements.
16743 (linearize_expr): Likewise.
16744 (do_reassoc): Renumber gimple statement UIDs.
16746 2013-05-17 Jan Hubicka <jh@suse.cz>
16748 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16750 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16751 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16753 (output_weakrefs): Update.
16755 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16756 Martin Jambor <mjambor@suse.cz>
16758 PR middle-end/57276
16759 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16760 value that corresponds to the given aggval is found in values vector.
16762 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16764 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16765 sse, sse2, sse3, ssse3 and sse4a flags to options.
16767 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16769 * gengtype-state.c: (s_expr_writer): New class, to handle
16770 prettifying of output layout of s-expressions.
16771 (state_writer): New class, to write out gtype.state.
16772 (state_written_type_count): Move this variable into member data of
16774 (s_expr_writer::s_expr_writer): New code: constructor for new class
16775 (state_writer::state_writer(): ditto
16776 (s_expr_writer::write_new_line): New function
16777 (s_expr_writer::write_any_indent): ditto
16778 (s_expr_writer::begin_s_expr): ditto
16779 (s_expr_writer::end_s_expr): ditto
16780 (write_state_fileloc): convert to method of state_writer...
16781 (state_writer:: write_state_fileloc): ...and use methods of
16782 s_expr_writer to write indentation into the gtype.state output file
16783 to visually represent the hierarchical structure of the list
16785 (write_state_fields): ditto, renaming to...
16786 (state_writer::write_state_fields)
16787 (write_state_a_string): ditto, renaming to...
16788 (state_writer::write_state_a_string)
16789 (write_state_string_option): ditto, renaming to...
16790 (state_writer::write_state_string_option)
16791 (write_state_type_option): ditto, renaming to...
16792 (state_writer::write_state_type_option)
16793 (write_state_nested_option): ditto, renaming to...
16794 (state_writer::write_state_nested_option)
16795 (write_state_option): ditto, renaming to...
16796 (state_writer::write_state_option)
16797 (write_state_options): ditto, renaming to...
16798 (state_writer::write_state_options)
16799 (write_state_lang_bitmap): ditto, renaming to...
16800 (state_writer::write_state_lang_bitmap)
16801 (write_state_version): ditto, renaming to...
16802 (state_writer::write_state_version)
16803 (write_state_scalar_type): ditto, renaming to...
16804 (state_writer::write_state_scalar_type)
16805 (write_state_string_type): ditto, renaming to...
16806 (state_writer::write_state_string_type)
16807 (write_state_undefined_type): ditto, renaming to...
16808 (state_writer::write_state_undefined_type)
16809 (write_state_struct_union_type): ditto, renaming to...
16810 (state_writer::write_state_struct_union_type)
16811 (write_state_struct_type): ditto, renaming to...
16812 (state_writer::write_state_struct_type)
16813 (write_state_user_struct_type): ditto, renaming to...
16814 (state_writer::write_state_user_struct_type)
16815 (write_state_lang_struct_type): ditto, renaming to...
16816 (state_writer::write_state_lang_struct_type)
16817 (write_state_param_struct_type): ditto, renaming to...
16818 (state_writer::write_state_param_struct_type)
16819 (write_state_pointer_type): ditto, renaming to...
16820 (state_writer::write_state_pointer_type)
16821 (write_state_array_type): ditto, renaming to...
16822 (state_writer::write_state_array_type)
16823 (write_state_gc_used): ditto, renaming to...
16824 (state_writer::write_state_gc_used)
16825 (write_state_common_type_content): ditto, renaming to...
16826 (state_writer::write_state_common_type_content)
16827 (write_state_type): ditto, renaming to...
16828 (state_writer::write_state_type)
16829 (write_state_pair_list): ditto, renaming to...
16830 (state_writer::write_state_pair_list)
16831 (write_state_pair): ditto, renaming to...
16832 (state_writer::write_state_pair)
16833 (write_state_typedefs): ditto, renaming to...
16834 (state_writer::write_state_typedefs)
16835 (write_state_structures): ditto, renaming to...
16836 (state_writer::write_state_structures)
16837 (write_state_param_structs): ditto, renaming to...
16838 (state_writer::write_state_param_structs)
16839 (write_state_variables): ditto, renaming to...
16840 (state_writer::write_state_variables)
16841 (write_state_srcdir): ditto, renaming to...
16842 (state_writer::write_state_srcdir)
16843 (write_state_files_list): ditto, renaming to...
16844 (state_writer::write_state_files_list)
16845 (write_state_languages): ditto, renaming to...
16846 (state_writer::write_state_languages)
16847 (write_state): create a state_writer instance and use it when
16848 writing out the state file
16850 2013-05-17 Mike Stump <mikestump@comcast.net>
16852 PR rtl-optimization/57304
16853 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16854 accessing DF_REF_REAL_LOC.
16856 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16858 PR rtl-optimization/57281
16859 PR rtl-optimization/57300
16860 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16861 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16862 what the other splitter did if the registers are dead.
16864 2013-05-17 Richard Biener <rguenther@suse.de>
16866 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16869 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16871 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16874 2013-05-17 Marek Polacek <polacek@redhat.com>
16876 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16877 length when doing non-zero store of storing '\0' to '\0'.
16879 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16881 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16882 vect_external_def oprnd1 with loop_vinfo, try to emit
16883 optional cast, negation and and stmts on the loop preheader
16884 edge instead of into the pattern def seq.
16886 PR tree-optimization/57051
16887 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16888 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16890 2013-05-16 Nick Clifton <nickc@redhat.com>
16892 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16893 (rl78_is_naked_func): New function.
16894 (rl78_expand_prologue): Skip prologue generation for naked functions.
16895 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16896 * doc/extend.texi (naked): Add RL78 to the list of processors
16897 that supports this attribute.
16899 2013-05-16 Jeff Law <law@redhat.com>
16901 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16903 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16905 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16906 cache parameters using detect_caches_amd also for CYRIX,
16907 NSC and TM2 signatures.
16909 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16910 Dzianis Kahanovich <mahatma@eu.by>
16914 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16915 VIA/Centaur processors and determine their cache parameters
16916 using detect_caches_amd.
16918 2013-05-16 Teresa Johnson <tejohnson@google.com>
16920 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16921 (rtl_verify_edges): New function.
16922 (rtl_verify_bb_insns): Ditto.
16923 (rtl_verify_bb_pointers): Ditto.
16924 (rtl_verify_bb_insn_chain): Ditto.
16925 (rtl_verify_fallthru): Ditto.
16926 (rtl_verify_bb_layout): Ditto.
16927 (rtl_verify_flow_info_1): Outline checks into new functions.
16928 (rtl_verify_flow_info): Ditto.
16930 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16932 * cfghooks.c (copy_bbs): Add update_dominance argument.
16933 * cfghooks.h (copy_bbs): Update prototype.
16934 * tree-cfg.c (gimple_duplicate_sese_region):
16935 Add update_dominance argument.
16936 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16937 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16938 gimple_duplicate_sese_region call.
16939 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16940 Update copy_bbs call.
16941 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16942 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16944 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16946 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16947 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16948 vect_recog_rotate_pattern.
16949 (vect_recog_rotate_pattern): New function.
16951 2013-05-16 Jason Merrill <jason@redhat.com>
16953 * Makefile.in (LLINKER): New variable.
16954 (mostlyclean): Remove link mutex.
16955 * configure.ac: Handle --enable-link-mutex.
16956 * lock-and-run.sh: New script.
16958 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16961 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16964 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16966 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16968 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16970 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16971 * config/arm/arm.c (next_consecutive_mem): New function.
16972 (gen_movmem_ldrd_strd): Likewise.
16973 * config/arm/arm.md (movmemqi): Update condition and code.
16974 (unaligned_loaddi, unaligned_storedi): New patterns.
16976 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16978 * config.gcc: Obsolete *-*-solaris2.9*.
16979 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16981 2013-05-16 Richard Biener <rguenther@suse.de>
16983 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16984 earlier, after GRAPHITE transforms and IV canonicalization.
16986 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16988 * omp-low.c (extract_omp_for_data): For collapsed loops,
16989 if at least one of the loops is known at compile time to
16990 iterate zero times, set count to 0.
16991 (expand_omp_regimplify_p): New function.
16992 (expand_omp_for_generic): For collapsed loops, if at least
16993 one of the loops isn't known to iterate at least once,
16994 add runtime check with setting count to 0.
16995 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16996 For unsigned types if it isn't known at compile time that
16997 the loop will iterate at least once, add runtime check to bypass
16998 the whole loop if initial condition isn't true.
17000 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17002 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17004 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17006 PR middle-end/57286
17007 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17008 transformations to avoid an infinite loop.
17010 2013-05-16 Marek Polacek <polacek@redhat.com>
17012 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17014 2013-05-15 Leif Ekblad <leif@rdos.net>
17016 * config/i386/i386.c (ix86_decompose_address): Use
17017 DEFAULT_TLS_SEG_REG to access TLS segment register.
17018 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17019 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17020 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17022 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17025 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17026 sibling calls to functions that would normally be lazily bound,
17027 unless $gp is call-clobbered.
17029 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17031 * config/i386/i386.c (ix86_option_override_internal): Update
17032 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17033 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17034 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17035 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17037 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17039 2013-05-15 Andreas Schwab <schwab@suse.de>
17041 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17042 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17045 2013-05-15 Teresa Johnson <tejohnson@google.com>
17047 * loop-unroll.c (report_unroll_peel): Check decision before
17048 emitting unroll/peel message.
17050 2013-05-15 Teresa Johnson <tejohnson@google.com>
17052 * function.h (has_bb_partition): New rtl_data flag.
17053 (bb_reorder_complete): Ditto.
17054 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17055 instead of flag_reorder_blocks_and_partition.
17056 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17057 with some enhancements.
17058 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17059 * bb-reorder.c (connect_traces): Check for has_bb_partition
17060 instead of flag_reorder_blocks_and_partition.
17061 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17062 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17063 verify_hot_cold_block_grouping.
17064 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17066 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17069 * config/arm/predicates.md (call_insn_operand): New predicate.
17070 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17071 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17072 if insn is not a tail call.
17073 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17075 * config/arm/arm.h (enum reg_class): New caller save register class.
17076 (REG_CLASS_NAMES): Likewise.
17077 (REG_CLASS_CONTENTS): Likewise.
17078 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17081 2013-05-15 Richard Biener <rguenther@suse.de>
17083 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17084 of MSG_OPTIMIZED_LOCATIONS.
17085 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17086 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17088 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17089 of MSG_OPTIMIZED_LOCATIONS.
17090 (execute_vect_slp): Likewise.
17091 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17092 (vect_create_cond_for_alias_checks): Likewise.
17093 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17094 (vect_recog_widen_mult_pattern): Likewise.
17095 (vect_recog_widen_sum_pattern): Likewise.
17096 (vect_recog_over_widening_pattern): Likewise.
17097 (vect_recog_widen_shift_pattern): Likewise.
17098 (vect_recog_vector_vector_shift_pattern): Likewise.
17099 (vect_recog_divmod_pattern): Likewise.
17100 (vect_recog_mixed_size_cond_pattern): Likewise.
17101 (vect_recog_bool_pattern): Likewise.
17102 (vect_pattern_recog_1): Likewise.
17104 2013-05-15 Martin Jambor <mjambor@suse.cz>
17106 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17107 non-functions to builtin_unreachable.
17108 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17109 correct when new direct edges were discovered.
17111 2013-05-15 Martin Jambor <mjambor@suse.cz>
17113 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17114 header, print symbol order instead of node uid, print more information
17115 about indirect edge targets.
17116 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17118 (ipa_make_edge_direct_to_target): Likewise.
17119 (remove_described_reference): Likewise.
17120 (propagate_controlled_uses): Likewise.
17121 (ipa_print_node_params): Also print symbol order.
17122 (ipcp_transform_function): Print symbol order instead of node uids.
17123 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17124 (cgraph_get_create_real_symbol_node): Likewise.
17125 * ipa-cp.c (print_lattice): Likewise.
17126 (print_all_lattices): Likewise.
17127 (determine_versionability): Likewise.
17128 (initialize_node_lattices): Likewise.
17129 (estimate_local_effects): Likewise.
17130 (update_profiling_info): Likewise.
17131 (create_specialized_node): Likewise.
17132 (perhaps_add_new_callers): Likewise.
17133 (decide_about_value): Likewise.
17134 (decide_whether_version_node): Likewise.
17135 (identify_dead_nodes): Likewise.
17136 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17137 (dump_inline_summary): Likewise.
17138 (estimate_node_size_and_time): Likewise.
17139 (inline_analyze_function): Likewise.
17140 * ipa-inline.c (report_inline_failed_reason): Likewise.
17141 (want_early_inline_function_p): Likewise.
17142 (edge_badness): Likewise.
17143 (update_edge_key): Likewise.
17144 (inline_small_functions): Likewise. Add dumping of order to two other
17146 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17147 instead of node uids.
17148 (propagate_pure_const): Likewise.
17149 (propagate_pure_const): Likewise.
17150 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17151 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17152 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17154 * tree-pretty-print.c (dump_function_header): Likewise.
17155 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17156 Print symbol order instead of node uids.
17158 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17160 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17161 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17163 2013-05-15 Richard Biener <rguenther@suse.de>
17165 PR tree-optimization/57275
17166 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17167 return value for fail to do runtime alias checks for gather loads.
17169 2013-05-15 Jan Hubicka <jh@suse.cz>
17173 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17174 weakrefs are not external.
17175 (lto_symtab_merge_decls): Fix thinko when dealing with
17176 non-lto_symtab decls.
17177 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17178 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17179 * varpool.c (dump_varpool_node): Dump more flags.
17181 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17183 * config/i386/i386.c (processor_alias_table): Add instruction
17184 FSGSBASE for AMD bdver3 architecture.
17186 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17188 * tree.c (warn_deprecated_use): Print file:line using locus color.
17189 * diagnostic.c (diagnostic_report_current_module): Print file:line
17190 and file:line:column using locus color.
17192 2013-05-14 Mike Stump <mikestump@comcast.net>
17194 * gdbinit.in: Add __null.
17196 2013-05-14 Mike Stump <mikestump@comcast.net>
17198 * recog.h: Rename struct recog_data to Recog_data.
17199 * recog.c: Likewise.
17200 * reload.c (can_reload_into): Likewise.
17201 * config/picochip/picochip.c: Likewise.
17203 2013-05-14 Mike Stump <mikestump@comcast.net>
17205 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17207 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17209 * resource.h (struct resources): Remove unch_memory member.
17210 (CLEAR_RESOURCE): Don't clear unch_memory.
17211 * resource.c (mark_referenced_resources): Don't set it.
17212 (mark_set_resources): Likewise.
17213 (mark_target_live_regs): Don't clear it.
17214 (init_resource_info): Likewise.
17215 * reorg.c (resource_conflicts_p): Don't compare it.
17216 (redundant_insn): Don't set it.
17218 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17220 * emit-rtl.c (next_label): Remove unused function.
17221 (skip_consecutive_labels, link_cc0_insns): Move to ...
17222 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17223 only place where these functions are used, and make them static.
17225 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17227 * fold-const.c (fold_negate_expr): Handle vectors.
17228 (fold_truth_not_expr): Make it static.
17229 (fold_invert_truthvalue): New static function.
17230 (invert_truthvalue_loc): Handle vectors. Do not call
17231 fold_truth_not_expr directly.
17232 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17233 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17234 (fold_comparison): Handle vector constants.
17235 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17236 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17237 * tree.h (fold_truth_not_expr): Remove declaration.
17239 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17241 * config/aarch64/aarch64-simd.md
17242 (aarch64_vcond_internal<mode>): Rename to...
17243 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17244 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17245 float modes. Clarify all iterator modes.
17246 (vcond<mode><mode>): Use new name for vcond expanders.
17247 (vcond<v_cmp_result><mode>): Likewise.
17248 (vcondu<mode><mode>: Likewise.
17249 * config/aarch64/iterators.md (VDQF_COND): New.
17251 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17254 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17255 variable for the shift amount. Check that we shift by non-negative
17258 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17261 * config/arm/arm.h (EPILOGUE_USES): Only return true
17262 for LR_REGNUM after epilogue_completed.
17264 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17266 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17267 is error_mark_node.
17269 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17272 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17273 and Solaris 11+/x86 with gld.
17274 * configure: Regenerate.
17276 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17278 * expmed.c (expand_shift_1): Canonicalize rotates by
17279 constant bitsize / 2 to bitsize - 1.
17280 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17281 case ROTATERT>: Likewise.
17284 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17286 * config/i386/i386.md (rotateinv): New code attr.
17287 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17288 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17289 roll $31, %eax, etc.
17291 2013-05-14 Richard Biener <rguenther@suse.de>
17293 PR middle-end/57235
17294 * tree-eh.c (sink_clobbers): Give up for successors with
17295 multiple predecessors and no virtual uses.
17297 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17299 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17300 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17302 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17304 PR middle-end/57251
17305 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17306 the case when both op0 and op1 have VOIDmode.
17308 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17310 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17311 in multiply-accumulate mode.
17313 2013-05-13 Guozhi Wei <carrot@google.com>
17315 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17317 2013-05-13 Kai Tietz <ktietz@redhat.com>
17320 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17321 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17322 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17323 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17324 (ix86_expand_prologue): Likewise.
17325 (ix86_expand_split_stack_prologue): Likewise.
17326 (legitimate_pic_address_disp_p): Likewise.
17327 (legitimize_pic_address): Likewise.
17328 (legitimize_tls_address): Likewise.
17329 (legitimize_pe_coff_symbol): Likewise.
17330 (output_pic_addr_const): Likewise.
17331 (construct_plt_address): Likewise.
17332 (ix86_expand_call): Likewise.
17333 (x86_output_mi_thunk): Likewise.
17334 (x86_function_profiler): Likewise.
17336 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17338 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17339 similar switch cases.
17340 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17341 (aarch64_simd_mov_to_<mode>low): Delete.
17342 (aarch64_simd_mov_to_<mode>high): Delete.
17343 (move_lo_quad_<mode>): Add w<-r alternative.
17344 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17345 (aarch64_simd_mov_from_*): Update type attribute.
17346 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17349 2013-05-13 Jan Hubicka <jh@suse.cz>
17351 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17352 * config/i386/i386.c (ix86_compute_frame_layout,
17353 ix86_expand_epilogue, emit_i387_cw_initialization,
17354 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17355 ix86_local_alignment): Fix use of size/speed predicates.
17357 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17359 PR tree-optimization/45216
17360 PR tree-optimization/57157
17361 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17362 the (-Y) & (B - 1) variant if OP is |.
17363 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17364 return shifted. Use (-op1) & (prec - 1) as other_amount
17365 instead of prec - op1.
17367 2013-05-13 Martin Jambor <mjambor@suse.cz>
17369 PR middle-end/42371
17370 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17371 (ipa_constant_data): New type.
17372 (ipa_jump_func): Use ipa_constant_data to hold information about
17373 constant jump functions.
17374 (ipa_get_jf_constant): Adjust to jump function type changes.
17375 (ipa_get_jf_constant_rdesc): New function.
17376 (ipa_param_descriptor): New field controlled_uses.
17377 (ipa_get_controlled_uses): New function.
17378 (ipa_set_controlled_uses): Likewise.
17379 * ipa-ref.h (ipa_find_reference): Declare.
17380 * ipa-prop.c (ipa_cst_ref_desc): New type.
17381 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17383 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17384 New parameter cs. Adjust all callers.
17385 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17386 (remove_described_reference): New function.
17387 (jfunc_rdesc_usable): Likewise.
17388 (try_make_edge_direct_simple_call): Decrement controlled use count,
17389 attempt to remove reference if it hits zero.
17390 (combine_controlled_uses_counters): New function.
17391 (propagate_controlled_uses): Likewise.
17392 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17393 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17394 (ipa_print_node_params): Print described use counter.
17395 (ipa_write_jump_function): Adjust to jump function type changes.
17396 (ipa_read_jump_function): New parameter CS, pass it to
17397 ipa_set_jf_constant. Adjust caller.
17398 (ipa_write_node_info): Stream controlled use count
17399 (ipa_read_node_info): Likewise.
17400 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17402 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17403 count. Remove cloning-added reference if it reaches zero.
17404 * ipa-ref.c (ipa_find_reference): New function.
17406 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17408 * config/i386/i386.c (processor_target_table): Modified default
17409 alignment values for AMD BD and BT architectures.
17411 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17413 * tree-vect-generic.c (uniform_vector_p): Move ...
17414 * tree.c (uniform_vector_p): ... here.
17415 * tree.h (uniform_vector_p): Declare it.
17416 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17419 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17421 PR tree-optimization/57230
17422 * tree-ssa-strlen.c (handle_char_store): Record length for
17423 array store from STRING_CST.
17425 PR tree-optimization/57230
17426 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17429 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17431 * config/epiphany/epiphany.c (epiphany_init): Check size of
17432 NUM_MODES_FOR_MODE_SWITCHING.
17433 (epiphany_expand_prologue):
17434 Remove CONFIG_REGNUM initial value handling code.
17435 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17436 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17437 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17438 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17439 Don't return 1 for FP_MODE_NONE.
17440 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17441 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17442 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17443 * config/epiphany/epiphany.md (save_config): New pattern.
17445 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17447 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17449 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17451 * config/i386/i386.md (memory): Handle sseishft1.
17452 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17453 (*vec_extractv2di_1): Ditto.
17455 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17457 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17460 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17462 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17463 Add mthumb/march=armv7-a multilib.
17464 Add mthumb/march=armv7-r multilib.
17465 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17467 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17469 * config/v850/t-rtems: Add more multilibs.
17471 2013-05-10 Richard Biener <rguenther@suse.de>
17473 PR tree-optimization/57214
17474 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17475 not propagate from SSA names that occur in abnormal PHI nodes.
17477 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17479 * stor-layout.c (element_precision): New function.
17480 * machmode.h (element_precision): Declare it.
17481 * tree.c (build_minus_one_cst): New function.
17482 (element_precision): Likewise.
17483 * tree.h (build_minus_one_cst): Declare new function.
17484 (element_precision): Likewise.
17485 * fold-const.c (operand_equal_p): Use element_precision.
17486 (fold_binary_loc): Handle vector types.
17487 * convert.c (convert_to_integer): Use element_precision.
17488 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17491 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17493 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17494 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17495 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17496 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17497 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17498 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17499 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17500 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17501 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17502 (Uuw8): New constraints.
17503 (Usb4): Move into alphabetical order.
17504 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17505 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17506 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17507 previously unnamed patterns.
17508 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17509 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17510 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17511 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17512 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17513 of specific lengths.
17515 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17517 * config/i386/i386.md (rotateinv): New code attr.
17518 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17519 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17520 roll $31, %eax, etc.
17522 PR tree-optimization/45216
17523 PR tree-optimization/57157
17524 * tree-ssa-forwprop.c (simplify_rotate): New function.
17525 (ssa_forward_propagate_and_combine): Call it.
17527 2013-05-10 Richard Biener <rguenther@suse.de>
17529 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17530 disable peeling when we version for aliasing.
17531 (vector_alignment_reachable_p): Honor explicit user alignment.
17532 (vect_supportable_dr_alignment): Likewise.
17533 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17534 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17535 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17536 then peeling to arrange for the cost-model check to come first.
17538 2013-05-10 Alan Modra <amodra@gmail.com>
17540 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17541 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17542 * configure: Regenerate.
17544 2013-05-10 Alan Modra <amodra@gmail.com>
17547 * varasm.c (default_elf_select_section): Move !DECL_P check..
17548 (get_named_section): ..to here before calling get_section_name.
17550 (default_section_type_flags): Add DECL_P check.
17551 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17552 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17554 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17556 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17557 When using gen_stack_adjust_str with a register offset, add a
17558 REG_FRAME_RELATED_EXPR note.
17560 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17562 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17563 (*vec_extractv4si_zext_mem): Ditto.
17564 (*vec_extractv2di): Add 0->x and x->x alternatives.
17565 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17566 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17568 2013-05-09 Jason Merrill <jason@redhat.com>
17570 N3639 C++1y VLA support
17571 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17574 * tree.c (build_constructor_va): New.
17575 * tree.h: Declare it.
17577 2013-05-09 Martin Jambor <mjambor@suse.cz>
17580 * gimple-fold.c (canonicalize_constructor_val): Call
17581 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17583 2013-05-09 Jan Hubicka <jh@suse.cz>
17584 Richard Biener <rguenther@suse.de>
17587 * symtab.c (symtab_make_decl_local): Do not add private names.
17589 2013-05-09 Jan Hubicka <jh@suse.cz>
17592 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17593 (unlink_from_assembler_name_hash): Likewise.
17594 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17595 symtab_unregister_node, symtab_initialize_asm_name_hash,
17596 change_decl_assembler_name): Update.
17598 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17600 * config/aarch64/aarch64.md: New movtf split.
17601 (*movtf_aarch64): Update.
17602 (aarch64_movdi_tilow): Handle TF modes and rename to
17603 aarch64_movdi_<mode>low.
17604 (aarch64_movdi_tihigh): Handle TF modes and rename to
17605 aarch64_movdi_<mode>high
17606 (aarch64_movtihigh_di): Handle TF modes and rename to
17607 aarch64_mov<mode>high_di
17608 (aarch64_movtilow_di): Handle TF modes and rename to
17609 aarch64_mov<mode>low_di
17610 (aarch64_movtilow_tilow): Remove spurious whitespace.
17611 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17613 (aarch64_print_operand): Update.
17615 2013-05-09 Alan Modra <amodra@gmail.com>
17617 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17619 * configure: Regenerate.
17621 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17623 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17624 splitter preparation statements.
17625 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17626 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17629 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17631 * gimple-ssa-strength-reduction.c (count_candidates): Change
17632 return value to int.
17633 (analyze_candidates_and_replace): Change type of length to int.
17635 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17637 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17638 (*vec_extract<mode>): Use VI12_128 mode iterator.
17639 (*vec_extract<mode>_mem): Ditto.
17640 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17643 2013-05-08 Diego Novillo <dnovillo@google.com>
17648 2012-08-17 Diego Novillo <dnovillo@google.com>
17651 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17652 * config.in: Regenerate.
17653 * configure: Regenerate.
17654 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17656 2013-05-08 Jan Hubicka <jh@suse.cz>
17659 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17660 * cgraph.h (symtab_node_base): Add unique_name.
17661 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17662 input_overwrite_node, input_varpool_node): Stream unique_name.
17663 * cgraphclones.c (cgraph_create_virtual_clone,
17664 cgraph_function_versioning): Set unique_name.
17665 * ipa.c (function_and_variable_visibility): Set unique_name.
17667 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17669 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17670 (alloc_cand_and_find_basis): Restrict conditional candidate
17671 processing to CAND_MULTs.
17673 2013-05-08 Jan Hubicka <jh@suse.cz>
17676 lto-symtab.c (lto_symtab_symbol_p): New function.
17677 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17678 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17679 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17680 Skip static symbols.
17682 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17684 PR tree-optimization/57200
17685 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17686 Only call inform if the preceding warning_at returns true.
17688 2013-05-07 Han Shen <shenhan@google.com>
17690 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17691 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17692 * common.opt (fstack-protector-strong): New option.
17693 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17694 * doc/invoke.texi (Optimization Options): Document
17695 "-fstack-protector-strong".
17696 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17698 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17700 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17702 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17704 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17705 it is not lhs of the insn.
17706 (setup_reg_equiv): Remove insn having equiv memory which it is not
17708 * lra-constraints.c (process_address): Try to improve generation
17709 code for address base + disp.
17710 (lra_constraints): Make correct the code for checking insn setting
17711 up backward equivalence. Remove insn only if it is in the init
17713 * lra-eliminations.c (update_reg_eliminate): Change return value.
17714 (lra_eliminate): Use the result.
17716 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17718 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17719 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17720 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17721 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17722 register target operands.
17723 (*vec_extractv8hi_sse2): New pattern.
17724 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17725 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17726 (*vec_extract<mode>_mem): New insn and split pattern.
17728 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17730 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17731 (TARGET_ASAN_SHADOW_OFFSET): Define.
17732 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17733 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17735 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17737 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17738 (incr_vec_index): Return -1 if increment not found.
17739 (create_add_on_incoming_edge): Assert if increment not found.
17740 (record_increment): Limit number of increments recorded.
17741 (all_phi_incrs_profitable): Return false if an increment not found.
17742 (replace_profitable_candidates): Don't process increments that were
17744 (analyze_candidates_and_replace): Limit size of incr_vec.
17746 2013-05-07 Richard Biener <rguenther@suse.de>
17748 * calls.c (special_function_p): setjmp-like functions are leaf.
17749 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17750 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17752 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17754 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17755 (aarch64_simd_mov<mode>): New expander.
17756 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17757 (aarch64_simd_mov_to_<mode>high): Likewise.
17758 (aarch64_simd_mov_from_<mode>low): Likewise.
17759 (aarch64_simd_mov_from_<mode>high): Likewise.
17760 (aarch64_dup_lane<mode>): Update.
17761 (aarch64_dup_lanedi): New instruction pattern.
17762 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17763 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17765 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17767 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17768 (replace_mult_candidate): Remove unnecessary argument; remove
17769 unnecessary parameter from call to introduce_cast_before_cand.
17770 (replace_unconditional_candidate): Remove unnecessary parameter
17771 from call to replace_mult_candidate.
17772 (replace_conditional_candidate): Likewise.
17773 (insert_initializers): Use make_temp_ssa_name.
17774 (introduce_cast_before_cand): Remove unnecessary argument; use
17775 make_temp_ssa_name.
17776 (replace_one_candidate): Remove unnecessary argument; remove
17777 unnecessary parameter from calls to introduce_cast_before_cand.
17778 (replace_profitable_candidates): Remove unnecessary parameters
17779 from calls to replace_one_candidate.
17781 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17783 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17784 phi def as possibly hiding a basis for a CAND_ADD whose operands
17785 have been commuted in the analysis.
17786 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17788 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17790 * config/aarch64/aarch64.md
17791 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17792 shift value between 0-4.
17794 2013-05-07 Richard Biener <rguenther@suse.de>
17796 * double-int.h (rshift): New overload.
17797 * double-int.c (rshift): New function.
17798 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17799 (create_reference_ops_from_ref): Remove.
17800 (vn_reference_insert): Use shared ops for constructing the
17801 reference and copy it.
17803 2013-05-07 Richard Biener <rguenther@suse.de>
17805 PR middle-end/57190
17806 * tree-eh.c (sink_clobbers): Properly propagate
17807 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17809 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17811 PR tree-optimization/57149
17812 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17813 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17814 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17815 uninit_undefined_value_p instead of ssa_undefined_value_p.
17818 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17819 for modifier == EXPAND_INITIALIZER.
17821 2013-05-07 Anton Blanchard <anton@samba.org>
17823 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17824 for powerpc64 little endian.
17825 * configure: Regenerate.
17827 2013-05-06 Graham Stott <grahams@btinternet.com>
17829 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17830 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17831 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17832 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17834 2013-05-06 Graham Stott <grahams@btinternet.com>
17836 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17837 codes which allow non-lvalues.
17839 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17841 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17842 components are all 1s.
17843 (integer_minus_onep): New function.
17844 * tree.h (integer_minus_onep): Declare it.
17845 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17846 integer_minus_onep instead of integer_all_onesp.
17848 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17851 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17852 variations of these patterns.
17854 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17856 * config/i386/i386.md (isa): Add x64_sse4 member.
17857 (enabled): Handle x64_sse4.
17858 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17859 instruction for 64bit SSE4_1 targets. Update insn attributes.
17860 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17861 instruction for SSE4_1 targets. Update insn attributes.
17862 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17863 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17864 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17866 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17867 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17868 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17870 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17873 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17875 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17877 * final.c (do_assembler_dialects): Don't handle curly braces and
17878 vertical bar escaped by % as dialect delimiters.
17879 (output_asm_insn): Print curly braces and vertical bar if escaped
17880 by % and ASSEMBLER_DIALECT defined.
17881 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17882 * doc/tm.texi: Regenerated.
17884 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17886 * config/mips/mips.c: Include tree-pass.h.
17887 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17888 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17889 (pass_mips_machine_reorg2): New machine specific pass.
17890 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17891 (mips_option_override): Register the new pass.
17892 * rtl.h (cleanup_barriers): Remove prototype.
17893 (dbr_schedule): Likewise.
17894 * jump.c (cleanup_barriers): Make static.
17895 * reorg.c (dbr_schedule): Likewise.
17897 2013-05-06 Richard Biener <rguenther@suse.de>
17899 PR tree-optimization/57185
17900 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17901 SSA names properly.
17903 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17906 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17907 between "first_arg" and "insn", not "last" and "insn".
17909 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17911 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17912 (find_candidates_in_block): Re-enable slsr_process_phi.
17913 (create_phi_basis): Fix double counting of candidate adjustment.
17915 2013-05-06 Richard Biener <rguenther@suse.de>
17917 PR middle-end/57147
17918 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17919 the edge is also fallthru, preserve it and just clear the
17921 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17922 also complex, preserve that and just clear the fallthru flag.
17923 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17924 update virtual operands.
17926 2013-05-06 Alan Modra <amodra@gmail.com>
17928 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17929 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17930 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17931 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17932 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17933 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17935 2013-05-06 Alan Modra <amodra@gmail.com>
17937 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17938 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17939 (DEFAULT_ASM_ENDIAN): Define.
17940 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17941 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17942 Update -K PIC clause from sysv4.h.
17943 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17944 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17946 2013-05-06 Alan Modra <amodra@gmail.com>
17948 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17949 twice for little-endian.
17950 (ashrdi3_no_power, ashrdi3): Support little-endian.
17952 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17955 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17956 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17958 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17959 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17960 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17963 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17964 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17966 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17967 * config/pa/pa.opt: Make mbig-switch a no-op.
17968 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17969 (CASE_VECTOR_MODE): Always return SImode.
17970 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17971 for the !TARGET_BIG_SWITCH case.
17972 * config/pa/pa-linux.h: Likewise.
17973 * config/pa/pa-openbsd.h: Likewise.
17974 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17975 * config/pa/pa.md (short_jump): Remove define_insn.
17976 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17977 (casesi0): Remove define_insn.
17978 (type): Remove btable_branch.
17979 (pa_combine_type): Likewise.
17980 (in_nullified_branch_delay): Likewise.
17981 (in_call_delay): Likewise.
17982 (define_delay): Likewise.
17983 (define_insn_reservation "Z3"): Likewise.
17984 (define_insn_reservation "Z4"): Likewise.
17985 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17986 (pa_adjust_insn_length): Remove adjustment for btable branches.
17987 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17990 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17992 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17993 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17994 Add m->r,x alternatives.
17995 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17996 splitters using SWI48x mode iterator.
17997 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17998 TARGET_64BIT. Add m->x alternative.
17999 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18000 Add o->x alternative. Enable for TARGET_SSE.
18001 (sse_storeq): Remove expander.
18002 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18003 with memory input operand.
18004 (*vec_extractv2di_1 splitter): New.
18005 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18006 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18008 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18010 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18011 (INT_LOWPART): Delete.
18012 (extract_MB): Adjust.
18013 (extract_ME): Adjust.
18014 (print_operand): Adjust.
18016 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18018 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18019 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18020 (reg_or_logical_cint_operand, easy_fp_constant,
18021 logical_const_operand): Delete "CONST_DOUBLE" case.
18022 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18023 "HOST_BITS_PER_WIDE_INT == 64" test.
18024 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18025 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18026 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18027 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18029 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18030 CONST_DOUBLE DImode/VOIDmode case.
18031 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18032 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18033 CONST_DOUBLE VOIDmode case.
18034 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18035 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18036 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18037 Delete CONST_DOUBLE case.
18038 (splitters for mov FMOVE64 const_double): Delete
18039 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18040 "HOST_BITS_PER_WIDE_INT >= 64" test.
18041 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18043 (mov DI const_double): Delete.
18045 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18047 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18048 on op shows all bits zero in mode of a lowpart subreg, return zero.
18050 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18053 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18054 to save TFmode registers and DImode to save TImode registers for
18055 caller save operations.
18056 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18057 mark being partially clobbered since they only use the first
18060 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18061 and TDmode only use the upper 64-bits of each VSX register.
18063 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18065 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18066 (find_candidates_in_block): Disable slsr_process_phi.
18068 2013-05-03 Guozhi Wei <carrot@google.com>
18070 * coverage.c (coverage_obj_init): Move the construction of gcov
18072 (build_init_ctor): ... here.
18074 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18076 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18077 (slsr_cand_d): Redefine def_phi.
18078 (stride_status, phi_adjust_status, count_phis_status): New enums.
18079 (find_phi_def): New.
18080 (find_basis_for_base_expr): New.
18081 (find_basis_for_candidate): Handle hidden bases.
18082 (alloc_cand_and_find_basis): Handle phi candidates.
18083 (slsr_process_phi): New.
18084 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18085 (create_mul_imm_cand): Likewise.
18086 (create_add_ssa_cand): Exclude phi base candidates.
18087 (create_add_imm_cand): Likewise.
18088 (slsr_process_cast): Likewise.
18089 (slsr_process_copy): Likewise.
18090 (find_candidates_in_block): Handle phi candidates.
18091 (dump_candidate): Likewise.
18092 (unconditional_cands): Delete.
18093 (unconditional_cands_with_known_stride_p): Delete.
18094 (phi_dependent_cand_p): New.
18095 (cand_increment): Handle phi-dependent candidates.
18096 (replace_dependent): Delete.
18097 (replace_mult_candidate): New.
18098 (replace_unconditional_candidate): New.
18099 (incr_vec_index): Move to avoid forward reference.
18100 (create_add_on_incoming_edge): New.
18101 (create_phi_basis): New.
18102 (replace_dependents): Delete.
18103 (replace_conditional_candidate): New.
18104 (phi_add_costs): New.
18105 (replace_uncond_cands_and_profitable_phis): New.
18106 (record_increment): Handle phi adjustments.
18107 (record_phi_increments): New.
18108 (record_increments): Handle phi adjustments.
18109 (phi_incr_cost): New.
18110 (lowest_cost_path): Handle phis.
18111 (total_savings): Likewise.
18112 (analyze_increments): Likewise.
18113 (ncd_with_phi): New.
18114 (ncd_of_cand_and_phis): New.
18115 (nearest_common_dominator_for_cands): Handle phi increments.
18116 (all_phi_incrs_profitable): New.
18117 (replace_profitable_candidates): Handle phi-dependent candidates.
18118 (analyze_candidates_and_replace): Likewise.
18120 2013-05-03 Teresa Johnson <tejohnson@google.com>
18123 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18124 do not exceed REG_BR_PROB_BASE.
18126 2013-05-03 Jeff Law <law@redhat.com>
18128 PR tree-optimization/57144
18129 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18130 operand of the condition will bit into the new type when eliminating
18131 a cast feeding a condition.
18133 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18135 PR rtl-optimization/57130
18136 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18137 of COMPARE as in_code to the recursive call if needed.
18139 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18141 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18142 (enabled): Handle new members.
18143 * config/i386/sse.md (*vec_concatv2si): Merge from
18144 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18145 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18147 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18149 PR tree-optimization/57027
18150 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18151 for fnms opportunity, check we got the prerequisite kind
18152 of tree / gimple before using accessor functions.
18154 2013-05-03 Richard Biener <rguenther@suse.de>
18156 * double-int.h (lshift): New overload without precision
18157 and arith argument.
18158 (operator *=, operator +=, operator -=): Move ...
18159 * double-int.c (operator *=, operator +=, operator -=): ... here
18160 and implement more efficiently.
18161 (mul_double_with_sign): Remove.
18162 (lshift_double): Adjust to take unsinged shift argument, push
18163 dispatching code to callers.
18164 (mul_double_wide_with_sign): Add early out for callers that
18165 are not interested in high parts or overflow.
18166 (lshift): New function.
18167 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18168 dispatch code here.
18169 (lrotate, rrotate): Use logical shifts.
18170 * expr.c (get_inner_reference): Use lshift.
18171 * fixed-value.c (do_fixed_divide): Likewise.
18172 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18173 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18174 (indirect_refs_may_alias_p): Likewise.
18175 (stmt_kills_ref_p_1): Likewise.
18177 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18179 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18181 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18183 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18184 scalar form of FABD instruction.
18186 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18188 * lra-constraints.c (process_alt_operands): Add checking alt
18189 number to choose the best alternative.
18191 2013-05-02 Richard Biener <rguenther@suse.de>
18193 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18194 bitmap and its handling.
18195 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18197 2013-05-02 Richard Biener <rguenther@suse.de>
18199 PR middle-end/57140
18200 * tree-inline.c (copy_loops): Properly handle removed loops.
18201 (copy_cfg_body): Mark destination loops for fixup if source
18202 loops needed fixup.
18204 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18207 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18208 generating simple_return for naked functions.
18210 2013-05-02 Martin Jambor <mjambor@suse.cz>
18212 PR middle-end/56988
18213 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18214 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18216 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18217 ipa_agg_replacement_value structures.
18218 (known_aggs_to_agg_replacement_list): Likewise.
18219 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18220 (read_agg_replacement_chain): Likewise.
18221 (ipcp_transform_function): Also check that by_ref flags match.
18223 2013-05-02 Richard Biener <rguenther@suse.de>
18225 * graphds.h (struct graph): Add obstack member.
18226 * graphds.c (new_graph): Initialize obstack and allocate
18228 (add_edge): Allocate edge from the obstack.
18229 (free_graph): Free the obstack instead of all edges and vertices.
18231 2013-05-02 Teresa Johnson <tejohnson@google.com>
18233 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18235 * cfg.c (update_bb_profile_for_threading): Ditto.
18236 * tree-inline.c (copy_bb): Ditto.
18237 (copy_edges_for_bb): Ditto.
18238 (initialize_cfun): Ditto.
18239 (copy_cfg_body): Ditto.
18240 (expand_call_inline): Ditto.
18241 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18242 (estimate_node_size_and_time): Ditto.
18243 (inline_merge_summary): Ditto.
18244 * cgraphclones.c (cgraph_clone_edge): Ditto.
18245 (cgraph_clone_node): Ditto.
18246 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18247 (compute_trg_info): Ditto.
18249 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18251 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18252 S reg when fp attribute set.
18253 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18255 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18257 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18259 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18260 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18261 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18263 2013-05-02 Richard Biener <rguenther@suse.de>
18265 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18266 (struct instantiate_cache_entry): New type.
18267 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18268 (struct instantiate_cache_type): New type.
18269 (set_instantiated_value, get_instantiated_value): Remove.
18270 (get_instantiated_value_entry): New function.
18271 (instantiate_scev_name): Use the new cache and adjust.
18272 (instantiate_scev_poly): Adjust.
18273 (instantiate_scev_binary): Likewise.
18274 (instantiate_array_ref): Likewise.
18275 (instantiate_scev_convert): Likewise.
18276 (instantiate_scev_not): Likewise.
18277 (instantiate_scev_3): Likewise.
18278 (instantiate_scev_2): Likewise.
18279 (instantiate_scev_r): Likewise.
18280 (instantiate_scev): Likewise.
18281 (resolve_mixers): Likewise.
18283 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18286 * lra-constraints.c (best_small_class_operands_num): Remove.
18287 (process_alt_operands): Remove small_class_operands_num. Take
18288 small classes operands into losers and only if the operand is not
18289 matched. Modify debugging output.
18290 (curr_insn_transform): Remove best_small_class_operands_num.
18293 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18295 * config/aarch64/aarch64-builtins.c
18296 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18297 * config/aarch64/aarch64-simd-builtins.def
18298 (reduc_splus_): Add new modes.
18299 (reduc_uplus_): New.
18300 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18301 (reduc_uplus_v4sf): Likewise.
18302 (reduc_splus_v4sf): Likewise.
18303 (aarch64_addv<mode>): Likewise.
18304 (reduc_uplus_<mode>): Likewise.
18305 (reduc_splus_<mode>): Likewise.
18306 (aarch64_addvv2di): Likewise.
18307 (reduc_uplus_v2di): Likewise.
18308 (reduc_splus_v2di): Likewise.
18309 (aarch64_addvv2si): Likewise.
18310 (reduc_uplus_v2si): Likewise.
18311 (reduc_splus_v2si): Likewise.
18312 (reduc_<sur>plus_<mode>): New.
18313 (reduc_<sur>plus_v2di): Likewise.
18314 (reduc_<sur>plus_v2si): Likewise.
18315 (reduc_<sur>plus_v4sf): Likewise.
18316 (aarch64_addpv4sf): Likewise.
18317 * config/aarch64/arm_neon.h
18318 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18319 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18320 add UNSPEC_SADDV, UNSPEC_UADDV.
18322 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18324 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18326 * config/aarch64/arm_neon.h
18327 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18329 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18331 * config/aarch64/aarch64-builtins
18332 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18334 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18336 * config/aarch64/aarch64-simd-builtins.def
18337 (reduc_smax_): New.
18338 (reduc_smin_): Likewise.
18339 (reduc_umax_): Likewise.
18340 (reduc_umin_): Likewise.
18341 (reduc_smax_nan_): Likewise.
18342 (reduc_smin_nan_): Likewise.
18345 (smax): Update for V2SF, V4SF and V2DF modes.
18348 (smin_nan): Likewise.
18349 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18350 (<su><maxmin><mode>3): ...This, refactor.
18351 (s<maxmin><mode>3): New.
18352 (<maxmin_uns><mode>3): Likewise.
18353 (reduc_<maxmin_uns>_<mode>): Refactor.
18354 (reduc_<maxmin_uns>_v4sf): Likewise.
18355 (reduc_<maxmin_uns>_v2si): Likewise.
18356 (aarch64_<fmaxmin><mode>: Remove.
18357 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18359 (vmin<q>_f<32,64>): Likewise.
18360 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18362 (su): Add mappings for smax, smin, umax, umin.
18364 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18365 (FMAXMIN): Rename as...
18366 (FMAXMIN_UNS): ...This.
18368 (fmaxminv): Likewise.
18369 (fmaxmin): Likewise.
18371 (maxmin_uns_op): Likewise.
18373 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18375 * config/aarch64/arm_neon.h
18376 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18377 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18378 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18380 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18382 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18383 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18385 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18387 * config/aarch64/aarch64-simd.md
18388 (vcond<mode>_internal): Handle special cases for constant masks.
18389 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18390 (vcondu<mode><mode>): Likewise.
18391 (vcond<v_cmp_result><mode>): New.
18393 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18395 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18396 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18397 * config/aarch64/aarch64-simd-builtins.def
18398 (cmeq): Update to BUILTIN_VALLDI.
18403 * config/aarch64/arm_neon.h
18404 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18405 to builtins or C as appropriate.
18407 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18409 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18411 (cmhi): Rename to...
18413 * config/aarch64/aarch64-simd.md
18414 (simd_mode): Add SF.
18415 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18416 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18417 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18418 (cstore<mode>_neg): ...This.
18419 * config/aarch64/iterators.md
18421 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18422 (COMPARISONS): New.
18423 (UCOMPARISONS): Likewise.
18424 (optab): Add missing comparisons.
18430 (VCMP_S): Likewise.
18431 (VCMP_U): Likewise.
18432 (V_cmp_result): Add DF, SF modes.
18433 (v_cmp_result): Likewise.
18435 (vmtype): Likewise.
18436 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18438 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18440 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18441 define_insn to define_insn_and_split.
18442 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18443 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18444 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18445 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18446 (thumb2_negscc): Likewise.
18448 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18450 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18452 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18454 * config/arm/thumb2.md: Remove trailing whitespaces.
18456 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18458 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18459 Use gen_int_mode rather than GEN_INT.
18461 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18463 * value-prof.c (stream_in_histogram_value): Remove the strayed
18466 2013-04-30 Richard Biener <rguenther@suse.de>
18468 PR middle-end/57122
18469 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18471 2013-04-30 Richard Biener <rguenther@suse.de>
18473 PR middle-end/57107
18474 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18476 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18478 PR rtl-optimization/56957
18479 PR rtl-optimization/57105
18480 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18481 variable. Use just INSN_UID for determining whether an insn
18482 should be only disconnected from the insn stream.
18483 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18485 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18487 PR tree-optimization/57104
18488 * tsan.c (instrument_expr): Don't instrument accesses to
18489 DECL_HARD_REGISTER VAR_DECLs.
18491 2013-04-30 Richard Biener <rguenther@suse.de>
18493 * function.h (loops_for_fn): New inline function.
18494 (set_loops_for_fn): Likewise.
18495 * cfgloop.h (place_new_loop): Add struct function parameter.
18496 (get_loop): Likewise.
18497 (get_loops): Likewise.
18498 (number_of_loops): Likewise.
18499 (fel_next): Adjust.
18500 (fel_init): Likewise.
18501 * cfg.c (get_loop_copy): Adjust.
18502 * cfgloop.c (flow_loops_dump): Likewise.
18503 (record_loop_exits): Likewise.
18504 (verify_loop_structure): Likewise.
18505 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18506 (estimate_reg_pressure_cost): Likewise.
18507 (mark_loop_exit_edges): Likewise.
18508 * cfgloopmanip.c (place_new_loop): Likewise.
18509 (add_loop): Likewise.
18510 (duplicate_loop): Likewise.
18511 * graph.c (draw_cfg_nodes): Likewise.
18512 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18513 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18514 (extract_affine_chrec): Likewise.
18515 (build_scop_iteration_domain): Likewise.
18516 * graphite.c (graphite_initialize): Likewise.
18517 * ira-build.c (create_loop_tree_nodes): Likewise.
18518 (more_one_region_p): Likewise.
18519 (rebuild_regno_allocno_maps): Likewise.
18520 (mark_loops_for_removal): Likewise.
18521 (mark_all_loops_for_removal): Likewise.
18522 (remove_unnecessary_regions): Likewise.
18523 (ira_build): Likewise.
18524 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18525 * loop-init.c (fix_loop_structure): Likewise.
18526 (gate_rtl_move_loop_invariants): Likewise.
18527 (gate_rtl_unswitch): Likewise.
18528 (gate_rtl_unroll_and_peel_loops): Likewise.
18529 (rtl_doloop): Likewise.
18530 * lto-streamer-in.c (input_cfg): Likewise.
18531 * lto-streamer-out.c (output_cfg): Likewise.
18532 * modulo-sched.c (sms_schedule): Likewise.
18533 * predict.c (tree_estimate_probability): Likewise.
18534 (tree_estimate_probability_driver): Likewise.
18535 (estimate_loops): Likewise.
18536 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18537 (move_sese_region_to_fn): Likewise.
18538 (debug_loop_num): Likewise.
18539 * tree-chrec.c (chrec_evaluate): Likewise.
18540 (hide_evolution_in_other_loops_than_loop): Likewise.
18541 (chrec_component_in_loop_num): Likewise.
18542 (reset_evolution_in_loop): Likewise.
18543 (evolution_function_is_invariant_rec_p): Likewise.
18544 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18545 * tree-inline.c (copy_loops): Likewise.
18546 (copy_cfg_body): Likewise.
18547 (tree_function_versioning): Likewise.
18548 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18549 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18551 (add_to_evolution_1): Likewise.
18552 (scev_const_prop): Likewise.
18553 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18554 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18555 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18556 (tree_ssa_lim_initialize): Likewise.
18557 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18558 (verify_loop_closed_ssa): Likewise.
18559 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18560 (tree_ssa_loop_im): Likewise.
18561 (tree_ssa_loop_unswitch): Likewise.
18562 (tree_vectorize): Likewise.
18563 (check_data_deps): Likewise.
18564 (tree_ssa_loop_ivcanon): Likewise.
18565 (tree_ssa_loop_bounds): Likewise.
18566 (tree_complete_unroll): Likewise.
18567 (tree_complete_unroll_inner): Likewise.
18568 (tree_parallelize_loops): Likewise.
18569 (tree_ssa_loop_prefetch): Likewise.
18570 (tree_ssa_loop_ivopts): Likewise.
18571 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18572 * tree-vectorizer.c (vectorize_loops): Likewise.
18574 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18576 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18577 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18578 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18579 with EABI_LINK_SPEC.
18581 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18584 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18587 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18590 * lra-constraints.c (process_alt_operands): Discourage a bit more
18591 using memory for pseudos. Print cost dump for alternatives.
18592 Modify cost values for conflicts with early clobbers.
18593 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18595 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18598 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18600 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18602 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18603 from/to S register.
18604 (movdi_aarch64): Support LDR/STR from/to D register.
18606 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18608 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18609 or higher by default.
18611 2013-04-29 Richard Biener <rguenther@suse.de>
18613 PR middle-end/57075
18614 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18615 even if not adding abnormal edges for calls that can make
18618 2013-04-29 Richard Biener <rguenther@suse.de>
18620 PR middle-end/57103
18621 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18623 (move_stmt_r): Remove redundant checking.
18625 2013-04-29 Teresa Johnson <tejohnson@google.com>
18628 * basic-block.h (apply_scale): New function.
18629 (apply_probability): Use apply_scale.
18630 * gimple-streamer-in.c (input_bb): Ditto.
18631 * lto-streamer-in.c (input_cfg): Ditto.
18632 * lto-cgraph.c (merge_profile_summaries): Ditto.
18633 * tree-optimize.c (execute_fixup_cfg): Ditto.
18634 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18635 (copy_edges_for_bb): Ditto.
18636 (copy_cfg_body): Ditto.
18638 2013-04-29 Tom de Vries <tom@codesourcery.com>
18640 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18641 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18642 (tail_merge_optimize): Handle current_loops == NULL.
18644 2013-04-26 Jeff Law <law@redhat.com>
18646 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18647 (simplify_cond_using_ranges): Generalize code to simplify
18648 COND_EXPRs where one argument is a constant and the other
18649 is an SSA_NAME created by an integral type conversion.
18651 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18653 * config/arm/arm.md (store_minmaxsi): Use only when
18654 optimize_insn_for_size_p.
18656 2013-04-29 Christian Bruel <christian.bruel@st.com>
18659 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18661 2013-04-29 Richard Biener <rguenther@suse.de>
18663 PR middle-end/57089
18664 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18665 loop tree make sure to schedule a fixup for the child as well.
18666 (expand_omp_for_generic): Properly add loops.
18667 (expand_omp_for_static_nochunk): Likewise.
18668 (expand_omp_for_static_chunk): Likewise.
18669 (expand_omp_for): For the degenerate case fixup loops.
18670 (expand_omp_sections): Fix default bb placement in loops.
18671 (expand_omp_atomic_pipeline): Properly add loops.
18673 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18675 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18677 2013-04-29 Tom de Vries <tom@codesourcery.com>
18679 * tree-ssa-tail-merge.c: Update header comment.
18681 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18683 * config/aarch64/arm_neon.h
18684 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18685 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18686 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18687 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18688 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18689 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18690 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18691 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18693 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18695 * config/aarch64/aarch64-simd.md
18696 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18697 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18698 fix_trunc, fixuns_trunc.
18699 (ftrunc<VDQF:mode>2): New.
18700 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18701 (fix_trunc_optab): New.
18703 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18705 * config/aarch64/aarch64-builtins.c
18706 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18707 iceilf, lround, iroundf.
18709 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18712 * config/i386/i386.h (enum ix86_tune_indices)
18713 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18714 New, split from X86_TUNE_INTER_UNIT_MOVES.
18715 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18716 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18717 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18718 (TARGET_INTER_UNIT_MOVES): Remove.
18719 * config/i386/i386.c (initial_ix86_tune_features): Update.
18720 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18721 (ix86_expand_convert_uns_didf_sse): Use
18722 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18723 (ix86_expand_vector_init_one_nonzero): Ditto.
18724 (ix86_expand_vector_init_interleave): Ditto.
18725 (inline_secondary_memory_needed): Return true for moves from SSE class
18726 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18727 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18728 * config/i386/constraints.md (Yi, Ym): Depend on
18729 TARGET_INTER_UNIT_MOVES_TO_VEC.
18730 (Yj, Yn): New constraints.
18731 * config/i386/i386.md (*movdi_internal): Change constraints of
18732 operand 1 from Yi to Yj and from Ym to Yn.
18733 (*movsi_internal): Ditto.
18734 (*movdf_internal): Ditto.
18735 (*movsf_internal): Ditto.
18736 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18737 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18738 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18739 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18740 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18741 * config/i386/sse.md (movdi_to_sse): Ditto.
18742 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18743 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18744 TARGET_INTER_UNIT_MOVES.
18745 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18746 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18747 instead of TARGET_INTER_UNIT_MOVES.
18748 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18749 operand 1 from Yi to Yj and from Ym to Yn.
18751 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18753 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18754 (float_truncate_hi_): Likewise.
18755 (float_extend_lo_): Likewise.
18756 (float_truncate_lo_): Likewise.
18757 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18758 (aarch64_float_extend_lo_v2df): Likewise.
18759 (vec_unpacks_hi_v4sf): Likewise.
18760 (aarch64_float_truncate_lo_v2sf): Likewise.
18761 (aarch64_float_truncate_hi_v4sf): Likewise.
18762 (vec_pack_trunc_v2df): Likewise.
18763 (vec_pack_trunc_df): Likewise.
18765 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18767 * config/aarch64/aarch64-builtins.c
18768 (aarch64_fold_builtin): Fold float conversions.
18769 * config/aarch64/aarch64-simd-builtins.def
18770 (floatv2si, floatv4si, floatv2di): New.
18771 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18772 * config/aarch64/aarch64-simd.md
18773 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18774 * config/aarch64/iterators.md (FLOATUORS): New.
18775 (optab): Add float, floatuns.
18776 (su_optab): Likewise.
18778 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18780 * config/aarch64/aarch64-builtins.c
18781 (aarch64_builtin_vectorized_function): Use new names for
18783 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18784 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18785 (fcvtzu): Split as...
18786 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18787 (fcvtas): Split as...
18788 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18789 (fcvtau): Split as...
18790 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18791 (fcvtps): Split as...
18792 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18793 (fcvtpu): Split as...
18794 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18795 (fcvtms): Split as...
18796 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18797 (fcvtmu): Split as...
18798 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18799 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18800 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18801 (lfrintnusf, lfrintnudf): Likewise.
18802 * config/aarch64/aarch64-simd.md
18803 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18805 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18806 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18807 (fcvt_pattern): Likewise.
18809 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18811 * config/aarch64/aarch64-simd.md
18812 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18813 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18815 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18817 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18818 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18819 (vrnd<a,m,n,p>_f32): Implement using builtins.
18820 (vrnd<i,x><q>_f<32, 64>): New.
18822 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18824 * config/aarch64/aarch64-builtins.c
18825 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18826 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18827 (frintz): Rename to...
18829 (frintp): Rename to...
18831 (frintm): Rename to...
18833 (frinti): Rename to...
18834 (nearbyint): ...this.
18835 (frintx): Rename to...
18837 (frinta): Rename to...
18839 * config/aarch64/aarch64-simd.md
18840 (aarch64_frint<frint_suffix><mode>): Delete.
18841 (<frint_pattern><mode>2): Convert to insn.
18842 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18843 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18844 (frint_pattern): Likewise.
18845 (frint_suffix): Likewise.
18847 2013-04-29 Richard Biener <rguenther@suse.de>
18849 PR tree-optimization/57081
18850 * loop-init.c: Include tree-flow.h.
18851 (loop_optimizer_finalize): Free number of iteration estimates.
18852 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18854 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18856 PR tree-optimization/57083
18857 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18858 non-singleton shift count range, zero extend low_bound for uns case.
18860 * config/i386/predicates.md (general_vector_operand): New predicate.
18861 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18862 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18863 if they aren't nonimmediate operands. If their original values
18864 satisfy const_vector_equal_evenodd_p, don't shift them.
18865 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18866 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18867 if not nonimmediate_operand.
18868 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18869 instead of register_operand.
18870 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18872 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18874 * stor-layout.c (finalize_size_functions): Allocate a structure and
18875 reset cfun before dumping the functions.
18877 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18879 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18882 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18883 use xop_pmacsdqh if uns_p.
18884 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18885 the immediate rotate count.
18887 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18889 * rtl.h (struct rtx_def): Add comment for field jump.
18890 (LRA_SUBREG_P): New macro.
18891 * recog.c (register_operand): Check LRA_SUBREG_P.
18892 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18894 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18895 (lra_final_code_change): Skip subreg change for operators.
18896 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18897 if there are no operand changes.
18898 * lra-constraints.c (curr_insn_set): New.
18899 (match_reload): Set LRA_SUBREG_P.
18900 (emit_spill_move): Ditto.
18901 (check_and_process_move): Use curr_insn_set. Process only single
18902 set insns. Don't initialize sec_mem_p and change_p.
18903 (simplify_operand_subreg): Use LRA_SUBREG_P.
18904 (reg_in_class_p): New function.
18905 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18906 of #ifdef. Add code to remove cycling.
18907 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18908 non-null disp. Reload inner instead of disp when base and index
18909 are null. Try to put lo_sum into register.
18910 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18911 (check_and_process_move): Move code for move cost check to
18912 simple_move_p. Remove equiv_substitution.
18913 (simple_move_p): New function.
18914 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18915 curr_insn_set. Call check_and_process_move only for single set
18916 insns. Use the new function. Move call of check_and_process_move
18917 after operand equiv substitution and address process.
18919 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18922 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18923 with nonlocal goto receivers or returns twice calls, ignore
18924 unininitialized values from abnormal edges to nl goto receiver
18925 or returns twice call.
18927 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18929 PR tree-optimization/57051
18930 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18931 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18934 2013-04-26 Richard Biener <rguenther@suse.de>
18936 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18937 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18938 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18939 (fixup_loop_arrays_after_move): New function.
18940 (move_sese_region_to_fn): Properly outline the loop tree parts
18941 of the SESE region.
18943 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18945 * config/i386/i386.md (type, unit): Fix long lines.
18947 2013-04-26 Richard Biener <rguenther@suse.de>
18949 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18950 (lto-streamer-out.o): Likewise.
18951 * cfgloop.c (init_loops_structure): Export, add struct function
18952 argument and adjust.
18953 (flow_loops_find): Adjust.
18954 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18955 (init_loops_structure): Declare.
18956 * lto-streamer-in.c: Include cfgloop.h.
18957 (input_cfg): Input the loop tree.
18958 * lto-streamer-out.c: Include cfgloop.h.
18959 (output_cfg): Output the loop tree.
18960 (output_struct_function_base): Do not drop PROP_loops.
18962 2013-03-26 Richard Biener <rguenther@suse.de>
18964 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18965 (pass_build_cfg): Provide PROP_loops.
18966 (move_sese_region_to_fn): Remove loops that are outlined into fn
18968 * tree-inline.c: Include cfgloop.h.
18969 (initialize_cfun): Do not drop PROP_loops.
18970 (copy_loops): New function.
18971 (copy_cfg_body): Copy loop structure.
18972 (tree_function_versioning): Initialize destination loop tree.
18973 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18974 (pass_parallelize_loops): Do IL verification.
18975 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18976 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18977 the CFG make sure we fixup loops as well.
18978 * tree-ssa-tail-merge.c: Include cfgloop.h.
18979 (replace_block_by): When merging loop latches mark loops for fixup.
18980 * lto-streamer-out.c (output_struct_function_base): Drop
18981 PROP_loops for now.
18982 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18983 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18984 * ipa-split.c: Include cfgloop.h.
18985 (split_function): Add the new return block to the loop tree root.
18986 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18987 whether we have removed the forwarder block.
18988 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18989 * cfgloop.h (place_new_loop): Declare.
18990 * cfgloopmanip.c (place_new_loop): Export.
18991 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18992 (tree-switch-conversion.o): Likewise.
18993 (tree-complex.o): Likewise.
18994 (tree-inline.o): Likewise.
18995 (tree-ssa-tailmerge.o): Likewise.
18996 (ipa-split.o): Likewise.
18997 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18998 (tree-ssa-copy.o): Likewise.
18999 * tree-switch-conversion.c: Include cfgloop.h
19000 (process_switch): If we emit a bit-test cascade, schedule loops
19002 * tree-complex.c: Include cfgloop.h.
19003 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19004 * asan.c: Include cfgloop.h.
19005 (create_cond_insert_point): Properly add new basic-blocks to
19006 loops, schedule loop fixup.
19007 * cfgloop.c (verify_loop_structure): Check that looks are not
19009 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19011 (expand_omp_for_generic): Likewise.
19012 (expand_omp_sections): Likewise.
19013 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19014 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19015 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19016 is initialized, not when loops are present.
19017 * tree-parloops.c (parallelize_loops): Remove checking here.
19018 * passes.c (init_optimization_passes): Schedule a copy-propagation
19019 pass before complete unrolling of inner loops.
19021 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19023 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19024 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19025 (colorize_init): Add argument to _WIN32 version.
19026 * toplev.c: Include diagnostic-color.h.
19027 (process_options): Default to -fdiagnostics-color=auto if
19028 GCC_COLORS env var is in the environment.
19029 * common.opt (fdiagnostics-color=): Add Var and Init.
19030 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19031 env var is in the environment, the default is auto rather than never.
19033 * diagnostic.h (file_name_as_prefix): Add context argument.
19034 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19035 the string as locus.
19036 * langhooks.c (lhd_print_error_function): Adjust caller.
19038 2013-04-25 Lawrence Crowl <crowl@google.com>
19040 * var-tracking.c (shared_hash_def::htab):
19041 Change type to hash_table. Update dependent calls and types.
19043 2013-04-25 Lawrence Crowl <crowl@google.com>
19045 * Makefile.in: Update as needed below.
19047 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19048 Move declaration to after the type's method definitons.
19050 * attribs.c (htab_t scoped_attributes::attribute_hash):
19051 Change type to hash_table. Update dependent calls and types.
19053 * bitmap.c (htab_t bitmap_desc_hash):
19054 Change type to hash_table. Update dependent calls and types.
19056 * cselib.c (htab_t cselib_hash_table):
19057 Change type to hash_table. Update dependent calls and types.
19059 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19060 (hash_string_slot_node): Move implementation into lto-streamer.h
19061 struct string_slot_hasher.
19062 (eq_string_slot_node): Likewise.
19064 * data-streamer-out.c: Update output_block::string_hash_table
19065 dependent calls and types.
19067 * dwarf2cfi.c (htab_t trace_index):
19068 Change type to hash_table. Update dependent calls and types.
19070 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19071 Change type to hash_table. Update dependent calls and types.
19072 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19073 (htab_t optimize_external_refs::map): Likewise.
19074 (htab_t output_comp_unit::extern_map): Likewise.
19075 (htab_t output_comdat_type_unit::extern_map): Likewise.
19076 (htab_t output_macinfo::macinfo_htab): Likewise.
19077 (htab_t optimize_location_lists::htab): Likewise.
19078 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19080 * except.c (htab_t ehspec_hash_type):
19081 Change type to hash_table. Update dependent calls and types.
19082 (assign_filter_values::ttypes): Likewise.
19083 (assign_filter_values::ehspec): Likewise.
19084 (sjlj_assign_call_site_values::ar_hash): Likewise.
19085 (convert_to_eh_region_ranges::ar_hash): Likewise.
19087 * gcse.c (htab_t pre_ldst_table):
19088 Change type to hash_table. Update dependent calls and types.
19090 * ggc-common.c (htab_t saving_htab):
19091 Change type to hash_table. Update dependent calls and types.
19092 (htab_t loc_hash): Likewise.
19093 (htab_t ptr_hash): Likewise.
19094 (call_count): Rename ggc_call_count.
19095 (call_alloc): Rename ggc_call_alloc.
19096 (loc_descriptor): Rename make_loc_descriptor.
19097 (add_statistics): Rename ggc_add_statistics.
19099 * ggc-common.c (saving_htab):
19100 Change type to hash_table. Update dependent calls and types.
19102 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19103 (push_gimplify_context): Likewise.
19104 (pop_gimplify_context): Likewise.
19105 (struct gimple_temp_hash_elt): Added.
19106 (struct gimplify_hasher): Likewise.
19107 (struct gimplify_ctx.temp_htab):
19108 Change type to hash_table. Update dependent calls and types.
19110 * gimple-fold.c: Include gimplify-ctx.h.
19112 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19113 Change type to hash_table. Update dependent calls and types.
19114 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19115 avoid potential global name collision.
19117 * gimplify.c: Include gimplify-ctx.h.
19118 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19119 (htab_t gimplify_ctx::temp_htab):
19120 Update dependent calls and types for new type hash_table.
19121 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19122 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19124 * gimplify-ctx.h: New.
19125 (struct gimple_temp_hash_elt): Move from gimplify.c.
19126 (class gimplify_hasher): New.
19127 (struct gimplify_ctx): Move from gimple.h.
19128 (htab_t gimplify_ctx::temp_htab):
19129 Change type to hash_table. Update dependent calls and types.
19131 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19132 (htab_t ivs_params::newivs_index):
19133 Change type to hash_table. Update dependent calls and types.
19134 (htab_t ivs_params::params_index): Likewise.
19135 (htab_t print_generated_program::params_index): Likewise.
19136 (htab_t gloog::newivs_index): Likewise.
19137 (htab_t gloog::params_index): Likewise.
19139 * graphite.c: Include graphite-htab.h.
19140 4htab_t graphite_transform_loops::bb_pbb_mapping):
19141 Change type to hash_table. Update dependent calls and types.
19143 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19144 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19145 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19147 * graphite-dependences.c: Include graphite-htab.h.
19148 (loop_is_parallel_p): Change hash table type of parameter.
19150 * graphite-htab.h: New.
19151 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19152 (extern find_pbb_via_hash): Move from graphite-poly.h.
19153 (extern loop_is_parallel_p): Move from graphite-poly.h.
19154 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19156 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19157 (extern loop_is_parallel_p): Move to graphite-htab.h.
19158 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19160 * haifa-sched.c (htab_t delay_htab):
19161 Change type to hash_table. Update dependent calls and types.
19162 (htab_t delay_htab_i2): Likewise.
19164 * ira-color.c (htab_t allocno_hard_regs_htab):
19165 Change type to hash_table. Update dependent calls and types.
19167 * ira-costs.c (htab_t cost_classes_htab):
19168 Change type to hash_table. Update dependent calls and types.
19170 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19171 Change type to hash_table. Update dependent calls and types.
19173 * loop-iv.c (htab_t bivs):
19174 Change type to hash_table. Update dependent calls and types.
19176 * loop-unroll.c (htab_t opt_info::insns_to_split):
19177 Change type to hash_table. Update dependent calls and types.
19178 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19180 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19181 (struct string_slot_hasher): New.
19182 (htab_t output_block::string_hash_table):
19183 Change type to hash_table. Update dependent calls and types.
19185 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19186 (htab_t file_name_hash_table):
19187 Change type to hash_table. Update dependent calls and types.
19189 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19192 * lto-streamer.c (htab_t tree_htab):
19193 Change type to hash_table. Update dependent calls and types.
19195 * omp-low.c: Include gimplify-ctx.h.
19197 * passes.c (htab_t name_to_pass_map):
19198 Change type to hash_table. Update dependent calls and types.
19199 (pass_traverse): Rename to passes_pass_traverse.
19201 * plugin.c (htab_t event_tab):
19202 Change type to hash_table. Update dependent calls and types.
19204 * postreload-gcse.c (htab_t expr_table):
19205 Change type to hash_table. Update dependent calls and types.
19206 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19208 * sese.c (debug_rename_map_1): Make extern.
19209 (htab_t copy_bb_and_scalar_dependences::rename_map):
19210 Change type to hash_table. Update dependent calls and types.
19212 * sese.h (extern debug_rename_map): Move to .c file.
19214 * store-motion.c (htab_t store_motion_mems_table):
19215 Change type to hash_table. Update dependent calls and types.
19217 * trans-mem.c (htab_t tm_new_mem_hash):
19218 Change type to hash_table. Update dependent calls and types.
19220 * tree-browser.c (htab_t TB_up_ht):
19221 Change type to hash_table. Update dependent calls and types.
19223 * tree-cfg.c (htab_t discriminator_per_locus):
19224 Change type to hash_table. Update dependent calls and types.
19226 * tree-complex.c: Include tree-hasher.h
19227 (htab_t complex_variable_components):
19228 Change type to hash_table. Update dependent calls and types.
19230 * tree-eh.c (htab_t finally_tree):
19231 Change type to hash_table. Update dependent calls and types.
19233 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19234 struct int_tree_hasher.
19235 (extern int_tree_map_eq): Likewise.
19236 (uid_decl_map_hash): Removed.
19237 (extern decl_tree_map_eq): Likewise.
19239 * tree-hasher.h: New.
19240 (struct int_tree_hasher): New.
19241 (typedef int_tree_htab_type): New.
19243 * tree-inline.c: Include gimplify-ctx.h.
19245 * tree-mudflap.c: Include gimplify-ctx.h.
19247 * tree-parloops.c: Include tree-hasher.h.
19248 (htab_t eliminate_local_variables_stmt::decl_address):
19249 Change type to hash_table. Update dependent calls and types.
19250 (htab_t separate_decls_in_region::decl_copies): Likewise.
19252 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19253 Change type to hash_table. Update dependent calls and types.
19255 * tree-sra.c (candidates):
19256 Change type to hash_table. Update dependent calls and types.
19258 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19260 (int_tree_map_hash): Likewise.
19262 * tree-ssa-dom.c (htab_t avail_exprs):
19263 Change type to hash_table. Update dependent calls and types.
19265 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19266 Change type to hash_table. Update dependent calls and types.
19268 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19269 Change type to hash_table. Update dependent calls and types.
19271 * tree-ssa-phiopt.c (seen_ssa_names):
19272 Change type to hash_table. Update dependent calls and types.
19274 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19275 Change type to hash_table. Update dependent calls and types.
19277 * tree-ssa-uncprop.c (equiv):
19278 Change type to hash_table. Update dependent calls and types.
19280 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19282 PR rtl-optimization/57003
19283 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19284 call note_stores with kill_clobbered_value callback again after
19285 killing regs_invalidated_by_call.
19287 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19289 * config/aarch64/aarch64-simd.md
19290 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19291 (aarch64_simd_bsl<mode>): Likewise.
19292 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19294 2013-04-25 Marek Polacek <polacek@redhat.com>
19296 PR tree-optimization/57066
19297 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19299 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19301 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19303 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19305 * config/aarch64/aarch64-builtins.c
19306 (aarch64_fold_builtin): New.
19307 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19308 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19309 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19310 * config/aarch64/arm_neon.h
19311 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19313 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19314 Tejas Belagod <tejas.belagod@arm.com>
19316 * config/aarch64/aarch64-builtins.c
19317 (aarch64_gimple_fold_builtin): New.
19318 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19319 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19320 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19321 (addvv4sf): Update.
19322 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19324 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19326 * config/aarch64/aarch64.md
19327 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19329 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19331 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19332 (*ngcsi_uxtw): New pattern.
19334 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19335 Julian Brown <julian@codesourcery.com>
19337 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19338 (TB_DREG): Add T_V4HF.
19339 (v4hf_UP): New macro.
19340 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19341 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19342 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19344 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19345 (arm_vector_mode_supported_p): Handle V4HF.
19346 (arm_mangle_map): Handle V4HFmode.
19347 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19348 * config/arm/arm_neon_builtins.def: Add entries for
19349 vcvtv4hfv4sf, vcvtv4sfv4hf.
19350 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19351 (neon_vcvtv4hfv4sf): Likewise.
19352 * config/arm/neon-gen.ml: Handle half-precision floating point
19354 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19355 * config/arm/arm_neon.h: Regenerate.
19356 * config/arm/neon.ml (type elts): Add F16.
19357 (type vectype): Add T_float16x4, T_floatHF.
19358 (type vecmode): Add V4HF.
19359 (type features): Add Requires_FP_bit feature.
19360 (elt_width): Handle F16.
19361 (elt_class): Likewise.
19362 (elt_of_class_width): Likewise.
19363 (mode_of_elt): Refactor.
19364 (type_for_elt): Handle F16, fix error messages.
19365 (vectype_size): Handle T_float16x4.
19366 (vcvt_sh): New function.
19367 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19368 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19369 (string_of_mode): Handle V4HF.
19370 * doc/arm-neon-intrinsics.texi: Regenerate.
19372 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19374 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19375 format specifier in 'X' case.
19377 2013-04-25 Alan Modra <amodra@gmail.com>
19380 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19381 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19382 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19383 Repeat for many other rotate/shift and mask patterns using subregs.
19384 Name lshiftrt insns.
19385 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19386 on WORDS_BIG_ENDIAN.
19388 2013-04-25 Alan Modra <amodra@gmail.com>
19390 * config.gcc: Support little-endian powerpc-linux targets.
19391 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19392 (LINK_OS_LINUX_SPEC): Define.
19393 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19394 Preserve MASK_LITTLE_ENDIAN.
19395 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19396 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19397 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19398 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19399 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19400 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19401 Correct fp word order for little-endian. Don't shift toc entries
19402 smaller than a word for little-endian.
19403 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19404 (bswapdi2 splits): Correct low-part subreg for little-endian.
19405 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19406 low/high where such is correct only for be.
19407 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19408 little-endian for -mcall-aixdesc.
19410 2013-04-25 Alan Modra <amodra@gmail.com>
19412 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19413 replace_equiv_address_nv.
19415 2013-04-25 Alan Modra <amodra@gmail.com>
19417 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19419 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19422 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19423 * rtl.h (struct rtx_def): ...
19425 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19427 PR rtl-optimizations/57046
19428 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19429 for multi-reg splits.
19431 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19433 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19435 2013-04-24 Sterling Augustine <saugustine@google.com>
19437 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19438 (comp_dir_string, debug_str_dwo_section): New.
19439 (DEBUG_STR_DWO_SECTION): Rename to ...
19440 (DEBUG_DWO_STR_SECTION): ... this.
19441 (DEBUG_NORM_STR_SECTION): Delete.
19442 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19443 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19444 (find_AT_string): Move most logic to ...
19445 (find_AT_string_in_table): ... here. New.
19446 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19447 add_skeleton_AT_string. Delete logic.
19448 (output_skeleton_debug_sections): Remove call to
19449 add_top_level_skeleton_die_attrs.
19450 (add_comp_dir_attribute): Move logic to comp_dir_string.
19451 (dwarf2out_init): Initialize debug_str_dwo_section.
19452 (output_indirect_string): Call find_string_form.
19453 (output_indirect_strings): Rewrite.
19454 (prune_unused_types): Empty skeleton_debug_str_hash.
19455 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19456 (dwarf2out_finish): Call output_indirect_strings.
19458 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19460 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19462 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19464 * rtl.h (struct rtx_def): Add comment for field jump.
19465 (LRA_SUBREG_P): New macro.
19466 * recog.c (register_operand): Check LRA_SUBREG_P.
19467 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19469 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19470 (lra_final_code_change): Skip subreg change for operators.
19471 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19472 if there are no operand changes.
19473 * lra-constraints.c (curr_insn_set): New.
19474 (match_reload): Set LRA_SUBREG_P.
19475 (emit_spill_move): Ditto.
19476 (check_and_process_move): Use curr_insn_set. Process only single
19477 set insns. Don't initialize sec_mem_p and change_p.
19478 (simplify_operand_subreg): Use LRA_SUBREG_P.
19479 (reg_in_class_p): New function.
19480 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19481 of #ifdef. Add code to remove cycling.
19482 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19483 non-null disp. Reload inner instead of disp when base and index
19484 are null. Try to put lo_sum into register.
19485 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19486 (check_and_process_move): Move code for move cost check to
19487 simple_move_p. Remove equiv_substitution.
19488 (simple_move_p): New function.
19489 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19490 curr_insn_set. Call check_and_process_move only for single set
19491 insns. Use the new function. Move call of check_and_process_move
19492 after operand equiv substitution and address process.
19494 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19496 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19497 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19498 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19500 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19502 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19504 2013-04-24 Marek Polacek <polacek@redhat.com>
19506 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19507 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19508 (select_loops_exit_conditions): Likewise.
19509 (number_of_iterations_for_all_loops): Likewise.
19510 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19511 (scev_analysis): Likewise.
19513 2013-04-02 Catherine Moore <clm@codesourcery.com>
19514 Chao-ying Fu <fu@mips.com>
19516 * config/mips/micromips.md (jraddiusp): New pattern.
19517 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19518 instruction if possible.
19520 2013-04-24 Alan Modra <amodra@gmail.com>
19522 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19524 2013-04-24 Julian Brown <julian@codesourcery.com>
19525 Chung-Lin Tang <cltang@codesourcery.com>
19527 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19528 dependency behavior in enumeration type DIE generation. Add TODO note
19529 to comments about future DW_FORM_sdata/udata re-work of related code.
19531 2013-04-23 Lawrence Crowl <crowl@google.com>
19533 * Makefile.in: Update as needed below.
19535 * hash-table.h (class hash_table):
19536 Correct many methods with parameter types compare_type to the correct
19537 value_type. (Correct code was unlikely to notice the change.)
19538 (hash_table::elements_with_deleted) New.
19539 (class hashtable::iterator): New.
19540 (hashtable::begin()): New.
19541 (hashtable::end()): New.
19542 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19544 * statistics.c (statistics_hashes):
19545 Change type to hash_table. Update dependent calls and types.
19547 * tree-into-ssa.c (var_infos):
19548 Change type to hash_table. Update dependent calls and types.
19550 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19551 Change type to hash_table. Update dependent calls and types.
19553 * tree-ssa-loop-im.c (struct mem_ref.refs):
19554 Change type to hash_table. Update dependent calls and types.
19556 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19557 Change type to hash_table. Update dependent calls and types.
19559 * tree-ssa-sccvn.c (vn_tables_s::nary):
19560 Change type to hash_table. Update dependent calls and types.
19561 (vn_tables_s::phis): Likewise.
19562 (vn_tables_s::references): Likewise.
19564 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19565 (vn_reference_eq): Update parameter and return types.
19567 * tree-ssa-structalias.c (pointer_equiv_class_table):
19568 Change type to hash_table. Update dependent calls and types.
19569 (location_equiv_class_table): Likewise.
19571 * tree-vect-data-refs.c: Consequential changes for making
19572 peeling a hash_table.
19574 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19575 (destroy_loop_vec_info): Dependent hash_table update.
19577 * tree-vectorizer.h (peeling_htab):
19578 Change type to hash_table. Update dependent calls and types.
19580 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19582 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19583 to check the register content is equal or not.
19584 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19585 to assign register content record.
19586 * lra-eliminations.c (update_reg_eliminate): Use
19587 lra_update_reg_val_offset to update register content offset.
19588 * lra-int.h (struct lra_reg): Add offset member.
19589 (lra_reg_val_equal_p): New static inline function.
19590 (lra_update_reg_val_offset): New static inline function.
19591 (lra_assign_reg_val): New static inline function.
19592 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19593 to assign register content record.
19594 (initialize_lra_reg_info_element): Initial offset to zero.
19596 2013-04-23 Catherine Moore <clm@codesourcery.com>
19598 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19599 operands. Record compression.
19601 2013-04-23 Xinliang David Li <davidxl@google.com>
19603 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19605 2013-04-23 Richard Biener <rguenther@suse.de>
19607 PR middle-end/57036
19608 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19609 parameter, only add abnormal goto edges from the copied body
19610 if the call could perform abnormal gotos.
19611 (copy_cfg_body): Adjust.
19613 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19615 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19617 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19619 * coretypes.h (gimple_stmt_iterator): Add struct to make
19622 2013-04-23 Richard Biener <rguenther@suse.de>
19624 PR tree-optimization/57026
19625 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19626 from SSA names occuring in abnormal PHI nodes.
19628 2013-04-22 Andi Kleen <ak@linux.intel.com>
19630 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19632 2013-04-22 Andi Kleen <ak@linux.intel.com>
19634 * lto/lto.c (print_lto_report_1): Declare early.
19635 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19637 2013-04-22 Andi Kleen <ak@linux.intel.com>
19639 * common.opt (-flto-report-wpa): Add.
19640 * doc/invoke.texi (-flto-report-wpa): Add.
19641 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19644 2013-04-22 Xinliang David Li <davidxl@google.com>
19646 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19647 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19648 * Makefile.in: New dependency
19650 David Daney <ddaney.cavm@gmail.com>
19652 * configure.ac (gcc_cv_as_micromips_support): Use the
19653 --fatal-warnings option.
19654 * configure: Regenerate.
19656 2013-04-22 Marek Polacek <polacek@redhat.com>
19659 * tsan.c (instrument_expr): Don't instrument expression
19660 in case its size is zero.
19662 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19666 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19668 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19670 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19672 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19673 (gimple_stmt_iterator): New typedef.
19674 * gimple.h (gimple_stmt_iterator): Rename to...
19675 (gimple_stmt_iterator_d): ... This.
19676 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19677 trees be valid for GIMPLE and GENERIC.
19678 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19679 * gimple-fold.c (gimple_fold_call): Call target hook
19680 gimple_fold_builtin.
19681 * hooks.c (hook_bool_gsiptr_false): New.
19682 * hooks.h (hook_bool_gsiptr_false): New.
19683 * target.def (fold_stmt): New.
19684 * doc/tm.texi: Regenerate.
19686 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19689 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19690 a set sp if no stack realignment.
19692 2013-04-22 Nick Clifton <nickc@redhat.com>
19694 * config.gcc (tilegx-linux): Extend extra_objs rather than
19696 (tilepro-linux): Likewise.
19698 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19700 * config/aarch64/aarch64-builtins.c
19702 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19703 (VAR<1-12>): Add MAP parameter.
19704 (BUILTIN_*): Likewise.
19705 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19706 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19707 (aarch64_ushl_n<mode>): Likewise.
19708 (aarch64_sshr_n<mode>): Likewise.
19709 (aarch64_ushr_n<mode>): Likewise.
19710 (aarch64_<maxmin><mode>): Likewise.
19711 (aarch64_sqrt<mode>): Likewise.
19712 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19713 (vshr<q>_n_*): Likewise.
19715 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19717 * config/aarch64/aarch64-builtins.c
19718 (aarch64_simd_builtin_type_mode): Handle SF types.
19720 (BUILTIN_GPF): Define.
19721 (aarch64_init_simd_builtins): Handle SF types.
19722 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19723 (frecps): Likewise.
19724 (frecpx): Likewise.
19725 * config/aarch64/aarch64-simd.md
19726 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19727 (aarch64_frecpe<mode>): New.
19728 (aarch64_frecps<mode>): Likewise.
19729 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19730 (v8type): Add frecp<esx>.
19731 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19732 (aarch64_frecps<mode>): Likewise.
19733 * config/aarch64/iterators.md (FRECP): New.
19734 (frecp_suffix): Likewise.
19735 * config/aarch64/arm_neon.h
19736 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19738 2013-04-22 Christian Bruel <christian.bruel@st.com>
19741 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19742 (REG_CLASS_NAMES): Idem.
19743 (REG_CLASS_CONTENTS): Idem.
19744 (REGCLASS_HAS_FP_REG): Idem.
19745 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19746 (sh_conditional_register_usage): Idem.
19748 2013-04-21 Jeff Law <law@redhat.com>
19750 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19751 (ssa_forward_propagate_and_combine): Use it.
19753 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19755 * lra.c: Update the flow chart diagram.
19757 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19759 PR rtl-optimization/56847
19760 * lra-constraints.c (process_alt_operands): Discourage alternative
19761 with non-matche doffsettable memory constraint fro memory with
19764 2013-04-19 Richard Biener <rguenther@suse.de>
19766 PR tree-optimization/56982
19767 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19769 * gimplify.c (gimplify_call_expr): Notice special calls.
19770 (gimplify_modify_expr): Likewise.
19771 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19772 abnormal control flow receivers.
19773 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19774 in the same way as cfun->has_nonlocal_labels.
19775 (gimple_purge_dead_abnormal_call_edges): Likewise.
19776 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19777 receivers start a basic-block.
19779 2013-04-19 Richard Biener <rguenther@suse.de>
19781 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19783 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19784 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19785 (SLP_TREE_LOAD_PERMUTATION): Add.
19786 (vect_transform_slp_perm_load): Adjust prototype.
19787 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19788 (vect_free_slp_instance): Likewise.
19789 (vect_create_new_slp_node): Likewise.
19790 (vect_supported_slp_permutation_p): Remove.
19791 (vect_slp_rearrange_stmts): Adjust.
19792 (vect_supported_load_permutation_p): Likewise. Inline
19793 vect_supported_slp_permutation_p here.
19794 (vect_analyze_slp_instance): Compute load permutations per
19795 slp node instead of per instance.
19796 (vect_get_slp_defs): Adjust.
19797 (vect_transform_slp_perm_load): Likewise.
19798 (vect_schedule_slp_instance): Remove redundant code.
19799 (vect_schedule_slp): Remove hack for PR56270, add it ...
19800 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19801 CSE loads for SLP. Adjust.
19803 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19805 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19806 spelling in two comments.
19808 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19811 * config/arm/arm.c (load_multiple_sequence): Require SP
19812 as base register for loads if SP is in the register list.
19814 2013-04-19 Martin Jambor <mjambor@suse.cz>
19816 PR tree-optimization/56718
19817 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19818 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19819 and made public. Adjusted all callers.
19820 (ipa_intraprocedural_devirtualization): New function.
19821 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19822 (ipa_intraprocedural_devirtualization): Likewise.
19823 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19825 2013-04-19 Richard Biener <rguenther@suse.de>
19827 PR tree-optimization/57000
19828 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19830 2013-04-19 Terry Guo <terry.guo@arm.com>
19832 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19834 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19835 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19836 (cortex_m4_fmacs): Use new reservations.
19837 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19839 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19841 PR rtl-optimization/56999
19842 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19844 (lra_coalesce): Remove split_origin_bitmap and related code.
19845 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19846 ranges if necessary.
19848 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19850 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19852 (ix86_expand_call): Remove clobbered_registers array and use
19853 x86_64_ms_sysv_extra_clobbered_registers instead.
19854 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19856 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19858 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19859 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19860 (*call_value_rex64_ms_sysv): Ditto.
19862 2013-04-18 Cary Coutant <ccoutant@google.com>
19864 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19865 parent instead of die_mark.
19867 2013-04-18 Diego Novillo <dnovillo@google.com>
19869 * gimple.c (create_gimple_tmp): New.
19870 (get_expr_type): New.
19871 (build_assign): New.
19872 (build_type_cast): New.
19873 * gimple.h (enum ssa_mode): Define.
19874 (gimple_seq_set_location): New.
19875 * asan.c (build_check_stmt): Change some gimple_build_* calls
19876 to use build_assign and build_type_cast.
19878 2013-04-18 Richard Biener <rguenther@suse.de>
19880 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19881 handle negative step. Remove redundant checks.
19882 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19883 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19884 for negative step and grouped loads fail to vectorize.
19886 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19888 * emit-rtl.c (reset_insn_used_flags): New function.
19889 (reset_all_used_flags): Use it.
19890 (verify_insn_sharing): New function.
19891 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19893 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19895 PR tree-optimization/56984
19896 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19897 and (x >> M) >= N don't register any assertion if N << M is the
19900 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19902 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19903 do not delete_insn insns that have not yet been emitted, only
19904 unlink them with remove_insn.
19905 * df-scan.c (df_insn_delete): Revert r197492.
19907 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19909 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19910 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19912 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19914 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19915 define_insn_and_split.
19916 (and_scc,ior_scc,negscc): Likewise.
19917 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19919 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19921 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19922 can benefit from using a sequence of LDRD instructions in epilogue
19923 instead of a single LDM instruction.
19925 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19928 * doc/extend.texi: Fix typo.
19930 2013-04-17 Richard Biener <rguenther@suse.de>
19932 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19933 (vect_build_slp_tree): ... here.
19934 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19935 match. Remove special-casing of mismatched complex loads.
19936 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19937 re-try the match with swapped commutative operands.
19938 (vect_supported_load_permutation_p): Remove special-casing of
19939 mismatched complex loads.
19940 (vect_analyze_slp_instance): Adjust.
19942 2013-04-17 Richard Biener <rguenther@suse.de>
19944 PR rtl-optimization/56921
19945 * cfgloop.h (struct loop): Add simple_loop_desc member.
19946 (struct niter_desc): Mark with GTY(()).
19947 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19948 * loop-iv.c (get_simple_loop_desc): Likewise.
19949 (free_simple_loop_desc): Likewise.
19952 2013-04-16 Richard Biener <rguenther@suse.de>
19954 PR rtl-optimization/56921
19955 * loop-init.c (pass_rtl_move_loop_invariants): Add
19956 TODO_do_not_ggc_collect to todo_flags_finish.
19957 (pass_rtl_unswitch): Same.
19958 (pass_rtl_unroll_and_peel_loops): Same.
19959 (pass_rtl_doloop): Same.
19961 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19963 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19964 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19965 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19967 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19968 * tree-streamer.c (record_common_node): Adjust reference in comment.
19970 2013-04-17 Terry Guo <terry.guo@arm.com>
19972 * config/arm/cortex-m4.md: Add a new bypass.
19974 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19976 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19978 (*subs_<optab><mode>_multp2): New pattern.
19979 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19980 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19982 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19984 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19985 (*subs_mul_imm_<mode>): New pattern.
19987 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19990 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19991 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19992 (vsx_movti_32bit): Change j->wa to O->wa.
19994 2013-04-16 Richard Biener <rguenther@suse.de>
19996 PR rtl-optimization/56921
19997 * loop-init.c (pass_rtl_move_loop_invariants): Add
19998 TODO_do_not_ggc_collect to todo_flags_finish.
19999 (pass_rtl_unswitch): Same.
20000 (pass_rtl_unroll_and_peel_loops): Same.
20001 (pass_rtl_doloop): Same.
20003 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20005 * config/arm/arm.c (emit_multi_reg_push): New declaration
20006 for an existing function.
20007 (arm_emit_strd_push): New function.
20008 (arm_expand_prologue): Used here.
20009 (arm_emit_ldrd_pop): New function.
20010 (arm_expand_epilogue): Used here.
20011 (arm_get_frame_offsets): Update condition.
20012 (arm_emit_multi_reg_pop): Add a special case for load of a single
20013 register with writeback.
20015 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20017 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20020 2013-04-16 Richard Biener <rguenther@suse.de>
20022 PR tree-optimization/56756
20023 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20024 (first_mem_ref_loc): New.
20025 (execute_sm): Place the load temporarily before a previous
20026 access instead of in the latch edge to ensure its SSA dependencies
20027 are defined at points dominating the load.
20029 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20031 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20032 correct fix by moving header and footer insn to the footer of
20033 the merged basic block. Clear BB_END of the merged-away block.
20035 PR middle-end/43631
20036 * emit-rtl.c (make_note_raw): New function.
20037 (link_insn_into_chain): New static inline function.
20038 (add_insn): Use it.
20039 (add_insn_before, add_insn_after): Factor insn chain linking code...
20040 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20041 using link_insn_into_chain.
20042 (note_outside_basic_block_p): New helper function for emit_note_after
20043 and emit_note_before.
20044 (emit_note_after): Use nobb variant of add_insn_after if the note
20045 should not be contained in a basic block.
20046 (emit_note_before): Use nobb variant of add_insn_before if the note
20047 should not be contained in a basic block.
20048 (emit_note_copy): Use make_note_raw.
20049 (emit_note): Likewise.
20050 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20051 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20052 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20053 the moved barrier the tail of the basic block it follows.
20054 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20056 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20058 PR tree-optimization/56962
20059 * gimple-ssa-strength-reduction.c (record_increment): Only set
20060 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20061 either rhs1 or rhs2 is equal to c->base_expr.
20063 2013-04-15 Richard Biener <rguenther@suse.de>
20065 PR tree-optimization/56933
20066 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20068 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20069 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20070 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20071 dependence check ...
20072 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20074 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20075 GROUP_READ_WRITE_DEPENDENCE.
20077 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20079 * emit-rtl.c (reset_all_used_flags): New function.
20080 (verify_rtl_sharing): Call reset_all_used_flags before and after
20081 performing the checks.
20083 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20085 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20086 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20087 * config/arm/constraints.md (De): New constraint.
20088 * config/arm/neon.md (anddi3_neon): Delete.
20089 (neon_vand<mode>): Expand to standard anddi3 pattern.
20090 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20091 Move earlier in the file.
20092 (neon_inv_logic_op2): Likewise.
20093 (arm_anddi_operand_neon): New predicate.
20095 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20097 * configure.ac (gcc_cv_ld_as_needed): Set
20098 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20099 Use -z ignore, -z record on *-*-solaris2*.
20100 (HAVE_LD_AS_NEEDED): Update comment.
20101 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20102 * configure: Regenerate.
20103 * config.in: Regenerate.
20104 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20105 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20106 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20107 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20108 equivalents. Fix markup.
20109 * doc/tm.texi: Regenerate.
20111 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20113 * config/i386/i386.opt: New option mstack-protector-guard=.
20114 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20115 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20116 TARGET_SSP_TLS_GUARD.
20117 * config/i386/i386.c (ix86_option_override_internal): Set
20118 ix86_stack_protector_guard.
20119 * config/i386/i386.md (stack_protect_set): Enable for
20120 TARGET_SSP_TLS_GUARD only.
20121 (stack_protect_set_<mode>): Ditto.
20122 (stack_protect_test): Ditto.
20123 (stack_protect_test_<mode>): Ditto.
20124 * doc/invoke.texi (i386 Option): Document.
20126 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20129 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20130 (S_MODES): Set H_MODE bit.
20131 (SF_MODES): Set only S_MODE and SF_MODE bits.
20132 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20133 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20134 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20135 <MODE_FLOAT>: Likewise.
20137 2013-04-15 Joey Ye <joey.ye@arm.com>
20139 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20141 2013-04-15 Joey Ye <joey.ye@arm.com>
20143 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20145 (thumb_far_jump_used_p): Count instruction size and set
20148 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20150 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20151 * resource.c (mark_target_live_regs): Reformat conditional block.
20153 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20155 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20156 notes, they are emitted only just before final.
20157 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20159 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20161 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20162 * cfgrtl.c (delete_insn): Call it here instead.
20163 * lra-spills.c (lra_final_code_change): Use delete_insn.
20164 * haifa-sched.c (sched_remove_insn): Likewise.
20165 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20166 returning to the nop pool.
20167 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20168 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20170 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20172 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20173 * doc/tm.texi: Regenerated.
20175 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20177 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20180 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20182 * df-core.c (df_find_def): Compare register numbers.
20183 (df_find_use): Likewise.
20185 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20188 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20189 lra_in_progress for return.
20191 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20193 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20194 define_insn into define_insn_and_split and emit movsicc patterns.
20196 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20198 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20200 2013-04-12 Richard Biener <rguenther@suse.de>
20202 * tree-pass.h (TODO_do_not_ggc_collect): New.
20203 * passes.c (execute_one_ipa_transform_pass): Honor
20204 TODO_do_not_ggc_collect.
20205 (execute_one_pass): Likewise.
20208 2013-04-10 Richard Biener <rguenther@suse.de>
20210 * passes.c (init_optimization_passes): Remove reload pass.
20211 * ira.c (do_reload): Merge into ...
20213 (rest_of_handle_reload): Remove.
20214 (pass_reload): Likewise.
20215 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20216 of reload for vzeroupper pass placement.
20218 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20220 PR tree-optimization/56918
20221 PR tree-optimization/56920
20222 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20223 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20224 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20225 use wide_mul_with_sign method.
20227 2013-04-12 Richard Biener <rguenther@suse.de>
20229 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20230 not be considered a gimple constant.
20232 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20234 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20235 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20238 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20239 Jakub Jelinek <jakub@redhat.com>
20241 * opts.c: Include diagnostic-color.h.
20242 (common_handle_option): Handle OPT_fdiagnostics_color_.
20243 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20244 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20245 (diagnostic-color.o): New.
20246 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20247 (diagnostic_color_rule): New enum.
20248 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20249 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20250 the location string.
20251 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20252 either NULL, or color kind.
20253 * diagnostic-color.c: New file.
20254 * diagnostic-color.h: New file.
20255 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20257 * doc/invoke.texi (-fdiagnostics-color): Document.
20258 * pretty-print.h (pp_show_color): Define.
20259 (struct pretty_print_info): Add show_color field.
20260 * diagnostic.c: Include diagnostic-color.h.
20261 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20262 macros. Colorize error:, warning: etc. strings and also the location
20264 (diagnostic_show_locus): Colorize the caret line.
20265 * pretty-print.c: Include diagnostic-color.h.
20266 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20267 inside of %< %> quotes or quoted through q format modifier.
20269 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20271 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20273 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20275 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20276 code in CC_NZ mode.
20277 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20280 2013-04-11 Marek Polacek <polacek@redhat.com>
20282 PR tree-optimization/48184
20283 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20285 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20287 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20288 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20289 (skip_simple_arithmetic): Tidy up.
20290 * tree.h (skip_simple_constant_arithmetic): Declare.
20292 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20294 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20296 2013-04-11 Richard Biener <rguenther@suse.de>
20298 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20299 generate vector constants.
20301 2013-04-11 Richard Biener <rguenther@suse.de>
20303 PR tree-optimization/56878
20304 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20305 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20307 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20308 Prefer to align the DR with the most invariant base address.
20310 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20312 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20315 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20317 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20318 floating-point vector comparisons against 0.
20320 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20322 PR tree-optimization/56899
20323 * fold-const.c (extract_muldiv_1): Apply distributive law
20324 only if TYPE_OVERFLOW_WRAPS (ctype).
20326 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20329 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20330 loading has side effect.
20332 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20334 * config/sparc/sparc.c: Include tree-pass.h.
20335 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20336 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20337 head of file. Change return type. Split off gate function.
20338 (sparc_gate_work_around_errata): New function.
20339 (pass_work_around_errata): New pass definition.
20340 (insert_pass_work_around_errata) New pass insert definition to
20341 insert pass_work_around_errata just after delayed-branch scheduling.
20342 (sparc_option_override): Insert the pass.
20343 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20345 2013-04-10 David S. Miller <davem@davemloft.net>
20347 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20348 or -mcpu=hypersparc.
20350 * target.def (cstore_mode): New hook.
20351 * target.h: Include insn-codes.h
20352 * targhooks.c: Likewise.
20353 (default_cstore_mode): New function.
20354 * targhooks.h: Declare it.
20355 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20356 * doc/tm.texi: Rebuild.
20357 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20358 target hook, rather than inspecting the insn_data.
20359 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20360 (TARGET_CSTORE_MODE): Redefine.
20361 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20363 * config/sparc/predicates.md (cstore_result_operand): New special
20365 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20366 Use it for operand 0.
20367 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20368 (*snesi_special): Likewise.
20369 (*snesi_zero): Likewise.
20370 (*seqsi_zero): Likewise.
20371 (*sltu_insn): Likewise.
20372 (*sgeu_insn): Likewise.
20373 (*seqdi_special): Make operand 0 and comparison operation be of
20375 (*snedi_special): Likewise.
20376 (*snedi_special_vis3): Likewise.
20377 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20378 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20379 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20380 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20381 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20382 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20383 (*sltu_extend_sp64): Likewise.
20384 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20385 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20386 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20387 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20388 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20390 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20392 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20393 (aarch64_start_file): Use the new function.
20395 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20396 Jason Merrill <jason@redhat.com>
20398 * common.opt: Add -gdwarf.
20399 * opts.c (common_handle_option): Handle it.
20400 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20402 2013-04-10 Richard Biener <rguenther@suse.de>
20404 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20405 (execute_one_ipa_transform_pass): But unconditionally here.
20406 (execute_one_pass): And here.
20407 (init_optimization_passes): Remove reload pass.
20408 * tree-pass.h (TODO_ggc_collect): Remove.
20409 (pass_reload): Likewise.
20410 * ira.c (do_reload): Merge into ...
20412 (rest_of_handle_reload): Remove.
20413 (pass_reload): Likewise.
20414 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20415 of reload for vzeroupper pass placement.
20416 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20417 and todo_flags_finish of all passes.
20419 2013-04-10 Richard Biener <rguenther@suse.de>
20421 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20422 first_const_oprnd field, rename first_def_type to first_op_type.
20423 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20424 (vect_get_and_check_slp_defs): Always use the type of the
20425 operand. Allow mixed vect_external_def, vect_constant_def types.
20426 (vect_get_constant_vectors): Handle mixed vect_external_def,
20427 vect_constant_def types.
20429 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20431 PR tree-optimization/55524
20432 * tree-ssa-math-opts.c
20433 (convert_mult_to_fma): Don't use an fms construct
20434 when we don't have an fms operation, but fnma, and it looks
20435 likely that we'll be able to use the latter.
20437 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20439 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20441 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20442 inline fail caused by overwritable functions.
20444 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20446 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20447 unnecessary bits in the constant power of two case.
20449 2013-04-10 Richard Biener <rguenther@suse.de>
20451 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20452 broken code swapping operands.
20453 (vect_build_slp_tree): Do not compute load permutations here.
20454 (vect_analyze_slp_instance): Compute load permutations here,
20455 after building the SLP tree.
20457 2013-04-09 Christian Bruel <christian.bruel@st.com>
20459 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20460 of next/prev_real_insn.
20462 2013-04-09 Jan Hubicka <jh@suse.cz>
20464 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20465 Drop aliased parameter.
20466 (function_and_variable_visibility): Do not handle alias pairs.
20467 * cgraph.c (varpool_externally_visible_p): Update prototype.
20468 * varpool.c (varpool_add_new_variable): Update.
20470 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20472 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20474 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20476 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20478 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20480 2013-04-09 Marek Polacek <polacek@redhat.com>
20482 PR tree-optimization/48762
20483 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20485 2013-04-09 Richard Biener <rguenther@suse.de>
20487 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20489 (vect_build_slp_tree): Likewise.
20490 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20491 calculating the cost of a SLP instance.
20492 (vect_analyze_slp_instance): Use it from here, after building
20495 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20497 PR middle-end/56883
20498 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20499 expand_omp_for_static_chunk): Use simple_p = true in
20500 force_gimple_operand_gsi calls when assigning to addressable decls.
20502 2013-04-09 Jeff Law <law@redhat.com>
20504 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20505 when the boolean was created by converting a wider object which
20506 had a boolean range.
20508 2013-04-09 Richard Biener <rguenther@suse.de>
20510 * tree-vectorizer.h (slp_void_p): Remove.
20511 (slp_tree): Typedef before _slp_tree declaration.
20512 (struct _slp_tree): Use a vector of slp_tree as children.
20513 (vect_get_place_in_interleaving_chain): Remove.
20514 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20516 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20518 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20519 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20520 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20521 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20522 Use slp_node instead of slp_void_p and adjust.
20524 2013-04-09 Richard Biener <rguenther@suse.de>
20526 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20527 work that is not necessary.
20529 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20531 PR tree-optimization/56854
20532 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20533 forward into clobber stmts if it would change MEM_REF lhs into
20536 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20538 * tree.c (type_hash_lookup, type_hash_add): Make static.
20539 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20541 2013-04-09 Richard Biener <rguenther@suse.de>
20543 * tree.h (unsave_expr_now): Remove.
20544 * tree-inline.c (mark_local_for_remap_r): Remove.
20545 (unsave_expr_1): Likewise.
20546 (unsave_r): Likewise.
20547 (unsave_expr_now): Likewise.
20548 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20549 (propagate_tree_value): Likewise.
20551 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20553 * doc/rtl.texi (sequence): Rewrite documentation to match the
20554 current use of SEQUENCE rtl objects.
20555 * rtl.def (SEQUENCE): Likewise.
20557 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20558 Update documentation.
20559 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20560 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20562 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20564 2013-04-08 Teresa Johnson <tejohnson@google.com>
20566 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20567 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20569 (estimate_edge_size_and_time): Add comment to suggest using rounding
20571 (estimate_node_size_and_time): Ditto.
20572 (remap_edge_change_prob): Use helper rounding divide methods.
20573 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20574 (gimple_mod_pow2_value_transform): Ditto.
20575 (gimple_mod_subtract_transform): Ditto.
20576 (gimple_ic_transform): Ditto.
20577 (gimple_stringops_transform): Ditto.
20578 * stmt.c (conditional_probability): Ditto.
20579 (emit_case_dispatch_table): Ditto.
20580 * lto-cgraph.c (merge_profile_summaries): Ditto.
20581 * tree-optimize.c (execute_fixup_cfg): Ditto.
20582 * cfgcleanup.c (try_forward_edges): Ditto.
20583 * cfgloopmanip.c (scale_loop_profile): Ditto.
20585 (duplicate_loop_to_header_edge): Ditto.
20586 (lv_adjust_loop_entry_edge): Ditto.
20587 * tree-vect-loop.c (vect_transform_loop): Ditto.
20588 * profile.c (compute_branch_probabilities): Ditto.
20589 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20590 * lto-streamer-in.c (input_cfg): Ditto.
20591 * gimple-streamer-in.c (input_bb): Ditto.
20592 * ipa-cp.c (update_profiling_info): Ditto.
20593 (update_specialized_profile): Ditto.
20594 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20595 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20597 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20598 (compute_trg_info): Ditto.
20599 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20600 (purge_dead_edges): Ditto.
20601 * loop-unswitch.c (unswitch_loop): Ditto.
20602 * cgraphclones.c (cgraph_clone_edge): Ditto.
20603 (cgraph_clone_node): Ditto.
20604 * tree-inline.c (copy_bb): Ditto.
20605 (copy_edges_for_bb): Ditto.
20606 (initialize_cfun): Ditto.
20607 (copy_cfg_body): Ditto.
20608 (expand_call_inline): Ditto.
20610 2013-04-08 Kai Tietz <ktietz@redhat.com>
20612 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20613 TARGET_CYGWIN64 by TARGET_64BIT.
20615 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20617 * config/epiphany/epiphany.md (GPR_1): New constant.
20618 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20619 * config/epiphany/epiphany.c (gen_compare_reg):
20620 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20621 is already in place.
20622 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20623 Don't require being called during rtl expansion; If y operlaps r0,
20625 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20626 (epiphany_expand_epilogue): Likewise.
20628 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20629 Don't use CC_FPmode for ORDERED / UNORDERED.
20630 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20632 * config/epiphany/constraints.md (CnL): New constraint.
20633 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20634 * config/epiphany/predicates.md (add_operand): Allow 1024.
20636 * config/epiphany/epiphany.md (logical_op): New code iterator.
20637 (op_mnc): New code attribute.
20638 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20639 (mov_f+1, mov_f+2): New peephole2 patterns.
20641 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20642 (cstoresi4): Also allow re-use of zero result when doing a NE
20643 comparison to a non-zero operand.
20644 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20646 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20647 Use gen_addsi3_i / gen_subsi3_i.
20649 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20653 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20654 contain anything but clobbers, at most one __builtin_stack_restore,
20655 optionally debug stmts and final resx, and if it has at least one
20656 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20657 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20658 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20659 which isn't defaut definition, remove them.
20660 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20661 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20662 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20663 with MEM_REF LHS with SSA_NAME address.
20665 2013-04-08 Jeff Law <law@redhat.com>
20667 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20669 2013-04-08 Richard Biener <rguenther@suse.de>
20671 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20673 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20674 determined vector type.
20675 (vect_analyze_data_refs): Likewise.
20676 (vect_get_new_vect_var): Adjust.
20677 (vect_create_destination_var): Preserve SSA name versions.
20678 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20679 not dump anything here.
20681 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20683 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20684 Add member lr_slot_known.
20685 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20687 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20688 Remove code that sets lr_slot_offset according to what a previous
20689 version of epiphany_emit_save_restore used to do.
20690 (epiphany_emit_save_restore): When doing an lr save or restore,
20691 set/verify lr_slot_known and lr_slot_offset.
20693 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20696 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20699 2013-04-08 Richard Biener <rguenther@suse.de>
20701 * alias.c (find_base_term): Fix thinko in previous change.
20703 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20705 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20706 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20707 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20708 if possible to compute val.
20709 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20710 For QImode integers don't require anything about precision. Use
20711 const_with_all_bytes_same to find out if the constant doesn't have
20712 repeated bytes in it.
20714 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20716 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20719 2013-04-08 Marek Polacek <polacek@redhat.com>
20721 PR rtl-optimization/48182
20722 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20725 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20728 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20729 nuses, make sure we have a label.
20731 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20734 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20735 (rs6000_emit_swdiv_low_precision): Remove.
20736 (rs6000_emit_swdiv): Rewrite to handle between one and four
20737 iterations of Newton-Raphson generally; modify required number of
20738 iterations for some cases.
20739 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20741 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20743 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20744 set-but-unused variable.
20746 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20747 basic blocks of released function bodies garbage-collectable.
20749 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20750 (struct rtl_opt_pass): Add TODO_df_finish.
20752 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20754 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20756 * config/arm/constraints.md (q): New constraint.
20757 * config/arm/ldrdstrd.md: New file.
20758 * config/arm/arm.md (ldrdstrd.md) New include.
20759 (arm_movdi): Use "q" instead of "r" constraint
20760 for double-word memory access.
20761 (movdf_soft_insn): Likewise.
20762 * config/arm/vfp.md (movdi_vfp): Likewise.
20763 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20764 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20765 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20766 (mem_ok_for_ldrd_strd): Likewise.
20767 (output_move_double): Update assertion.
20769 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20771 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20773 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20775 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20776 define_insn_and_split.
20777 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20779 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20781 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20782 define_insn_and_split.
20783 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20784 (shiftsi3_compare): New pattern.
20785 (rrx): New pattern.
20786 * config/arm/unspecs.md (UNSPEC_RRX): New.
20788 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20790 * config/arm/arm.md (negdi_extendsidi): New pattern.
20791 (negdi_zero_extendsidi): Likewise.
20793 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20795 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20796 define_insn_and_split.
20797 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20798 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20800 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20802 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20803 define_insn_and_split.
20804 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20805 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20807 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20809 * config/arm/arm.md (subsi3_carryin): New pattern.
20810 (subsi3_carryin_const): Likewise.
20811 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20812 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20814 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20816 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20818 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20820 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20821 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20823 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20825 * config/arm/arm.c (arm_expand_builtin): Change fcode
20826 type to unsigned int.
20828 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20830 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20832 2013-04-04 Ian Lance Taylor <iant@google.com>
20834 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20837 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20839 PR middle-end/56729
20840 * df-scan.c (df_insn_delete): Disable failing assert.
20842 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20844 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20845 New function prototype.
20846 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20847 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20848 (arm_builtin_vectorized_function): New function.
20850 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20852 * config/arm/arm_neon_builtins.def: New file.
20853 * config/arm/arm.c (neon_builtin_data): Move contents to
20854 arm_neon_builtins.def.
20855 (enum arm_builtins): Include neon builtin definitions.
20856 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20857 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20859 2013-04-04 Marek Polacek <polacek@redhat.com>
20861 PR tree-optimization/48186
20862 * predict.c (maybe_hot_frequency_p): Return false if
20863 HOT_BB_FREQUENCY_FRACTION is 0.
20864 (cgraph_maybe_hot_edge_p): Likewise.
20866 2013-04-04 Richard Biener <rguenther@suse.de>
20868 PR tree-optimization/56826
20869 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20872 2013-04-04 Richard Biener <rguenther@suse.de>
20874 PR tree-optimization/56213
20875 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20876 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20878 2013-04-04 Richard Biener <rguenther@suse.de>
20880 PR tree-optimization/56837
20881 * tree-loop-distribution.c (classify_partition): For non-zero
20882 values require that the value has the same precision as its
20883 mode to be useful as memset value.
20885 2013-04-03 Nick Clifton <nickc@redhat.com>
20887 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20888 (fmssf4): Use fmsf.s on E3V5 architectures.
20889 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20890 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20892 2013-04-03 Jeff Law <law@redhat.com>
20894 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20895 (lra-eliminations.o): Likewise.
20897 2013-04-03 Teresa Johnson <tejohnson@google.com>
20899 * gcov-io.c (compute_working_sets): Moved most of body of old
20900 compute_working_sets here from profile.c.
20901 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20902 (gcov_working_set_t): Moved typedef here from basic-block.h
20903 (compute_working_set): Declare.
20904 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20905 (get_working_sets): Renamed from compute_working_set,
20906 replace most of body with call to new compute_working_sets.
20907 (get_exec_counts): Replace call to compute_working_sets
20908 to get_working_sets.
20909 * profile.h (get_working_sets): Renamed from compute_working_set.
20910 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20911 to get_working_sets.
20912 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20913 * gcov-dump.c (dump_working_sets): New function.
20915 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20917 * hwint.c (sext_hwi, zext_hwi): New functions.
20918 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20919 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20920 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20921 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20922 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20923 (sext_hwi, zext_hwi): New functions.
20925 2013-04-03 Jeff Law <law@redhat.com>
20927 PR tree-optimization/56799
20928 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20929 back test for widening conversion erroneously dropped in prior change.
20931 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20934 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20935 instead of next_real_insn.
20937 2013-04-03 Marek Polacek <polacek@redhat.com>
20940 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20942 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20945 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20947 (thumb1_output_casesi): Likewise.
20948 (thumb2_output_casesi): Likewise.
20950 2013-04-03 Richard Biener <rguenther@suse.de>
20952 PR tree-optimization/56817
20953 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20955 (tree_unroll_loops_completely_1): ... new function to manually
20956 walk the loop tree, properly defering outer loops of unrolled
20957 loops to later iterations.
20959 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20961 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20962 (vectorizable_load): Likewise.
20963 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20964 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20966 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20968 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20971 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20973 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20975 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20977 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20979 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20981 PR tree-optimization/56790
20982 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20985 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20987 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20989 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20990 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20993 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20996 * tree.h (force_folding_builtin_constant_p): New decl.
20997 * builtins.c (force_folding_builtin_constant_p): New variable.
20998 (fold_builtin_constant_p): Fold immediately also if
20999 force_folding_builtin_constant_p.
21001 2013-04-03 Richard Biener <rguenther@suse.de>
21003 PR tree-optimization/56812
21004 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21005 DRs of the same interleaving chain are independent.
21007 2013-04-02 Jason Merrill <jason@redhat.com>
21009 * gdbinit.in (pbb): Use debug fn.
21011 2013-04-02 Lawrence Crowl <crowl@google.com>
21013 * sese.h (struct ivtype_map_elt_s): Remove unused.
21014 (extern debug_ivtype_map): Remove unused.
21015 (extern eq_ivtype_map_elts): Remove unused.
21016 * sese.c (debug_ivtype_map): Removed unused.
21017 (debug_ivtype_map_1): Removed unused.
21018 (debug_ivtype_elt): Remove unused.
21019 (eq_ivtype_map_elts): Remove unused.
21022 2013-04-02 Kai Tietz <ktietz@redhat.com>
21025 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21026 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21027 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21029 (legitimize_pe_coff_symbol): Likewise.
21030 (is_imported_p): New helper-function.
21031 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21032 for Windows x64 targets.
21033 (ix86_expand_prologue): Optimize for pe-coff targets.
21034 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21035 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21036 medium/large code-model.
21037 (legitimize_pic_address): Likewise.
21038 (legitimize_tls_address): Likewise.
21039 (ix86_expand_call): Likewise.
21040 (x86_output_mi_thunk): Likewise.
21041 (get_dllimport_decl): Add new beimport argument.
21042 (construct_plt_address): Don't assert for x64 pe-coff targets.
21043 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21045 (SYMBOL_FLAG_STUBVAR): New macro.
21046 (SYMBOL_REF_STUBVAR_P): Likewise.
21047 * config/i386/winnt.c (stub_list): New structure.
21048 (stub_head): New local variable.
21049 (i386_pe_record_stub): New function.
21050 (i386_pe_file_end): Emit refptr-stubs.
21052 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21054 PR rtl-optimization/56745
21055 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21056 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21059 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21060 and both of them are MEM_REFs, just compare first argument for
21061 equality and attempt to deal even with differing offsets.
21064 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21065 of gimple_clobber_p to be MEM_REF.
21066 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21067 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21068 after gimplification.
21069 * asan.c (get_mem_ref_of_assignment): Don't instrument
21070 gimple_clobber_p stmts.
21071 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21072 gimple_clobber_p stmt if they have MEM_REF lhs and
21073 are dead because of another gimple_clobber_p stmt.
21074 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21075 gimple_clobber_p stmts like debug stmts.
21076 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21077 that refer to unused VAR_DECLs or uninitialized values.
21078 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21079 gimple_clobber_p stmts if they refer to removed parameters.
21080 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21083 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21085 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21086 using SWI48 mode attribute.
21088 2013-04-02 Wei Mi <wmi@google.com>
21090 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21091 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21092 *<rotate_insn><mode>3_mask in i386.md.
21094 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21096 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21098 2013-04-02 Richard Biener <rguenther@suse.de>
21100 PR tree-optimization/56778
21101 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21102 Runtime alias tests are not supported for gather loads.
21103 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21104 stmts referenced from SSA operands before updating SSA form.
21106 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21107 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21109 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21110 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21111 * config/arm/cortex-a53.md: New file.
21112 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21113 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21114 * config/arm/arm.c (arm_issue_rate): Likewise.
21115 * config/arm/arm-tune.md: Regenerate
21116 * config/arm/arm-tables.opt: Regenerate.
21117 * config/arm/arm-cores.def: Add cortex-a53.
21119 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21121 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21124 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21126 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21127 scalar load/store operations using B/H registers.
21128 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21130 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21132 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21134 * config/aarch64/aarch64.c
21135 (aarch64_simd_scalar_immediate_valid_for_move): New.
21136 * config/aarch64/aarch64-protos.h
21137 (aarch64_simd_scalar_immediate_valid_for_move): New.
21138 * config/aarch64/constraints.md (Dh, Dq): New.
21139 * config/aarch64/iterators.md (hq): New.
21141 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21143 * reorg.c (get_branch_condition): Deal with conditional returns.
21144 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21146 2013-04-01 Wei Mi <wmi@google.com>
21148 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21149 Truncate operand 2 using %b asm operand modifier.
21150 (*<shift_insn><mode>3_mask): Ditto.
21151 (*<rotate_insn><mode>3_mask): Ditto.
21153 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21155 PR middle-end/56798
21156 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21158 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21160 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21162 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21164 2013-03-30 Lawrence Crowl <crowl@google.com>
21166 * dse.c (clear_alias_sets): Remove never set.
21167 (disqualified_clear_alias_sets): Remove never set.
21168 (clear_alias_mode_pool): Remove never set.
21169 (dse_step0): Remove condition that is never true.
21170 (canon_address): Remove condition that is never true.
21171 (dse_step7): Remove condition that is never true.
21172 (rest_of_handle_dse): Remove condition that is never true.
21173 (rest_of_handle_dse::did_global): Remove never read from above.
21174 (dse_step2_spill): Remove never called from above.
21175 (dse_step5_spill): Remove never called from above.
21177 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21179 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21180 JUMP_TABLE_DATA changes.
21181 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21182 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21183 (Insns) <jump_table_data>: New entry.
21184 * doc/tm.texi: Regenerate.
21186 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21188 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21189 for table jump at the end of a basic block using tablejump_p.
21190 * targhooks.c (default_invalid_within_doloop): Likewise.
21191 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21192 target hook implementation that is identical to the default hook.
21193 (rs6000_invalid_within_doloop): Remove.
21195 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21196 unused variable from tablejump_p call.
21198 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21199 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21200 (INSN_DELETED_P): Likewise.
21201 (emit_jump_table_data): New prototype.
21202 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21203 after 4th as unused.
21204 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21205 * sched-vis.c (print_insn): Likewise.
21206 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21207 insn for compatibility with back ends that use next_active_insn to
21208 identify jump table data.
21209 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21210 (remove_insn): Likewise.
21211 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21213 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21214 (emit_jump_table_data): New function.
21216 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21217 basic block, a JUMP_TABLE_DATA never is.
21218 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21219 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21220 off from code handling real insns.
21221 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21222 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21223 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21224 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21225 is not a NONDEBUG_INSN_P.
21226 * ira-costs.c (scan_one_insn): Likewise.
21227 * jump.c (mark_all_labels): Likewise.
21228 (mark_jump_label_1): Likewise.
21229 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21230 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21232 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21233 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21234 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21235 JUMP_TABLE_DATA_P insns.
21236 (calculate_elim_costs_all_insns): Likewise.
21237 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21238 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21239 (delete_output_reload): Code style fixups.
21240 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21241 insn flags on this non-insn.
21242 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21243 as scheduling barriers, for pre-change compatibility.
21244 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21245 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21247 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21248 redundant JUMP_TABLE_DATA_P test.
21249 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21250 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21251 (frv_for_each_packet): Likewise.
21252 * config/i386/i386.c (min_insn_size): Likewise.
21253 (ix86_avoid_jump_mispredicts): Likewise.
21254 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21255 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21256 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21257 (mips16_insn_length): Robustify.
21258 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21259 (mips16_split_long_branches): Likewise.
21260 * config/pa/pa.c (pa_combine_instructions): Likewise.
21261 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21262 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21263 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21264 as contributing to pool range lengths.
21265 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21266 Remove redundant JUMP_TABLE_DATA_P test.
21267 (sh_loop_align): Likewise.
21268 (split_branches): Likewise.
21269 (sh_insn_length_adjustment): Likewise.
21270 * config/spu/spu.c (get_branch_target): Likewise.
21272 2013-03-29 Jan Hubicka <jh@suse.cz>
21274 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21275 gcov streaming; stream hot bb threshold to ltrans.
21276 * predict.c (get_hot_bb_threshold): Break out from ....
21277 (maybe_hot_count_p): ... here.
21278 (set_hot_bb_threshold): New function.
21279 * lto-section-in.c (lto_section_name): Add profile.
21280 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21281 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21282 and data-streamer.h
21283 (histogram_entry): New structure.
21284 (histogram, histogram_pool): New global vars.
21285 (histogram_hash): New structure.
21286 (histogram_hash::hash): New method.
21287 (histogram_hash::equal): Likewise.
21288 (account_time_size): New function.
21289 (cmp_counts): New function.
21290 (dump_histogram): New function.
21291 (ipa_profile_generate_summary): New function.
21292 (ipa_profile_write_summary): New function.
21293 (ipa_profile_read_summary): New function.
21294 (ipa_profile): Decide on threshold.
21295 (pass_ipa_profile): Add ipa_profile_write_summary and
21296 ipa_profile_read_summary.
21297 * Makefile.in (ipa.o): Update dependencies.
21298 * lto-streamer.h (LTO_section_ipa_profile): New section.
21300 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21302 * tree.h (VAR_P): New.
21304 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21307 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21309 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21311 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21312 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21313 (control_flow_insn_p): Likewise.
21314 * cfgrtl.c (duplicate_insn_chain): Likewise.
21315 * final.c (get_attr_length_1): Likewise.
21316 (shorten_branches): Likewise.
21317 (final_scan_insn): Likewise.
21318 * function.c (instantiate_virtual_regs): Likewise.
21319 * gcse.c (insert_insn_end_basic_block): Likewise.
21320 * ira-costs.c (scan_one_insn): Likewise.
21321 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21322 * lra.c (check_rtl): Likewise.
21323 * reload1.c (elimination_costs_in_insn): Likewise.
21324 * reorg.c (follow_jumps): Likewise.
21326 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21327 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21328 (thumb_far_jump_used_p): Likewise.
21329 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21330 (workaround_speculation): Likewise.
21331 (add_sched_insns_for_speculation): Likewise.
21332 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21333 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21334 (frv_for_each_packet): Likewise.
21335 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21336 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21337 (final_emit_insn_group_barriers): Likewise.
21338 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21339 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21340 (mips16_insn_length): Likewise.
21341 * config/pa/pa.c (pa_reorg): Likewise.
21342 (pa_combine_instructions): Likewise.
21343 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21344 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21345 (sh_reorg): Likewise.
21346 (split_branches): Likewise.
21347 * config/spu/spu.c (get_branch_target): Likewise.
21349 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21352 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21354 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21355 Fix declaration name.
21357 2013-03-28 Lawrence Crowl <crowl@google.com>
21359 * graphds.h (struct graph.indicies): Remove unused.
21360 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21361 (SCOP_ORIGINAL_PDDRS): Remove unused.
21362 * sese.h (extern insert_loop_close_phis): Removed unused.
21363 (extern insert_guard_phis): Removed unused.
21364 (extern ivtype_map_elt_info): Removed unused.
21365 (new_ivtype_map_elt): Removed unused.
21366 * sese.c (ivtype_map_elt_info): Removed unused.
21368 2013-03-28 Lawrence Crowl <crowl@google.com>
21370 * Makefile.in: Add several missing include dependences.
21372 (test-dump.o): New. This object is not added to any executable,
21373 but is present for ad-hoc testing.
21375 (debug (const bitmap_head_def &)): New.
21376 (debug (const bitmap_head_def *)): New.
21378 (extern debug (const bitmap_head_def &)): New.
21379 (extern debug (const bitmap_head_def *)): New.
21381 (debug (edge_def &)): New.
21382 (debug (edge_def *)): New.
21384 (debug (basic_block_def &)): New.
21385 (debug (basic_block_def *)): New.
21387 (dump_node (const_tree, int, FILE *)): Correct source file.
21389 (debug (die_struct &)): New.
21390 (debug (die_struct *)): New.
21392 (extern debug (die_struct &)): New.
21393 (extern debug (die_struct *)): New.
21394 * gimple-pretty-print.c
21395 (debug (gimple_statement_d &)): New.
21396 (debug (gimple_statement_d *)): New.
21397 * gimple-pretty-print.h
21398 (extern debug (gimple_statement_d &)): New.
21399 (extern debug (gimple_statement_d *)): New.
21401 (debug (ira_allocno_copy &)): New.
21402 (debug (ira_allocno_copy *)): New.
21403 (debug (ira_allocno &)): New.
21404 (debug (ira_allocno *)): New.
21406 (extern debug (ira_allocno_copy &)): New.
21407 (extern debug (ira_allocno_copy *)): New.
21408 (extern debug (ira_allocno &)): New.
21409 (extern debug (ira_allocno *)): New.
21411 (debug (live_range &)): New.
21412 (debug (live_range *)): New.
21414 (debug (lra_live_range &)): New.
21415 (debug (lra_live_range *)): New.
21417 (debug (lra_live_range &)): New.
21418 (debug (lra_live_range *)): New.
21420 (debug (omega_pb_d &)): New.
21421 (debug (omega_pb_d *)): New.
21423 (extern debug (omega_pb_d &)): New.
21424 (extern debug (omega_pb_d *)): New.
21426 (debug (const rtx_def &)): New.
21427 (debug (const rtx_def *)): New.
21429 (debug_tree (tree): Move within file.
21430 (debug_raw (const tree_node &)): New.
21431 (debug_raw (const tree_node *)): New.
21432 (dump_tree_via_hooks (const tree_node *, int)): New.
21433 (debug (const tree_node &)): New.
21434 (debug (const tree_node *)): New.
21435 (debug_verbose (const tree_node &)): New.
21436 (debug_verbose (const tree_node *)): New.
21437 (debug_head (const tree_node &)): New.
21438 (debug_head (const tree_node *)): New.
21439 (debug_body (const tree_node &)): New.
21440 (debug_body (const tree_node *)): New.
21441 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21442 (debug (vec<tree, va_gc> &)): New.
21443 (debug (vec<tree, va_gc> *)): New.
21445 (extern debug (const rtx_def &)): New.
21446 (extern debug (const rtx_def *)): New.
21448 (debug_raw (simple_bitmap_def &)): New.
21449 (debug_raw (simple_bitmap_def *)): New.
21450 (debug (simple_bitmap_def &)): New.
21451 (debug (simple_bitmap_def *)): New.
21453 (extern debug (simple_bitmap_def &)): New.
21454 (extern debug (simple_bitmap_def *)): New.
21455 (extern debug_raw (simple_bitmap_def &)): New.
21456 (extern debug_raw (simple_bitmap_def *)): New.
21458 (debug (vinsn_def &)): New.
21459 (debug (vinsn_def *)): New.
21460 (debug_verbose (vinsn_def &)): New.
21461 (debug_verbose (vinsn_def *)): New.
21462 (debug (expr_def &)): New.
21463 (debug (expr_def *)): New.
21464 (debug_verbose (expr_def &)): New.
21465 (debug_verbose (expr_def *)): New.
21466 (debug (vec<rtx> &)): New.
21467 (debug (vec<rtx> *)): New.
21469 (extern debug (vinsn_def &)): New.
21470 (extern debug (vinsn_def *)): New.
21471 (extern debug_verbose (vinsn_def &)): New.
21472 (extern debug_verbose (vinsn_def *)): New.
21473 (extern debug (expr_def &)): New.
21474 (extern debug (expr_def *)): New.
21475 (extern debug_verbose (expr_def &)): New.
21476 (extern debug_verbose (expr_def *)): New.
21477 (extern debug (vec<rtx> &)): New.
21478 (extern debug (vec<rtx> *)): New.
21480 (_list_iter_cond_expr): Make inline instead of static.
21482 (debug (sreal &)): New.
21483 (debug (sreal *)): New.
21485 (extern debug (sreal &)): New.
21486 (extern debug (sreal *)): New.
21488 (extern debug_raw (const tree_node &)): New.
21489 (extern debug_raw (const tree_node *)): New.
21490 (extern debug (const tree_node &)): New.
21491 (extern debug (const tree_node *)): New.
21492 (extern debug_verbose (const tree_node &)): New.
21493 (extern debug_verbose (const tree_node *)): New.
21494 (extern debug_head (const tree_node &)): New.
21495 (extern debug_head (const tree_node *)): New.
21496 (extern debug_body (const tree_node &)): New.
21497 (extern debug_body (const tree_node *)): New.
21498 (extern debug (vec<tree, va_gc> &)): New.
21499 (extern debug (vec<tree, va_gc> *)): New.
21501 (debug (struct loop &)): New.
21502 (debug (struct loop *)): New.
21503 (debug_verbose (struct loop &)): New.
21504 (debug_verbose (struct loop *)): New.
21505 * tree-dump.c: Add header dependence.
21507 (extern debug (struct loop &)): New.
21508 (extern debug (struct loop *)): New.
21509 (extern debug_verbose (struct loop &)): New.
21510 (extern debug_verbose (struct loop *)): New.
21512 (debug (data_reference &)): New.
21513 (debug (data_reference *)): New.
21514 (debug (vec<data_reference_p> &)): New.
21515 (debug (vec<data_reference_p> *)): New.
21516 (debug (vec<ddr_p> &)): New.
21517 (debug (vec<ddr_p> *)): New.
21519 (extern debug (data_reference &)): New.
21520 (extern debug (data_reference *)): New.
21521 (extern debug (vec<data_reference_p> &)): New.
21522 (extern debug (vec<data_reference_p> *)): New.
21523 (extern debug (vec<ddr_p> &)): New.
21524 (extern debug (vec<ddr_p> *)): New.
21526 (debug (pt_solution &)): New.
21527 (debug (pt_solution *)): New.
21529 (extern debug (pt_solution &)): New.
21530 (extern debug (pt_solution *)): New.
21532 (debug (_var_map &)): New.
21533 (debug (_var_map *)): New.
21534 (debug (tree_live_info_d &)): New.
21535 (debug (tree_live_info_d *)): New.
21537 (extern debug (_var_map &)): New.
21538 (extern debug (_var_map *)): New.
21539 (extern debug (tree_live_info_d &)): New.
21540 (extern debug (tree_live_info_d *)): New.
21542 2013-03-28 Jan Hubicka <jh@suse.cz>
21544 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21546 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21548 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21549 record only when desired or required.
21551 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21553 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21554 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21556 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21558 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21559 (*andsi3_compare0_uxtw): New pattern.
21560 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21561 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21563 2013-03-28 Jan Hubicka <jh@suse.cz>
21565 * data-streamer-in.c (streamer_read_gcov_count): New function.
21566 * gimple-streamer-out.c: Include value-prof.h.
21567 (output_gimple_stmt): Output histogram.
21568 (output_bb): Use streamer_write_gcov_count.
21569 * value-prof.c: Include data-streamer.h
21570 (dump_histogram_value): Add HIST_TYPE_MAX.
21571 (stream_out_histogram_value): New function.
21572 (stream_in_histogram_value): New function.
21573 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21574 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21575 * data-streamer-out.c (streamer_write_gcov_count): New function.
21576 (streamer_write_gcov_count_stream): New function.
21577 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21578 (lto_output_node): Likewise.
21579 (input_node, input_edge): Likewise.
21580 * lto-streamer-out.c (output_cfg): Update streaming.
21581 * lto-streamer-in.c (input_cfg): Likewise.
21582 * data-streamer.h (streamer_write_gcov_count,
21583 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21584 * gimple-streamer-in.c: Include value-prof.h
21585 (input_gimple_stmt): Input histograms.
21586 (input_bb): Update profile streaming.
21588 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21590 * genmodes.c (emit_max_int): New function.
21591 (emit_insn_modes_h): Added call to emit_max_function.
21592 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21594 * machmode.def: Fixed comment.
21596 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21598 * combine.c (try_combine): Removed useless assert.
21599 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21601 2013-03-28 Marek Polacek <polacek@redhat.com>
21602 Richard Biener <rguenther@suse.de>
21604 PR tree-optimization/56695
21605 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21606 build signed result of a vector comparison.
21607 * tree-cfg.c (verify_gimple_comparison): Check that a result
21608 of a vector comparison has signed type.
21610 2013-03-28 Richard Biener <rguenther@suse.de>
21612 PR tree-optimization/37021
21613 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21614 do not restrict gaps between groups.
21615 * tree-vect-stmts.c (vectorizable_load): Properly account for
21616 a gap between groups.
21618 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21620 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21621 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21624 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21626 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21627 * doc/extend.texi (Named Address Spaces): Ditto.
21628 (Variable Attributes): Ditto.
21630 2013-03-27 Kai Tietz <ktietz@redhat.com>
21632 * config.build: Add support for cygwin x64 target.
21633 * config.gcc: Likewise.
21634 * config.host: Likewise.
21635 * configure.ac: Likewise
21636 * configure: Regenerated.
21638 2013-03-27 Kai Tietz <ktietz@redhat.com>
21640 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21641 * config/i386/t-cygwin-w64: New file.
21642 * config/i386/cygwin-w64.h: New file.
21643 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21644 and add support for x64-cygwin target.
21645 (CPP_SPEC): Likewise.
21646 (CXX_WRAP_SPEC_LIST): Undefine before define.
21647 (LIBGCJ_SONAME): Use 15 as version.
21649 2013-03-27 Richard Biener <rguenther@suse.de>
21651 PR tree-optimization/56716
21652 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21653 dumping for ref nodes.
21655 2013-03-27 Martin Jambor <mjambor@suse.cz>
21657 PR tree-optimization/55334
21658 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21659 restricted pointers to arrays.
21661 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21663 * Makefile.in (.SUFFIXES): Add .cc.
21664 (.c.o): Apply same recipe for implicit rule .cc.o.
21666 2013-03-27 Richard Biener <rguenther@suse.de>
21668 PR tree-optimization/37021
21669 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21670 REALPART/IMAGPART_EXPRs around the supported refs.
21671 * tree-ssa-structalias.c (find_func_aliases): Assume that
21672 floating-point values are not used to transfer pointers.
21674 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21676 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21677 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21678 * doc/tm.texi: Regenerate.
21679 * targhooks.h (default_has_ifunc_p): New.
21680 * targhooks.c (default_has_ifunc_p): Ditto.
21681 * config/linux-protos.h: New file.
21682 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21683 hook for linux which disables support of indirect functions in android.
21684 * config/linux-android.c: New file.
21685 * config/t-linux-android.c: Ditto.
21686 * config.gcc: Added new object file linux-android.o.
21687 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21688 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21689 * varasm.c (do_assemble_alias): Likewise.
21690 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21691 doesn't support indirect functions.
21692 * configure: Regenerate.
21694 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21697 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21698 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21701 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21703 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21705 2013-03-27 Terry Guo <terry.guo@arm.com>
21707 * config/arm/arm-cores.def: Added core cortex-r7.
21708 * config/arm/arm-tune.md: Regenerated.
21709 * config/arm/arm-tables.opt: Regenerated.
21710 * doc/invoke.texi: Added entry for core cortex-r7.
21712 2013-03-27 Walter Lee <walt@tilera.com>
21714 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21715 double-decrement of next_scratch_regno.
21717 2013-03-27 Walter Lee <walt@tilera.com>
21719 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21721 (insn_v1mulus): Ditto.
21722 (insn_v2muls): Ditto.
21724 2013-03-27 Walter Lee <walt@tilera.com>
21726 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21727 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21729 2013-03-27 Walter Lee <walt@tilera.com>
21731 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21732 (*sibcall_value): Ditto.
21734 2013-03-27 Walter Lee <walt@tilera.com>
21736 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21737 (insn_mnz_v8qi): ... this ...
21738 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21740 (insn_v<n>mnz): Replaced by ...
21741 (insn_v1mnz): ... this ...
21742 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21744 (insn_mz_<mode>): Replaced by ...
21745 (insn_mz_v8qi): ... this ...
21746 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21748 (insn_v<n>mz): Replaced by ...
21749 (insn_v1mz): ... this ...
21750 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21753 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21755 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21757 2013-03-26 Roland McGrath <mcgrathr@google.com>
21759 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21760 than fprintf with a non-constant, non-format string.
21762 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21764 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21765 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21766 operand 0 predicate.
21767 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21768 attribute. Use general_x64nomem_operand as operand 1 predicate.
21769 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21770 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21771 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21772 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21773 (mov<mode>_insv_1): Remove expander. Merge insn with
21774 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21775 Use general_x64nomem_operand as operand 1 predicate.
21776 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21777 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21778 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21779 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21780 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21781 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21782 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21783 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21784 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21785 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21786 (general_x64nomem_operand): Ditto.
21788 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21790 * config/rtems.opt: Add -pthread option.
21792 2013-03-26 Richard Biener <rguenther@suse.de>
21794 * alias.c (find_base_term): Avoid redundant and not used recursion.
21795 (base_alias_check): Get the initial base term from the caller.
21796 (true_dependence_1): Compute and pass base terms to base_alias_check.
21797 (write_dependence_p): Likewise.
21798 (may_alias_p): Likewise.
21800 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21802 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21803 PC-relative load in SI modes and above only.
21805 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21807 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21808 * config/arm/iwmmxt.md (WCGR0): Update.
21809 (WCGR1, WCGR2, WCGR3): Likewise.
21811 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21813 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21814 Use x64 and nox64 isa attributes.
21816 2013-03-26 Richard Biener <rguenther@suse.de>
21818 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21819 alignment computations and rely on get_object_alignment_1
21820 for the !TYPE_P case.
21821 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21823 2013-03-26 Walter Lee <walt@tilera.com>
21825 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21826 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21828 2013-03-25 Jeff Law <law@redhat.com>
21830 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21831 check for INTEGRAL_TYPE_P that was missing due to checking in
21832 wrong version of prior patch.
21834 2013-03-25 Walter Lee <walt@tilera.com>
21836 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21837 TILEGX_INSN_SHUFFLEBYTES1.
21838 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21840 (tilegx_builtins): Ditto.
21841 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21843 2013-03-25 Walter Lee <walt@tilera.com>
21845 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21846 (floatunssisf2): New pattern.
21847 (floatsidf2): New pattern.
21848 (floatunssidf2): New pattern.
21850 2013-03-25 Walter Lee <walt@tilera.com>
21852 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21853 tests for constraint J, K, N, P.
21855 2013-03-25 Walter Lee <walt@tilera.com>
21857 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21858 Use indirect/pcrel encoding.
21859 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21862 2013-03-25 Steve Ellcey <sellcey@mips.com>
21864 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21865 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21866 * config/mips/mips.c (mips_option_override): Set IMADD default.
21867 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21868 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21869 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21870 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21871 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21873 2013-03-25 Jeff Law <law@redhat.com>
21875 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21876 slightly to avoid creating and folding useless trees. Simplify
21877 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21879 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21881 * config/i386/i386.md (*zero_extendsidi2): Merge with
21882 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21883 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21884 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21885 targets. Clarify comment.
21887 2013-03-25 Martin Jambor <mjambor@suse.cz>
21889 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21890 pass-through jump functions differently.
21891 (ipa_read_jump_function): Likewise. Also use setter functions to set
21894 2013-03-25 Martin Jambor <mjambor@suse.cz>
21896 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21897 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21899 (ipa_get_indirect_edge_target): New function.
21900 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21901 ipa_get_indirect_edge_target. Update all callers.
21902 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21903 ipa_get_indirect_edge_target_1 instead of calling
21904 ipa_get_indirect_edge_target.
21905 (create_specialized_node): Pass aggvlas to
21906 ipcp_discover_new_direct_edges.
21908 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21910 * config/arm/arm.md (f_sels, f_seld): New types.
21911 (*cmov<mode>): New pattern.
21912 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21915 2013-03-25 Kai Tietz <ktietz@redhat.com>
21917 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21918 POSIX-printf for mingw-hosted builds.
21920 2013-03-25 Richard Biener <rguenther@suse.de>
21922 PR middle-end/56694
21923 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21924 must-not-throw stmt location.
21926 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21928 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21929 Emit load-acquire versions when acq is true.
21930 (arm_emit_store_exclusive): Add rel parameter.
21931 Emit store-release versions when rel is true.
21932 (arm_split_compare_and_swap): Use acquire-release instructions
21934 of barriers when appropriate.
21935 (arm_split_atomic_op): Likewise.
21936 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21937 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21938 (VUNSPEC_SLX): Likewise.
21939 (VUNSPEC_LDA): Likewise.
21940 (VUNSPEC_STL): Likewise.
21941 * config/arm/sync.md (atomic_load<mode>): New pattern.
21942 (atomic_store<mode>): Likewise.
21943 (arm_load_acquire_exclusive<mode>): Likewise.
21944 (arm_load_acquire_exclusivesi): Likewise.
21945 (arm_load_acquire_exclusivedi): Likewise.
21946 (arm_store_release_exclusive<mode>): Likewise.
21948 2013-03-25 Catherine Moore <clm@codesourcery.com>
21950 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21951 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21952 * config/mip/predicates.md (lwsp_swsp_operand,
21953 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21954 sb16_operand, db4_operand, db7_operand, ib3_operand,
21955 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21956 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21957 andi16_operand): New predicates.
21958 * config/mips/mips.md (compression): New attribute.
21959 (enabled): New attribute.
21960 (length): Consider compression in computing length.
21961 (shift_compression): New code attribute.
21962 (*add<mode>3): New operands. Record compression.
21963 (sub<mode>3): Likewise.
21964 (one_cmpl<mode>2): Likewise.
21965 (*and<mode>3): Likewise.
21966 (*ior<mode>3): Likewise.
21967 (unnamed pattern for xor): Likewise.
21968 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21969 (*<optab><mode>3): Likewise.
21970 (*mov<mode>_internal: Likewise.
21971 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21972 (mips_unsigned_immediate_p): New.
21973 (umips_lwsp_swsp_address_p): New.
21974 (m16_based_address_p): New.
21975 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21976 (mips_unsigned_immediate_p): New prototype.
21977 (lwsp_swsp_address_p): New prototype.
21978 (m16_based_address_p): New prototype.
21979 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21980 (mips_signed_immediate_p): New function.
21981 (m16_based_address_p): New function.
21982 (lwsp_swsp_address_p): New function.
21983 (mips_print_operand_punctuation): Recognize short delay slot insns
21984 for microMIPS.add<mode>3"
21986 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21989 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21990 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21992 2013-03-25 Richard Biener <rguenther@suse.de>
21994 PR tree-optimization/56689
21995 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21998 2013-03-25 Richard Biener <rguenther@suse.de>
22000 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22002 (memory_references): Likewise.
22003 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22004 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22005 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22006 (gather_mem_refs_in_loops): Fold into ...
22007 (analyze_memory_references): ... this. Move initialization
22008 to tree_ssa_lim_initialize.
22009 (fill_always_executed_in): Rename to ...
22010 (fill_always_executed_in_1): ... this.
22011 (fill_always_executed_in): Move contains_call computation to
22012 this new function from ...
22013 (tree_ssa_lim_initialize): ... here.
22014 (tree_ssa_lim): Call fill_always_executed_in.
22016 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22018 * postreload.c (reload_combine): Fix code detecting returns.
22020 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22022 * function.c (emit_use_return_register_into_block): On cc0 targets,
22023 do not emit the sequence between cc0 setter and user.
22025 2013-03-25 Kai Tietz <ktietz@redhat.com>
22027 * config/i386/predicates.md (local_symbolic_operand): Interpret
22028 dll-imported symbols as none-local.
22030 2013-03-25 Richard Biener <rguenther@suse.de>
22032 * tree-ssa-loop-im.c (struct depend): Remove.
22033 (struct lim_aux_data): Make depends a vec of gimples.
22034 (free_lim_aux_data): Adjust.
22035 (add_dependency): Likewise.
22036 (set_level): Likewise.
22038 2013-03-25 Richard Biener <rguenther@suse.de>
22040 PR middle-end/56434
22041 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22042 the pointer returned by calls with ECF_MALLOC set.
22044 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22046 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22048 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22050 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22051 using MMXMODE mode iterator.
22052 (*move<mode>_internal): Merge with *movv2sf_internal and
22053 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22055 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22057 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22058 (record_last_mem_set_info): Likewise.
22060 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22061 of XNEWVEC followed by memset.
22062 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22064 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22066 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22067 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22068 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22069 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22070 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22071 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22072 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22073 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22074 BARRIER_P instead of GET_CODE.
22076 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22078 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22079 inaccuracy in the probing code.
22081 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22082 (ctrapdi4): Likewise.
22084 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22086 * calls.c (expand_call): Add missing guard to code handling return
22087 of non-BLKmode structures in MSB.
22088 * function.c (expand_function_end): Likewise.
22090 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22092 * combine.c (try_combine): Adjust comment. Do not add the set of
22093 insn #0 if the destination indirectly is set or dies in insn #2.
22094 Tidy up code to distribute a new note.
22096 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22098 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22099 also for alternatives 16 and 17.
22101 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22103 * config/i386/sse.md (*mov<mode>_internal): Merge with
22104 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22105 Emit insn template depending on type attribute. Use
22106 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22107 movd instead of movq mnemonic for interunit moves. Rewrite mode
22108 attribute calculation. Remove unit attribute calculation.
22109 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22110 Set prefix_data16 attribute for DImode ssemov types.
22111 Use Ym instead of y for SSE-MMX conversion alternatives.
22112 Reorder operand constraints.
22114 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22116 * df.h (df_insn_delete): Adjust prototype.
22117 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22118 and let it decide whether mark the basic block dirty.
22119 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22120 * df-scan.c (df_insn_info_delete): New helper function, split
22121 off from df_insn_delete.
22122 (df_scan_free_bb_info): Use it.
22123 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22125 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22126 that the insn is actually an insn and it has a non-NULL basic block.
22127 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22129 2013-03-22 Richard Biener <rguenther@suse.de>
22131 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22133 (mem_ref_alloc): Do not allocate them.
22134 (refs_independent_p): Do not query or maintain a cache.
22136 2013-03-22 Richard Biener <rguenther@suse.de>
22138 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22139 (gather_mem_refs_in_loops): Do not compute it.
22140 (analyze_memory_references): Do not allocate it.
22141 (tree_ssa_lim_finalize): Do not free it.
22142 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22144 2013-03-22 Richard Biener <rguenther@suse.de>
22146 * is-a.h (as_a): Use gcc_checking_assert.
22148 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22150 * config/aarch64/aarch64.c (aarch64_print_operand): New
22151 format specifier for printing a constant in hex.
22152 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22153 format specifier for printing second operand.
22155 2013-03-22 Richard Biener <rguenther@suse.de>
22157 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22159 (gather_mem_refs_in_loops): Perform store accumulation here.
22160 (create_vop_ref_mapping_loop): Remove.
22161 (create_vop_ref_mapping): Likewise.
22162 (analyze_memory_references): Initialize refs_stored_in_loop.
22163 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22164 (record_indep_loop): Remove.
22165 (record_dep_loop): New function.
22166 (ref_indep_loop_p_1): Adjust to only walk over references
22167 in the loop, not its subloops.
22168 (ref_indep_loop_p): Rename to ...
22169 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22170 maintaining a more fine-grained cache.
22171 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22172 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22174 2013-03-22 Richard Biener <rguenther@suse.de>
22176 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22177 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22178 aggregate mem_ref_loc.
22179 (free_mem_ref_locs): Inline into ...
22180 (memref_free): ... this and adjust.
22181 (mem_ref_alloc): Adjust.
22182 (mem_ref_locs_alloc): Remove.
22183 (record_mem_ref_loc): Adjust.
22184 (get_all_locs_in_loop): Rewrite into ...
22185 (for_all_locs_in_loop): ... this iterator.
22186 (rewrite_mem_ref_loc): New functor.
22187 (rewrite_mem_refs): Use for_all_locs_in_loop.
22188 (sm_set_flag_if_changed): New functor.
22189 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22190 (ref_always_accessed): New functor.
22191 (ref_always_accessed_p): Use for_all_locs_in_loop.
22193 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22195 * tree-pass.h (PROP_gimple_lvec): New.
22196 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22197 (init_optimization_passes): Move pass_lower_vector.
22198 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22200 (pass_lower_vector): Provide PROP_gimple_lvec.
22201 (pass_lower_vector_ssa): Likewise.
22202 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22204 2013-03-21 Mark Wielaard <mjw@redhat.com>
22206 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22208 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22210 * config/i386/i386.md (*movdi_internal): Disparage slightly
22211 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22212 conversion alternatives.
22214 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22216 PR middle-end/48087
22217 * diagnostic.def (DK_WERROR): New kind.
22218 * diagnostic.h (werrorcount): Define.
22219 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22220 promoted to DK_ERROR, increment DK_WERROR counter instead of
22222 * toplev.c (toplev_main): Call print_ignored_options even if
22223 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22224 even if just werrorcount is non-zero.
22227 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22229 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22230 (string_cst_pool_decl): New function.
22231 (optimize_one_addr_into_implicit_ptr): New function.
22232 (resolve_addr_in_expr): Optimize DWARF location expression
22233 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22234 which doesn't live in memory, but has DW_AT_location or
22235 DW_AT_const_value, or refers to a string literal, into
22236 DW_OP_GNU_implicit_pointer.
22237 (optimize_location_into_implicit_ptr): New function.
22238 (resolve_addr): If removing DW_AT_location of a variable because
22239 it was DW_OP_addr of address of the variable, but the variable doesn't
22240 live in memory, try to emit const value attribute for the initializer.
22242 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22244 * tree.h (VECTOR_TYPE_P): New macro.
22245 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22246 TYPE_MODE): Use it.
22247 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22248 VEC_COND_EXPR cannot be lvalues.
22249 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22251 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22253 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22254 Restrict the transformation to equal modes.
22256 2013-03-21 Richard Biener <rguenther@suse.de>
22258 PR tree-optimization/39326
22259 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22260 (MEM_ANALYZABLE): Adjust.
22261 (record_mem_ref_loc): Move bitmap ops ...
22262 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22263 unanalyzable refs, do not record locations for it.
22264 (analyze_memory_references): Allocate ref zero as shared
22266 (refs_independent_p): Do not test for unanalyzed mems here.
22267 (ref_indep_loop_p_1): Special-case disambiguation against
22268 the unanalyzed ref.
22269 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22271 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22273 * config/arm/arm-protos.h (tune_params): Add
22274 prefer_neon_for_64bits field.
22275 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22276 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22277 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22278 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22279 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22280 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22281 (arm_option_override): Handle -mneon-for-64bits new option.
22282 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22283 (prefer_neon_for_64bits): Declare new variable.
22284 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22285 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22286 (arch_enabled): Handle new arch types. Remove support for onlya8
22288 (one_cmpldi2): Use new arch names.
22289 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22290 * config/arm/arm.opt (mneon-for-64bits): Add option.
22291 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22292 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22293 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22295 * doc/invoke.texi (-mneon-for-64bits): Document.
22297 2013-03-21 Richard Biener <rguenther@suse.de>
22299 PR tree-optimization/39326
22300 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22301 (sort_bbs_in_loop_postorder_cmp): New function.
22302 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22304 2013-03-21 Richard Biener <rguenther@suse.de>
22306 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22307 (vect_insert_into_interleaving_chain): Likewise.
22308 (vect_drs_dependent_in_basic_block): Inline ...
22309 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22311 (vect_analyze_data_ref_dependence): ... here. Simplify.
22312 (vect_check_interleaving): Simplify.
22313 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22314 (vect_slp_analyze_data_ref_dependences): ... this new function.
22315 (dr_group_sort_cmp): New function.
22316 (vect_analyze_data_ref_accesses): Compute data-reference groups
22317 here instead of in vect_analyze_data_ref_dependence. Use
22318 a more efficient algorithm.
22319 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22320 vect_slp_analyze_data_ref_dependences. Call
22321 vect_analyze_data_ref_accesses earlier.
22322 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22323 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22324 (vect_slp_analyze_data_ref_dependences): New prototype.
22326 2013-03-21 Richard Biener <rguenther@suse.de>
22328 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22329 ref is stored in the loop.
22330 (find_refs_for_sm): Walk only over all stores.
22331 (store_motion_loop): Allocate from lim_bitmap_obstack.
22332 (store_motion): Likewise.
22334 2013-03-21 Richard Biener <rguenther@suse.de>
22336 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22337 Update virtual SSA form.
22339 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22341 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22342 * configure: Regenerate.
22343 * config.in: Regenerate.
22344 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22345 if !HAVE_LD_EH_FRAME_CIEV3.
22347 2013-03-21 Richard Biener <rguenther@suse.de>
22349 * tree-cfg.c (verify_expr_no_block): New function.
22350 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22351 nor DECL_VALUE_EXPR have locations with associated blocks.
22352 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22353 (clear_unused_block_pointer): Remove code dealing with
22354 blocks in DECL_DEBUG_EXPR locations.
22356 2013-03-21 Richard Biener <rguenther@suse.de>
22358 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22359 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22360 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22361 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22362 instead of DECL_DEBUG_EXPR_IS_FROM.
22363 * gimplify.c (gimplify_modify_expr): Likewise.
22364 * tree-cfg.c (verify_expr_location_1): Likewise.
22365 * tree-complex.c (create_one_component_var): Likewise.
22366 * tree-sra.c (create_access_replacement): Likewise.
22367 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22368 (clear_unused_block_pointer): Likewise.
22369 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22370 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22371 * var-tracking.c (var_debug_decl): Likewise.
22372 (track_expr_p): Likewise.
22373 * tree-inline.c (add_local_variables): Likewise. Set
22374 DECL_HAS_DEBUG_EXPR_P after copying it.
22375 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22376 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22378 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22381 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22382 * configure: Regenerate.
22383 * config.in: Regenerate.
22384 * config/i386/i386.md (*movdf_internal): Use
22385 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22386 movd instead of movq mnemonic for interunit moves.
22387 (*movdi_internal): Ditto.
22389 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22391 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22392 (abd<mode>_3): New pattern.
22393 (aba<mode>_3): New pattern.
22394 (fabd<mode>_3): New pattern.
22396 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22398 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22399 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22400 occurrence of REGISTER_PREFIX as its empty string.
22402 2013-03-20 Jeff Law <law@redhat.com>
22404 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22405 addititional equivalences for equality comparisons between an SSA_NAME
22406 and a constant where the SSA_NAME was set from a widening conversion.
22408 2013-03-20 Walter Lee <walt@tilera.com>
22410 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22412 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22414 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22415 depending on type attribute.
22416 (*movti_internal): Ditto.
22417 (*movtf_internal): Ditto.
22418 (*movxf_internal): Ditto.
22419 (*movdf_internal): Ditto.
22420 (*movsf_internal): Ditto.
22422 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22424 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22425 maybe_vex for sselog1 and ssemov types.
22426 (*movdi_internal): Reorder operand constraints.
22427 (*movsi_internal): Ditto. Set prefix attribute to
22428 maybe_vex for sselog1 and ssemov types.
22429 (*movtf_internal): Set prefix attribute to maybe_vex
22430 for sselog1 and ssemov types.
22431 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22432 DImode ssemov types. Reorder operand constraints.
22433 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22434 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22435 attribute for SImode ssemov types. Reorder operand constraints.
22437 2013-03-20 Martin Jambor <mjambor@suse.cz>
22439 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22440 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22442 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22444 * config/rs6000/predicates.md (indexed_address, update_address_mem
22445 update_indexed_address_mem): New predicates.
22446 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22447 attribute for load/store instructions.
22448 * config/rs6000/dfp.md (movsd_store): Likewise.
22449 (movsd_load): Likewise.
22450 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22451 (unnamed HI->DI extend define_insn): Likewise.
22452 (unnamed SI->DI extend define_insn): Likewise.
22453 (unnamed QI->SI extend define_insn): Likewise.
22454 (unnamed QI->HI extend define_insn): Likewise.
22455 (unnamed HI->SI extend define_insn): Likewise.
22456 (unnamed HI->SI extend define_insn): Likewise.
22457 (extendsfdf2_fpr): Likewise.
22458 (movsi_internal1): Likewise.
22459 (movsi_internal1_single): Likewise.
22460 (movhi_internal): Likewise.
22461 (movqi_internal): Likewise.
22462 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22463 attribute for load/store instructions.
22464 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22466 (mov<mode>_softfloat): Likewise.
22467 (mov<mode>_hardfloat32): Likewise.
22468 (mov<mode>_hardfloat64): Likewise.
22469 (mov<mode>_softfloat64): Likewise.
22470 (movdi_internal32): Likewise.
22471 (movdi_internal64): Likewise.
22472 (probe_stack_<mode>): Likewise.
22474 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22476 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22477 floating point, and decimal floating point to reload iterator.
22479 * config/rs6000/constraints.md (wl constraint): New constraints to
22480 return FLOAT_REGS if certain options are used to reduce the number
22481 of separate patterns that exist in the file.
22482 (wx constraint): Likewise.
22483 (wz constraint): Likewise.
22485 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22486 -mdebug=reg, print wg, wl, wx, and wz constraints.
22487 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22488 Initialize the reload functions for 64-bit binary/decimal floating
22490 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22491 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22492 create the buffer on the stack to overcome not having a 32-bit
22494 (rs6000_emit_move): Likewise.
22495 (rs6000_secondary_memory_needed_rtx): Likewise.
22496 (rs6000_alloc_sdmode_stack_slot): Likewise.
22497 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22498 via xxlxor, just like DFmode 0.0.
22500 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22501 define as 1 if we are running on a power7 or newer.
22502 (enum r6000_reg_class_enum): Add new constraints.
22504 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22505 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22506 with other moves by using conditional constraits (wg). Use LFIWZX
22507 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22508 (movsd splitter): Likewise.
22509 (movsd_hardfloat): Likewise.
22510 (movsd_softfloat): Likewise.
22512 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22513 binary and decimal floating point moves.
22514 (fmove_ok): New attributes to combine binary and decimal floating
22515 point moves, and to combine power6x (mfpgpr) moves along normal
22517 (real_value_to_target): Likewise.
22518 (f32_lr): Likewise.
22519 (f32_lm): Likewise.
22520 (f32_li): Likewise.
22521 (f32_sr): Likewise.
22522 (f32_sm): Likewise.
22523 (f32_si): Likewise.
22524 (movsf): Combine binary and decimal floating point moves. Combine
22525 power6x (mfpgpr) moves with other moves by using conditional
22526 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22527 (mov<mode> for SFmode/SDmode); Likewise.
22528 (SFmode/SDmode splitters): Likewise.
22529 (movsf_hardfloat): Likewise.
22530 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22531 (movsf_softfloat): Likewise.
22532 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22534 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22535 wx and wz constraints.
22537 * config/rs6000/constraints.md (wg constraint): New constraint to
22538 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22540 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22543 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22544 -mdebug=reg, print wg, wl, wx, and wz constraints.
22545 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22546 Initialize the reload functions for 64-bit binary/decimal floating
22548 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22549 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22550 create the buffer on the stack to overcome not having a 32-bit
22552 (rs6000_emit_move): Likewise.
22553 (rs6000_secondary_memory_needed_rtx): Likewise.
22554 (rs6000_alloc_sdmode_stack_slot): Likewise.
22555 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22556 via xxlxor, just like DFmode 0.0.
22558 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22559 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22560 with other moves by using conditional constraits (wg). Use LFIWZX
22561 and STFIWX for loading SDmode on power7.
22562 (movdd splitters): Likewise.
22563 (movdd_hardfloat32): Likewise.
22564 (movdd_softfloat32): Likewise.
22565 (movdd_hardfloat64_mfpgpr): Likewise.
22566 (movdd_hardfloat64): Likewise.
22567 (movdd_softfloat64): Likewise.
22569 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22570 64-bit binary and decimal floating point moves.
22571 (FMOVE64X): Likewise.
22572 (movdf): Combine 64-bit binary and decimal floating point moves.
22573 Combine power6x (mfpgpr) moves with other moves by using
22574 conditional constraits (wg).
22575 (mov<mode> for DFmode/DDmode): Likewise.
22576 (DFmode/DDmode splitters): Likewise.
22577 (movdf_hardfloat32): Likewise.
22578 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22579 (movdf_softfloat32): Likewise.
22580 (movdf_hardfloat64_mfpgpr): Likewise.
22581 (movdf_hardfloat64): Likewise.
22582 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22583 (movdf_softfloat64): Likewise.
22584 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22585 (reload_<mode>_load): Move to later in the file so they aren't in
22586 the middle of the floating point move insns.
22587 (reload_<mode>_store): Likewise.
22589 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22592 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22593 constraint if -mdebug=reg.
22594 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22595 Enable using dd reload support if needed.
22597 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22598 binary and decimal floating point moves in rs6000.md.
22599 (movtd_internal): Likewise.
22601 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22602 decimal floating point moves.
22604 (movtf_internal): Likewise.
22605 (mov<mode>_internal, TDmode/TFmode): Likewise.
22606 (movtf_softfloat): Likewise.
22607 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22609 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22610 movdi_internal64, using wg constraint for move direct operations.
22611 (movdi_internal64): Likewise.
22613 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22614 MODES_TIEABLE_P for selected modes. Print the numerical value of
22615 the various virtual registers. Use GPR/FPR first/last values,
22616 instead of hard coding the register numbers. Print which modes
22617 have reload functions registered.
22618 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22619 settings before/after setting cpu, target and subtarget settings.
22620 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22621 and for secondary reload failures in rs6000_secondary_reload_inner.
22622 (rs6000_secondary_reload_fail): Likewise.
22623 (rs6000_secondary_reload_inner): Likewise.
22625 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22626 macros for first/last GPR and FPR registers.
22627 (LAST_GPR_REGNO): Likewise.
22628 (FIRST_FPR_REGNO): Likewise.
22629 (LAST_FPR_REGNO): Likewise.
22631 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22632 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22633 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22634 (vcond<mode><mode>): Likewise.
22635 (vcondu<mode><mode>): Likewise.
22636 (vector_gtu<mode>): Likewise.
22637 (vector_gte<mode>): Likewise.
22638 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22639 to prevent the compiler from converting DImode operations to TImode.
22640 (ior<mode>3): Likewise.
22641 (and<mode>3): Likewise.
22642 (one_cmpl<mode>2): Likewise.
22643 (nor<mode>3): Likewise.
22644 (andc<mode>3): Likewise.
22646 * config/rs6000/constraints.md (wt constraint): New constraint
22647 that returns VSX_REGS if TImode is allowed in VSX registers.
22649 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22650 constant under VSX.
22652 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22653 similar to TImode, but it is restricted to being in the GPRs.
22655 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22656 TImode to occupy a single VSX register.
22658 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22659 -mvsx-timode for power7/power8.
22660 (power7 cpu): Likewise.
22661 (power8 cpu): Likewise.
22663 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22664 sure that TFmode/TDmode take up two registers if they are ever
22665 allowed in the upper VSX registers.
22666 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22668 (rs6000_init_hard_regno_mode_ok): Likewise.
22669 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22670 constraint. Print if LRA is turned on.
22671 (rs6000_option_override_internal): Give an error if -mvsx-timode
22672 and VSX is not enabled.
22673 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22674 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22675 to reg+offset addressing. Use PTImode when checking offset
22676 addresses for validity.
22677 (reg_offset_addressing_ok_p): Likewise.
22678 (rs6000_legitimate_offset_address_p): Likewise.
22679 (rs6000_legitimize_address): Likewise.
22680 (rs6000_legitimize_reload_address): Likewise.
22681 (rs6000_legitimate_address_p): Likewise.
22682 (rs6000_eliminate_indexed_memrefs): Likewise.
22683 (rs6000_emit_move): Likewise.
22684 (rs6000_secondary_reload): Likewise.
22685 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22686 reloads to fpr registers to continue to use reg+offset addressing,
22687 but 64-bit reloads to altivec registers need reg+reg addressing.
22688 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22689 it. Treat LO_SUM like a PLUS operation.
22690 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22691 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22692 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22693 registers to share a register with a smaller sized type, since VSX
22694 puts scalars in the upper 64-bits.
22695 (print_operand): Add support for PTImode.
22696 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22697 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22698 registers, but don't have arithmetic support.
22699 (rs6000_memory_move_cost): Add test for VSX.
22700 (rs6000_opt_masks): Add -mvsx-timode.
22702 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22705 (VSr): Use wt constraint for TImode.
22706 (VSv): Drop TImode support.
22707 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22708 (vsx_movti_64bit): Likewise.
22709 (vsx_movti_32bit): Likewise.
22710 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22711 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22712 one '?' on the appropriate output constraint. Do not allow TImode
22713 logical operations on 32-bit systems.
22714 (vsx_ior<mode>3): Likewise.
22715 (vsx_xor<mode>3): Likewise.
22716 (vsx_one_cmpl<mode>2): Likewise.
22717 (vsx_nor<mode>3): Likewise.
22718 (vsx_andc<mode>3): Likewise.
22719 (vsx_concat_<mode>): Likewise.
22720 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22722 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22723 OPTION_MASK_VSX_TIMODE.
22724 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22725 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22727 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22728 (TI2 iterator): New iterator for TImode, PTImode.
22729 (wd mode attribute): Add values for vector types.
22730 (movti_string): Replace TI move operations with operations for TImode
22731 and PTImode. Add support for TImode being allowed in VSX registers.
22732 (mov<mode>_string, TImode/PTImode): Likewise.
22733 (movti_ppc64): Likewise.
22734 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22735 (TI mode splitters): Likewise.
22737 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22740 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22742 PR tree-optimization/56355
22743 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22744 Also handle integers with undefined overflow.
22746 2013-03-20 Catherine Moore <clm@codesourcery.com>
22747 Maciej W. Rozycki <macro@codesourcery.com>
22748 Tom de Vries <tom@codesourcery.com>
22749 Nathan Sidwell <nathan@codesourcery.com>
22750 Iain Sandoe <iain@codesourcery.com>
22751 Nathan Froyd <froydnj@codesourcery.com>
22752 Chao-ying Fu <fu@mips.com>
22754 * doc/extend.texi: (micromips, nomicromips, nocompression):
22755 Document new function attributes.
22756 * doc/invoke.texi (minterlink-compressed, mmicromips,
22757 m14k, m14ke, m14kec): Document new options.
22758 (minterlink-mips16): Update documentation.
22759 * doc/md.texi (ZC, ZD): Document new constraints.
22760 * configure.ac (gcc_cv_as_micromips): Check if linker
22761 supports the .set micromips directive.
22762 * configure: Regenerate.
22763 * config.in: Regenerate.
22764 * config/mips/mips-tables.opt: Regenerate.
22765 * config/mips/micromips.md: New file.
22766 * constraints.md (ZC, ZD): New constraints.
22767 * config/mips/predicates.md (movep_src_register): New predicate.
22768 (movep_src_operand): New predicate.
22769 (non_volatile_mem_operand): New predicate.
22770 * config/mips/mips.md (multimem): New type.
22771 (length): Differentiate between 17-bit and 18-bit branch offsets.
22772 (MOVEP1, MOVEP2): New mode iterator.
22773 (mov_<load>l): Use ZC constraint.
22774 (mov_<load>r): Likewise.
22775 (mov_<store>l): Likewise.
22776 (mov_<store>r): Likewise.
22777 (*branch_equality<mode>_inverted): Add microMIPS support.
22778 (*branch_equality<mode>): Likewise.
22779 (*jump_absolute): Likewise.
22780 (indirect_jump_<mode>): Likewise.
22781 (tablejump_<mode>): Likewise.
22782 (<optab>_internal): Likewise.
22783 (sibcall_internal): Likewise.
22784 (sibcall_value_internal): Likewise.
22785 (prefetch): Use constraint ZD.
22786 * config/mips/mips.opt (minterlink-compressed): New option.
22787 (minterlink-mips16): Now an alias for minterlink-compressed.
22788 (mmicromips): New option.
22789 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22790 (compare_and_swap_12): Likewise.
22791 (sync_add<mode>): Likewise.
22792 (sync_<optab>_12): Likewise.
22793 (sync_old_<optab>_12): Likewise.
22794 (sync_new_<optab>_12): Likewise.
22795 (sync_nand_12): Likewise.
22796 (sync_old_nand_12): Likewise.
22797 (sync_new_nand_12): Likewise.
22798 (sync_sub<mode>): Likewise.
22799 (sync_old_add<mode>): Likewise.
22800 (sync_old_sub<mode>): Likewise.
22801 (sync_new_add<mode>): Likewise.
22802 (sync_new_sub<mode>): Likewise.
22803 (sync_<optab><mode>): Likewise.
22804 (sync_old_<optab><mode>): Likewise.
22805 (sync_new_<optab><mode>): Likewise.
22806 (sync_nand<mode>): Likewise.
22807 (sync_old_nand<mode>): Likewise.
22808 (sync_new_nand<mode>): Likewise.
22809 (sync_lock_test_and_set<mode>): Likewise.
22810 (test_and_set_12): Likewise.
22811 (atomic_compare_and_swap<mode>): Likewise.
22812 (atomic_exchange<mode>_llsc): Likewise.
22813 (atomic_fetch_add<mode>_llsc): Likewise.
22814 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22815 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22816 (umips_save_restore_pattern_p): Likewise.
22817 (umips_load_store_pair_p): Likewise.
22818 (umips_output_load_store_pair): Likewise.
22819 (umips_movep_target_p): Likewise.
22820 (umips_12bit_offset_address_p): Likewise.
22821 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22822 (mips_base_mips16): Rename this...
22823 (mips_base_compression_flags): ...to this. Update all uses.
22824 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22825 (mips_mips16_decl_p): Delete.
22826 (mips_nomips16_decl_p): Delete.
22827 (mips_get_compress_on_flags): New function.
22828 (mips_get_compress_off_flags): New function.
22829 (mips_get_compress_mode): New function.
22830 (mips_get_compress_on_name): New function.
22831 (mips_get_compress_off_name): New function.
22832 (mips_insert_attributes): Support multiple compression types.
22833 (mips_merge_decl_attributes): Likewise.
22834 (umips_12bit_offset_address_p): New function.
22835 (mips_start_function_definition): Emit .set micromips directive.
22836 (mips_call_may_need_jalx_p): New function.
22837 (mips_function_ok_for_sibcall): Add microMIPS support.
22838 (mips_print_operand_punctuation): Support short delay slots and
22840 (umips_swm_mask, umips_swm_encoding): New.
22841 (umips_build_save_restore): New function.
22842 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22843 (was_mips16_p): Remove.
22844 (old_compression_mode): New.
22845 (mips_set_compression_mode): New function.
22846 (mips_set_current_function): Add microMIPS support.
22847 (mips_option_override): Likewise.
22848 (umips_save_restore_pattern_p): New function.
22849 (umips_output_save_restore): New function.
22850 (umips_load_store_pair_p_1): New function.
22851 (umips_load_store_pair_p): New function.
22852 (umips_output_load_store_pair_1): New function.
22853 (umips_output_load_store_pair): New function.
22854 (umips_movep_target_p) New function.
22855 (mips_prepare_pch_save): Add microMIPS support.
22856 * config/mips/mips.h (TARGET_COMPRESSION): New.
22857 (TARGET_CPU_CPP_BUILTINS): Update macro
22858 to use new compression flags and to support microMIPS.
22859 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22860 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22861 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22862 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22863 (ASM_SPEC): Support mmicromips and mno-micromips.
22864 (M16STORE_REG_P): New macro.
22865 (MIPS_CALL): Support TARGET_MICROMIPS.
22866 (MICROMIPS_J): New macro.
22867 (mips_base_mips16): Rename this...
22868 (mips_base_compression_flags): ...to this.
22869 (UMIPS_12BIT_OFFSET_P): New macro.
22870 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22871 (MULTILIB_DIRNAMES): Likewise.
22872 2013-03-20 Richard Biener <rguenther@suse.de>
22874 PR tree-optimization/56661
22875 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22876 the result does not have to be distinct.
22878 2013-03-20 Richard Biener <rguenther@suse.de>
22880 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22883 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22884 Steven Bosscher <steven@gcc.gnu.org>
22886 PR rtl-optimization/56605
22887 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22889 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22892 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22893 that require movd instead of movq.
22895 2013-03-20 Richard Biener <rguenther@suse.de>
22897 * tree-ssa-structalias.c (struct variable_info): Add pointer
22898 to the first field of an aggregate with sub-vars. Make
22899 this and the pointer to the next subfield its ID.
22900 (vi_next): New function.
22901 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22902 storedanything_id, integer_id): Increment by one.
22903 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22904 get_call_clobber_vi): Adjust.
22905 (solution_set_expand): Simplify and speedup.
22906 (solution_set_add): Inline into ...
22907 (set_union_with_increment): ... this. Adjust accordingly.
22908 (do_sd_constraint): Likewise.
22909 (do_ds_constraint): Likewise.
22910 (do_complex_constraint): Simplify.
22911 (build_pred_graph): Adjust.
22912 (solve_graph): Likewise. Simplify and speedup.
22913 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22914 get_constraint_for_component_ref, get_constraint_for_1,
22915 first_vi_for_offset, first_or_preceding_vi_for_offset,
22916 create_function_info_for, create_variable_info_for_1,
22917 create_variable_info_for, intra_create_variable_infos): Adjust.
22918 (init_base_vars): Push NULL for ID zero.
22919 (compute_points_to_sets): Adjust.
22921 2013-03-20 Richard Biener <rguenther@suse.de>
22923 * cfgloop.c (verify_loop_structure): Streamline and avoid
22924 ICEing on corrupt loop tree.
22925 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22928 2013-03-20 Richard Biener <rguenther@suse.de>
22930 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22931 check whether an SSA update is needed.
22933 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22935 * config/mips/constraints.md (T): Rename to...
22939 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22940 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22942 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22944 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22945 (*subsi3_carryin_uxtw): Likewise.
22947 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22949 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22950 (*rorsi3_insn_uxtw): Likewise.
22952 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22954 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22955 (*extrsi5_insn_uxtw): Likewise.
22957 2013-03-19 Richard Biener <rguenther@suse.de>
22959 PR tree-optimization/56273
22960 * passes.c (init_optimization_passes): Move second VRP after DOM.
22962 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22964 * config/i386/i386.md (*movti_internal): Merge from
22965 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22966 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22967 nox64 isa attributes.
22969 2013-03-18 Richard Biener <rguenther@suse.de>
22971 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22973 (merge_node_constraints): Likewise.
22974 (build_succ_graph): Likewise.
22975 (valid_graph_edge): Inline into single caller.
22976 (unify_nodes): Likewise. Use bitmap_set_bit return value
22978 (scc_visit): Fix formatting and variable use.
22979 (do_sd_constraint): Use gcc_checking_assert.
22980 (do_ds_constraint): Likewise.
22981 (do_complex_constraint): Likewise.
22982 (condense_visit): Likewise. Cleanup.
22983 (dump_pred_graph): New function.
22984 (perform_var_substitution): Dump the pred-graph before
22985 variable substitution.
22986 (find_equivalent_node): Use gcc_checking_assert.
22987 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22989 2013-03-18 Richard Biener <rguenther@suse.de>
22991 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22992 Remove cond_expr_stmt_list argument and do not gimplify the
22994 (vect_loop_versioning): Adjust.
22995 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22996 Cleanup to use less temporaries.
22997 (vect_create_data_ref_ptr): Cleanup.
22999 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23001 PR tree-optimization/56635
23002 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23003 require types_compatible_p types.
23005 2013-03-18 Nick Clifton <nickc@redhat.com>
23007 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23008 spurious backslash.
23010 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23011 Add missing line to comment describing function.
23013 2013-03-18 Richard Biener <rguenther@suse.de>
23015 PR tree-optimization/56210
23016 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23017 Handle string / character search functions.
23018 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23020 2013-03-18 Richard Biener <rguenther@suse.de>
23022 PR middle-end/56483
23023 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23024 and implement properly.
23025 * gimple.h (gimple_cond_single_var_p): Remove.
23027 2013-03-18 Richard Biener <rguenther@suse.de>
23029 * tree-data-ref.h (find_data_references_in_loop): Declare.
23030 * tree-data-ref.c (get_references_in_stmt): Use a stack
23031 vector pre-allocated in the callers.
23032 (find_data_references_in_stmt): Adjust.
23033 (graphite_find_data_references_in_stmt): Likewise.
23034 (create_rdg_vertices): Likewise.
23035 (find_data_references_in_loop): Export.
23036 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23037 Compute dependences here...
23038 (vect_analyze_data_refs): ...not here. When we encounter
23039 a non-vectorizable data reference in basic-block vectorization
23040 truncate the data reference vector. Do not bother to
23041 fixup data-dependence information for gather loads.
23042 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23043 of data references, as reported.
23045 2013-03-18 Richard Biener <rguenther@suse.de>
23047 PR tree-optimization/3713
23048 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23049 has_constants and expr.
23050 (stmt_has_constants): Properly valueize SSA names when deciding
23051 whether the stmt has constants.
23053 2013-03-18 Richard Biener <rguenther@suse.de>
23055 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23056 whole function when there is nothing to do.
23057 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23058 * tree-vectorizer.c (vectorize_loops): Update virtual and
23059 loop-closed SSA once.
23060 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23062 2013-03-18 Richard Biener <rguenther@suse.de>
23064 PR middle-end/56113
23065 * domwalk.c (bb_postorder): New global static.
23066 (cmp_bb_postorder): New function.
23067 (walk_dominator_tree): Replace scheme imposing an order for
23068 visiting dominator sons by one sorting them at the time they
23069 are pushed on the stack.
23071 2013-03-18 Richard Biener <rguenther@suse.de>
23073 PR tree-optimization/39326
23074 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23075 (struct mem_ref): Replace mem member with ao_ref typed member.
23076 (MEM_ANALYZABLE): Adjust.
23077 (memref_eq): Likewise.
23078 (mem_ref_alloc): Likewise.
23079 (gather_mem_refs_stmt): Likewise.
23080 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23081 (execute_sm_if_changed_flag_set): Adjust.
23082 (execute_sm): Likewise.
23083 (ref_always_accessed_p): Likewise.
23084 (refs_independent_p): Likewise.
23085 (can_sm_ref_p): Likewise.
23087 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23090 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23091 return 1 even for !unsignedp.
23093 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23095 * config/i386/i386.md (isa): Add x64 and nox64.
23096 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23097 (*pushtf): Enable *roF alternative for x64 isa only.
23098 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23099 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23100 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23101 (*movtf_internal): Merge from *movtf_internal_rex64 and
23102 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23103 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23104 nox64 isa attributes.
23105 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23106 nox64 isa attributes.
23107 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23109 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23111 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23113 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23116 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23117 class names. Remove trailing comma after "ALL_REGS".
23119 2013-03-16 Jan Hubicka <jh@suse.cz>
23121 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23122 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23123 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23124 of cgraph_get_create_node.
23125 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23127 2013-03-16 Jason Merrill <jason@redhat.com>
23130 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23131 with DW_AT_default_value.
23133 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23135 * BASE-VER: Set to 4.9.0.
23137 2013-03-14 Andi Kleen <ak@linux.intel.com>
23140 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23141 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23142 Document _x* TSX intrinsics.
23144 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23145 David Holsgrove <david.holsgrove@xilinx.com>
23147 * configure.ac: Add MicroBlaze TLS support detection.
23148 * configure: Regenerate.
23149 * config/microblaze/microblaze-protos.h
23150 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23151 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23152 * config/microblaze/microblaze.c (microblaze_address_type): Add
23153 ADDRESS_TLS and tls_reloc address types.
23154 (microblaze_address_info): Add tls_reloc.
23155 (TARGET_HAVE_TLS): Define.
23156 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23157 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23158 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23159 load_tls_operand, microblaze_call_tls_get_addr,
23160 microblaze_legitimize_tls_address): New functions.
23161 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23162 (get_base_reg): Use microblaze_tls_symbol_p.
23163 (microblaze_classify_address): Handle TLS.
23164 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23165 label_mentioned_p and microblaze_tls_referenced_p.
23166 (microblaze_legitimize_address): Handle TLS.
23167 (microblaze_address_insns): Handle ADDRESS_TLS.
23168 (pic_address_needs_scratch): Handle TLS.
23169 (print_operand_address): Handle TLS.
23170 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23171 (microblaze_expand_move): Handle TLS.
23172 (microblaze_legitimate_constant_p): Check
23173 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23174 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23175 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23176 (PIC_OFFSET_TABLE_REGNUM): Set.
23177 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23178 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23179 (addsi3, movsi_internal2, movdf_internal): Update constraints
23180 * config/microblaze/predicates.md (arith_plus_operand): Define
23181 (move_operand): Redefine as move_src_operand,
23182 check microblaze_tls_referenced_p.
23184 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23186 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23187 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23189 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23191 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23194 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23196 PR tree-optimization/53265
23197 * common.opt (Waggressive-loop-optimizations): New option.
23198 * tree-ssa-loop-niter.c: Include tree-pass.h.
23199 (do_warn_aggressive_loop_optimizations): New function.
23200 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23201 if number_of_latch_executions returned constant.
23202 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23203 early. If number_of_latch_executions returned constant, set
23204 nb_iterations_upper_bound back to it.
23205 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23207 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23208 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23210 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23211 (MULTILIB_OSDIRNAMES): Set.
23212 * genmultilib: If defaultosdirname doesn't start with :: , set
23213 defaultosdirname2 instead, clear it and emit two . multilib_raw
23214 entries instead of just one.
23216 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23218 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23219 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23220 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23221 (SUBTARGET_OVERRIDE_OPTIONS): New.
23223 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23226 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23227 (musermode): Convert to Var(TARGET_USERMODE).
23228 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23229 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23230 * config/sh/sh.c (sh_option_override): Use
23231 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23232 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23234 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23236 (udivsi3_i4_single, divsi3_i4_single): Use
23237 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23239 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23241 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23244 2013-03-13 Richard Biener <rguenther@suse.de>
23246 PR tree-optimization/56608
23247 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23248 calls when vectorizing basic-blocks.
23250 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23253 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23256 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23258 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23260 2013-03-11 Jan Hubicka <jh@suse.cz>
23263 * lto-streamer-out.c (output_symbol_p): Skip references from
23264 constructors of external variables.
23266 2013-03-11 Jan Hubicka <jh@suse.cz>
23268 PR middle-end/56571
23269 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23271 * emit-rtl.c (verify_rtx_sharing): Likewise.
23272 (copy_insn_1): Likewise.
23273 * rtl.c (copy_rtx): Likewise.
23275 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23278 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23279 output_operand_lossage message.
23281 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23284 * arm.c (shift_op): Validate RTL pattern on the fly.
23285 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23288 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23291 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23292 new call patterns that clobber %fr12.
23293 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23294 split and postreload patterns.
23295 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23296 registers %fr12 and %fr12R as call used.
23298 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23300 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23301 (canon_address, record_store, replace_read, check_mem_read_rtx,
23302 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23303 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23304 rest_of_handle_dse): Likewise.
23306 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23308 PR middle-end/56524
23309 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23311 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23312 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23313 (save_optabs_if_changed): Replace with...
23314 (init_tree_optimization_optabs): ...this.
23315 * optabs.c (save_optabs_if_changed): Rename to...
23316 (init_tree_optimization_optabs): ...this. Take the optimization node
23317 as argument. Do nothing if the base optabs are already correct.
23318 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23319 to recompute optabs.
23320 * function.h (function): Remove optabs field.
23321 * function.c (invoke_set_current_function_hook): Call
23322 init_tree_optimization_optabs. Use the result to initialize
23325 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23327 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23328 if GTMA_HAS_NO_INSTRUMENTATION.
23329 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23330 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23331 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23332 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23333 GTMA_HAS_NO_INSTRUMENTATION.
23335 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23337 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23340 2013-03-08 Marek Polacek <polacek@redhat.com>
23341 Jakub Jelinek <jakub@redhat.com>
23343 PR tree-optimization/56478
23344 * predict.c (is_comparison_with_loop_invariant_p): Change the
23345 type of loop_step to tree.
23346 (predict_loops): Adjust.
23347 (predict_iv_comparison): Perform the computations on double_ints.
23349 2013-03-08 Richard Biener <rguenther@suse.de>
23351 PR tree-optimization/56570
23352 * tree-cfg.c (verify_expr_location_1): Verify locations for
23354 * tree-sra.c (create_access_replacement): Strip locations
23355 from DECL_DEBUG_EXPRs.
23357 2013-03-08 Richard Biener <rguenther@suse.de>
23359 * tree-inline.c (expand_call_inline): Do not associate
23360 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23361 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23363 2013-03-08 Richard Biener <rguenther@suse.de>
23365 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23366 or block changes with -Og. Fix for location / block encoding
23367 changes and PHI arguments with locations.
23369 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23371 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23373 (struct output_info): Likewise.
23374 (register_overhead): Remove bad gcc_assert.
23375 (bitmap_find_bit): If there is only a single bitmap element, do not
23376 count a miss as a search.
23377 (print_statistics): Update for counter type changes.
23378 (dump_bitmap_statistics): Likewise. Print headers such that they
23379 are properly lined up with the printed counters.
23381 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23383 PR tree-optimization/56559
23384 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23385 check that it has only a single use.
23387 2013-03-07 Richard Biener <rguenther@suse.de>
23389 * doc/invoke.texi (fwhole-program): Discourage use in combination
23392 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23394 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23396 PR tree-optimization/56539
23397 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23398 instead of GSI_CONTINUE_LINKING as last argument to
23399 force_gimple_operand_gsi. Adjust function comment.
23401 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23404 PR middle-end/56548
23405 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23406 promoted mode, convert the result back to the original mode.
23408 2013-03-06 Richard Biener <rguenther@suse.de>
23410 PR middle-end/56294
23411 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23412 (insert_updated_phi_nodes_compare_uids): New function.
23413 (update_ssa): Sort symbols_to_rename after UID before
23414 traversing it to insert PHI nodes.
23416 2013-03-06 Richard Biener <rguenther@suse.de>
23418 PR middle-end/50494
23419 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23420 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23423 2013-02-13 Richard Biener <rguenther@suse.de>
23426 * varasm.c (output_constant_def_1): Get the decl representing
23427 the constant as argument.
23428 (output_constant_def): Wrap output_constant_def_1.
23429 (make_decl_rtl): Use output_constant_def_1 with the decl
23430 representing the constant.
23431 (build_constant_desc): Optionally re-use a decl already
23432 representing the constant.
23433 (tree_output_constant_def): Adjust.
23435 2013-03-06 Joey Ye <joey.ye@arm.com>
23438 * gcc.c (convert_white_space): New function.
23439 (main): Handles white space in function name.
23441 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23444 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23445 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23446 to SH_DIV_CALL_TABLE for TARGET_SH2.
23447 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23449 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23450 call-table options.
23452 2013-03-05 Sterling Augustine <saugustine@google.com>
23453 Cary Coutant <ccoutant@google.com>
23456 * dwarf2out.c (resolve_addr): Don't call
23457 remove_loc_list_addr_table_entries a second time for the same
23460 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23463 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23464 (avoid_complex_debug_insns): New function.
23465 (expand_debug_locations): Call it.
23467 PR rtl-optimization/56484
23468 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23469 lifetimes of hard registers on small register class machines.
23471 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23473 * config/microblaze/microblaze-protos.h: Rename
23474 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23475 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23477 (microblaze_fast_interrupt_function_p): New function.
23478 (microblaze_is_interrupt_handler): Rename to
23479 microblaze_is_interrupt_variant and add fast_interrupt check.
23480 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23481 (save_restore_insns): Likewise.
23482 (compute_frame_size): Likewise.
23483 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23484 (microblaze_globalize_label): Likewise.
23485 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23486 * config/microblaze/microblaze.md: Use wrapper
23487 microblaze_is_interrupt_variant.
23489 2013-03-05 Kai Tietz <ktietz@redhat.com>
23491 * sdbout.c (sdbout_one_type): Switch to current function's section
23492 supporting cold/hot.
23494 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23496 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23499 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23501 PR middle-end/56461
23502 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23503 if VALGRIND_GET_VBITS is defined, temporarily make object
23504 memory all defined, and restore previous valgrind addressability
23505 and definability afterwards. Free this_object at the end.
23507 PR middle-end/56461
23508 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23509 right before calling lra_create_live_ranges, also call it
23510 when clearing live_p. Only call lra_clear_live_ranges
23511 at the end if live_p.
23513 PR middle-end/56461
23514 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23516 2013-03-05 Richard Biener <rguenther@suse.de>
23518 PR tree-optimization/56521
23519 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23522 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23525 * except.h (remove_unreachable_eh_regions): New prototype.
23526 * except.c (remove_eh_handler_splicer): New function, split out
23527 of remove_eh_handler.
23528 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23529 warning about running it on many EH regions one at a time.
23530 (remove_unreachable_eh_regions_worker): New function, walk the
23531 EH tree in depth-first order and remove non-marked regions.
23532 (remove_unreachable_eh_regions): New function.
23533 * tree-eh.c (mark_reachable_handlers): New function, split out
23534 from remove_unreachable_handlers.
23535 (remove_unreachable_handlers): Use mark_reachable_handlers and
23536 remove_unreachable_eh_regions.
23537 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23538 and remove_unreachable_eh_regions.
23540 2013-03-05 Richard Biener <rguenther@suse.de>
23542 PR middle-end/56525
23543 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23544 not freeing them until the end.
23546 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23548 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23550 2013-03-05 Richard Biener <rguenther@suse.de>
23552 PR tree-optimization/56270
23553 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23554 of loads after scheduling an SLP instance.
23556 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23558 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23560 (check_gcc_parallelize): Run guality.exp as a separate job from
23561 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23562 struct-layout-1.exp with stackalign.exp.
23564 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23566 PR middle-end/56461
23567 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23568 load_index sbitmap even if some bit in it isn't set.
23570 PR middle-end/56461
23571 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23572 (discover_iteration_bound_by_body_walk): Change queues to
23573 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23574 spelling in comment. Call safe_push on queues[bound_index] directly.
23575 Release queues[queue_index] in every iteration unconditionally.
23576 Release bounds vector.
23578 PR middle-end/56461
23579 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23580 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23581 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23584 2013-03-05 Richard Biener <rguenther@suse.de>
23587 * tree-inline.c (remap_blocks_to_null): New function.
23588 (expand_call_inline): When expanding a call stmt without
23589 an associated block inline remap all callee blocks to NULL.
23591 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23593 PR rtl-optimization/56494
23594 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23595 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23596 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23598 PR middle-end/56461
23599 * sel-sched-ir.c (free_sched_pools): Release
23600 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23601 if succs_info_pool.max_top isn't -1.
23604 * opts.c (opts_obstack, opts_concat): Moved to...
23605 * opts-common.c (opts_obstack, opts_concat): ... here.
23607 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23609 PR middle-end/56461
23610 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23612 2013-03-04 Martin Jambor <mjambor@suse.cz>
23614 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23615 all appropriate places.
23617 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23619 PR tree-optimization/56424
23620 * ipa-split.c (split_function): Do not set the RSO flag if result is
23621 not by reference and its type is a register type.
23623 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23625 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23626 (microblaze_legitimate_pic_operand): Likewise
23627 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23628 new function microblaze_legitimate_pic_operand
23629 * config/microblaze/microblaze-protos.h
23630 (microblaze_legitimate_pic_operand): Declare.
23632 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23634 * config/microblaze/predicates.md (call_insn_simple_operand):
23635 New predicate for supported rtx code types.
23636 * config/microblaze/microblaze.md (call_internal1): Use
23637 call_insn_simple_operand predicate.
23639 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23641 PR middle-end/56461
23642 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23643 partitions.ordered_remove.
23645 PR middle-end/56461
23646 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23647 vec_oprnds0.create (1) for modifier == NONE.
23649 PR middle-end/56461
23650 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23651 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23652 vec_oprnds1 right before pushing anything to it for
23655 PR middle-end/56461
23656 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23657 set nbbs to 0 instead of having separate code path.
23658 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23659 instead of false as last argument if returning NULL.
23661 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23663 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23664 the attribute is now called "target" instead of "option".
23665 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23666 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23667 attribute/pragma name for TARGET_OPTION_VALID_P and
23668 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23669 * doc/tm.texi: Regenerated.
23671 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23673 * config/microblaze/microblaze.c:
23674 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23675 * config/microblaze/microblaze.h: Add -mxl-reorder to
23677 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23678 instructions emitted if TARGET_REORDER.
23679 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23680 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23683 2013-03-01 Xinliang David Li <davidxl@google.com>
23685 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23688 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23690 PR middle-end/56461
23691 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23692 vector even when returning true. Fix up function comment formatting.
23694 PR middle-end/56461
23695 * ira-build.c (ira_loop_nodes_count): New variable.
23696 (create_loop_tree_nodes): Initialize it.
23697 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23699 PR middle-end/56461
23700 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23701 method on dr_chain and result_chain.
23702 * tree-vect-stmts.c (vectorizable_store): Only call
23703 result_chain.create if j == 0.
23705 PR middle-end/56461
23706 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23707 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23708 before overwriting it.
23710 2013-03-01 Tobias Burnus <burnus@net-b.de>
23712 * doc/extended.texi (C Extensions): Change order in @menu
23714 (Other MIPS Built-in Functions): Move last MIPS entry before
23715 "picoChip Built-in Functions".
23716 (SH Built-in Functions): Move after RX Built-in Functions.
23717 * doc/gcc.texi (Introduction): Change order in @menu
23719 * doc/md.texi (Constraints): Ditto.
23720 * gty.texi (Type Information): Ditto.
23721 (User-provided marking routines for template types): Make
23723 * doc/invoke.texi (AArch64 Options): Move before
23724 "Adapteva Epiphany Options".
23726 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23727 Jakub Jelinek <jakub@redhat.com>
23730 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23731 no_address_safety_analysis attribute.
23732 * doc/extend.texi (no_address_safety_attribute): Rename to
23733 no_sanitize_address attribute, mention no_address_safety_analysis
23734 attribute as deprecated alias.
23736 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23738 PR middle-end/56461
23739 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23740 type to vec<vec<tree> > *.
23741 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23742 to be vec<tree> instead of vec<tree> *, set vec_defs
23743 to vNULL and call vec_defs.create (number_of_vects), adjust other
23745 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23746 vectorizable_condition): Adjust vect_get_slp_defs callers.
23748 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23750 * config/aarch64/aarch64.c
23751 (aarch64_float_const_representable): Remove unused variable.
23753 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23755 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23757 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23759 * config/aarch64/aarch64-builtins.c
23760 (aarch64_init_simd_builtins): Make static.
23762 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23764 * config/aarch64/aarch64.c
23765 (aarch64_simd_make_constant): Make static.
23767 2013-02-28 Martin Jambor <mjambor@suse.cz>
23769 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23770 with no initialization to the RHS of debug statements.
23772 2013-02-28 Martin Jambor <mjambor@suse.cz>
23774 PR tree-optimization/56294
23775 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23777 (get_access_replacement): Do not call create_access_replacement.
23778 Assert a replacement exists.
23779 (get_repl_default_def_ssa_name): Create the replacement declaration
23782 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23784 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23785 final_end_function.
23787 2013-02-28 Marek Polacek <polacek@redhat.com>
23789 PR rtl-optimization/56466
23790 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23791 if we're changing a loop.
23792 (peel_loops_completely): Likewise.
23794 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23797 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23799 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23802 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23803 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23804 INTX_FTYPE_FX, FX_FTYPE_INTX.
23805 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23807 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23809 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23810 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23811 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23812 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23813 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23814 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23815 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23816 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23817 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23818 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23819 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23820 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23821 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23822 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23823 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23824 (avrxmega6): Increase max flash segments from 5 to 6.
23825 * config/avr/t-multilib: Regenerate.
23826 * config/avr/avr-tables.opt: Regenerate.
23827 * doc/avr-mmcu.texi: Regenerate.
23829 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23831 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23832 (avr_device_to_arch): Rename to avr_device_to_ld.
23833 (avr_device_to_as): New prototype.
23834 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23835 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23836 * config/avr/driver-avr.c (avr_device_to_as): New.
23837 (avr_device_to_arch): Rename to avr_device_to_ld.
23839 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23841 PR middle-end/56461
23842 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23843 method on dr_chain and result_chain.
23845 PR middle-end/56461
23846 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23847 pointer_set_destroy on not_executed_last_iteration.
23849 PR middle-end/56461
23850 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23852 PR middle-end/56461
23853 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23854 FOR_EACH_DEFINED_FUNCTION when freeing state.
23856 PR middle-end/56461
23857 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23859 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23862 PR middle-end/56461
23863 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23864 known_aggs[i].items and release known_aggs vector.
23866 PR middle-end/56461
23867 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23869 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23871 * config/microblaze/microblaze.c (microblaze_emit_compare):
23872 Use xor for EQ/NE comparisions.
23873 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23874 (cbranchsf4): Adjust operator to comparison_operator.
23876 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23878 PR middle-end/56461
23879 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23881 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23882 vec_safe_push, always update *slot.
23883 (redirect_edge_var_map_clear): Use vec_free.
23884 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23885 (free_var_map_entry): Use vec_free.
23886 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23887 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23889 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23891 PR middle-end/45472
23892 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23893 when the may_trap_p bit of the exprs being merged differs.
23894 Reorder tests for speculativeness in the logical and operator.
23896 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23898 * incpath.c (add_standard_paths): Use reconcat instead of concat
23899 where appropriate and avoid leaking memory.
23901 * opts.h: Include obstack.h.
23902 (opts_concat): New prototype.
23903 (opts_obstack): New declaration.
23904 * opts.c (opts_concat): New function.
23905 (opts_obstack): New variable.
23906 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23907 (finish_options): Use opts_concat instead of concat
23908 and XOBNEWVEC instead of XNEWVEC.
23909 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23910 generate_option): Likewise.
23911 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23912 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23915 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23916 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23918 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23920 PR middle-end/56461
23921 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23923 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23925 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23926 (arm_block_move_unaligned_straight): Likewise.
23927 (arm_adjust_block_mem): Likewise.
23929 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23932 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23933 temp, cond and label.
23934 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23937 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23938 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23941 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23942 of prologue/epilogue functions.
23945 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23946 Remove unused variable cfa_offset.
23947 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23950 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23951 type promotion to unsigned.
23954 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23955 for HOST_WIDE_INT of 32 bit / same size as int.
23956 (arm_block_move_unaligned_straight): Likewise.
23957 (arm_adjust_block_mem): Likewise.
23960 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23963 2013-02-26 Marek Polacek <polacek@redhat.com>
23965 PR tree-optimization/56426
23966 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23968 2013-02-26 Richard Biener <rguenther@suse.de>
23971 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23972 unused variable loops.
23974 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23976 PR tree-optimization/56448
23977 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23978 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23979 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23980 later operands of the references, or even first operand for
23981 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23983 PR tree-optimization/56443
23984 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23985 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23986 to type_for_mode langhook.
23988 2013-02-25 Matt Turner <mattst88@gmail.com>
23990 * doc/invoke.texi: Document r4700.
23992 2013-02-25 Richard Biener <rguenther@suse.de>
23994 PR tree-optimization/56175
23995 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23997 (simplify_bitwise_binary): ... here. Also guard the conversion
23998 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24000 2013-02-25 Catherine Moore <clm@codesourcery.com>
24003 2013-02-24 Catherine Moore <clm@codesourcery.com>
24004 Maciej W. Rozycki <macro@codesourcery.com>
24005 Tom de Vries <tom@codesourcery.com>
24006 Nathan Sidwell <nathan@codesourcery.com>
24007 Iain Sandoe <iain@codesourcery.com>
24008 Nathan Froyd <froydnj@codesourcery.com>
24009 Chao-ying Fu <fu@mips.com>
24011 * doc/extend.texi: (micromips, nomicromips, nocompression):
24012 Document new function attributes.
24013 * doc/invoke.texi (minterlink-compressed, mmicromips,
24014 m14k, m14ke, m14kec): Document new options.
24015 (minterlink-mips16): Update documentation.
24016 * doc/md.texi (ZC, ZD): Document new constraints.
24017 * configure.ac (gcc_cv_as_micromips): Check if linker
24018 supports the .set micromips directive.
24019 * configure: Regenerate.
24020 * config.in: Regenerate.
24021 * config/mips/mips-tables.opt: Regenerate.
24022 * config/mips/micromips.md: New file.
24023 * constraints.md (ZC, AD): New constraints.
24024 * config/mips/predicates.md (movep_src_register): New predicate.
24025 (movep_src_operand): New predicate.
24026 (non_volatile_mem_operand): New predicate.
24027 * config/mips/mips.md (multimem): New type.
24028 (length): Differentiate between 17-bit and 18-bit branch offsets.
24029 (MOVEP1, MOVEP2): New mode iterator.
24030 (mov_<load>l): Use ZC constraint.
24031 (mov_<load>r): Likewise.
24032 (mov_<store>l): Likewise.
24033 (mov_<store>r): Likewise.
24034 (*branch_equality<mode>_inverted): Add microMIPS support.
24035 (*branch_equality<mode>): Likewise.
24036 (*jump_absolute): Likewise.
24037 (indirect_jump_<mode>): Likewise.
24038 (tablejump_<mode>): Likewise.
24039 (<optab>_internal): Likewise.
24040 (sibcall_internal): Likewise.
24041 (sibcall_value_internal): Likewise.
24042 (prefetch): Use constraint ZD.
24043 * config/mips/mips.opt (minterlink-compressed): New option.
24044 (minterlink-mips16): Now an alias for minterlink-compressed.
24045 (mmicromips): New option.
24046 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24047 (compare_and_swap_12): Likewise.
24048 (sync_add<mode>): Likewise.
24049 (sync_<optab>_12): Likewise.
24050 (sync_old_<optab>_12): Likewise.
24051 (sync_new_<optab>_12): Likewise.
24052 (sync_nand_12): Likewise.
24053 (sync_old_nand_12): Likewise.
24054 (sync_new_nand_12): Likewise.
24055 (sync_sub<mode>): Likewise.
24056 (sync_old_add<mode>): Likewise.
24057 (sync_old_sub<mode>): Likewise.
24058 (sync_new_add<mode>): Likewise.
24059 (sync_new_sub<mode>): Likewise.
24060 (sync_<optab><mode>): Likewise.
24061 (sync_old_<optab><mode>): Likewise.
24062 (sync_new_<optab><mode>): Likewise.
24063 (sync_nand<mode>): Likewise.
24064 (sync_old_nand<mode>): Likewise.
24065 (sync_new_nand<mode>): Likewise.
24066 (sync_lock_test_and_set<mode>): Likewise.
24067 (test_and_set_12): Likewise.
24068 (atomic_compare_and_swap<mode>): Likewise.
24069 (atomic_exchange<mode>_llsc): Likewise.
24070 (atomic_fetch_add<mode>_llsc): Likewise.
24071 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24072 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24073 (umips_save_restore_pattern_p): Likewise.
24074 (umips_load_store_pair_p): Likewise.
24075 (umips_output_load_store_pair): Likewise.
24076 (umips_movep_target_p): Likewise.
24077 (umips_12bit_offset_address_p): Likewise.
24078 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24079 (mips_base_mips16): Rename this...
24080 (mips_base_compression_flags): ...to this. Update all uses.
24081 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24082 (mips_mips16_decl_p): Delete.
24083 (mips_nomips16_decl_p): Delete.
24084 (mips_get_compress_on_flags): New function.
24085 (mips_get_compress_off_flags): New function.
24086 (mips_get_compress_mode): New function.
24087 (mips_get_compress_on_name): New function.
24088 (mips_get_compress_off_name): New function.
24089 (mips_insert_attributes): Support multiple compression types.
24090 (mips_merge_decl_attributes): Likewise.
24091 (umips_12bit_offset_address_p): New function.
24092 (mips_start_function_definition): Emit .set micromips directive.
24093 (mips_call_may_need_jalx_p): New function.
24094 (mips_function_ok_for_sibcall): Add microMIPS support.
24095 (mips_print_operand_punctuation): Support short delay slots and
24097 (umips_swm_mask, umips_swm_encoding): New.
24098 (umips_build_save_restore): New function.
24099 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24100 (was_mips16_p): Remove.
24101 (old_compression_mode): New.
24102 (mips_set_compression_mode): New function.
24103 (mips_set_current_function): Add microMIPS support.
24104 (mips_option_override): Likewise.
24105 (umips_save_restore_pattern_p): New function.
24106 (umips_output_save_restore): New function.
24107 (umips_load_store_pair_p_1): New function.
24108 (umips_load_store_pair_p): New function.
24109 (umips_output_load_store_pair_1): New function.
24110 (umips_output_load_store_pair): New function.
24111 (umips_movep_target_p) New function.
24112 (mips_prepare_pch_save): Add microMIPS support.
24113 * config/mips/mips.h (TARGET_COMPRESSION): New.
24114 (TARGET_CPU_CPP_BUILTINS): Update macro
24115 to use new compression flags and to support microMIPS.
24116 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24117 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24118 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24119 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24120 (ASM_SPEC): Support mmicromips and mno-micromips.
24121 (M16STORE_REG_P): New macro.
24122 (MIPS_CALL): Support TARGET_MICROMIPS.
24123 (MICROMIPS_J): New macro.
24124 (mips_base_mips16): Rename this...
24125 (mips_base_compression_flags): ...to this.
24126 (UMIPS_12BIT_OFFSET_P): New macro.
24127 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24128 (MULTILIB_DIRNAMES): Likewise.
24130 2013-02-25 Tom de Vries <tom@codesourcery.com>
24132 PR rtl-optimization/56131
24133 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24134 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24135 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24137 2013-02-25 Tobias Burnus <burnus@net-b.de>
24139 * doc/invoke.texi (-fsanitize=): Move from optimization
24140 to debugging options.
24142 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24144 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24146 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24147 Alexander Monakov <amonakov@ispras.ru>
24149 PR middle-end/56077
24150 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24151 flush pending lists also on non-jumps. Adjust comment.
24153 2013-02-24 Catherine Moore <clm@codesourcery.com>
24154 Maciej W. Rozycki <macro@codesourcery.com>
24155 Tom de Vries <tom@codesourcery.com>
24156 Nathan Sidwell <nathan@codesourcery.com>
24157 Iain Sandoe <iain@codesourcery.com>
24158 Nathan Froyd <froydnj@codesourcery.com>
24159 Chao-ying Fu <fu@mips.com>
24161 * doc/extend.texi: (micromips, nomicromips, nocompression):
24162 Document new function attributes.
24163 * doc/invoke.texi (minterlink-compressed, mmicromips,
24164 m14k, m14ke, m14kec): Document new options.
24165 (minterlink-mips16): Update documentation.
24166 * doc/md.texi (ZC, ZD): Document new constraints.
24167 * configure.ac (gcc_cv_as_micromips): Check if linker
24168 supports the .set micromips directive.
24169 * configure: Regenerate.
24170 * config.in: Regenerate.
24171 * config/mips/mips-tables.opt: Regenerate.
24172 * config/mips/micromips.md: New file.
24173 * constraints.md (ZC, AD): New constraints.
24174 * config/mips/predicates.md (movep_src_register): New predicate.
24175 (movep_src_operand): New predicate.
24176 (non_volatile_mem_operand): New predicate.
24177 * config/mips/mips.md (multimem): New type.
24178 (length): Differentiate between 17-bit and 18-bit branch offsets.
24179 (MOVEP1, MOVEP2): New mode iterator.
24180 (mov_<load>l): Use ZC constraint.
24181 (mov_<load>r): Likewise.
24182 (mov_<store>l): Likewise.
24183 (mov_<store>r): Likewise.
24184 (*branch_equality<mode>_inverted): Add microMIPS support.
24185 (*branch_equality<mode>): Likewise.
24186 (*jump_absolute): Likewise.
24187 (indirect_jump_<mode>): Likewise.
24188 (tablejump_<mode>): Likewise.
24189 (<optab>_internal): Likewise.
24190 (sibcall_internal): Likewise.
24191 (sibcall_value_internal): Likewise.
24192 (prefetch): Use constraint ZD.
24193 * config/mips/mips.opt (minterlink-compressed): New option.
24194 (minterlink-mips16): Now an alias for minterlink-compressed.
24195 (mmicromips): New option.
24196 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24197 (compare_and_swap_12): Likewise.
24198 (sync_add<mode>): Likewise.
24199 (sync_<optab>_12): Likewise.
24200 (sync_old_<optab>_12): Likewise.
24201 (sync_new_<optab>_12): Likewise.
24202 (sync_nand_12): Likewise.
24203 (sync_old_nand_12): Likewise.
24204 (sync_new_nand_12): Likewise.
24205 (sync_sub<mode>): Likewise.
24206 (sync_old_add<mode>): Likewise.
24207 (sync_old_sub<mode>): Likewise.
24208 (sync_new_add<mode>): Likewise.
24209 (sync_new_sub<mode>): Likewise.
24210 (sync_<optab><mode>): Likewise.
24211 (sync_old_<optab><mode>): Likewise.
24212 (sync_new_<optab><mode>): Likewise.
24213 (sync_nand<mode>): Likewise.
24214 (sync_old_nand<mode>): Likewise.
24215 (sync_new_nand<mode>): Likewise.
24216 (sync_lock_test_and_set<mode>): Likewise.
24217 (test_and_set_12): Likewise.
24218 (atomic_compare_and_swap<mode>): Likewise.
24219 (atomic_exchange<mode>_llsc): Likewise.
24220 (atomic_fetch_add<mode>_llsc): Likewise.
24221 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24222 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24223 (umips_save_restore_pattern_p): Likewise.
24224 (umips_load_store_pair_p): Likewise.
24225 (umips_output_load_store_pair): Likewise.
24226 (umips_movep_target_p): Likewise.
24227 (umips_12bit_offset_address_p): Likewise.
24228 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24229 (mips_base_mips16): Rename this...
24230 (mips_base_compression_flags): ...to this. Update all uses.
24231 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24232 (mips_mips16_decl_p): Delete.
24233 (mips_nomips16_decl_p): Delete.
24234 (mips_get_compress_on_flags): New function.
24235 (mips_get_compress_off_flags): New function.
24236 (mips_get_compress_mode): New function.
24237 (mips_get_compress_on_name): New function.
24238 (mips_get_compress_off_name): New function.
24239 (mips_insert_attributes): Support multiple compression types.
24240 (mips_merge_decl_attributes): Likewise.
24241 (umips_12bit_offset_address_p): New function.
24242 (mips_start_function_definition): Emit .set micromips directive.
24243 (mips_call_may_need_jalx_p): New function.
24244 (mips_function_ok_for_sibcall): Add microMIPS support.
24245 (mips_print_operand_punctuation): Support short delay slots and
24247 (umips_swm_mask, umips_swm_encoding): New.
24248 (umips_build_save_restore): New function.
24249 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24250 (was_mips16_p): Remove.
24251 (old_compression_mode): New.
24252 (mips_set_compression_mode): New function.
24253 (mips_set_current_function): Add microMIPS support.
24254 (mips_option_override): Likewise.
24255 (umips_save_restore_pattern_p): New function.
24256 (umips_output_save_restore): New function.
24257 (umips_load_store_pair_p_1): New function.
24258 (umips_load_store_pair_p): New function.
24259 (umips_output_load_store_pair_1): New function.
24260 (umips_output_load_store_pair): New function.
24261 (umips_movep_target_p) New function.
24262 (mips_prepare_pch_save): Add microMIPS support.
24263 * config/mips/mips.h (TARGET_COMPRESSION): New.
24264 (TARGET_CPU_CPP_BUILTINS): Update macro
24265 to use new compression flags and to support microMIPS.
24266 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24267 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24268 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24269 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24270 (ASM_SPEC): Support mmicromips and mno-micromips.
24271 (M16STORE_REG_P): New macro.
24272 (MIPS_CALL): Support TARGET_MICROMIPS.
24273 (MICROMIPS_J): New macro.
24274 (mips_base_mips16): Rename this...
24275 (mips_base_compression_flags): ...to this.
24276 (UMIPS_12BIT_OFFSET_P): New macro.
24277 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24278 (MULTILIB_DIRNAMES): Likewise.
24280 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24283 * target-globals.c (save_target_globals): For init_reg_sets and
24284 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24286 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24288 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24289 * config/aarch64/t-aarch64
24290 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24292 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24294 PR inline-asm/56148
24295 * lra-constraints.c (process_alt_operands): Reload operand
24296 conflicting with earlier clobber only if no more other conflicting
24299 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24302 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24303 if not linking a shared library.
24305 2013-02-22 Seth LaForge <sethml@google.com>
24307 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24309 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24311 * config/arm/arm.md (split for extendsidi): Update condition.
24312 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24313 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24314 (qhs_zextenddi_cstr): Likewise.
24316 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24318 PR middle-end/56420
24319 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24320 avoid signed wrapping.
24321 (expand_mult): Handle properly multiplication by
24322 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24323 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24324 in the compiler if coeff is HOST_WIDE_INT_MIN.
24325 (expand_divmod): Don't make ext_op1 static, change it's type to
24326 uhwi. Avoid undefined behavior in -INTVAL (op1).
24328 PR rtl-optimization/50339
24329 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24331 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24332 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24333 into splitting_ashiftrt field.
24334 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24336 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24339 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24341 PR middle-end/56108
24342 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24343 are sure to go irrevocable.
24345 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24347 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24348 scalars are valid operands.
24350 2013-02-21 Martin Jambor <mjambor@suse.cz>
24352 PR tree-optimization/56310
24353 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24354 only matching indices and non-negative final offsets.
24355 (intersect_aggregates_with_edge): Pass src_idx to
24356 agg_replacements_to_vector. Pass src_idx insstead of index to
24357 intersect_with_agg_replacements.
24359 2013-02-21 Martin Jambor <mjambor@suse.cz>
24361 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24362 instead of hard-wired defaults.
24364 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24366 * doc/invoke.texi (MIPS Options): Update documentation of the
24367 floating-point multiply-accumulate instruction restrictions.
24369 2013-02-21 Kostya Serebryany <kcc@google.com>
24371 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24372 asan_shadow_offset on x86_64 linux.
24374 2013-02-21 Richard Biener <rguenther@suse.de>
24376 PR tree-optimization/56415
24378 2013-02-11 Richard Biener <rguenther@suse.de>
24380 PR tree-optimization/56273
24381 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24384 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24387 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24390 PR inline-asm/56405
24391 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24392 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24394 2013-02-20 Jan Hubicka <jh@suse.cz>
24396 PR tree-optimization/56265
24397 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24398 when target is referenced for first time.
24400 2013-02-20 Richard Biener <rguenther@suse.de>
24402 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24403 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24404 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24405 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24406 not return anything.
24407 (rename_ssa_copies): Do not remove unused locals.
24408 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24409 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24410 * passes.c (execute_function_todo): Do not schedule unused locals
24411 removal if cleanup_tree_cfg did something.
24412 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24413 about the number of removed locals.
24415 2013-02-20 Richard Biener <rguenther@suse.de>
24417 PR tree-optimization/56398
24418 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24420 2013-02-20 Martin Jambor <mjambor@suse.cz>
24422 PR tree-optimization/55334
24423 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24424 restricted pointers to arrays.
24426 2013-02-20 Richard Biener <rguenther@suse.de>
24427 Jakub Jelinek <jakub@redhat.com>
24429 PR tree-optimization/56396
24430 * tree-ssa-ccp.c (n_const_val): New static variable.
24431 (get_value): Return NULL for SSA names we don't have a lattice
24433 (ccp_initialize): Initialize n_const_val.
24434 * tree-ssa-copy.c (n_copy_of): New static variable.
24435 (init_copy_prop): Initialize n_copy_of.
24436 (get_value): Return NULL_TREE for SSA names we don't have a
24439 2013-02-20 Martin Jambor <mjambor@suse.cz>
24441 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24443 2013-02-20 Richard Biener <rguenther@suse.de>
24445 * genpreds.c (write_lookup_constraint): Do not compare first
24446 letter of the constraint again.
24448 2013-02-20 Richard Biener <rguenther@suse.de>
24450 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24452 (get_use_iv_cost): Terminate hashtable walk when coming across
24455 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24457 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24458 reassociation for avx2 targets.
24460 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24462 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24463 Add version check for v8.10.a to enable microblaze_has_clz
24464 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24465 version and TARGET_PATTERN_COMPARE check
24466 * config/microblaze/microblaze.md: New clzsi2 instruction
24468 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24470 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24471 function before branching.
24473 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24475 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24477 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24479 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24481 PR middle-end/55889
24482 * sel-sched.c: Include ira.h.
24483 (implicit_clobber_conflict_p): New function.
24484 (moveup_expr): Use it.
24485 * Makefile.in (sel-sched.o): Depend on ira.h.
24487 2013-02-19 Richard Biener <rguenther@suse.de>
24489 PR tree-optimization/56384
24490 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24491 (vn_hash_type): Split out from ...
24492 (vn_hash_constant_with_type): ... here.
24493 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24494 (vn_phi_eq): Compare types from vn_phi_s structure.
24495 (vn_phi_lookup): Populate vn_phi_s type.
24496 (vn_phi_insert): Likewise.
24498 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24500 PR tree-optimization/56350
24501 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24502 if haven't found reduction or nested cycle operand, rather than
24503 asserting we must find it.
24505 PR tree-optimization/56381
24506 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24509 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24510 Jakub Jelinek <jakub@redhat.com>
24513 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24514 (swap_optab_enable): Same.
24515 (init_all_optabs): Use argument instead of global.
24516 * tree.h (struct tree_optimization_option): New field target_optabs.
24517 * expr.h (init_all_optabs): Add argument to prototype.
24518 (TREE_OPTIMIZATION_OPTABS): New.
24519 (save_optabs_if_changed): Protoize.
24520 * optabs.h: Declare this_fn_optabs.
24521 * optabs.c (save_optabs_if_changed): New.
24522 Declare this_fn_optabs.
24523 (init_optabs): Add argument to init_all_optabs() call.
24524 * function.c (invoke_set_current_function_hook): Handle per
24526 * function.h (struct function): New field optabs.
24527 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24528 optimization_current_node has changed.
24529 * target-globals.h (save_target_globals_default_opts): Protoize.
24530 * target-globals.c (save_target_globals_default_opts): New.
24532 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24535 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24536 registers %fr12 and %fr12R as call used.
24539 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24540 and HImode, require all displacements to be an integer multiple of
24542 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24543 only allow QImode and HImode when reload is in progress and strict is
24544 true. Likewise for symbolic addresses. Use base14_operand to check
24545 displacements in REG+BASE addresses.
24547 2013-02-18 Richard Biener <rguenther@suse.de>
24549 PR tree-optimization/56366
24550 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24551 handle sign-conversion of outer-loop initial induction value.
24553 2013-02-18 Richard Biener <rguenther@suse.de>
24555 PR middle-end/56349
24556 * cfghooks.c (merge_blocks): If we merge a latch into another
24557 block adjust references to it.
24558 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24559 (verify_loop_structure): Verify that a recorded latch is in fact
24562 2013-02-18 Richard Biener <rguenther@suse.de>
24564 PR tree-optimization/56321
24565 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24566 order SSA name release and virtual operand unlinking.
24568 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24570 * config/microblaze/microblaze.md (save_stack_block): Define.
24571 (restore_stack_block): Likewise.
24573 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24575 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24576 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24577 * config/microblaze/microblaze.c (microblaze_option_override):
24578 Bail out early for PIC modes when target does not support PIC.
24580 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24582 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24583 Replace with a microblaze version.
24584 (microblaze_trampoline_init): Adapt for microblaze.
24585 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24588 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24589 Dodji Seketeli <dodji@redhat.com>
24592 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24593 (instrument_mem_region_access): Do not forget to always put
24594 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24595 0) statement, even for cases where either 'base' or 'base + len'
24596 are not instrumented -- because they have been previously
24597 instrumented. Simplify the logic by putting all the statements
24598 instrument 'base + len' inside a sequence, and then insert that
24599 sequence right before the current insertion point. Then, to
24600 instrument 'base + len', just get an iterator on that statement.
24601 And do not forget to update the pointer to iterator the function
24602 received as argument.
24604 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24606 PR rtl-optimization/56348
24607 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24609 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24611 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24612 (clean_graph_dump_file): Pass base to start_graph_dump.
24614 2013-02-14 Richard Henderson <rth@redhat.com>
24617 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24619 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24621 * collect2-aix.h: Define F_LOADONLY.
24623 2013-02-14 Richard Biener <rguenther@suse.de>
24626 * varasm.c (output_constant_def_1): Get the decl representing
24627 the constant as argument.
24628 (output_constant_def): Wrap output_constant_def_1.
24629 (make_decl_rtl): Use output_constant_def_1 with the decl
24630 representing the constant.
24631 (build_constant_desc): Optionally re-use a decl already
24632 representing the constant.
24633 (tree_output_constant_def): Adjust.
24635 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24638 * asan.c (instrument_builtin_call): Really put the length of the
24639 second source argument into src1_len.
24641 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24643 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24644 argument. If it is false, don't create edge from then_bb to
24646 (insert_if_then_before_iter): Pass true to it.
24647 (build_check_stmt): Pass false to it.
24648 (transform_statements): Flush hash table only on extended basic
24649 block boundaries, rather than at the beginning of every bb.
24650 Don't flush hash table on nonfreeing_call_p calls.
24651 * tree-flow.h (nonfreeing_call_p): New prototype.
24652 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24654 2013-02-13 David S. Miller <davem@davemloft.net>
24656 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24658 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24661 * ira.c (max_regno_before_ira): Move from ...
24663 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24666 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24668 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24670 2013-02-13 Richard Biener <rguenther@suse.de>
24673 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24674 globals in MEM_REFs.
24676 2013-02-13 Richard Biener <rguenther@suse.de>
24678 * loop-init.c (loop_optimizer_init): Clear loop state when
24679 re-initializing preserved loops.
24680 * loop-unswitch.c (unswitch_single_loop): Return whether
24681 we unswitched the loop. Do not verify loop state here.
24682 (unswitch_loops): When we unswitched a loop discover new loops.
24684 2013-02-13 Kostya Serebryany <kcc@google.com>
24686 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24688 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24690 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24692 Avoid instrumenting duplicated memory access in the same basic block
24693 * Makefile.in (asan.o): Add new dependency on hash-table.h
24694 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24695 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24696 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24697 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24698 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24699 (get_mem_ref_of_assignment): New functions.
24700 (get_mem_refs_of_builtin_call): Extract from
24701 instrument_builtin_call and tweak a little bit to make it fit with
24703 (instrument_builtin_call): Use the new
24704 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24705 of is_gimple_builtin_call.
24706 (instrument_derefs, instrument_mem_region_access): Insert the
24707 instrumented memory reference into the hash table.
24708 (maybe_instrument_assignment): Renamed instrument_assignment into
24709 this, and change it to advance the iterator when instrumentation
24710 actually happened and return true in that case. This makes it
24711 homogeneous with maybe_instrument_assignment, and thus give a
24712 chance to callers to be more 'regular'.
24713 (transform_statements): Clear the memory reference hash table
24714 whenever we enter a new BB, when we cross a function call, or when
24715 we are done transforming statements. Use
24716 maybe_instrument_assignment instead of instrumentation. No more
24717 need to special case maybe_instrument_assignment and advance the
24718 iterator after calling it; it's now handled just like
24719 maybe_instrument_call. Update comment.
24721 2013-02-13 Richard Biener <rguenther@suse.de>
24723 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24724 Fix loop discovery code.
24726 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24728 PR inline-asm/56148
24729 * lra-constraints.c (process_alt_operands): Match early clobber
24730 operand with itself. Check conflicts with earlyclobber only if
24731 the operand is not reloaded. Prefer to reload conflicting operand
24732 if earlyclobber and matching operands are the same.
24734 2013-02-12 Richard Biener <rguenther@suse.de>
24737 * lto-streamer-out.c (write_symbol): Do not output symbols
24738 for hard register variables.
24740 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24743 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24744 (umulsidi3_insn, mulsidi3_insn): New insns.
24746 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24748 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24749 (struct tune_params): Add vec_costs field.
24750 * config/arm/arm.c (arm_builtin_vectorization_cost)
24751 (arm_add_stmt_cost): New functions.
24752 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24753 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24754 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24755 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24756 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24757 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24758 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24760 2013-02-12 Richard Biener <rguenther@suse.de>
24763 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24764 decls again if possible.
24766 2013-02-12 Richard Biener <rguenther@suse.de>
24768 PR middle-end/56288
24769 * tree-ssa.c (verify_ssa_name): Fix check, move
24770 SSA_NAME_IN_FREE_LIST check up.
24772 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24773 Steven Bosscher <steven@gcc.gnu.org>
24775 PR rtl-optimization/56151
24776 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24777 equal to op0 or op1, and last_insn pattern is CODE operation
24778 with MEM dest and one of the operands matches that MEM.
24780 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24782 * doc/extend.texi: Document Function Multiversioning and "default"
24783 parameter string to target attribute.
24784 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24785 target attribute parameter is "default".
24786 (ix86_compare_version_priority): Remove checks for target attribute.
24787 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24788 Remove check for target attribute equal to NULL. Add assert.
24789 (ix86_generate_version_dispatcher_body): Change error to sorry.
24791 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24792 Jack Howarth <howarth@bromo.med.uc.edu>
24793 Patrick Marlier <patrick.marlier@gmail.com>
24796 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24797 define ENDFILE_SPEC as TM_DESTRUCTOR.
24798 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24800 2013-02-11 Alexander Potapenko <glider@google.com>
24801 Jack Howarth <howarth@bromo.med.uc.edu>
24802 Jakub Jelinek <jakub@redhat.com>
24805 * config/darwin.c (cdtor_record): Rename ctor_record.
24806 (sort_cdtor_records): Rename sort_ctor_records.
24807 (finalize_dtors): New routine to sort destructors by
24808 priority before use in assemble_integer.
24809 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24811 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24813 PR rtl-optimization/56275
24814 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24815 offset is non-negative and less than cmode size before
24816 calling simplify_subreg.
24818 2013-02-11 Richard Biener <rguenther@suse.de>
24820 PR tree-optimization/56264
24821 * cfgloop.h (fix_loop_structure): Adjust prototype.
24822 * loop-init.c (fix_loop_structure): Return the number of
24823 newly discovered loops.
24824 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24825 are discovered, do a full loop-closed SSA rewrite.
24827 2013-02-11 Richard Biener <rguenther@suse.de>
24829 PR tree-optimization/56273
24830 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24832 (check_array_ref): Fix missing newline in dumps.
24833 (search_for_addr_array): Likewise.
24835 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24837 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24839 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24842 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24844 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24846 PR rtl-optimization/56246
24847 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24849 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24850 constraints are satisfied.
24852 2013-02-08 Jeff Law <law@redhat.com>
24855 * emit-rtl.c (reg_is_parm_p): New function.
24856 * regs.h (reg_is_parm_p): New prototype.
24857 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24858 callee-clobbered registers.
24860 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24863 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24864 If there is no implicit builtin declaration, just return NULL.
24866 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24868 * config/i386/sse.md (FMAMODEM): New mode iterator.
24869 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24870 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24872 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24874 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24875 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24876 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24878 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24880 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24881 (microblaze*-*-elf): Likewise.
24882 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24884 * config/microblaze/microblaze-c.c: Add builtin defines for
24885 _LITTLE_ENDIAN and _BIG_ENDIAN.
24886 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24887 add to TARGET_DEFAULT flags.
24888 Expand ASM_SPEC and LINK_SPEC.
24889 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24890 * config/microblaze/microblaze.md: Update extendsidi2 and
24891 movdi_internal instructions to use low-order / high-order reg
24893 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24894 options and inversemask / mask of LITTLE_ENDIAN.
24895 * config/microblaze/t-microblaze: Expand multilib options to
24896 include mlittle-endian (le) and update exceptions patterns.
24898 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24900 PR rtl-optimization/56195
24901 * lra-constraints.c (get_reload_reg): Don't reuse regs
24902 if they have smaller mode than requested, if they have
24903 wider mode than requested, try to return a SUBREG.
24905 PR tree-optimization/56250
24906 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24907 if type is unsigned and code isn't MULT_EXPR.
24909 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24911 PR tree-optimization/56064
24912 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24913 bits according to mode.
24914 * fixed-value.h (fixed_from_double_int)
24915 (const_fixed_from_double_int): Adjust comments.
24917 2013-02-08 Richard Biener <rguenther@suse.de>
24920 * lto-streamer.h (struct data_in): Remove current_file, current_line
24921 and current_col members.
24922 * lto-streamer-out.c (lto_output_location): Stream changed bits
24923 en-block for efficiency.
24924 * lto-streamer-in.c (clear_line_info): Remove.
24925 (lto_input_location): Cache current file, line and column
24926 globally via local statics. Read changed bits en-block.
24927 (input_function): Do not call clear_line_info.
24928 (lto_read_body): Likewise.
24929 (lto_input_toplevel_asms): Likewise.
24931 2013-02-08 Michael Matz <matz@suse.de>
24933 PR tree-optimization/52448
24934 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24935 (nt_call_phase): New static.
24936 (add_or_mark_expr): Only mark accesses with newer phase than any
24938 (nonfreeing_call_p): New.
24939 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24940 (nt_fini_block): Keep blocks marked as visited.
24941 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24943 2013-02-08 Richard Biener <rguenther@suse.de>
24945 * ira.c (ira): Free broken dominator information.
24947 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24949 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24951 2013-02-08 Marek Polacek <polacek@redhat.com>
24953 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24955 2013-02-08 Richard Biener <rguenther@suse.de>
24957 PR middle-end/56181
24958 * cfgloop.h (flow_loops_find): Adjust.
24959 (bb_loop_header_p): Declare.
24960 * cfgloop.c (bb_loop_header_p): New function split out from ...
24961 (flow_loops_find): ... here. Adjust function signature,
24962 support incremental loop structure update.
24963 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24964 * cfgloopmanip.c (fix_loop_structure): Move ...
24965 * loop-init.c (fix_loop_structure): ... here.
24966 (apply_loop_flags): Split out from ...
24967 (loop_optimizer_init): ... here.
24968 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24969 in incremental mode, only remove dead loops here.
24971 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24974 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24975 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24976 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24977 (*round<mode>3.libgcc): New insns for fixed-modes.
24978 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24979 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24980 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24981 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24982 implementations. Define to __builtin_avr_absFX,
24983 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24984 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24985 __builtin_avr_countlsFX, respectively.
24986 * config/avr/avr-c.c (target.h): Include it.
24987 (enum avr_builtin_id): New enum.
24988 (avr_resolve_overloaded_builtin): New static function.
24989 (avr_register_target_pragmas): Use it to set
24990 targetm.resolve_overloaded_builtin.
24991 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24992 tree nodes used by DEF_BUILTIN.
24993 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24994 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24995 <AVR_BUILTIN_xxBITS>: Same.
24997 2013-02-08 Richard Biener <rguenther@suse.de>
24999 * cfgloop.c (verify_loop_structure): Properly handle
25000 a loop exiting to another loop header.
25001 * ira-int.h (ira_loops): Remove.
25002 * ira.c (ira_loops): Remove.
25003 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25004 (do_reload): Use loop_optimizer_finalize.
25005 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25006 number_of_loops to access the loop tree.
25007 (more_one_region_p): Likewise.
25008 (finish_loop_tree_nodes): Likewise.
25009 (rebuild_regno_allocno_maps): Likewise.
25010 (mark_loops_for_removal): Likewise.
25011 (mark_all_loops_for_removal): Likewise.
25012 (remove_unnecessary_regions): Likewise.
25013 (ira_build): Likewise.
25014 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25016 2013-02-08 Richard Biener <rguenther@suse.de>
25018 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25019 * ipa-pure-const.c (analyze_function): Avoid calling
25020 mark_irreducible_loops twice.
25021 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25023 2013-02-07 David S. Miller <davem@davemloft.net>
25025 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25027 * var-tracking.c (vt_add_function_parameter): Test the presence of
25028 HAVE_window_save properly and do not remap argument registers when
25029 we have a leaf function.
25031 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25034 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25036 * config/i386/i386.c (ix86_print_operand): Ditto.
25038 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25040 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25042 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25044 PR rtl-optimization/56225
25045 * lra-constraints.c (process_alt_operands): Check that reload hard
25046 reg can hold value for strict_low_part.
25048 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25051 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25052 dwarf2out_end_function.
25053 (in_first_function_p, maybe_at_text_label_p,
25054 first_loclabel_num_not_at_text_label): New variables.
25055 (dwarf2out_var_location): In the first function find out
25056 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25057 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25060 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25062 PR rtl-optimization/56178
25063 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25064 SUBREG of a register. Tidy up related block of code.
25065 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25066 note if the source is a register or a SUBREG of a register.
25068 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25071 * config/rs6000/rs6000.md (ptrm): New mode attr.
25072 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25073 call_value_indirect_aix<pttrsize>,
25074 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25077 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25079 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25080 if -bnortl. Convert to strcmp and strncmp.
25082 2013-02-07 Alan Modra <amodra@gmail.com>
25085 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25086 addresses won't wrap when offsetting.
25087 (rs6000_secondary_reload): Provide secondary reloads needed for
25088 wrapping LO_SUM addresses.
25090 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25092 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25093 MACH, just __MACH__.
25095 2013-02-06 Richard Biener <rguenther@suse.de>
25097 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25098 instead of calling fix_loop_structure.
25100 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25102 PR middle-end/56217
25103 * omp-low.c (use_pointer_for_field): Return false if
25104 lower_send_shared_vars doesn't generate any copy-out code.
25106 2013-02-06 Tom de Vries <tom@codesourcery.com>
25108 PR rtl-optimization/56131
25109 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25110 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25111 of the label is NULL. Add comment.
25113 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25115 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25118 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25119 (STATIC_LIBTSAN_LIBS): Likewise.
25120 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25121 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25122 is defined, don't add anything else beyond that.
25123 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25124 (LINK_COMMAND_SPEC): Use them.
25126 PR tree-optimization/56205
25127 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25128 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25129 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25131 2013-02-05 Richard Biener <rguenther@suse.de>
25133 PR tree-optimization/53342
25134 PR tree-optimization/53185
25135 * tree-vectorizer.h (vect_check_strided_load): Remove.
25136 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25137 not disallow peeling for vectorized strided loads.
25138 (vect_check_strided_load): Make static and simplify.
25139 (vect_analyze_data_refs): Adjust.
25140 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25141 correctly when vectorizing strided loads.
25143 2013-02-05 Richard Biener <rguenther@suse.de>
25145 * doc/install.texi: Refer to ISL, not PPL.
25147 2013-02-05 Jan Hubicka <jh@suse.cz>
25149 PR tree-optimization/55789
25150 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25152 2013-02-05 Jan Hubicka <jh@suse.cz>
25154 PR tree-optimization/55789
25155 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25156 the dead call anyway.
25158 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25161 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25163 2013-02-04 Alexander Potapenko <glider@google.com>
25164 Jack Howarth <howarth@bromo.med.uc.edu>
25165 Jakub Jelinek <jakub@redhat.com>
25168 * config/darwin.c (sort_ctor_records): Stabilized qsort
25169 on constructor priority by using original position.
25170 (finalize_ctors): New routine to sort constructors by
25171 priority before use in assemble_integer.
25172 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25174 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25177 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25178 about visibility on artificial decls.
25179 * config/sol2.c (solaris_assemble_visibility): Likewise.
25181 2013-02-04 Kai Tietz <ktietz@redhat.com>
25184 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25185 argument and improve checking of return-argument types for 16-byte
25187 (ix86_function_value_1): Add additional valtype argument on call
25188 of function_value_64.
25189 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25190 handling infunction_value_64 function.
25192 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25194 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25196 2013-02-04 Richard Biener <rguenther@suse.de>
25198 PR tree-optimization/56188
25199 * tree-ssa-structalias.c (label_visit): Consider case with
25200 initially non-empty points-to set.
25201 (perform_var_substitution): Dump node mapping and clean up.
25203 2013-02-04 Richard Guenther <rguenther@suse.de>
25206 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25207 node prevail as last resort.
25208 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25209 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25211 2013-02-04 Richard Biener <rguenther@suse.de>
25213 PR tree-optimization/56113
25214 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25216 (equiv_class_lookup_or_add): ... this.
25217 (label_visit): Adjust and fix error in previous patch.
25218 (perform_var_substitution): Adjust.
25220 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25222 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25223 * config/sh/sh4-300.md: Likewise.
25224 * config/sh/sh4a.md: Likewise.
25225 * config/sh/constraints.md: Likewise.
25226 * config/sh/sh.md: Likewise.
25227 * config/sh/netbsd-elf.h: Likewise.
25228 * config/sh/predicates.md: Likewise.
25229 * config/sh/sh-protos.h: Likewise.
25230 * config/sh/ushmedia.h: Likewise.
25231 * config/sh/linux.h: Likewise.
25232 * config/sh/sh.c: Likewise.
25233 * config/sh/superh.h: Likewise.
25234 * config/sh/elf.h: Likewise.
25235 * config/sh/sh4.md: Likewise.
25236 * config/sh/sh.h: Likewise.
25238 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25240 * config/pa/constraints.md: Adjust unused letters. Change "T"
25241 constraint to match_test floating_point_store_memory_operand().
25242 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25243 (base14_operand): New.
25244 (floating_point_store_memory_operand): New.
25245 (integer_store_memory_operand): Revise to use base14_operand and
25246 reg_plus_base_memory_operand.
25247 (move_dest_operand): Allow symbolic_memory_operands.
25248 (symbolic_memory_operand): Check for LO_SOM.
25249 (symbolic_operand): Change default case to break.
25250 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25251 CONST_DOUBLE values to be reloaded by putting them into memory when
25252 the destination is a floating point register.
25253 (movdf): Remove code to handle CONST_DOUBLE.
25255 (reload_indf_r1): New.
25256 (reload_insf_r1): New.
25257 Consistently use "Q" and "T" constraints with integer and floating
25258 point move instructions, respectively.
25259 (movdi): Remove FAIL.
25260 Change predicate for source operand unamed DImode move from
25261 general_operand to move_src_operand.
25262 (umulsidi3): Change predicate for destination operand to
25264 Likewise for similar unamed patterns.
25265 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25266 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25267 (hppa_legitimize_address): Simplify mask calculation.
25268 (pa_emit_move_sequence): Revised handling of secondary reloads from
25269 REG+D addresses for floating point loads and stores. Directly handle
25270 loading CONST0_RTX (mode) to a floating point register.
25271 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25272 to floating point registers. Don't restrict secondary reloads to
25273 floating point registers to integer modes. Revise some comments and
25275 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25276 (pa_legitimate_address_p): New.
25277 (pa_legitimize_reload_address): New.
25278 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25279 (STRICT_REG_OK_FOR_BASE_P): New.
25280 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25281 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25283 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25284 Andrew Dixie <andrewd@gentrack.com>
25286 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25289 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25291 * expmed.c (extract_bit_field_1): Pass the full width of the
25292 structure to get_best_reg_extraction_insn.
25294 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25297 * configure.ac (use_cxa_atexit): Add AIX.
25298 * configure: Regenerate.
25300 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25302 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25305 * final.c (need_profile_function): New variable.
25306 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25307 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25308 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25309 notes, targetm.asm_out.function_prologue doesn't emit anything,
25310 HAVE_prologue and profiler should be emitted before prologue,
25311 set need_profile_function instead of emitting it.
25312 (final_scan_insn): If need_profile_function, emit
25313 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25314 NOTE_INSN_FUNCTION_BEG note.
25316 2013-02-01 Richard Henderson <rth@redhat.com>
25318 * config/rs6000/rs6000.md (smulditi3): New.
25321 * config/alpha/alpha.md (umulditi3): New.
25323 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25325 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25326 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25328 2013-02-01 Richard Biener <rguenther@suse.de>
25330 PR tree-optimization/56113
25331 * tree-ssa-structalias.c (label_visit): Reduce work for
25332 single-predecessor nodes.
25334 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25336 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25337 range isn't testing for zero.
25339 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25341 PR middle-end/56113
25342 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25344 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25345 Nick Clifton <nickc@redhat.com>
25347 * config/v850/constraints.md (Q): Define as a memory constraint.
25348 * config/v850/predicates.md (label_ref_operand): New predicate.
25349 (e3v5_shift_operand): New predicate.
25350 (ior_operator): New predicate.
25351 * config/v850/t-v850: Add e3v5 multilib.
25352 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25353 (v850_gen_movdi): Prototype.
25354 * config/v850/v850.c: Add support for e3v5 architecture.
25355 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25357 (construct_save_jarl): Add e3v5 long JARL support.
25358 (v850_adjust_insn_length): New function. Adjust length of call
25359 insns when using e3v5 instructions.
25360 (v850_gen_movdi): New function: Generate instructions to move a
25362 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25363 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25364 (TARGET_USE_FPU): Enable for e3v5.
25365 (CONST_OK_FOR_W): New macro.
25366 (ADJUST_INSN_LENGTH): Define.
25367 * config/v850/v850.md (UNSPEC_LOOP): Define.
25368 (attr cpu): Add v850e3v5.
25369 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25370 (movdi): New pattern.
25371 (movdi_internal): New pattern.
25372 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25373 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25374 (cstoresf4): Likewise.
25375 (cstoredf4): Likewise.
25376 (insv): New pattern.
25377 (rotlso3_a): New pattern.
25378 (rotlsi3_b): New pattern
25379 (rotlsi3_v850e3v5): New pattern.
25380 (doloop_begin): New pattern.
25381 (fix_loop_counter): New pattern.
25382 (doloop_end): New pattern.
25383 (branch_normal): Add e3v5 long branch support.
25384 (branch_invert): Likewise.
25385 (branch_z_normal): Likewise.
25386 (branch_z_invert): Likewise.
25387 (branch_nz_normal): Likewise.
25388 (branch_nz_invert): Likewise.
25389 (call_internal_short): Add e3v5 register-indirect JARL support.
25390 (call_internal_long): Likewise.
25391 (call_value_internal_short): Likewise.
25392 (call_value_internal_long): Likewise.
25393 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25394 (mloop): New option.
25395 * config.gcc: Add support for configuring v840e3v5 target.
25396 * doc/invoke.texi: Document new v850 specific command line options.
25398 2013-01-31 Paul Koning <ni1d@arrl.net>
25402 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25403 children if parent is a class.
25404 (prune_unused_types_prune): Don't add DW_AT_declaration.
25406 2013-01-31 Richard Biener <rguenther@suse.de>
25408 PR tree-optimization/56157
25409 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25410 match up operand with SLP child.
25412 2013-01-31 Jason Merrill <jason@redhat.com>
25415 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25416 parameters the first time.
25417 (gen_scheduled_generic_parms_dies): Check completeness here.
25419 2013-01-31 Richard Biener <rguenther@suse.de>
25421 PR middle-end/53073
25422 * common.opt (faggressive-loop-optimizations): New flag,
25423 enabled by default.
25424 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25425 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25426 infer_loop_bounds_from_undefined by it.
25428 2013-01-31 Richard Biener <rguenther@suse.de>
25430 PR tree-optimization/56150
25431 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25432 visit virtual operands.
25433 (find_uses_to_rename_bb): Likewise.
25435 2013-01-31 Richard Biener <rguenther@suse.de>
25437 PR tree-optimization/56150
25438 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25439 mixed store non-store stmts.
25441 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25444 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25445 LIBASAN_EARLY_SPEC is defined.
25446 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25447 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25449 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25452 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25453 invalid args instead of ICEing on it.
25454 (ix86_valid_target_attribute_tree): Return error_mark_node if
25455 ix86_valid_target_attribute_inner_p failed.
25456 (ix86_valid_target_attribute_p): Return false only if
25457 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25458 target("default") attribute.
25459 (sorted_attr_string): Change argument from const char * to tree,
25460 merge in all target attribute arguments rather than just one.
25461 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25462 instead of free. Avoid using strcat.
25463 (ix86_mangle_function_version_assembler_name): Mangle
25464 target("default") as if no target attribute is present. Adjust
25465 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25466 instead of xmalloc and XDELETEVEC instead of free.
25467 (ix86_function_versions): Don't return true if one of the decls
25468 doesn't have target attribute. If they don't and one of the decls
25469 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25470 sorted_attr_string caller. Use XDELETEVEC instead of free.
25471 (ix86_supports_function_versions): Remove.
25472 (make_name): Fix up formatting.
25473 (make_dispatcher_decl): Remove resolver_name and its initialization.
25474 Avoid leaking memory.
25475 (is_function_default_version): Return true if there is
25476 target("default") attribute rather than no target attribute at all.
25477 (make_resolver_func): Avoid leaking memory.
25478 (ix86_generate_version_dispatcher_body): Likewise.
25479 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25480 * target.def (supports_function_versions): Remove.
25481 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25482 * doc/tm.texi: Regenerated.
25484 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25486 PR rtl-optimization/56144
25487 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25488 for values with side effects.
25490 2013-01-30 Richard Biener <rguenther@suse.de>
25492 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25493 (sparseset_pop): Likewise.
25494 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25495 to be able to use quick_push in the worker loop.
25497 2013-01-30 Marek Polacek <polacek@redhat.com>
25499 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25501 2013-01-30 Richard Biener <rguenther@suse.de>
25504 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25506 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25508 PR tree-optimization/56064
25509 * fixed-value.c (fixed_from_double_int): New function.
25510 * fixed-value.h (fixed_from_double_int): New prototype.
25511 (const_fixed_from_double_int): New static inline function.
25512 * fold-const.c (native_interpret_fixed): New static function.
25513 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25514 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25515 (native_encode_fixed): New static function.
25516 (native_encode_expr) <FIXED_CST>: Use it.
25517 (native_interpret_int): Move double_int worker code to...
25518 * double-int.c (double_int::from_buffer): ...this new static method.
25519 * double-int.h (double_int::from_buffer): Prototype it.
25521 2013-01-30 Richard Biener <rguenther@suse.de>
25523 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25524 New pointer-map and obstack.
25525 (init_alias_vars): Allocate pointer-map and obstack.
25526 (delete_points_to_sets): Free them.
25527 (find_what_var_points_to): Cache result.
25528 (find_what_p_points_to): Adjust for changed interface of
25529 find_what_var_points_to.
25530 (compute_points_to_sets): Likewise.
25531 (ipa_pta_execute): Likewise.
25533 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25535 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25536 * configure: Regenerate.
25537 * config.in: Regenerate.
25538 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25539 #nobits/#progbits if supported.
25541 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25544 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25545 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25546 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25548 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25550 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25551 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25553 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25555 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25556 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25558 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25560 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25562 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25563 * config/arm/cortex-a7.md: New bypasses using
25564 arm_mac_accumulator_is_result.
25566 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25568 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25569 (cortex_a7_neon_mla): Likewise.
25570 (cortex_a7_fpfmad): New reservation.
25571 (cortex_a7_fpmacs): Use ffmas and update required units.
25572 (cortex_a7_fpmuld): Update required units and latency.
25573 (cortex_a7_fpmacd): Likewise.
25574 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25575 (cortex_a7_neon). Likewise.
25576 (bypass) Update participating units.
25578 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25580 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25581 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25583 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25584 (vfp_fmul): Use ffmad.
25585 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25586 (cortex_r4_fmacd): Use ffmad.
25587 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25588 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25589 (cortex_a9_fmacd): Use ffmad.
25590 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25591 (cortex_a8_vfp_macd): Use ffmad.
25592 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25593 (cortex_a5_fpmacd): Use ffmad.
25594 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25595 (cortex_a15_vfp_macd): Use ffmad.
25596 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25598 2013-01-29 Jason Merrill <jason@redhat.com>
25601 * varasm.c (default_assemble_visibility): Don't warn about
25602 visibility on artificial decls.
25604 2013-01-29 Richard Biener <rguenther@suse.de>
25606 PR tree-optimization/56113
25607 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25609 (label_visit): Free duplicate bitmaps and record the leader instead.
25610 (perform_var_substitution): Adjust.
25612 2013-01-29 Richard Biener <rguenther@suse.de>
25614 PR tree-optimization/55270
25615 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25616 the CFG, schedule loops for fixup.
25618 2013-01-29 Nick Clifton <nickc@redhat.com>
25620 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25623 2013-01-28 Leif Ekblad <leif@rdos.net>
25625 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25626 * config/i386/i386.h (TARGET_RDOS): New macro.
25627 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25628 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25629 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25630 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25631 DEFAULT_LARGE_SECTION_THRESHOLD.
25632 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25633 * config/i386/rdos.h: New file.
25634 * config/i386/rdos64.h: New file.
25636 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25639 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25642 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25644 PR rtl-optimization/56117
25645 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25646 call cselib_lookup_from_insn on the MEM before calling
25647 add_insn_mem_dependence.
25649 2013-01-28 Richard Biener <rguenther@suse.de>
25651 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25652 to a stmt that didn't have one.
25653 (copy_phis_for_bb): Likewise for PHI arguments.
25654 (copy_debug_stmt): Likewise for debug stmts.
25656 2013-01-28 Richard Biener <rguenther@suse.de>
25658 PR tree-optimization/56034
25659 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25660 (partition_builtin_p): Adjust.
25661 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25662 it is the last partition.
25663 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25664 up the vertex for the definition.
25665 (classify_partition): Classify whether a partition is a
25666 PKIND_REDUCTION, thus has uses outside of the loop.
25667 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25668 Merge all PKIND_REDUCTION partitions into the last partition.
25669 (tree_loop_distribution): Seed partitions from reductions as well.
25671 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25673 PR tree-optimization/56125
25674 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25675 pow(x,c) into sqrt(x) * powi(x, n/2) or
25676 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25677 optimizing for size.
25678 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25679 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25682 PR tree-optimization/56094
25683 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25684 to UNKNOWN_LOCATION while gimplifying expr.
25686 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25689 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25690 operand 0 in movabs insn template for -masm=intel asm alternative.
25691 (*movabs<mode>_2): Ditto for operand 1.
25693 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25696 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25699 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25701 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25704 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25706 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25707 Undef to avoid warning.
25709 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25711 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25712 * configure: Regenerate.
25714 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25716 PR tree-optimization/56098
25717 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25718 for stmts with volatile ops.
25719 (cond_store_replacement): Don't optimize if assign has volatile ops.
25720 (cond_if_else_store_replacement_1): Don't optimize if either
25721 then_assign or else_assign have volatile ops.
25722 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25725 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25727 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25729 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25731 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25732 missing ':' in asm example.
25734 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25736 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25737 entries into lane and laneq entries.
25738 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25739 Remove AdvSIMD scalar modes.
25740 (aarch64_sq<r>dmulh_laneq<mode>): New.
25741 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25743 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25744 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25745 * config/aarch64/iterators.md (VCOND): New.
25748 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25751 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25752 Add NULL LIBNAME argument to existing definitions.
25753 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25754 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25755 * config/avr/avr.c (DEF_BUILTIN): Same.
25756 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25757 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25758 implementation is available (DECL_ASSEMBLER_NAME is set).
25759 (avr_fold_absfx): New static function.
25760 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25761 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25762 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25763 AVR_BUILTIN_ABSLLK.
25764 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25765 (abshk, absk, abslk, absllk): Provide as static inline functions.
25767 2013-01-25 Marek Polacek <polacek@redhat.com>
25769 PR tree-optimization/56035
25770 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25772 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25774 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25775 (*movtf_internal_rex64): Add (!o,C) alternative
25776 (*movxf_internal_rex64): Ditto.
25777 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25779 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25781 * doc/invoke.texi: fix typo.
25782 * doc/objc.texi: fix typo.
25784 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25786 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25787 for the first two alternatives.
25789 2013-01-24 Diego Novillo <dnovillo@google.com>
25791 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25792 (ggc-zone.o): Remove.
25793 * configure.ac: Remove option --with-gc.
25794 * configure: Re-generate.
25795 * doc/install.texi: Remove documentation for --with-gc.
25796 * gengtype.c (write_enum_defn): Remove. Update all users.
25797 (write_Types_process_field): Remove generation of gt_e_* argument.
25798 (output_type_enum): Remove. Update all users.
25799 (write_enum_defn): Remove. Update all users.
25800 (enum alloc_zone): Remove. Update all users.
25801 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25802 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25803 Update all callers.
25804 (struct ptr_data): Remove field TYPE. Update all users.
25805 (gt_pch_note_object): Remove argument TYPE. Update all users.
25806 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25808 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25809 (struct alloc_zone): Remove.
25810 (ggc_internal_alloc_zone_stat): Remove.
25811 (ggc_internal_cleared_alloc_zone_stat): Remove.
25812 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25813 (ggc_pch_count_object): Remove last argument. Update all users.
25814 (ggc_pch_alloc_object): Remove last argument. Update all users.
25815 (struct alloc_zone): Remove.
25816 * ggc-zone.c: Remove.
25817 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25818 (struct alloc_zone): Remove.
25819 (ggc_alloc_typed_stat): Remove.
25820 (ggc_alloc_typed): Remove.
25821 (ggc_splay_alloc): Remove first argument.
25822 (rtl_zone): Remove. Update all users.
25823 (tree_zone): Remove. Update all users.
25824 (tree_id_zone): Remove. Update all users.
25825 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25826 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25827 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25828 * tree-ssanames.c: Remove references to zone allocator in comments.
25830 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25832 * config/avr/avr.c (avr_out_fract): Make register numbers that
25833 might be outside of source operand signed.
25835 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25837 * config/i386/constraints.md (Yf): New constraint.
25838 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25839 of f constraint to conditionaly disable x87 register preferences.
25840 (*movdf_internal): Ditto.
25841 (*movsf_internal): Ditto.
25843 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25845 PR inline-asm/55934
25846 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25847 that have operands with impossible constraints.
25848 Add a FIXME for a speed-up opportunity.
25849 * lra-constraints.c (process_alt_operands): Verify that a class
25850 selected from constraints on asms is valid for the operand mode.
25851 (curr_insn_transform): Remove incorrect comment.
25853 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25855 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25856 TOC operand is a valid symbol ref in the constant pool.
25858 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25860 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25862 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25865 * config/avr/stdfix.h: New file.
25866 * t-avr (stdfix-gcc.h): New rule to build it.
25867 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25869 2013-01-23 Kostya Serebryany <kcc@google.com>
25871 * config/darwin.h: remove dependency on
25872 CoreFoundation (asan on Mac OS).
25874 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25877 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25878 instead of cmpdi_operand for first comparison operand.
25879 Don't assert that comparison operands aren't both constants.
25881 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25883 * doc/install.texi (Downloading the Source): Update references to
25884 downloading separate components.
25886 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25888 * doc/extend.texi (__int128): Improve grammar.
25890 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25893 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25894 alternative to (o,r).
25895 (*movdi_internal_rex64): Remove (!o,n) alternative.
25896 (DImode immediate->memory splitter): Remove.
25897 (DImode immediate->memory peephole2): Remove.
25898 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25899 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25900 alternative to (!o,*r).
25901 (*movtf_internal_sse): New pattern.
25902 (*movxf_internal_rex64): New pattern.
25903 (*movxf_internal): Disable for TARGET_64BIT.
25904 (*movdf_internal_rex64): Remove (!o,F) alternative.
25906 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25908 PR middle-end/56074
25909 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25910 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25911 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25912 stmt locations where LOCATION_LOCUS of the stmt location is
25913 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25916 * config/i386/i386.md (UNSPEC_STOS): New.
25917 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25918 *strsetqi_1): Add UNSPEC_STOS.
25920 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25923 * doc/invoke.texi: Remove left over -Wsynth example.
25925 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25927 PR tree-optimization/56051
25928 * fold-const.c (fold_binary_loc): Don't fold
25929 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25930 a narrowing conversion, or widening conversion from signed
25933 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25935 PR rtl-optimization/56023
25936 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25937 dependent on debug instruction.
25939 2013-01-21 Martin Jambor <mjambor@suse.cz>
25941 PR middle-end/56022
25942 * function.c (allocate_struct_function): Call
25943 invoke_set_current_function_hook earlier.
25945 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25947 * reload1.c (init_reload): Only initialize reload_obstack
25948 during the first call.
25950 2013-01-21 Marek Polacek <polacek@redhat.com>
25952 * cfgloop.c (verify_loop_structure): Fix up grammar.
25954 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25956 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25957 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25959 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25962 * config/arm/marvell-pj4.md: Update copyright year.
25963 Fix up use of alu to alu_reg and simple_alu_imm.
25965 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25967 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25969 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25972 * lra-constraints.c (curr_insn_transform): Don't reuse original
25973 insn for secondary memory move when memory mode should be different.
25975 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25977 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25978 atomic_storedi_1): New patterns.
25980 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25982 btver2 pipeline descriptions.
25983 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25985 * config/i386/i386.md (btver2_decode): New type attributes.
25986 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25988 * config/i386/btver2.md: New file describing btver2 pipelines.
25990 2013-01-19 Andrew Pinski <apinski@cavium.com>
25992 PR tree-optimization/52631
25993 * tree-ssa-sccvn (visit_use): Before looking up the original
25994 statement, try looking up the simplified expression.
25996 2013-01-19 Anthony Green <green@moxielogic.com>
25998 * config/moxie/moxie.c (moxie_expand_prologue): Set
25999 current_function_static_stack_size.
26001 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26003 PR tree-optimization/56029
26004 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26005 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26007 2013-01-18 Sharad Singhai <singhai@google.com>
26009 PR tree-optimization/55995
26010 * dumpfile.c (dump_loc): Print location only if available.
26011 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26013 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26016 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26017 secondary memory move.
26018 (inherit_reload_reg): Use rclass instead of cl for
26019 check_secondary_memory_needed_p.
26021 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26023 PR middle-end/56015
26024 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26025 the case where writing real complex part of target modifies op1.
26027 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26029 * config/aarch64/aarch64-simd.md
26030 (aarch64_vcond_internal<mode>): Handle unordered cases.
26031 * config/aarch64/iterators.md (v_cmp_result): New.
26033 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26034 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26036 * config/arm/marvell-pj4.md: New file.
26037 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26038 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26039 (generic_vfp): Likewise.
26040 * config/arm/arm-cores.def: Add marvell-pj4.
26041 * config/arm/arm-tune.md: Regenerate.
26042 * config/arm/arm-tables.opt: Regenerate.
26043 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26044 * doc/invoke.texi: Document marvell-pj4.
26046 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26048 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26050 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26055 * var-tracking.c (negative_power_of_two_p): New.
26056 (global_get_addr_cache, local_get_addr_cache): New.
26057 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26058 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26060 (vt_stack_offset_p): Remove.
26061 (vt_canon_true_dep): Always canonicalize loc's address.
26062 (clobber_overlapping_mems): Make sure we have a MEM.
26063 (local_get_addr_clear_given_value): New.
26064 (val_reset): Clear local cached entries.
26065 (compute_bb_dataflow): Create and release the local cache.
26066 Disable duplicate MEMs clobbering.
26067 (emit_notes_in_bb): Clobber MEMs likewise.
26068 (vt_emit_notes): Create and release the local cache.
26069 (vt_initialize, vt_finalize): Create and release the global
26070 cache, respectively.
26071 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26073 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26075 PR libmudflap/53359
26076 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26077 not found in the symtab.
26079 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26082 PR rtl-optimization/55547
26083 PR rtl-optimization/53827
26086 * alias.c (offset_overlap_p): New, factored out of...
26087 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26088 the conservative special case for symbolic constants. Don't
26089 adjust zero sizes on alignment.
26091 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26093 PR rtl-optimization/52573
26094 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26095 REG_UNUSED for the same register.
26097 2013-01-17 Richard Biener <rguenther@suse.de>
26098 Marek Polacek <polacek@redhat.com>
26100 PR rtl-optimization/55833
26101 * loop-unswitch.c (unswitch_loops): Move loop verification...
26102 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26103 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26104 Set it to true when we're removing a loop from hierarchy tree in
26105 an irreducible region.
26106 (fix_bb_placements): Adjust caller.
26107 (fix_loop_placements): Likewise.
26109 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26111 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26112 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26113 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26114 Remove ID. Adjust comments.
26115 * config/avr/avr-c.c (avr_builtin_name): Remove.
26116 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26117 * config/avr/avr.c (avr_tolower): New static function.
26118 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26119 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26120 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26123 2013-01-17 Jan Hubicka <jh@suse.cz>
26125 PR tree-optimization/55273
26126 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26128 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26131 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26132 store through atomic_store<mode>_1.
26133 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26135 2013-01-17 Martin Jambor <mjambor@suse.cz>
26137 PR tree-optimizations/55264
26138 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26139 for virtual methods.
26140 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26141 virtual methods before inlining is over.
26142 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26144 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26147 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26149 PR rtl-optimization/56005
26150 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26151 pending reads for prefetch.
26153 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26155 * config/aarch64/aarch64.md
26156 (*cstoresi_neg_uxtw): New pattern.
26157 (*cmovsi_insn_uxtw): New pattern.
26158 (*<optab>si3_uxtw): New pattern.
26159 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26160 (*<optab>si3_insn_uxtw): New pattern.
26161 (*bswapsi2_uxtw): New pattern.
26163 2013-01-16 Richard Biener <rguenther@suse.de>
26165 * tree-inline.c (tree_function_versioning): Remove set but
26166 never used variable.
26168 2013-01-16 Richard Biener <rguenther@suse.de>
26170 PR tree-optimization/55964
26171 * tree-flow.h (rename_variables_in_loop): Remove.
26172 (rename_variables_in_bb): Likewise.
26173 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26174 (copy_loop_before): Adjust and delete update-ssa status.
26175 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26176 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26177 (rename_variables_in_loop): Remove.
26178 (slpeel_update_phis_for_duplicate_loop): Likewise.
26179 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26180 use available cfg machinery instead of duplicating it.
26181 Update PHI nodes and perform poor-mans SSA update here.
26182 (slpeel_tree_peel_loop_to_edge): Adjust.
26184 2013-01-16 Richard Biener <rguenther@suse.de>
26186 PR tree-optimization/54767
26187 PR tree-optimization/53465
26188 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26189 (vrp_visit_phi_node): For PHI arguments coming via backedges
26190 drop all symbolical range information.
26191 (execute_vrp): Compute backedges.
26193 2013-01-16 Richard Biener <rguenther@suse.de>
26195 * doc/install.texi: Update CLooG and ISL requirements to
26198 2013-01-16 Christian Bruel <christian.bruel@st.com>
26201 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26202 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26203 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26205 2013-01-16 DJ Delorie <dj@redhat.com>
26207 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26208 (UNSPECV_SP_SWITCH_E): New.
26209 (sp_switch_1): Change to an unspec.
26210 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26213 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26215 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26216 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26217 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26218 (expand_mem_thread_fence): Ditto.
26219 (expand_mem_signal_fence): Ditto.
26220 (expand_atomic_load): Ditto.
26221 (expand_atomic_store): Ditto.
26223 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26225 PR rtl-optimization/55547
26226 PR rtl-optimization/53827
26229 * alias.c (memrefs_conflict_p): Set sizes to negative after
26232 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26235 * function.c (thread_prologue_and_epilogue_insns): Always
26236 add crtl->drap_reg to set_up_by_prologue.set, even if
26237 stack_realign_drap is false.
26239 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26241 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26242 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26243 *call): Fix indention.
26245 2013-01-15 Tom de Vries <tom@codesourcery.com>
26248 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26251 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26253 PR rtl-optimization/55153
26254 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26256 2013-01-15 Martin Jambor <mjambor@suse.cz>
26258 PR tree-optimization/55920
26259 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26260 accesses as grp_to_be_debug_replaced.
26262 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26264 PR tree-optimization/55920
26265 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26266 there is non-useless type conversion needed from debug rhs to lhs,
26267 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26269 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26270 Mikael Pettersson <mikpe@it.uu.se>
26273 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26275 (ASM_OUTPUT_CASE_LABEL): Remove.
26276 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26277 * final.c (shorten_branches): Update alignment of labels before
26278 jump tables if CASE_VECTOR_SHORTEN_MODE.
26280 2013-01-15 Richard Biener <rguenther@suse.de>
26283 * system.h: Do not include gmp.h for building host tools.
26285 2013-01-15 Richard Biener <rguenther@suse.de>
26287 PR middle-end/55882
26288 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26289 account for bitpos when computing alignment.
26291 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26293 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26294 (ix86_target_macros_internal): Likewise.
26296 * config/i386/i386.c (m_CORE2I7): Removed.
26297 (m_CORE_HASWELL): New macro.
26298 (m_CORE_ALL): Likewise.
26299 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26300 (initial_ix86_arch_features): Likewise.
26301 (processor_target_table): Initializations for Core avx2.
26302 (cpu_names): New names "core-avx2".
26303 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26304 PROCESSOR_CORE_HASWELL.
26305 (ix86_issue_rate): New case.
26306 (ia32_multipass_dfa_lookahead): Likewise.
26307 (ix86_sched_init_global): Likewise.
26309 * config/i386/i386.h (TARGET_HASWELL): New macro.
26310 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26311 (processor_type): New PROCESSOR_HASWELL.
26313 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26315 PR tree-optimization/55955
26316 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26317 *SHIFT_EXPR and *ROTATE_EXPR codes.
26319 PR tree-optimization/48766
26320 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26321 -ftrapv disable -fwrapv.
26323 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26326 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26327 etc. to 1 and not to __flash.
26328 Use LL suffix for __INT24_MAX__ with -mint8.
26329 Use ULL suffix for __UINT24_MAX__ with -mint8.
26331 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26333 * config/avr/avr-arch.h
26334 (struct base_arch_s): Use typedef avr_arch_t instead.
26335 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26336 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26337 * config/avr/avr.c: Same.
26338 * config/avr/avr-devices.c: Same.
26339 * config/avr/driver-avr.c: Same.
26340 * config/avr/gen-avr-mmcu-texi.c: Same.
26341 * config/avr/avr-mcus.def: Adjust comment.
26343 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26345 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26346 * config/aarch64/iterators.md (VALLDI): New.
26348 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26349 Andi Kleen <ak@linux.intel.com>
26352 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26353 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26356 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26358 * config/avr/avr-stdint.h: Remove trailing blanks.
26359 * config/avr/avr-log.h: Same.
26360 * config/avr/avr-arch.h: Same.
26361 * config/avr/avr-devices.c: Same.
26362 * config/avr/avr-dimode.md: Same.
26363 * config/avr/predicates.md: Same.
26364 * config/avr/avr-c.c: Same. And fix typo.
26366 * config/avr/avr-protos.h: Same. And:
26367 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26368 (init_cumulative_args): Rename to avr_init_cumulative_args.
26369 (expand_prologue): Rename to avr_expand_prologue.
26370 (expand_epilogue): Rename to avr_expand_epilogue.
26371 (adjust_insn_length): Rename to avr_adjust_insn_length.
26372 (notice_update_cc): Rename to avr_notice_update_cc.
26373 (final_prescan_insn): Rename to avr_final_prescan_insn.
26374 * config/avr/avr.c: Same.
26375 * config/avr/avr.h: Same.
26376 * config/avr/avr.md: Remove trailing blanks.
26377 (prologue): Use avr_expand_prologue.
26378 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26380 2013-01-14 Richard Biener <rguenther@suse.de>
26382 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26383 verify_location, collect_subblocks): New functions.
26384 (verify_gimple_in_cfg): Verify that locations only reference
26385 BLOCKs in the functions BLOCK tree.
26387 2013-01-14 Richard Biener <rguenther@suse.de>
26389 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26391 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26393 (insert_out_of_ssa_copy_on_edge): Likewise.
26394 (rewrite_close_phi_out_of_ssa): Likewise.
26395 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26397 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26398 propagated constants.
26399 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26402 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26404 * config/avr/avr-modes.def: Add GPL copyright notice.
26406 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26408 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26409 MEMMODEL_MASK to determine memory model.
26410 (atomic_store<mode>): Ditto from operands[2].
26411 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26413 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26416 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26417 (fold_gimple_assign): Don't call unshare_expr here.
26418 (fold_ctor_reference): Call unshare_expr.
26420 2013-01-13 Terry Guo <terry.guo@arm.com>
26422 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26423 * doc/fragments.texi: Document MULTILIB_REUSE.
26424 * gcc.c (multilib_reuse): New internal spec.
26425 (set_multilib_dir): Also search multilib from multilib_reuse.
26426 * genmultilib (tmpmultilib3): Refactor code.
26427 (tmpmultilib4): Ditto.
26428 (multilib_reuse): New multilib argument.
26430 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26432 * Makefile.in: Update copyright.
26434 2013-01-12 Tom de Vries <tom@codesourcery.com>
26436 PR middle-end/55890
26437 * calls.c (expand_call): Check if arg_nr is valid.
26439 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26441 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26442 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26443 documentation. Add missing '__' in front of
26444 __builtin_ia32_packssdw256.
26446 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26449 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26450 NO_REGS for larl operands.
26451 (s390_reload_larl_operand): Use s390_load_address instead of
26454 2013-01-11 Richard Biener <rguenther@suse.de>
26456 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26457 (verify_node_sharing): ... here.
26458 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26460 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26462 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26463 Substitute TREECHECKING.
26464 * configure: Regenerate.
26465 * Makefile.in (TREECHECKING): New.
26467 2013-01-11 Richard Guenther <rguenther@suse.de>
26469 PR tree-optimization/44061
26470 * tree-vrp.c (extract_range_basic): Compute zero as
26471 value-range for __builtin_constant_p of function parameters.
26473 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26475 Update copyright years.
26477 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26479 PR rtl-optimization/55672
26480 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26481 const to be eliminable.
26483 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26485 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26486 * configure: Regenerate.
26488 2013-01-10 Richard Biener <rguenther@suse.de>
26490 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26492 2013-01-10 Richard Biener <rguenther@suse.de>
26495 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26496 locations for virtual PHI arguments.
26497 (rewrite_update_phi_arguments): Likewise.
26499 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26501 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26504 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26506 PR tree-optimization/55921
26507 * tree-complex.c (expand_complex_asm): New function.
26508 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26510 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26513 * config/s390/s390.c (s390_symref_operand_p)
26514 (s390_loadrelative_operand_p): Merge the two functions.
26515 (s390_check_qrst_address, print_operand_address): Add parameters
26516 to s390_loadrelative_operand_p invokation.
26517 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26518 (s390_reload_larl_operand, s390_secondary_reload): Use
26519 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26520 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26522 2013-01-09 Mike Stump <mikestump@comcast.net>
26524 * dse.c (record_store): Remove unnecessary assert.
26526 2013-01-09 Jan Hubicka <jh@suse.cz>
26528 PR tree-optimization/55569
26529 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26530 * cfgloop.h (scale_loop_profile): Likewise.
26532 2013-01-09 Jan Hubicka <jh@suse.cz>
26535 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26537 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26539 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26541 PR middle-end/55114
26542 * expr.h (maybe_emit_group_store): Declare.
26543 * expr.c (maybe_emit_group_store): New function.
26544 * builtins.c (expand_builtin_int_roundingfn): Call it.
26545 (expand_builtin_int_roundingfn_2): Likewise.
26547 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26549 PR rtl-optimization/55829
26550 * lra-constraints.c (match_reload): Add code for absent output.
26551 (curr_insn_transform): Add code for reloads of matched inputs
26554 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26556 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26557 attribute of movddup insn to DF.
26558 (*vec_interleave_lowv2df): Ditto.
26559 (vec_dupv2df): Ditto.
26561 2013-01-09 Jan Hubicka <jh@suse.cz>
26563 PR tree-optimiation/55875
26564 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26565 EVERY_ITERATION parameter.
26566 (number_of_iterations_exit): Check if exit is executed every iteration.
26567 (idx_infer_loop_bounds): Similarly here.
26568 (n_of_executions_at_most): Simplify
26569 to only test for cases where statement is dominated by the
26570 particular bound; handle correctly the "postdominance" test.
26571 (scev_probably_wraps_p): Use max loop iterations info
26572 as a global bound first.
26574 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26575 Nick Clifton <nickc@redhat.com>
26577 * config/v850/v850.md (cbranchsf4): New pattern.
26578 (cstoresf4): New pattern.
26579 (cbranchdf4): New pattern.
26580 (cstoredf4): New pattern.
26581 (movsicc): Disallow floating point comparisons.
26582 (cmpsf_le_insn): Fix order of operators.
26583 (cmpsf_lt_insn): Likewise.
26584 (cmpsf_eq_insn): Likewise.
26585 (cmpdf_le_insn): Likewise.
26586 (cmpdf_lt_insn): Likewise.
26587 (cmpdf_eq_insn): Likewise.
26588 (cmpsf_ge_insn): Use LE comparison.
26589 (cmpdf_ge_insn): Likewise.
26590 (cmpsf_gt_insn): Use LT comparison.
26591 (cmpdf_gt_insn): Likewise.
26592 (cmpsf_ne_insn): Delete pattern.
26593 (cmpdf_ne_insn): Delete pattern.
26594 * config/v850/v850.c (v850_gen_float_compare): Use
26595 gen_cmpdf_eq_insn for NE comparison.
26596 (v850_float_z_comparison_operator)
26597 (v850_float_nz_comparison_operator): Move from here ...
26598 * config/v850/predicates.md: ... to here. Move GT and GE
26599 comparisons into v850_float_z_comparison_operator.
26600 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26602 (v850_float_nz_comparison_operator): Likewise.
26604 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26606 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26607 with calls to gen_insvsi/gen_insvdi.
26609 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26611 * config/i386/i386.c (initial_ix86_tune_features): Set up
26612 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26614 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26615 Jakub Jelinek <jakub@redhat.com>
26617 PR tree-optimization/48189
26618 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26619 If nitercst is 0, don't predict the exit edge.
26621 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26623 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26624 in asm_fprintf with reg_names.
26625 (aarch64_print_operand_address): Likewise.
26626 (aarch64_return_addr): Likewise.
26627 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26629 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26631 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26632 (INT_U6_BITS): Likewise.
26633 * config/pa/predicates.md (uint6_operand): New predicate.
26634 (shift5_operand, shift6_operand): Likewise.
26635 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26637 (lshrdi3): Use shift6_operand.
26638 (shrpsi4, shrpdi4): New insn patterns.
26639 (extzv): Delete expander.
26640 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26641 predicates in unamed zero extract patterns. Tighten common constraint.
26642 (extv): Delete expander.
26643 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26644 predicates in unamed sign extract patterns. Tighten common constraint.
26645 (insv): Delete expander.
26646 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26647 predicates in unamed insert patterns. Tighten common constraint.
26648 Change uint32_operand predicate to uint6_operand predicate in unamed
26649 DImode pattern to insert constant values of type 1...1xxxx.
26651 2013-01-04 Jan Hubicka <jh@suse.cz>
26653 PR tree-optimization/55823
26654 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26657 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26658 Uros Bizjak <ubizjak@gmail.com>
26660 PR rtl-optimization/55845
26661 * df-problems.c (can_move_insns_across): Stop scanning at
26662 volatile_insn_p source instruction or give up if
26663 across_from .. across_to range contains any volatile_insn_p
26666 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26668 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26669 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26671 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26672 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26674 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26677 * asan.c (asan_clear_shadow): New function.
26678 (asan_emit_stack_protection): Use it.
26680 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26682 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26683 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26684 with tab instead of space.
26686 2013-01-08 Nick Clifton <nickc@redhat.com>
26688 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26689 register bank 0 at the start of an interrupt handler.
26690 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26693 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26695 * config/aarch64/aarch64-simd.md
26696 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26697 (aarch64_simd_bsl): Likewise.
26698 (aarch64_vcond_internal<mode>): Likewise.
26699 (vcond<mode><mode>): Likewise.
26700 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26701 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26703 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26705 * config/aarch64/aarch64-builtins.c
26706 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26708 2013-01-08 Martin Jambor <mjambor@suse.cz>
26711 * tree-sra.c (analyze_access_subtree): Return true also after
26712 potentially creating a debug-only replacement.
26714 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26716 PR middle-end/55890
26717 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26719 PR tree-optimization/54120
26720 * tree-vrp.c (range_fits_type_p): Don't allow
26721 src_precision < precision from signed vr to unsigned_p
26722 if vr->min or vr->max is negative.
26723 (simplify_float_conversion_using_ranges): Test can_float_p
26724 against CODE_FOR_nothing.
26726 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26727 Richard Biener <rguenther@suse.de>
26729 PR middle-end/55851
26730 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26731 types instead of just INTEGER_TYPE types.
26733 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26735 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26738 2013-01-07 Steve Ellcey <sellcey@mips.com>
26741 * config/mips/mips.opt: Change mad to mmad to match documentation.
26743 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26746 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26747 .progmemx.data now.
26749 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26752 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26753 (avr_addrspace_t): Add .section_name field.
26754 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26756 (avr_addrspace): Same. Initialize .section_name. Remove last
26757 NULL entry. Put __memx into .progmemx.data.
26758 (progmem_section_prefix): Remove.
26759 (avr_asm_init_sections): No need to initialize progmem_section.
26760 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26761 section name prefix.
26762 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26763 retrieve the progmem section.
26764 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26765 boundary to run over avr_addrspace[].
26766 (avr_register_target_pragmas): Ditto.
26768 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26770 * varasm.c (output_constant_def_contents): For asan_protect_global
26771 protected strings, adjust DECL_ALIGN if needed, before testing for
26773 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26774 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26776 (output_object_block): For asan protected decls, emit asan padding
26777 after their contents.
26778 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26779 (asan_finish_file): Test it here instead.
26781 2013-01-07 Nick Clifton <nickc@redhat.com>
26782 Matthias Klose <doko@debian.org>
26783 Doug Kwan <dougkwan@google.com>
26784 H.J. Lu <hongjiu.lu@intel.com>
26787 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26789 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26791 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26793 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26795 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26797 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26800 * doc/install.texi (Cross-Compiler-Specific Options): Document
26803 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26805 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26806 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26807 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26808 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26809 vqmovun_high_s64): Fix source operand number and update copyright.
26811 2013-01-07 Richard Biener <rguenther@suse.de>
26813 PR middle-end/55890
26814 * gimple.h (gimple_call_builtin_p): New overload.
26815 * gimple.c (validate_call): New function.
26816 (gimple_call_builtin_p): Likewise.
26817 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26818 Use gimple_call_builtin_p.
26819 (find_func_clobbers): Likewise.
26820 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26821 (strlen_optimize_stmt): Likewise.
26823 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26825 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26826 (vld1q_dup_*): Likewise.
26827 (vld1_*): Likewise.
26828 (vld1q_*): Likewise.
26829 (vld1_lane_*): Likewise.
26830 (vld1q_lane_*): Likewise.
26832 2013-01-07 Richard Biener <rguenther@suse.de>
26834 * lto-streamer.h (LTO_minor_version): Bump to 2.
26836 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26838 * config/aarch64/aarch64-protos.h
26839 (aarch64_const_double_zero_rtx_p): Rename to...
26840 (aarch64_float_const_zero_rtx_p): ...this.
26841 (aarch64_float_const_representable_p): New.
26842 (aarch64_output_simd_mov_immediate): Likewise.
26843 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26844 move immediate case.
26845 * config/aarch64/aarch64.c
26846 (aarch64_const_double_zero_rtx_p): Rename to...
26847 (aarch64_float_const_zero_rtx_p): ...this.
26848 (aarch64_print_operand): Allow printing of new constants.
26849 (aarch64_valid_floating_const): New.
26850 (aarch64_legitimate_constant_p): Check for valid floating-point
26852 (aarch64_simd_valid_immediate): Likewise.
26853 (aarch64_vect_float_const_representable_p): New.
26854 (aarch64_float_const_representable_p): Likewise.
26855 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26856 (aarch64_output_simd_mov_immediate): New.
26857 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26858 (*movdf_aarch64): Likewise.
26859 * config/aarch64/constraints.md (Ufc): New.
26860 (Y): call aarch64_float_const_zero_rtx.
26861 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26863 2013-01-07 Richard Biener <rguenther@suse.de>
26865 PR tree-optimization/55888
26866 PR tree-optimization/55862
26867 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26868 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26869 not if it is contained therein.
26871 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26873 * config/avr/t-avr: Typo.
26875 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26878 * config/avr/t-avr: Don't automatically rebuild
26879 $(srcdir)/config/avr/t-multilib
26880 $(srcdir)/config/avr/avr-tables.opt
26881 $(srcdir)/doc/avr-mmcu.texi
26882 (avr-mcus): New phony target to build them on request.
26883 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26884 * avr/avr-mcus.def: Adjust comments.
26886 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26888 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26890 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26892 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26894 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26896 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26898 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26900 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26901 to generate profiling.
26902 * config/rs6000/aix64.h (LIB_SPEC): Same.
26904 2013-01-04 Andrew Pinski <apinski@cavium.com>
26906 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26908 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26910 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26912 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26914 (ix86_expand_move): Ditto.
26915 (ix86_zero_extend_to_Pmode): Ditto.
26916 (ix86_expand_call): Ditto.
26917 (ix86_expand_special_args_builtin): Ditto.
26918 (ix86_expand_builtin): Ditto.
26920 2013-01-04 Richard Biener <rguenther@suse.de>
26922 PR tree-optimization/55862
26923 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26924 translating them through PHI nodes.
26926 2013-01-04 Martin Jambor <mjambor@suse.cz>
26928 PR tree-optimization/55755
26929 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26930 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26932 2013-01-04 Marek Polacek <polacek@redhat.com>
26934 PR middle-end/55859
26935 * opts.c (default_options_optimization): Clarify error message.
26937 2013-01-04 Richard Biener <rguenther@suse.de>
26939 PR middle-end/55863
26940 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26943 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26946 * config/pa/pa.md (movsi): Revert previous change.
26947 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26950 2013-01-03 Richard Henderson <rth@redhat.com>
26952 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26953 after eliminating TLS symbols.
26955 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26958 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26959 * graphite-poly.c (debug_gmp_value): Likewise.
26961 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26964 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26965 selected code model, define __code_mode_small__, __code_model_medium__,
26966 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26967 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26968 xchg temporary register with %k. Declare temporary register as
26970 [__x86_64__]: For medium and large code models, preserve %rbx register.
26972 2013-01-03 Richard Biener <rguenther@suse.de>
26974 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26975 (dump_subscript): Adjust.
26976 (finalize_ddr_dependent): Do not dump redundant info.
26977 (analyze_siv_subscript): Adjust.
26978 (subscript_dependence_tester): Likewise.
26979 (compute_affine_dependence): Likewise.
26981 2013-01-03 Richard Biener <rguenther@suse.de>
26984 2013-01-03 Richard Biener <rguenther@suse.de>
26986 PR tree-optimization/55857
26987 * tree-vect-stmts.c (vectorizable_load): Do not setup
26988 re-alignment for invariant loads.
26990 2013-01-02 Richard Biener <rguenther@suse.de>
26992 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26993 invariant load do not generate a vector load from the scalar location.
26995 2013-01-03 Richard Biener <rguenther@suse.de>
26997 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26998 for not vectorizing.
26999 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27000 not build INDIRECT_REFs, call get_name once only.
27001 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27002 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27004 2013-01-03 Richard Biener <rguenther@suse.de>
27006 PR tree-optimization/55857
27007 * tree-vect-stmts.c (vectorizable_load): Do not setup
27008 re-alignment for invariant loads.
27010 2013-01-03 Richard Biener <rguenther@suse.de>
27013 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27014 prefer a built-in decl.
27016 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27018 * gcc.c (process_command): Update copyright notice dates.
27019 * gcov.c (print_version): Likewise.
27020 * gcov-dump.c (print_version): Likewise.
27022 PR rtl-optimization/55838
27023 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27024 iv0.step, iv1.step and step.
27026 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27027 Marc Glisse <marc.glisse@inria.fr>
27029 PR tree-optimization/55832
27030 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27031 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27032 integer_{one,zero}_node.
27034 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27037 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27038 * var-tracking.c (reverse_op): Don't add reverse ops to
27039 VALUEs that have already
27040 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27042 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27044 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27046 2013-01-02 Teresa Johnson <tejohnson@google.com>
27048 * dumpfile.c (dump_loc): Print filename with location.
27049 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27050 new location_t parameter to emit complete unroll message with
27051 new dump framework.
27052 (canonicalize_loop_induction_variables): Compute loops location
27053 and pass to try_unroll_loop_completely.
27054 * loop-unroll.c (report_unroll_peel): New function.
27055 (peel_loops_completely): Use new dump format with location
27056 for main dumpfile message, and invoke report_unroll_peel on success.
27057 (decide_unrolling_and_peeling): Ditto.
27058 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27059 by report_unroll_peel.
27060 (decide_peel_completely): Ditto.
27061 (decide_unroll_constant_iterations): Ditto.
27062 (decide_unroll_runtime_iterations): Ditto.
27063 (decide_peel_simple): Ditto.
27064 (decide_unroll_stupid): Ditto.
27065 * cfgloop.c (get_loop_location): New function.
27066 * cfgloop.h (get_loop_location): Declare.
27068 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27070 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27073 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27075 PR middle-end/55198
27076 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27077 BLKmode objects when EXPAND_MEMORY is specified.
27079 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27081 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27083 (fold_builtin_cpu): Do not share cpu model decls across statements.
27085 2013-01-02 Jason Merrill <jason@redhat.com>
27088 * tree.c (build_array_type_1): Revert earlier change.
27090 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27092 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27094 * config/aarch64/aarch64-tune.md: Re-generate.
27096 2013-01-02 Richard Biener <rguenther@suse.de>
27098 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27099 invariant load do not generate a vector load from the scalar location.
27101 2013-01-02 Richard Biener <rguenther@suse.de>
27104 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27105 * configure: Regenerate.
27107 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27109 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27110 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27111 (expand_builtin_int_roundingfn_2): Keep the original target around
27112 for the fallback case.
27114 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27116 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27117 to be clear for sign changes.
27119 2013-01-01 Jan Hubicka <jh@suse.cz>
27121 * ipa-inline-analysis.c: Fix formatting.
27123 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27125 PR tree-optimization/55831
27126 * tree-vect-loop.c (get_initial_def_for_induction): Use
27127 gsi_after_labels instead of gsi_start_bb.
27129 Copyright (C) 2013 Free Software Foundation, Inc.
27131 Copying and distribution of this file, with or without modification,
27132 are permitted in any medium without royalty provided the copyright
27133 notice and this notice are preserved.